Cnp3bis7dfkcompressed PDF

Computer Networking : Principles,
Protocols and Practice

Release
Olivier Bonaventure
September 16, 2015

Contents
1 Preface 3
1.1 About the author . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Part 1: Principles 5
2.1 Connecting two hosts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Building a network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.4 The transport layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.5 Naming and addressing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
2.6 Sharing resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
2.7 The reference models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
3 Part 2: Protocols 109

3.1 The application layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
3.2 The Domain Name System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
3.3 Electronic mail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
3.4 The HyperText Transfer Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
3.5 Remote Procedure Calls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
3.6 Internet transport protocols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
3.7 The User Datagram Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
3.8 The Transmission Control Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
3.9 The Stream Control Transmission Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
3.10 Congestion control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
3.11 The network layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
3.12 The IPv6 subnet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
3.13 Routing in IP networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
3.14 Intradomain routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
3.15 Interdomain routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
3.16 Datalink layer technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
4 Part 3: Practice 227

4.1 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
4.2 Reliable transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
4.3 Open questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
4.4 Discussion questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
4.5 Serving applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
4.6 Sharing resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
4.7 Application layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
i
4.8 Configuring DNS and HTTP servers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
4.9 Experimenting with Internet transport protocols . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
4.10 Injecting TCP segments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
4.11 Experimenting with Internet congestion control . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
4.12 Configuring IPv6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
4.13 Inter-domain routing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
4.14 IP Address Assignment Methods and Intradomain Routing . . . . . . . . . . . . . . . . . . . . . 282
4.15 Inter-domain routing and BGP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
4.16 Local Area Networks: The Spanning Tree Protocol and Virtual LANs . . . . . . . . . . . . . . . 297
5 Appendices 303
5.1 Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
5.2 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
5.3 Indices and tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
Bibliography 309
Index 323
ii
Computer Networking : Principles, Protocols and Practice, Release
Contents 1
2 Contents
CHAPTER 1
Preface
This is the current draft of the second edition of the Computer Networking : Principles, Protocols and Practice.
The document is updated every week.
The first edition of this ebook has been written by Olivier Bonaventure. Laurent Vanbever, Virginie Van den
Schriek, Damien Saucez and Mickael Hoerdt have contributed to exercises. Pierre Reinbold designed the icons
used to represent switches and Nipaul Long has redrawn many figures in the SVG format. Stephane Bortzmeyer
sent many suggestions and corrections to the text. Additional information about the textbook is available at
http://inl.info.ucl.ac.be/CNP3
Note: Computer Networking : Principles, Protocols and Practice, (c) 2011, Olivier Bonaventure, Universite
catholique de Louvain (Belgium) and the collaborators listed above, used under a Creative Commons Attribution
(CC BY) license made possible by funding from The Saylor Foundations Open Textbook Challenge in order to
be incorporated into Saylor.org collection of open courses available at http://www.saylor.org. Full license terms
may be viewed at : http://creativecommons.org/licenses/by/3.0/
1.1 About the author
Olivier Bonaventure is currently professor at Universite catholique de Louvain (Belgium) where he leads the
IP Networking Lab and is vice-president of the ICTEAM institute. His research has been focused on Internet
protocols for more than twenty years. Together with his Ph.D. students, he has developed traffic engineering
techniques, performed various types of Internet measurements, improved the performance of routing protocols
such as BGP and IS-IS and participated to the development of new Internet protocols including shim6, LISP and
Multipath TCP. He frequently contributes to standardisation within the IETF. He was on the editorial board of
IEEE/ACM Transactions on Networking and is Education Director of ACM SIGCOMM.
3
4 Chapter 1. Preface
CHAPTER 2
Part 1: Principles
2.1 Connecting two hosts
Warning: This is an unpolished draft of the second edition of this ebook. If you find any error or have sugges-
tions to improve the text, please create an issue via https://github.com/obonaventure/cnp3/issues?milestone=1
The first step when building a network, even a worldwide network such as the Internet, is to connect two hosts
together. This is illustrated in the figure below.
Figure 2.1: Connecting two hosts together
To enable the two hosts to exchange information, they need to be linked together by some kind of physical media.
Computer networks have used various types of physical media to exchange information, notably :
electrical cable. Information can be transmitted over different types of electrical cables. The most common
ones are the twisted pairs (that are used in the telephone network, but also in enterprise networks) and the
coaxial cables (that are still used in cable TV networks, but are no longer used in enterprise networks).
Some networking technologies operate over the classical electrical cable.
optical fiber. Optical fibers are frequently used in public and enterprise networks when the distance be-
tween the communication devices is larger than one kilometer. There are two main types of optical fibers
: multimode and monomode. Multimode is much cheaper than monomode fiber because a LED can be
used to send a signal over a multimode fiber while a monomode fiber must be driven by a laser. Due to the
different modes of propagation of light, monomode fibers are limited to distances of a few kilometers while
multimode fibers can be used over distances greater than several tens of kilometers. In both cases, repeaters
can be used to regenerate the optical signal at one endpoint of a fiber to send it over another fiber.
wireless. In this case, a radio signal is used to encode the information exchanged between the communi-
cating devices. Many types of modulation techniques are used to send information over a wireless channel
5
and there is lot of innovation in this field with new techniques appearing every year. While most wireless
networks rely on radio signals, some use a laser that sends light pulses to a remote detector. These optical
techniques allow to create point-to-point links while radio-based techniques, depending on the directionality
of the antennas, can be used to build networks containing devices spread over a small geographical area.
2.1.1 The physical layer
These physical media can be used to exchange information once this information has been converted into a suitable
electrical signal. Entire telecommunication courses and textbooks are devoted to the problem of converting analog
or digital information into an electrical signal so that it can be transmitted over a given physical link. In this book,
we only consider two very simple schemes that allow to transmit information over an electrical cable. This enables
us to highlight the key problems when transmitting information over a physical link. We are only interested in
techniques that allow to transmit digital information through the wire and will focus on the transmission of bits,
i.e. either 0 or 1.
Note: Bit rate

In computer networks, the bit rate of the physical layer is always expressed in bits per second. One Mbps is one
million bits per second and one Gbps is one billion bits per second. This is in contrast with memory specifica-
tions that are usually expressed in bytes (8 bits), KiloBytes ( 1024 bytes) or MegaBytes (1048576 bytes). Thus
transferring one MByte through a 1 Mbps link lasts 8.39 seconds.
Bit rate Bits per second
1 Kbps 103
1 Mbps 106
1 Gbps 109
1 Tbps 1012
To understand some of the principles behind the physical transmission of information, let us consider the simple
case of an electrical wire that is used to transmit bits. Assume that the two communicating hosts want to transmit
one thousand bits per second. To transmit these bits, the two hosts can agree on the following rules :
On the sender side :
set the voltage on the electrical wire at +5V during one millisecond to transmit a bit set to 1
set the voltage on the electrical wire at -5V during one millisecond to transmit a bit set to 0
On the receiver side :
every millisecond, record the voltage applied on the electrical wire. If the voltage is set to +5V,
record the reception of bit 1. Otherwise, record the reception of bit 0
This transmission scheme has been used in some early networks. We use it as a basis to understand how hosts com-
municate. From a Computer Science viewpoint, dealing with voltages is unusual. Computer scientists frequently
rely on models that enable them to reason about the issues that they face without having to consider all implemen-
tation details. The physical transmission scheme described above can be represented by using a time-sequence
diagram.
A time-sequence diagram describes the interactions between communicating hosts. By convention, the communi-
cating hosts are represented in the left and right parts of the diagram while the electrical link occupies the middle
of the diagram. In such a time-sequence diagram, time flows from the top to the bottom of the diagram. The trans-
mission of one bit of information is represented by three arrows. Starting from the left, the first horizontal arrow
represents the request to transmit one bit of information. This request is represented by using a primitive which can
be considered as a kind of procedure call. This primitive has one parameter (the bit being transmitted) and a name
(DATA.request in this example). By convention, all primitives that are named something.request correspond to a
request to transmit some information. The dashed arrow indicates the transmission of the corresponding electrical
signal on the wire. Electrical and optical signals do not travel instantaneously. The diagonal dashed arrow indi-
cates that it takes some time for the electrical signal to be transmitted from Host A to Host B. Upon reception of the
electrical signal, the electronics on Host Bs network interface detects the voltage and converts it into a bit. This
bit is delivered as a DATA.indication primitive. All primitives that are named something.indication correspond
6 Chapter 2. Part 1: Principles

to the reception of some information. The dashed lines also represents the relationship between two (or more)
primitives. Such a time-sequence diagram provides information about the ordering of the different primitives, but
the distance between two primitives does not represent a precise amount of time.
Host A Physical link Host B

DATA.req(0)
0
DATA.ind(0)
Time-sequence diagrams are usual when trying to understand the characteristics of a given communication
scheme. When considering the above transmission scheme, is it useful to evaluate whether this scheme allows
the two communicating hosts to reliably exchange information ? A digital transmission will be considered as
reliable when a sequence of bits that is transmitted by a host is received correctly at the other end of the wire. In
practice, achieving perfect reliability when transmitting information using the above scheme is difficult. Several
problems can occur with such a transmission scheme.
The first problem is that electrical transmission can be affected by electromagnetic interferences. These inter-
ferences can have various sources including natural phenomenons like thunderstorms, variations of the magnetic
field, but also can be caused by interference with other electrical signals such as interference from neighboring
cables, interferences from neighboring antennas, ... Due to all these interferences, there is unfortunately no guar-
antee that when a host transmit one bit on a wire, the same bit is received at the other end. This is illustrated in the
figure below where a DATA.request(0) on the left host leads to a Data.indication(1) on the right host.

DATA.req(0)
DATA.ind(1)
With the above transmission scheme, a bit is transmitted by setting the voltage on the electrical cable to a specific
value during some period of time. We have seen that due to electromagnetic interferences, the voltage measured
by the receiver can differ from the voltage set by the transmitter. This is the main cause of transmission errors.
However, this is not the only type of problem that can occur. Besides defining the voltages for bits 0 and 1, the
above transmission scheme also specifies the duration of each bit. If one million bits are sent every second, then
each bit lasts 1 microsecond. On each host, the transmission (resp. the reception) of each bit is triggered by a local
clock having a 1 MHz frequency. These clocks are the second source of problems when transmitting bits over
a wire. Although the two clocks have the same specification, they run on different hosts, possibly at a different
temperature and with a different source of energy. In practice, it is possible that the two clocks do not operate at
exactly the same frequency. Assume that the clock of the transmitting host operates at exactly 1000000 Hz while
the receiving clock operates at 999999 Hz. This is a very small difference between the two clocks. However,
when using the clock to transmit bits, this difference is important. With its 1000000 Hz clock, the transmitting
host will generate one million bits during a period of one second. During the same period, the receiving host
will sense the wire 999999 times and thus will receive one bit less than the bits originally transmitted. This small
difference in clock frequencies implies that bits can disappear during their transmission on an electrical cable.
This is illustrated in the figure below.

DATA.req(0)
DATA.ind(0)
DATA.req(0)
DATA.req(1)
DATA.ind(1)
A similar reasoning applies when the clock of the sending host is slower than the clock of the receiving host. In
this case, the receiver will sense more bits than the bits that have been transmitted by the sender. This is illustrated
in the figure below where the second bit received on the right was not transmitted by the left host.
2.1. Connecting two hosts 7


DATA.req(0)
DATA.ind(0)
DATA.ind(0)
DATA.req(1)
DATA.ind(1)
From a Computer Science viewpoint, the physical transmission of information through a wire is often considered
as a black box that allows to transmit bits. This black box is often referred to as the physical layer service
and is represented by using the DATA.request and DATA.indication primitives introduced earlier. This physical
layer service facilitates the sending and receiving of bits. This service abstracts the technological details that are
involved in the actual transmission of the bits as an electromagnetic signal. However, it is important to remember
that the physical layer service is imperfect and has the following characteristics :
the Physical layer service may change, e.g. due to electromagnetic interferences, the value of a bit being
transmitted
the Physical layer service may deliver more bits to the receiver than the bits sent by the sender
the Physical layer service may deliver fewer bits to the receiver than the bits sent by the sender
Many other types of encodings have been defined to transmit information over an electrical cable. All physical
layers are able to send and receive physical symbols that represent values 0 and 1. However, for various reasons
that are outside the scope of this chapter, several physical layers exchange other physical symbols as well. For
example, the Manchester encoding used in several physical layers can send four different symbols. The Manch-
ester encoding is a differential encoding scheme in which time is divided into fixed-length periods. Each period is
divided in two halves and two different voltage levels can be applied. To send a symbol, the sender must set one
of these two voltage levels during each half period. To send a 1 (resp. 0), the sender must set a high (resp. low)
voltage during the first half of the period and a low (resp. high) voltage during the second half. This encoding
ensures that there will be a transition at the middle of each period and allows the receiver to synchronise its clock
to the senders clock. Apart from the encodings for 0 and 1, the Manchester encoding also supports two additional
symbols : InvH and InvB where the same voltage level is used for the two half periods. By definition, these two
symbols cannot appear inside a frame which is only composed of 0 and 1. Some technologies use these special
symbols as markers for the beginning or end of frames.
Figure 2.2: Manchester encoding
Figure 2.3: The Physical layer
All the functions related to the physical transmission or information through a wire (or a wireless link) are usually
known as the physical layer. The physical layer allows thus two or more entities that are directly attached to the

same transmission medium to exchange bits. Being able to exchange bits is important as virtually any information
can be encoded as a sequence of bits. Electrical engineers are used to processing streams of bits, but computer
scientists usually prefer to deal with higher level concepts. A similar issue arises with file storage. Storage devices
such as hard-disks also store streams of bits. There are hardware devices that process the bit stream produced by
a hard-disk, but computer scientists have designed filesystems to allow applications to easily access such storage
devices. These filesystems are typically divided into several layers as well. Hard-disks store sectors of 512 bytes
or more. Unix filesystems group sectors in larger blocks that can contain data or inodes representing the structure
of the filesystem. Finally, applications manipulate files and directories that are translated in blocks, sectors and
eventually bits by the operating system.
Computer networks use a similar approach. Each layer provides a service that is built above the underlying layer
and is closer to the needs of the applications. The datalink layer builds upon the service provided by the physical
layer. We will see that it also contains several functions.
2.1.2 The datalink layer
Computer scientists are usually not interested in exchanging bits between two hosts. They prefer to write software
that deals with larger blocks of data in order to transmit messages or complete files. Thanks to the physical layer
service, it is possible to send a continuous stream of bits between two hosts. This stream of bits can include logical
blocks of data, but we need to be able to extract each block of data from the bit stream despite the imperfections
of the physical layer. In many networks, the basic unit of information exchanged between two directly connected
hosts is often called a frame. A frame can be defined has a sequence of bits that has a particular syntax or structure.
We will see examples of such frames later in this chapter.
To enable the transmission/reception of frames, the first problem to be solved is how to encode a frame as a
sequence of bits, so that the receiver can easily recover the received frame despite the limitations of the physical
layer.
If the physical layer were perfect, the problem would be very simple. We would simply need to define how to
encode each frame as a sequence of consecutive bits. The receiver would then easily be able to extract the frames
from the received bits. Unfortunately, the imperfections of the physical layer make this framing problem slightly
more complex. Several solutions have been proposed and are used in practice in different network technologies.
Framing
The framing problem can be defined as : How does a sender encode frames so that the receiver can efficiently
extract them from the stream of bits that it receives from the physical layer.
A first solution to this problem is to require the physical layer to remain idle for some time after the transmission of
each frame. These idle periods can be detected by the receiver and serve as a marker to delineate frame boundaries.
Unfortunately, this solution is not acceptable for two reasons. First, some physical layers cannot remain idle and
always need to transmit bits. Second, inserting an idle period between frames decreases the maximum bit rate that
can be achieved.
Note: Bit rate and bandwidth

Bit rate and bandwidth are often used to characterize the transmission capacity of the physical service. The original
definition of bandwidth, as listed in the Webster dictionary is a range of radio frequencies which is occupied by
a modulated carrier wave, which is assigned to a service, or over which a device can operate. This definition
corresponds to the characteristics of a given transmission medium or receiver. For example, the human ear is able
to decode sounds in roughly the 0-20 KHz frequency range. By extension, bandwidth is also used to represent
the capacity of a communication system in bits per second. For example, a Gigabit Ethernet link is theoretically
capable of transporting one billion bits per second.
Given that multi-symbol encodings cannot be used by all physical layers, a generic solution which can be used
with any physical layer that is able to transmit and receive only bits 0 and 1 is required. This generic solution is
called stuffing and two variants exist : bit stuffing and character stuffing. To enable a receiver to easily delineate

the frame boundaries, these two techniques reserve special bit strings as frame boundary markers and encode the
frames so that these special bit strings do not appear inside the frames.
Bit stuffing reserves the 01111110 bit string as the frame boundary marker and ensures that there will never be
six consecutive 1 symbols transmitted by the physical layer inside a frame. With bit stuffing, a frame is sent as
follows. First, the sender transmits the marker, i.e. 01111110. Then, it sends all the bits of the frame and inserts
an additional bit set to 0 after each sequence of five consecutive 1 bits. This ensures that the sent frame never
contains a sequence of six consecutive bits set to 1. As a consequence, the marker pattern cannot appear inside the
frame sent. The marker is also sent to mark the end of the frame. The receiver performs the opposite to decode a
received frame. It first detects the beginning of the frame thanks to the 01111110 marker. Then, it processes the
received bits and counts the number of consecutive bits set to 1. If a 0 follows five consecutive bits set to 1, this bit
is removed since it was inserted by the sender. If a 1 follows five consecutive bits sets to 1, it indicates a marker if
it is followed by a bit set to 0. The table below illustrates the application of bit stuffing to some frames.
Original frame Transmitted frame
0001001001001001001000011 01111110000100100100100100100001101111110
0110111111111111111110010 01111110011011111011111011111011001001111110
01111110 0111111001111101001111110
For example, consider the transmission of 0110111111111111111110010. The sender will first send the 01111110
marker followed by 011011111. After these five consecutive bits set to 1, it inserts a bit set to 0 followed by 11111.
A new 0 is inserted, followed by 11111. A new 0 is inserted followed by the end of the frame 110010 and the
01111110 marker.
Bit stuffing increases the number of bits required to transmit each frame. The worst case for bit stuffing is of course
a long sequence of bits set to 1 inside the frame. If transmission errors occur, stuffed bits or markers can be in
error. In these cases, the frame affected by the error and possibly the next frame will not be correctly decoded by
the receiver, but it will be able to resynchronize itself at the next valid marker.
Bit stuffing can be easily implemented in hardware. However, implementing it in software is difficult given the
complexity of performing bit manipulations in software. Software implementations prefer to process characters
than bits, software-based datalink layers usually use character stuffing. This technique operates on frames that
contain an integer number of characters. In computer networks, characters are usually encoded by relying on
the ASCII table. This table defines the encoding of various alphanumeric characters as a sequence of bits. RFC
20 provides the ASCII table that is used by many protocols on the Internet. For example, the table defines the
following binary representations :
A : 1000011 b
0 : 0110000 b
z : 1111010 b
@ : 1000000 b
space : 0100000 b
In addition, the ASCII table also defines several non-printable or control characters. These characters were de-
signed to allow an application to control a printer or a terminal. These control characters include CR and LF, that
are used to terminate a line, and the BEL character which causes the terminal to emit a sound.
NUL: 0000000 b
BEL: 0000111 b
CR : 0001101 b
LF : 0001010 b
DLE: 0010000 b
STX: 0000010 b
ETX: 0000011 b
Some characters are used as markers to delineate the frame boundaries. Many character stuffing techniques use
the DLE, STX and ETX characters of the ASCII character set. DLE STX (resp. DLE ETX) is used to mark the

beginning (end) of a frame. When transmitting a frame, the sender adds a DLE character after each transmitted
DLE character. This ensures that none of the markers can appear inside the transmitted frame. The receiver
detects the frame boundaries and removes the second DLE when it receives two consecutive DLE characters. For
example, to transmit frame 1 2 3 DLE STX 4, a sender will first send DLE STX as a marker, followed by 1 2 3
DLE. Then, the sender transmits an additional DLE character followed by STX 4 and the DLE ETX marker.
Original frame Transmitted frame
1234 DLE STX 1 2 3 4 DLE ETX
1 2 3 DLE STX 4 DLE STX 1 2 3 DLE DLE STX 4 DLE ETX
DLE STX DLE ETX DLE STX DLE DLE STX DLE DLE ETX** DLE ETX
Character stuffing , like bit stuffing, increases the length of the transmitted frames. For character stuffing, the worst
frame is a frame containing many DLE characters. When transmission errors occur, the receiver may incorrectly
decode one or two frames (e.g. if the errors occur in the markers). However, it will be able to resynchronise itself
with the next correctly received markers.
Bit stuffing and character stuffing allow to recover frames from a stream of bits or bytes. This framing mechanism
provides a richer service than the physical layer. Through the framing service, one can send and receive complete
frames. This framing service can also be represented by using the DATA.request and DATA.indication primitives.
This is illustrated in the figure below, assuming hypothetical frames containing four useful bit and one bit of
framing for graphical reasons.
Framing-A Phys-A Phys-B Framing-B

DATA.req(1...1)
DATA.req(0)
0
DATA.ind(0)
DATA.req(1)
1
DATA.ind(1)
DATA.req(1)
1
DATA.ind(1)
DATA.req(0)
0
DATA.ind(0)
DATA.ind(1...1)
We can now build upon the framing mechanism to allow the hosts to exchange frames containing an integer
number of bits or bytes. Once the framing problem has been solved, we can focus on designing a technique that
allows to reliably exchange frames.
Recovering from transmission errors
In this section, we develop a reliable datalink protocol running above the physical layer service. To design this
protocol, we first assume that the physical layer provides a perfect service. We will then develop solutions to
recover from the transmission errors.
The datalink layer is designed to send and receive frames on behalf of a user. We model these interactions by using
the DATA.req and DATA.ind primitives. However, to simplify the presentation and to avoid confusion between a
DATA.req primitive issued by the user of the datalink layer entity, and a DATA.req issued by the datalink layer
entity itself, we will use the following terminology :
the interactions between the user and the datalink layer entity are represented by using the classical
DATA.req and the DATA.ind primitives

the interactions between the datalink layer entity and the framing sublayer are represented by using send
instead of DATA.req and recvd instead of DATA.ind
When running on top of a perfect framing sublayer, a datalink entity can simply issue a send(SDU) upon arrival of
a DATA.req(SDU) 1 .Similarly, the receiver issues a DATA.ind(SDU) upon receipt of a recvd(SDU). Such a simple
protocol is sufficient when a single SDU is sent. This is illustrated in the figure below.
Host A Host B
DATA.req(SDU)
Frame(SDU)
DATA.ind(SDU)
Unfortunately, this is not always sufficient to ensure a reliable delivery of the SDUs. Consider the case where a
client sends tens of SDUs to a server. If the server is faster that the client, it will be able to receive and process all
the segments sent by the client and deliver their content to its user. However, if the server is slower than the client,
problems may arise. The datalink entity contains buffers to store SDUs that have been received as a Data.request
but have not yet been sent. If the application is faster than the physical link, the buffer may become full. At this
point, the operating system suspends the application to let the datalink entity empty its transmission queue. The
datalink entity also uses a buffer to store the received frames that have not yet been processed by the application.
If the application is slow to process the data, this buffer may overflow and the datalink entity will not able to
accept any additional frame. The buffers of the datalink entity have a limited size and if they overflow, the arriving
frames will be discarded, even if they are correct.
To solve this problem, a reliable protocol must include a feedback mechanism that allows the receiver to inform
the sender that it has processed a frame and that another one can be sent. This feedback is required even though
there are no transmission errors. To include such a feedback, our reliable protocol must process two types of
frames :
data frames carrying a SDU
control frames carrying an acknowledgment indicating that the previous frames was processed correctly
These two types of frames can be distinguished by dividing the frame in two parts :
the header that contains one bit set to 0 in data frames and set to 1 in control frames
the payload that contains the SDU supplied by the application
The datalink entity can then be modelled as a finite state machine, containing two states for the receiver and two
states for the sender. The figure below provides a graphical representation of this state machine with the sender
above and the receiver below.
The above FSM shows that the sender has to wait for an acknowledgement from the receiver before being able to
transmit the next SDU. The figure below illustrates the exchange of a few frames between two hosts.
Host A Host B
DATA.req(a)
D(a)
DATA.ind(a)
C(OK)
DATA.req(b)
D(b)
DATA.ind(b)
C(OK)
Note: Services and protocols

An important aspect to understand before studying computer networks is the difference between a service and a
protocol. In order to understand the difference between the two, it is useful to start with real world examples. The
traditional Post provides a service where a postman delivers letters to recipients. The Post defines precisely which
types of letters (size, weight, etc) can be delivered by using the Standard Mail service. Furthermore, the format
1 SDU is the acronym of Service Data Unit. We use it as a generic term to represent the data that is transported by a protocol.

Figure 2.4: Finite state machine of the simplest reliable protocol
of the envelope is specified (position of the sender and recipient addresses, position of the stamp). Someone who
wants to send a letter must either place the letter at a Post Office or inside one of the dedicated mailboxes. The
letter will then be collected and delivered to its final recipient. Note that for the regular service the Post usually
does not guarantee the delivery of each particular letter, some letters may be lost, and some letters are delivered to
the wrong mailbox. If a letter is important, then the sender can use the registered service to ensure that the letter
will be delivered to its recipient. Some Post services also provide an acknowledged service or an express mail
service that is faster than the regular service.
Reliable data transfer on top of an imperfect link
The datalink layer must deal with the transmission errors. In practice, we mainly have to deal with two types of
errors in the datalink layer :
Frames can be corrupted by transmission errors
Frames can be lost or unexpected frames can appear
A first glance, loosing frames might seem strange on single link. However, if we take framing into account,
transmission errors can affect the frame delineation mechanism and make the frame unreadable. For the same
reason, a receiver could receive two (likely invalid) frames after a sender has transmitted a single frame.
To deal with these types of imperfections, reliable protocols rely on different types of mechanisms. The first
problem is transmission errors. Data transmission on a physical link can be affected by the following errors :
random isolated errors where the value of a single bit has been modified due to a transmission error
random burst errors where the values of n consecutive bits have been changed due to transmission errors
random bit creations and random bit removals where bits have been added or removed due to transmission
errors
The only solution to protect against transmission errors is to add redundancy to the frames that are sent. Informa-
tion Theory defines two mechanisms that can be used to transmit information over a transmission channel affected
by random errors. These two mechanisms add redundancy to the transmitted information, to allow the receiver to
detect or sometimes even correct transmission errors. A detailed discussion of these mechanisms is outside the
scope of this chapter, but it is useful to consider a simple mechanism to understand its operation and its limitations.

Besides framing, datalink layers also include mechanisms to detect and sometimes even recover from transmission
errors. To allow a receiver to detect transmission errors, a sender must add some redundant information as an error
detection code to the frame sent. This error detection code is computed by the sender on the frame that it transmits.
When the receiver receives a frame with an error detection code, it recomputes it and verifies whether the received
error detection code matches the computer error detection code. If they match, the frame is considered to be valid.
Many error detection schemes exist and entire books have been written on the subject. A detailed discussion of
these techniques is outside the scope of this book, and we will only discuss some examples to illustrate the key
principles.
To understand error detection codes, let us consider two devices that exchange bit strings containing N bits. To
allow the receiver to detect a transmission error, the sender converts each string of N bits into a string of N+r
bits. Usually, the r redundant bits are added at the beginning or the end of the transmitted bit string, but some
techniques interleave redundant bits with the original bits. An error detection code can be defined as a function
that computes the r redundant bits corresponding to each string of N bits. The simplest error detection code is the
parity bit. There are two types of parity schemes : even and odd parity. With the even (resp. odd) parity scheme,
the redundant bit is chosen so that an even (resp. odd) number of bits are set to 1 in the transmitted bit string of
N+r bits. The receiver can easily recompute the parity of each received bit string and discard the strings with an
invalid parity. The parity scheme is often used when 7-bit characters are exchanged. In this case, the eighth bit is
often a parity bit. The table below shows the parity bits that are computed for bit strings containing three bits.
3 bits string Odd parity Even parity
000 1 0
001 0 1
010 0 1
100 0 1
111 0 1
110 1 0
101 1 0
011 1 0
The parity bit allows a receiver to detect transmission errors that have affected a single bit among the transmitted
N+r bits. If there are two or more bits in error, the receiver may not necessarily be able to detect the transmission
error. More powerful error detection schemes have been defined. The Cyclical Redundancy Checks (CRC) are
widely used in datalink layer protocols. An N-bits CRC can detect all transmission errors affecting a burst of
less than N bits in the transmitted frame and all transmission errors that affect an odd number of bits. Additional
details about CRCs may be found in [Williams1993].
It is also possible to design a code that allows the receiver to correct transmission errors. The simplest error
correction code is the triple modular redundancy (TMR). To transmit a bit set to 1 (resp. 0), the sender transmits
111 (resp. 000). When there are no transmission errors, the receiver can decode 111 as 1. If transmission errors
have affected a single bit, the receiver performs majority voting as shown in the table below. This scheme allows
the receiver to correct all transmission errors that affect a single bit.
Received bits Decoded bit
000 0
001 0
010 0
100 0
111 1
110 1
101 1
011 1
Other more powerful error correction codes have been proposed and are used in some applications. The Hamming
Code is a clever combination of parity bits that provides error detection and correction capabilities.
Reliable protocols use error detection schemes, but none of the widely used reliable protocols rely on error cor-
rection schemes. To detect errors, a frame is usually divided into two parts :
a header that contains the fields used by the reliable protocol to ensure reliable delivery. The header contains
a checksum or Cyclical Redundancy Check (CRC) [Williams1993] that is used to detect transmission errors

a payload that contains the user data

Some headers also include a length field, which indicates the total length of the frame or the length of the payload.
The simplest error detection scheme is the checksum. A checksum is basically an arithmetic sum of all the bytes
that a frame is composed of. There are different types of checksums. For example, an eight bit checksum can be
computed as the arithmetic sum of all the bytes of (both the header and trailer of) the frame. The checksum is
computed by the sender before sending the frame and the receiver verifies the checksum upon frame reception. The
receiver discards frames received with an invalid checksum. Checksums can be easily implemented in software,
but their error detection capabilities are limited. Cyclical Redundancy Checks (CRC) have better error detection
capabilities [SGP98], but require more CPU when implemented in software.
Note: Checksums, CRCs, ...

Most of the protocols in the TCP/IP protocol suite rely on the simple Internet checksum in order to verify that a
received packet has not been affected by transmission errors. Despite its popularity and ease of implementation,
the Internet checksum is not the only available checksum mechanism. Cyclical Redundancy Checks (CRC) are
very powerful error detection schemes that are used notably on disks, by many datalink layer protocols and file
formats such as zip or png. They can easily be implemented efficiently in hardware and have better error-detection
capabilities than the Internet checksum [SGP98] . However, CRCs are sometimes considered to be too CPU-
intensive for software implementations and other checksum mechanisms are preferred. The TCP/IP community
chose the Internet checksum, the OSI community chose the Fletcher checksum [Sklower89] . Nowadays there are
efficient techniques to quickly compute CRCs in software [Feldmeier95]
Since the receiver sends an acknowledgement after having received each data frame, the simplest solution to deal
with losses is to use a retransmission timer. When the sender sends a frame, it starts a retransmission timer. The
value of this retransmission timer should be larger than the round-trip-time, i.e. the delay between the transmis-
sion of a data frame and the reception of the corresponding acknowledgement. When the retransmission timer
expires, the sender assumes that the data segment has been lost and retransmits it. This is illustrated in the figure
Host A Host B
DATA.req(a)
start timer D(a)
DATA.ind(a)
C(OK)
cancel timer
DATA.req(b)
start timer D(b)
timer expires
D(b)
DATA.ind(b)
C(OK)
below.
Unfortunately, retransmission timers alone are not sufficient to recover from losses. Let us consider, as
an example, the situation depicted below where an acknowledgement is lost. In this case, the sender re-
transmits the data segment that has not been acknowledged. Unfortunately, as illustrated in the figure be-
low, the receiver considers the retransmission as a new segment whose payload must be delivered to its

Host A Host B
DATA.req(a)
start timer D(a)
DATA.ind(a)
C(OK)
cancel timer
DATA.req(b)
start timer D(b)
DATA.ind(b)
C(OK)
timer expires
D(b)
DATA.ind(b) !!!!!
C(OK)
user.
To solve this problem, datalink protocols associate a sequence number to each data frame. This sequence number
is one of the fields found in the header of data frames. We use the notation D(x,...) to indicate a data frame whose
sequence number field is set to value x. The acknowledgements also contain a sequence number indicating the data
frames that it is acknowledging. We use OKx to indicate an acknowledgement frame that confirms the reception
of D(x,...). The sequence number is encoded as a bit string of fixed length. The simplest reliable protocol is the
Alternating Bit Protocol (ABP).
The Alternating Bit Protocol uses a single bit to encode the sequence number. It can be implemented easily. The
sender and the receiver only require a four-state Finite State Machine.
Figure 2.5: Alternating bit protocol : Sender FSM
The initial state of the sender is Wait for D(0,...). In this state, the sender waits for a Data.request. The first
data frame that it sends uses sequence number 0. After having sent this frame, the sender waits for an OK0
acknowledgement. A frame is retransmitted upon expiration of the retransmission timer or if an acknowledgement
with an incorrect sequence number has been received.
The receiver first waits for D(0,...). If the frame contains a correct CRC, it passes the SDU to its user and sends
OK0. If the frame contains an invalid CRC, it is immediately discarded. Then, the receiver waits for D(1,...). In
this state, it may receive a duplicate D(0,...) or a data frame with an invalid CRC. In both cases, it returns an OK0
frame to allow the sender to recover from the possible loss of the previous OK0 frame.
Note: Dealing with corrupted frames

The receiver FSM of the Alternating bit protocol discards all frames that contain an invalid CRC. This is the safest
approach since the received frame can be completely different from the frame sent by the remote host. A receiver

Figure 2.6: Alternating bit protocol : Receiver FSM
should not attempt at extracting information from a corrupted frame because it cannot know which portion of the
frame has been affected by the error.
The figure below illustrates the operation of the alternating bit protocol.
Host A Host B
DATA.req(a)
start timer D(0,a)
DATA.ind(a)
C(OK0)
cancel timer
DATA.req(b)
start timer
D(1,b)
DATA.ind(b)
C(OK1)
cancel timer
DATA.req(c)
start timer D(0,c)
DATA.ind(c)
C(OK0)
cancel timer
The Alternating Bit Protocol can recover from the losses of data or control frames. This is
illustrated in the two figures below. The first figure shows the loss of one data segment.

Host A Host B
DATA.req(a)
start timer D(0,a)
DATA.ind(a)
C(OK0)
cancel timer
DATA.req(b)
start timer D(1,b)
timer expires
D(1,b)
DATA.ind(b)
C(OK1)
And
Host A Host B
DATA.req(a)
start timer D(0,a)
DATA.ind(a)
C(OK0)
cancel timer
DATA.req(b)
start timer D(1,b)
DATA.ind(b)
C(OK1)
timer expires
D(1,b)
Duplicate frame
ignored
C(OK1)
the loss of one control frame.
The Alternating Bit Protocol can recover from transmission errors and frame losses. However, it has one im-
portant drawback. Consider two hosts that are directly connected by a 50 Kbits/sec satellite link that has a 250
milliseconds propagation delay. If these hosts send 1000 bits frames, then the maximum throughput that can be
achieved by the alternating bit protocol is one frame every 20 + 250 + 250 = 520 milliseconds if we ignore the
transmission time of the acknowledgement. This is less than 2 Kbits/sec !
Go-back-n and selective repeat
To overcome the performance limitations of the alternating bit protocol, reliable protocols rely on pipelining. This
technique allows a sender to transmit several consecutive frames without being forced to wait for an acknowledge-
ment after each frame. Each data frame contains a sequence number encoded in an n bits field.
Pipelining allows the sender to transmit frames at a higher rate. However this higher transmission rate may
overload the receiver. In this case, the frames sent by the sender will not be correctly received by their final
destination. The reliable protocols that rely on pipelining allow the sender to transmit W unacknowledged frames
before being forced to wait for an acknowledgement from the receiving entity.
This is implemented by using a sliding window. The sliding window is the set of consecutive sequence numbers
that the sender can use when transmitting frames without being forced to wait for an acknowledgement. The figure

Figure 2.7: Pipelining improves the performance of reliable protocols
below shows a sliding window containing five segments (6,7,8,9 and 10). Two of these sequence numbers (6 and
7) have been used to send frames and only three sequence numbers (8, 9 and 10) remain in the sliding window.
The sliding window is said to be closed once all sequence numbers contained in the sliding window have been
used.
The figure below illustrates the operation of the sliding window. It uses a sliding window of three frames. The
sender can thus transmit three frames before being forced to wait for an acknowledgement. The sliding window
moves to the higher sequence numbers upon the reception of each acknowledgement. When the first acknowl-
edgement (OK0) is received, it allows the sender to move its sliding window to the right and sequence number 3
becomes available. This sequence number is used later to transmit the frame containing d.
In practice, as the frame header includes an n bits field to encode the sequence number, only the sequence numbers
between 0 and 2n 1 can be used. This implies that, during a long transfer, the same sequence number will be
used for different frames and the sliding window will wrap. This is illustrated in the figure below assuming that
2 bits are used to encode the sequence number in the frame header. Note that upon reception of OK1, the sender
slides its window and can use sequence number 0 again.
Unfortunately, frame losses do not disappear because a reliable protocol uses a sliding window. To recover from
losses, a sliding window protocol must define :
a heuristic to detect frame losses
a retransmission strategy to retransmit the lost frames
The simplest sliding window protocol uses the go-back-n recovery. Intuitively, go-back-n operates as follows.
A go-back-n receiver is as simple as possible. It only accepts the frames that arrive in-sequence. A go-back-n
receiver discards any out-of-sequence frame that it receives. When go-back-n receives a data frame, it always re-
turns an acknowledgement containing the sequence number of the last in-sequence frame that it has received. This
acknowledgement is said to be cumulative. When a go-back-n receiver sends an acknowledgement for sequence
number x, it implicitly acknowledges the reception of all frames whose sequence number is earlier than x. A key
advantage of these cumulative acknowledgements is that it is easy to recover from the loss of an acknowledge-
ment. Consider for example a go-back-n receiver that received frames 1, 2 and 3. It sent OK1, OK2 and OK3.
Unfortunately, OK1 and OK2 were lost. Thanks to the cumulative acknowledgements, when the receiver receives
OK3, it knows that all three frames have been correctly received.
The figure below shows the FSM of a simple go-back-n receiver. This receiver uses two variables : lastack and
next. next is the next expected sequence number and lastack the sequence number of the last data frame that has

Figure 2.8: The sliding window
Figure 2.9: Sliding window example

Figure 2.10: Utilisation of the sliding window with modulo arithmetic
been acknowledged. The receiver only accepts the frame that are received in sequence. maxseq is the number of
different sequence numbers (2n ).
A go-back-n sender is also very simple. It uses a sending buffer that can store an entire sliding window of
frames 2 . The frames are sent with increasing sequence numbers (modulo maxseq). The sender must wait for
an acknowledgement once its sending buffer is full. When a go-back-n sender receives an acknowledgement, it
removes from the sending buffer all the acknowledged frames and uses a retransmission timer to detect frame
losses. A simple go-back-n sender maintains one retransmission timer per connection. This timer is started when
the first frame is sent. When the go-back-n sender receives an acknowledgement, it restarts the retransmission
timer only if there are still unacknowledged frames in its sending buffer. When the retransmission timer expires,
the go-back-n sender assumes that all the unacknowledged frames currently stored in its sending buffer have been
lost. It thus retransmits all the unacknowledged frames in the buffer and restarts its retransmission timer.
The operation of go-back-n is illustrated in the figure below. In this figure, note that upon reception of the out-
of-sequence frame D(2,c), the receiver returns a cumulative acknowledgement C(OK,0) that acknowledges all the
frames that have been received in sequence. The lost frame is retransmitted upon the expiration of the retransmis-
sion timer.
The main advantage of go-back-n is that it can be easily implemented, and it can also provide good performance
when only a few frames are lost. However, when there are many losses, the performance of go-back-n quickly
drops for two reasons :
the go-back-n receiver does not accept out-of-sequence frames
the go-back-n sender retransmits all unacknowledged frames once it has detected a loss
Selective repeat is a better strategy to recover from losses. Intuitively, selective repeat allows the receiver to accept
out-of-sequence frames. Furthermore, when a selective repeat sender detects losses, it only retransmits the frames
that have been lost and not the frames that have already been correctly received.
A selective repeat receiver maintains a sliding window of W frames and stores in a buffer the out-of-sequence
frames that it receives. The figure below shows a five-frame receive window on a receiver that has already received
frames 7 and 9.
2 The size of the sliding window can be either fixed for a given protocol or negotiated during the connection establishment phase. Some
protocols allow to change the maximum window size during the data transfert. We will explain these techniques with real protocols later.

Figure 2.11: Go-back-n : receiver FSM
Figure 2.12: Go-back-n : sender FSM

Figure 2.13: Go-back-n : example
Figure 2.14: The receiving window with selective repeat

A selective repeat receiver discards all frames having an invalid CRC, and maintains the variable lastack as
the sequence number of the last in-sequence frame that it has received. The receiver always includes the value
of lastack in the acknowledgements that it sends. Some protocols also allow the selective repeat receiver to
acknowledge the out-of-sequence frames that it has received. This can be done for example by placing the list of
the correctly received, but out-of-sequence frames in the acknowledgements together with the lastack value.
When a selective repeat receiver receives a data frame, it first verifies whether the frame is inside its receiving
window. If yes, the frame is placed in the receive buffer. If not, the received frame is discarded and an acknowl-
edgement containing lastack is sent to the sender. The receiver then removes all consecutive frames starting at
lastack (if any) from the receive buffer. The payloads of these frames are delivered to the user, lastack and the
receiving window are updated, and an acknowledgement acknowledging the last frame received in sequence is
sent.
The selective repeat sender maintains a sending buffer that can store up to W unacknowledged frames. These
frames are sent as long as the sending buffer is not full. Several implementations of a selective repeat sender
are possible. A simple implementation associates one retransmission timer to each frame. The timer is started
when the frame is sent and cancelled upon reception of an acknowledgement that covers this frame. When a
retransmission timer expires, the corresponding frame is retransmitted and this retransmission timer is restarted.
When an acknowledgement is received, all the frames that are covered by this acknowledgement are removed
from the sending buffer and the sliding window is updated.
The figure below illustrates the operation of selective repeat when frames are lost. In this figure, C(OK,x) is used
to indicate that all frames, up to and including sequence number x have been received correctly.
Figure 2.15: Selective repeat : example
Pure cumulative acknowledgements work well with the go-back-n strategy. However, with only cumulative ac-
knowledgements a selective repeat sender cannot easily determine which frames have been correctly received after
a data frame has been lost. For example, in the figure above, the second C(OK,0) does not inform explicitly the
sender of the reception of D(2,c) and the sender could retransmit this frame although it has already been received.
A possible solution to improve the performance of selective repeat is to provide additional information about the
received frames in the acknowledgements that are returned by the receiver. For example, the receiver could add
in the returned acknowledgement the list of the sequence numbers of all frames that have already been received.
Such acknowledgements are sometimes called selective acknowledgements. This is illustrated in the figure above.
In the figure above, when the sender receives C(OK,0,[2]), it knows that all frames up to and including D(0,...)
have been correctly received. It also knows that frame D(2,...) has been received and can cancel the retransmission

timer associated to this frame. However, this frame should not be removed from the sending buffer before the
reception of a cumulative acknowledgement (C(OK,2) in the figure above) that covers this frame.
Note: Maximum window size with go-back-n and selective repeat

A reliable protocol that uses n bits to encode its sequence number can send up to 2n successive frames. However, to
ensure a reliable delivery of the frames, go-back-n and selective repeat cannot use a sending window of 2n frames.
Consider first go-back-n and assume that a sender sends 2n frames. These frames are received in-sequence by the
destination, but all the returned acknowledgements are lost. The sender will retransmit all frames. These frames
will all be accepted by the receiver and delivered a second time to the user. It is easy to see that this problem
can be avoided if the maximum size of the sending window is 2n 1 frames. A similar problem occurs with
selective repeat. However, as the receiver accepts out-of-sequence frames, a sending window of 2n 1 frames
is not sufficient to ensure a reliable delivery. It can be easily shown that to avoid this problem, a selective repeat
n
sender cannot use a window that is larger than 22 frames.
Reliable protocols often need to send data in both directions. To reduce the overhead caused by the acknowl-
edgements, most reliable protocols use piggybacking. Thanks to this technique, a datalink entity can place the
acknowledgements and the receive window that it advertises for the opposite direction of the data flow inside the
header of the data frames that it sends. The main advantage of piggybacking is that it reduces the overhead as it is
not necessary to send a complete frame to carry an acknowledgement. This is illustrated in the figure below where
the acknowledgement number is underlined in the data frames. Piggybacking is only used when data flows in both
directions. A receiver will generate a pure acknowledgement when it does not send data in the opposite direction
as shown in the bottom of the figure.
Figure 2.16: Piggybacking example
2.2 Building a network
In the previous section, we have explained how reliable protocols allow hosts to exchange data reliably even if the
underlying physical layer is imperfect and thus unreliable. Connecting two hosts together through a wire is the
2.2. Building a network 25

first step to build a network. However, this is not sufficient. Hosts usually need to interact with remote hosts that
are not directly connected through a direct physical layer link. This can be achieved by adding one layer above
the datalink layer : the network layer.
The main objective of the network layer is to allow endsystems, connected to different networks, to exchange
information through intermediate systems called router. The unit of information in the network layer is called a
packet.
R3 B
A R1 R2
Before explaining the network layer in detail, it is useful to remember the characteristics of the service provided
by the datalink layer. There are many variants of the datalink layer. Some provide a reliable service while others
do not provide any guarantee of delivery. The reliable datalink layer services are popular in environments such
as wireless networks were transmission errors are frequent. On the other hand, unreliable services are usually
used when the physical layer provides an almost reliable service (i.e. only a negligible fraction of the frames are
affected by transmission errors). Such almost reliable services are frequently in wired and optical networks. In
this chapter, we will assume that the datalink layer service provides an almost reliable service since this is both
the most general one and also the most widely deployed one.
Figure 2.17: The point-to-point datalink layer
There are two main types of datalink layers. The simplest datalink layer is when there are only two communicating
systems that are directly connected through the physical layer. Such a datalink layer is used when there is a point-
to-point link between the two communicating systems. The two systems can be endsystems or routers. PPP
(Point-to-Point Protocol), defined in RFC 1661, is an example of such a point-to-point datalink layer. Datalink
layers exchange frames and a datalink frame sent by a datalink layer entity on the left is transmitted through the
physical layer, so that it can reach the datalink layer entity on the right. Point-to-point datalink layers can either
provide an unreliable service (frames can be corrupted or lost) or a reliable service (in this case, the datalink layer
includes retransmission mechanisms). .. The unreliable service is frequently used above physical layers (e.g.
optical fiber, twisted pairs) having a low bit error ratio while reliability mechanisms are often used in wireless
networks to recover locally from transmission errors.
The second type of datalink layer is the one used in Local Area Networks (LAN). Conceptually, a LAN is a set of
communicating devices such that any two devices can directly exchange frames through the datalink layer. Both
endsystems and routers can be connected to a LAN. Some LANs only connect a few devices, but there are LANs
that can connect hundreds or even thousands of devices. In this chapter, we focus on the utilization of point-to-
point datalink layers. We will describe later the organisation and the operation of Local Area Networks and their
impact on the network layer.
Even if we only consider the point-to-point datalink layers, there is an important characteristics of these layers that
we cannot ignore. No datalink layer is able to send frames of unlimited size. Each datalink layer is characterized
by a maximum frame size. There are more than a dozen different datalink layers and unfortunately most of them
use a different maximum frame size. This heterogeneity in the maximum frame sizes will cause problems when
we will need to exchange data between hosts attached to different types of datalink layers.
As a first step, let us assume that we only need to exchange small amount of data. In this case, there is no issue
with the maximum length of the frames. However, there are other more interesting problems that we need to
tackle. To understand these problems, let us consider the network represented in the figure below.

A R1 R3 R5 B
R2 R4
This network contains two types of devices. The end hosts, represented as a small workstation and the routers,
represented as boxes with three arrows. An endhost is a device which is able to send and receive data for its own
usage in contrast with routers that most of the time forward data towards their final destination. Routers have
multiple links to neighboring routers or endhosts. Endhosts are usually attached via a single link to the network.
Nowadays, with the growth of wireless networks, more and more endhosts are equipped with several physical
interfaces. These endhosts are often called multihomed. Still, using several interfaces at the same time often leads
to practical issues that are beyond the scope of this document. For this reason, we will only consider single-homed
hosts in this ebook.
To understand the key principles behind the operation of a network, let us analyse all the operations that need to
be performed to allow host A in the above network to send one byte to host B. Thanks to the datalink layer used
above the A-R1 link, host A can easily send a byte to router R1 inside a frame. However, upon reception of this
frame, router R1 needs to understand that the byte is destined to host B and not to itself. This is the objective of
the network layer.
The network layer enables the transmission of information between hosts that are not directly connected through
intermediate routers. This transmission is carried out by putting the information to be transmitted inside a data
structure which is called a packet. Like a frame that contains useful data and control information, a packet also
contains useful data and control information. An important issue in the network layer is the ability to identify a
node (host or router) inside the network. This identification is performed by associating an address to each node.
An address is usually represented as a sequence of bits. Most networks use fixed-length addresses. At this stage,
let us simply assume that each of the nodes in the above network has an address which corresponds to the binary
representation on its name on the figure.
To send one byte of information to host B, host A needs to place this information inside a packet. In addition to the
data being transmitted, the packet must also contain either the addresses of the source and the destination nodes
or information that indicates the path that needs to be followed to reach the destination.
There are two possible organisations for the network layer :
datagram
virtual circuits
2.2.1 The datagram organisation
The first and most popular organisation of the network layer is the datagram organisation. This organisation is
inspired by the organisation of the postal service. Each host is identified by a network layer address. To send
information to a remote host, a host creates a packet that contains :
the network layer address of the destination host
its own network layer address
the information to be sent
To understand the datagram organisation, let us consider the figure below. A network layer address, represented
by a letter, has been assigned to each host and router. To send some information to host J, host A creates a packet
containing its own address, the destination address and the information to be exchanged.
With the datagram organisation, routers use hop-by-hop forwarding. This means that when a router receives a
packet that is not destined to itself, it looks up the destination address of the packet in its forwarding table. A
forwarding table is a data structure that maps each destination address (or set of destination addresses) to the

Figure 2.18: A simple internetwork
outgoing interface over which a packet destined to this address must be forwarded to reach its final destination.
The router consults its forwarding table for each packet that it handles.
The figure illustrates some possible forwarding tables in this network. By inspecting the forwarding tables of the
different routers, one can find the path followed by packets sent from a source to a particular destination. In the
example above, host A sends its packet to router R1. R1 consults its routing table and forwards the packet towards
R2. Based on its own routing table, R2 decides to forward the packet to R5 that can deliver it to its destination.
Thus, the path from A to J is A -> R1 -> R2 -> R5 -> J.
The computation of the forwarding tables of all the routers inside a network is a key element for the correct
operation of the network. This computation can be carried out in different ways and it is possible to use both
distributed and centralized algorithms. These algorithms provide different performance, may lead to different
types of paths, but their composition must lead to valid path.
In a network, a path can be defined as the list of all intermediate routers for a given source destination pair. For a
given source/destination pair, the path can be derived by first consulting the forwarding table of the router attached
to the source to determine the next router on the path towards the chosen destination. Then, the forwarding table
of this router is queried for the same destination... The queries continue until the destination is reached. In a
network that has valid forwarding tables, all the paths between all source/destination pairs contain a finite number
of intermediate routers. However, if forwarding tables have not been correctly computed, two types of invalid path
can occur.
A path may lead to a black hole. In a network, a black hole is a router that receives packets for at least one given
source/destination pair but does not have any entry inside its forwarding table for this destination. Since it does
not know how to reach the destination, the router cannot forward the received packets and must discard them. Any
centralized or distributed algorithm that computes forwarding tables must ensure that there are not black holes
inside the network.
A second type of problem may exist in networks using the datagram organisation. Consider a path that contains
a cycle. For example, router R1 sends all packets towards destination D via router R2, router R2 forwards these
packets to router R3 and finally router R3s forwarding table uses router R1 as its nexthop to reach destination D.
In this case, if a packet destined to D is received by router R1, it will loop on the R1 -> R2 -> R3 -> R1 cycle and
will never reach its final destination. As in the black hole case, the destination is not reachable from all sources in
the network. However, in practice the loop problem is worse than the black hole problem because when a packet is
caught in a forwarding loop, it unnecessarily consumes bandwidth. In the black hole case, the problematic packet
is quickly discarded. We will see later that network layer protocols include techniques to minimize the impact of
such forwarding loops.
Any solution which is used to compute the forwarding tables of a network must ensure that all destinations are
reachable from any source. This implies that it must guarantee the absence of black holes and forwarding loops.

The forwarding tables and the precise format of the packets that are exchanged inside the network are part of
the data plane of the network. This data plane contains all the protocols and algorithms that are used by hosts
and routers to create and process the packets that contain user data. On high-end routers, the data plane is often
implemented in hardware for performance reasons.
Besides the data plane, a network is also characterized by its control plane. The control plane includes all the
protocols and algorithms (often distributed) that are used to compute the forwarding tables that are installed on
all routers inside the network. While there is only one possible data plane for a given networking technology,
different networks using the same technology may use different control planes. The simplest control plane for
a network is always to compute manually the forwarding tables of all routers inside the network. This simple
control plane is sufficient when the network is (very) small, usually up to a few routers.
In most networks, manual forwarding tables are not a solution for two reasons. First, most networks are too large
to enable a manual computation of the forwarding tables. Second, with manually computed forwarding tables,
it is very difficult to deal with link and router failures. Networks need to operate 24h a day, 365 days per year.
During the lifetime of a network, many events can affect the routers and links that it contains. Link failures are
regular events in deployed networks. Links can fail for various reasons, including electromagnetic interference,
fiber cuts, hardware or software problems on the terminating routers, ... Some links also need to be added to the
network or removed because their utilisation is too low or their cost is too high. Similarly, routers also fail. There
are two types of failures that affect routers. A router may stop forwarding packets due to hardware or software
problem (e.g. due to a crash of its operating system). A router may also need to be halted from time to time (e.g.
to upgrade its operating system to fix some bugs). These planned and unplanned events affect the set of links and
routers that can be used to forward packets in the network. Still, most network users expect that their network will
continue to correctly forward packets despite all these events. With manually computed forwarding tables, it is
usually impossible to precompute the forwarding tables while taking into account all possible failure scenarios.
An alternative to manually computed forwarding tables is to use a network management platform that tracks the
network status and can push new forwarding tables on the routers when it detects any modification to the network
topology. This solution gives some flexibility to the network managers in computing the paths inside their network.
However, this solution only works if the network management platform is always capable of reaching all routers
even when the network topology changes. This may require a dedicated network that allows the management
platform to push information on the forwarding tables.
Another interesting point that is worth being discussed is when the forwarding tables are computed. A widely
used solution is to compute the entries of the forwarding tables for all destinations on all routers. This ensures that
each router has a valid route towards each destination. These entries can be updated when an event occurs and the
network topology changes. A drawback of this approach is that the forwarding tables can become large in large
networks since each router must maintain one entry for each destination at all times inside its forwarding table.
Some networks use the arrival of packets as the trigger to compute the corresponding entries in the forwarding
tables. Several technologies have been built upon this principle. When a packet arrives, the router consults its
forwarding table to find a path towards the destination. If the destination is present in the forwarding table, the
packet is forwarded. Otherwise, the router needs to find a way to forward the packet and update its forwarding
table.
Computing forwarding tables
Several techniques to update the forwarding tables upon the arrival of a packet have been used in deployed net-
works. In this section, we briefly present the principles that underly three of these techniques.
The first technique assumes that the underlying network topology is a tree. A tree is the simplest network to be
considered when forwarding packets. The main advantage of using a tree is that there is only one path between
any pair of nodes inside the network. Since a tree does not contain any cycle, it is impossible to have forwarding
loops in a tree-shaped network.
In a tree-shaped network, it is relatively simple for each node to automatically compute its forwarding table by
inspecting the packets that it receives. For this, each node uses the source and destination addresses present inside
each packet. The source address allows to learn the location of the different sources inside the network. Each
source has a unique address. When a node receives a packet over a given interface, it learns that the source
(address) of this packet is reachable via this interface. The node maintains a data structure that maps each known
source address to an incoming interface. This data structure is often called the port-address table since it indicates

the interface (or port) to reach a given address. Learning the location of the sources is not sufficient, nodes also
need to forward packets towards their destination. When a node receives a packet whose destination address is
already present inside its port-address table, it simply forwards the packet on the interface listed in the port-address
table. In this case, the packet will follow the port-address table entries in the downstream nodes and will reach
the destination. If the destination address is not included in the port-address table, the node simply forwards the
packet on all its interfaces, except the interface from which the packet was received. Forwarding a packet over
all interfaces is usually called broadcasting in the terminology of computer networks. Sending the packet over all
interfaces except one is a costly operation since the packet will be sent over links that do not reach the destination.
Given the tree-shape of the network, the packet will explore all downstream branches of the tree and will thus
finally reach its destination. In practice, the broadcasting operation does not occur too often and its cost is limited.
To understand the operation of the port-address table, let us consider the example network shown in the figure
below. This network contains three hosts : A, B and C and five nodes, R1 to R5. When the network boots, all the
forwarding tables of the nodes are empty.

R1
R2 R3
R4 R5
C
Host A sends a packet towards B. When receiving this packet, R1 learns that A is reachable via its North interface.
Since it does not have an entry for destination B in its port-address table, it forwards the packet to both R2 and
R3. When R2 receives the packet, it updates its own forwarding table and forward the packet to C. Since C is not
the intended recipient, it simply discards the received packet. Node R3 also received the packet. It learns that A is
reachable via its North interface and broadcasts the packet to R4 and R5. R5 also updates its forwarding table and
finally forwards it to destination B.Let us now consider what happens when B sends a reply to A. R5 first learns
that B is attached to its South port. It then consults its port-address table and finds that A is reachable via its North
interface. The packet is then forwarded hop-by-hop to A without any broadcasting. If C sends a packet to B, this

packet will reach R1 that contains a valid forwarding entry in its forwarding table.
By inspecting the source and destination addresses of packets, network nodes can automatically derive their for-
warding tables. As we will discuss later, this technique is used in Ethernet networks. Despite being widely used,
it has two important drawbacks. First, packets sent to unknown destinations are broadcasted in the network even
if the destination is not attached to the network. Consider the transmission of ten packets destined to Z in the
network above. When a node receives a packet towards this destination, it can only broadcast the packet. Since
Z is not attached to the network, no node will ever receive a packet whose source is Z to update its forwarding
table. The second and more important problem is that few networks have a tree-shaped topology. It is interesting
to analyze what happens when a port-address table is used in a network that contains a cycle. Consider the simple
network shown below with a single host.

R1
R2
R3
Assume that the network has started and all port-station and forwarding tables are empty. Host A sends a packet
towards B. Upon reception of this packet, R1 updates its port-address table. Since B is not present in the port-
address table, the packet is broadcasted. Both R2 and R3 receive a copy of the packet sent by A. They both update

their port-address table. Unfortunately, they also both broadcast the received packet. B receives a first copy of the
packet, but R3 and R2 receive it again. R3 will then broadcast this copy of the packet to B and R1 while R2 will
broadcast its copy to R1. Although B has already received two copies of the packet, it is still inside the network
and will continue to loop. Due to the presence of the cycle, a single packet towards an unknown destination
generates copies of this packet that loop and will saturate the network bandwidth. Network operators who are
using port-address tables to automatically compute the forwarding tables also use distributed algorithms to ensure
that the network topology is always a tree.
Another technique can be used to automatically compute forwarding tables. It has been used in interconnecting
Token Ring networks and in some wireless networks. Intuitively, Source routing enables a destination to auto-
matically discover the paths from a given source towards itself. This technique requires nodes to change some
information inside some packets. For simplicity, let us assume that the data plane supports two types of packets :
the data packets
the control packets
Data packets are used to exchange data while control packets are used to discover the paths between endhosts.
With Source routing, network nodes can be kept as simple as possible and all the complexity is placed on the
endhosts. This is in contrast with the previous technique where the nodes had to maintain a port-address and
a forwarding table while the hosts simply sent and received packets. Each node is configured with one unique
address and there is one identifier per outgoing link. For simplicity and to avoid cluttering the figures with those
identifiers, we will assume that each node uses as link identifiers north, west, south, ... In practice, a node would
associate one integer to each outgoing link.

R1
R2
R3
R4
2.2. Building a network B 35

In the network above, node R2 is attached to two outgoing links. R2 is connected to both R1 and R3. R2 can
easily determine that it is connected to these two nodes by exchanging packets with them or observing the packets
that it receives over each interface. Assume for example that when a host or node starts, it sends a special control
packet over each of its interfaces to advertise its own address to its neighbors. When a host or node receives such a
packet, it automatically replies with its own address. This exchange can also be used to verify whether a neighbor,
either node or host, is still alive. With source routing, the data plane packets include a list of identifiers. This list
is called a source route and indicates the path to be followed by the packet as a sequence of link identifiers. When
a node receives such a data plane packet, it first checks whether the packets destination is direct neighbor. In this
case, the packet is forwarded to the destination. Otherwise, the node extracts the next address from the list and
forwards it to the neighbor. This allows the source to specify the explicit path to be followed for each packet. For
example, in the figure above there are two possible paths between A and B. To use the path via R2, A would send a
packet that contains R1,R2,R3 as source route. To avoid going via R2, A would place R1,R3 as the source route in
its transmitted packet. If A knows the complete network topology and all link identifiers, it can easily compute the
source route towards each destination. If needed, it could even use different paths, e.g. for redundancy, to reach a
given destination. However, in a real network hosts do not usually have a map of the entire network topology.
In networks that rely on source routing, hosts use control packets to automatically discover the best path(s). In
addition to the source and destination addresses, control packets contain a list that records the intermediate nodes.
This list is often called the record route because it allows to record the route followed by a given packet. When a
node receives a control packet, it first checks whether its address is included in the record route. If yes, the control
packet is silently discarded. Otherwise, it adds its own address to the record route and forwards the packet to all
its interfaces, except the interface over which the packet has been received. Thanks to this, the control packet will
be able to explore all paths between a source and a given destination.
For example, consider again the network topology above. A sends a control packet towards B. The initial record
route is empty. When R1 receives the packet, it adds its own address to the record route and forwards a copy to R2
and another to R3. R2 receives the packet, adds itself to the record route and forwards it to R3. R3 receives two
copies of the packet. The first contains the [R1,R2] record route and the second [R1]. In the end, B will receive
two control packets containing [R1,R2,R3,R4] and [R1,R3,R4] as record routes. B can keep these two paths or
select the best one and discard the second. A popular heuristic is to select the record route of the first received
packet as being the best one since this likely corresponds to the shortest delay path.
With the received record route, B can send a data packet to A. For this, it simply reverses the chosen record route.
However, we still need to communicate the chosen path to A. This can be done by putting the record route inside
a control packet which is sent back to A over the reverse path. An alternative is to simply send a data packet back
to A. This packet will travel back to A. To allow A to inspect the entire path followed by the data packet, its source
route must contain all intermediate routers when it is received by A. This can be achieved by encoding the source
route using a data structure that contains an index and the ordered list of node addresses. The index always points
to the next address in the source route. It is initialized at 0 when a packet is created and incremented by each
intermediate node.
Flat or hierarchical addresses
The last, but important, point to discuss about the data plane of the networks that rely on the datagram mode is
their addressing scheme. In the examples above, we have used letters to represent the addresses of the hosts and
network nodes. In practice, all addresses are encoded as a bit string. Most network technologies use a fixed size
bit string to represent source and destination address. These addresses can be organized in two different ways.
The first organisation, which is the one that we have implicitly assumed until now, is the flat addressing scheme.
Under this scheme, each host and network node has a unique address. The unicity of the addresses is important for
the operation of the network. If two hosts have the same address, it can become difficult for the network to forward
packets towards this destination. Flat addresses are typically used in situations where network nodes and hosts
need to be able to communicate immediately with unique addresses. These flat addresses are often embedded
inside the hardware of network interface cards. The network card manufacturer creates one unique address for
each interface and this address is stored in the read-only memory of the interface. An advantage of this addressing
scheme is that it easily supports ad-hoc and mobile networks. When a host moves, it can attach to another network
and remain confident that its address is unique and enables it to communicate inside the new network.
With flat addressing the lookup operation in the forwarding table can be implemented as an exact match. The
forwarding table contains the (sorted) list of all known destination addresses. When a packet arrives, a network

node only needs to check whether this address is part of the forwarding table or not. In software, this is an
O(log(n)) operation if the list is sorted. In hardware, Content Addressable Memories can perform this lookup
operation efficiently, but their size is usually limited.
A drawback of the flat addressing scheme is that the forwarding tables grow linearly with the number of hosts and
nodes in the network. With this addressing scheme, each forwarding table must contain an entry that points to
every address reachable inside the network. Since large networks can contain tens of millions or more of hosts,
this is a major problem on network nodes that need to be able to quickly forward packets. As an illustration, it is
interesting to consider the case of an interface running at 10 Gbps. Such interfaces are found on high-end servers
and in various network nodes today. Assuming a packet size of 1000 bits, a pretty large and conservative number,
such interface must forward ten million packets every second. This implies that a network node that receives
packets over such a link must forward one 1000 bits packet every 100 nanoseconds. This is the same order of
magnitude as the memory access times of old DRAMs.
A widely used alternative to the flat addressing scheme is the hierarchical addressing scheme. This addressing
scheme builds upon the fact that networks usually contain much more hosts than network nodes. In this case, a
first solution to reduce the size of the forwarding tables is to create a hierarchy of addresses. This is the solution
chosen by the post office were addresses contain a country, sometimes a state or province, a city, a street and
finally a street number. When an enveloppe is forwarded by a postoffice in a remote country, it only looks at
the destination country, while a post office in the same province will look at the city information. Only the post
office responsible for a given city will look at the street name and only the postman will use the street number.
Hierarchical addresses provide a similar solution for network addresses. For example, the address of an Internet
host attached to a campus network could contain in the high-order bits an identification of the Internet Service
Provider (ISP) that serves the campus network. Then, a subsequent block of bits identifies the campus network
which is one of the customers from the ISP. Finally, the low order bits of the address identify the host in the
campus network.
This hierarchical allocation of addresses can be applied in any type of network. In practice, the allocation of
the addresses must follow the network topology. Usually, this is achieved by dividing the addressing space in
consecutive blocks and then allocating these blocks to different parts of the network. In a small network, the
simplest solution is to allocate one block of addresses to each network node and assign the host addresses from
the attached node.

R1
R2
R3
R4
38 B Chapter 2. Part 1: Principles

In the above figure, assume that the network uses 16 bits addresses and that the prefix 01001010 has been assigned
to the entire network. Since the network contains four routers, the network operator could assign one block
of sixty-four addresses to each router. R1 would use address 0100101000000000 while A could use address
0100101000000001. R2 could be assigned all adresses from 0100101001000000 to 0100101001111111. R4
could then use 0100101011000000 and assign 0100101011000001 to B. Other allocation schemes are possible.
For example, R3 could be allocated a larger block of addresses than R2 and R4 could use a sub-block from R3 s
address block.
The main advantage of hierarchical addresses is that it is possible to significantly reduce the size of the forwarding
tables. In many networks, the number of nodes can be several orders of magnitude smaller than the number of
hosts. A campus network may contain a few dozen of network nodes for thousands of hosts. The largest Internet
Services Providers typically contain no more than a few tens of thousands of network nodes but still serve tens or
hundreds of millions of hosts.
Despite their popularity, hierarchical addresses have some drawbacks. Their first drawback is that a lookup in
the forwarding table is more complex than when using flat addresses. For example, on the Internet, network
nodes have to perform a longest-match to forward each packet. This is partially compensated by the reduction in
the size of the forwarding tables, but the additional complexity of the lookup operation has been a difficulty to
implement hardware support for packet forwarding. A second drawback of the utilisation of hierarchical addresses
is that when a host connects for the first time to a network, it must contact one network node to determine its own
address. This requires some packet exchanges between the host and some network nodes. Furthermore, if a host
moves and is attached to another network node, its network address will change. This can be an issue with some
mobile hosts.
Dealing with heterogeneous datalink layers
Sometimes, the network layer needs to deal with heterogenous datalink layers. For example, two hosts connected
to different datalink layers exchange packets via routers that are using other types of datalink layers. Thanks to
the network layer, this exchange of packets is possible provided that each packet can be placed inside a datalink
layer frame before being transmitted. If all datalink layers support the same frame size, this is simple. When a
node receives a frame, it decapsulate the packet that it contains, checks the header and forwards it, encapsulated
inside another frame, to the outgoing interface. Unfortunately, the encapsulation operation is not always possible.
Each datalink layer is characterized by the maximum frame size that it supports. Datalink layers typically support
frames containing up to a few hundreds or a few thousands of bytes. The maximum frame size that a given datalink
layer supports depends on its underlying technology and unfortunately, most datalink layers support a different
maximum frame size. This implies that when a host sends a large packet inside a frame to its nexthop router, there
is a risk that this packet will have to traverse a link that is not capable of forwarding the packet inside a single
frame. In principle, there are three possibilities to solve this problem. We will discuss them by considering a
simpler scenario with two hosts connected to a router as shown in the figure below.
Max. Max. Max.

1000 bytes 500 bytes 1000 bytes
R1 R2
A B
Considering in the network above that host A wants to send a 900 bytes packet (870 bytes of payload and 30 bytes
of header) to host B via router R1. Host A encapsulates this packet inside a single frame. The frame is received by
router R1 which extracts the packet. Router R1 has three possible options to process this packet.
1. The packet is too large and router R1 cannot forward it to router R2. It rejects the packet and
sends a control packet back to the source (host A) to indicate that it cannot forward packets
longer than 500 bytes (minus the packet header). The source will have to react to this control
packet by retransmitting the information in smaller packets.
2. The network layer is able to fragment a packet. In our example, the router could fragment the

packet in two parts. The first part contains the beginning of the payload and the second the end.
There are two possible ways to achieve this fragmentation.
1. Router R1 fragments the packet in two fragments before transmitting them to router R2. Router
R2 reassembles the two packet fragments in a larger packet before transmitting them on the link
towards host B.
2. Each of the packet fragments is a valid packet that contains a header with the source (host A)
and destination (host B) addresses. When router R2 receives a packet fragment, it treats this
packet as a regular packet and forwards it to its final destination (host B). Host B reassembles
the received fragments.
These three solutions have advantages and drawbacks. With the first solution, routers remain simple and do
not need to perform any fragmentation operation. This is important when routers are implemented mainly in
hardware. However, hosts are more complex since they need to store the packets that they produce if they need
to pass through a link that does not support large packets. This increases the buffering required on the end hosts.
Furthermore, a single large packet may potentially need to be retransmitted several times. Consider for example a
network similar to the one shown above but with four routers. Assume that the link R1->R2 supports 1000 bytes
packets, link R2->R3 800 bytes packets and link R3->R4 600 bytes packets. A host attached to R1 that sends large
packet will have to first try 1000 bytes, then 800 bytes and finally 600 bytes. Fortunately, this scenario does not
occur very often in practice and this is the reason why this solution is used in real networks.
Fragmenting packets on a per-link basis, as presented for the second solution, can minimize the transmission
overhead since a packet is only fragmented on the links where fragmentation is required. Large packets can
continue to be used downstream of a link that only accepts small packets. However, this reduction of the overhead
comes with two drawbacks. First, fragmenting packets, potentially on all links, increases the processing time
and the buffer requirements on the routers. Second, this solution leads to a longer end-to-end delay since the
downstream router has to reassemble all the packet fragments before forwarding the packet.
The last solution is a compromise between the two others. Routers need to perform fragmentation but they do not
need to reassemble packet fragments. Only the hosts need to have buffers to reassemble the received fragments.
This solution has a lower end-to-end delay and requires fewer processing time and memory on the routers.
The first solution to the fragmentation problem presented above suggests the utilization of control packets to
inform the source about the reception of a too long packet. This is only one of the functions that are performed by
the control protocol in the network layer. Other functions include :
sending a control packet back to the source if a packet is received by a router that does not have a valid entry
in its forwarding table
sending a control packet back to the source if a router detects that a packet is looping inside the network
verifying that packets can reach a given destination
We will discuss these functions in more details when we will describe the protocols that are used in the network
layer of the TCP/IP protocol suite.
2.2.2 Virtual circuit organisation
The second organisation of the network layer, called virtual circuits, has been inspired by the organisation of
telephone networks. Telephone networks have been designed to carry phone calls that usually last a few minutes.
Each phone is identified by a telephone number and is attached to a telephone switch. To initiate a phone call, a
telephone first needs to send the destinations phone number to its local switch. The switch cooperates with the
other switches in the network to create a bi-directional channel between the two telephones through the network.
This channel will be used by the two telephones during the lifetime of the call and will be released at the end of
the call. Until the 1960s, most of these channels were created manually, by telephone operators, upon request of
the caller. Todays telephone networks use automated switches and allow several channels to be carried over the
same physical link, but the principles remain roughly the same.
In a network using virtual circuits, all hosts are also identified with a network layer address. However, packet
forwarding is not performed by looking at the destination address of each packet. With the virtual circuit organ-

isation, each data packet contains one label 3 . A label is an integer which is part of the packet header. Network
nodes implement label switching to forward labelled data packet. Upon reception of a packet, a network nodes
consults its label forwarding table to find the outgoing interface for this packet. In contrast with the datagram
mode, this lookup is very simple. The label forwarding table is an array stored in memory and the label of the
incoming packet is the index to access this array. This implies that the lookup operation has an O(1) complexity
in contrast with other packet forwarding techniques. To ensure that on each node the packet label is an index in
the label forwarding table, each network node that forwards a packet replaces the label of the forwarded packet
with the label found in the label forwarding table. Each entry of the label forwarding table contains two pieces of
information :
the outgoing interface for the packet
the label for the outgoing packet
For example, consider the label forwarding table of a network node below.
index outgoing interface label
0 South 7
1 none none
2 West 2
3 East 2
If this node receives a packet with label=2, it forwards the packet on its West interface and sets the label of the
outgoing packet to 2. If the received packets label is set to 3, then the packet is forwarded over the East interface
and the label of the outgoing packet is set to 2. If a packet is received with a label field set to 1, the packet is
discarded since the corresponding label forwarding table entry is invalid.
Label switching enables a full control over the path followed by packets inside the network. Consider the network
below and assume that we want to use two virtual circuits : R1->R3->R4->R2->R5 and R2->R1->R3->R4->R5.
R1
R3 R2
R4
R5
To create these virtual circuits, we need to configure the label forwarding tables of all network nodes. For
simplicity, assume that a label forwarding table only contains two entries. Assume that R5 wants to receive the
packets from the virtual circuit created by R1 (resp. R2) with label=1 (label=0). R4 could use the following label
forwarding table:
0 ->R2 1
1 ->R5 0
Since a packet received with label=1 must be forwarded to R5 with label=1, R2s label forwarding table could
contain :
0 none none
1 ->R5 1
3 We will see later a more detailed description of Multiprotocol Label Switching, a networking technology that is capable of using one or
more labels.

Two virtual circuits pass through R3. They both need to be forwarded to R4, but R4 expects label=1 for packets
belonging to the virtual circuit originated by R2 and label=0 for packets belonging to the other virtual circuit. R3
could choose to leave the labels unchanged.
0 ->R4 0
1 ->R4 1
With the above label forwarding table, R1 needs to originate the packets that belong to the R1->R3->R4->R2->R5
with label=1. The packets received from R2 and belonging to the R2->R1->R3->R4->R5 would then use label=0
on the R1-R3 link. R1 s label forwarding table could be built as follows :
0 ->R3 0
1 none 1
The figure below shows the path followed by the packets on the R1->R3->R4->R2->R5 path in red with on each
arrow the label used in the packets.
R1
R3 R2
0 0
R4 1
R5
Multi-Protocol Label Switching (MPLS) is the example of a deployed networking technology that relies on label
switching. MPLS is more complex than the above description because it has been designed to be easily integrated
with datagram technologies. However, the principles remain. Asynchronous Transfer Mode (ATM) and Frame
Relay are other examples of technologies that rely on label switching.
Nowadays, most deployed networks rely on distributed algorithms, called routing protocols, to compute the for-
warding tables that are installed on the network nodes. These distributed algorithms are part of the control plane.
They are usually implemented in software and are executed on the main CPU of the network nodes. There are two
main families of routing protocols : distance vector routing and link state routing. Both are capable of discovering
autonomously the network and react dynamically to topology changes.
2.2.3 The control plane
One of the objectives of the control plane in the network layer is to maintain the routing tables that are used on all
routers. As indicated earlier, a routing table is a data structure that contains, for each destination address (or block
of addresses) known by the router, the outgoing interface over which the router must forward a packet destined to
this address. The routing table may also contain additional information such as the address of the next router on
the path towards the destination or an estimation of the cost of this path.
In this section, we discuss the main techniques that can be used to maintain the forwarding tables in a network.
Distance vector routing
Distance vector routing is a simple distributed routing protocol. Distance vector routing allows routers to auto-
matically discover the destinations reachable inside the network as well as the shortest path to reach each of these

destinations. The shortest path is computed based on metrics or costs that are associated to each link. We use
l.cost to represent the metric that has been configured for link l on a router.
Each router maintains a routing table. The routing table R can be modelled as a data structure that stores, for each
known destination address d, the following attributes :
R[d].link is the outgoing link that the router uses to forward packets towards destination d
R[d].cost is the sum of the metrics of the links that compose the shortest path to reach destination d
R[d].time is the timestamp of the last distance vector containing destination d
A router that uses distance vector routing regularly sends its distance vector over all its interfaces. The distance
vector is a summary of the routers routing table that indicates the distance towards each known destination. This
distance vector can be computed from the routing table by using the pseudo-code below.
Every N seconds:
v=Vector()
for d in R[]:
# add destination d to vector
v.add(Pair(d,R[d].cost))
for i in interfaces
# send vector v on this interface
send(v,interface)
When a router boots, it does not know any destination in the network and its routing table only contains itself. It
thus sends to all its neighbours a distance vector that contains only its address at a distance of 0. When a router
receives a distance vector on link l, it processes it as follows.
# V : received Vector
# l : link over which vector is received
def received(V,l):
# received vector from link l
for d in V[]
if not (d in R[]) :
# new route
R[d].cost=V[d].cost+l.cost
R[d].link=l
R[d].time=now
else :
# existing route, is the new better ?
if ( ((V[d].cost+l.cost) < R[d].cost) or ( R[d].link == l) ) :
# Better route or change to current route
R[d].cost=V[d].cost+l.cost
R[d].link=l
R[d].time=now
The router iterates over all addresses included in the distance vector. If the distance vector contains an address
that the router does not know, it inserts the destination inside its routing table via link l and at a distance which is
the sum between the distance indicated in the distance vector and the cost associated to link l. If the destination
was already known by the router, it only updates the corresponding entry in its routing table if either :
the cost of the new route is smaller than the cost of the already known route ( (V[d].cost+l.cost) < R[d].cost)
the new route was learned over the same link as the current best route towards this destination ( R[d].link
== l)
The first condition ensures that the router discovers the shortest path towards each destination. The second condi-
tion is used to take into account the changes of routes that may occur after a link failure or a change of the metric
associated to a link.
To understand the operation of a distance vector protocol, let us consider the network of five routers shown below.
Assume that A is the first to send its distance vector [A=0].

B and D process the received distance vector and update their routing table with a route towards A.

Figure 2.19: Operation of distance vector routing in a simple network
D sends its distance vector [D=0,A=1] to A and E. E can now reach A and D.
C sends its distance vector [C=0] to B and E
E sends its distance vector [E=0,D=1,A=2,C=1] to D, B and C. B can now reach A, C, D and E
B sends its distance vector [B=0,A=1,C=1,D=2,E=1] to A, C and E. A, B, C and E can now reach all
destinations.
A sends its distance vector [A=0,B=1,C=2,D=1,E=2] to B and D.
At this point, all routers can reach all other routers in the network thanks to the routing tables shown in the figure
below.
Figure 2.20: Routing tables computed by distance vector in a simple network
To deal with link and router failures, routers use the timestamp stored in their routing table. As all routers send
their distance vector every N seconds, the timestamp of each route should be regularly refreshed. Thus no route
should have a timestamp older than N seconds, unless the route is not reachable anymore. In practice, to cope
with the possible loss of a distance vector due to transmission errors, routers check the timestamp of the routes
stored in their routing table every N seconds and remove the routes that are older than 3 N seconds. When a
router notices that a route towards a destination has expired, it must first associate an cost to this route and send
its distance vector to its neighbours to inform them. The route can then be removed from the routing table after
some time (e.g. 3 N seconds), to ensure that the neighbouring routers have received the bad news, even if some
distance vectors do not reach them due to transmission errors.
Consider the example above and assume that the link between routers A and B fails. Before the failure, A used B
to reach destinations B, C and E while B only used the A-B link to reach A. The affected entries timeout on routers
A and B and they both send their distance vector.
A sends its distance vector [A = 0, B = , C = , D = 1, E = ]. D knows that it cannot reach B
anymore via A
D sends its distance vector [D = 0, B = , A = 1, C = 2, E = 1] to A and E. A recovers routes towards
C and E via D.
B sends its distance vector [B = 0, A = , C = 1, D = 2, E = 1] to E and C. D learns that there is no
route anymore to reach A via B.

E sends its distance vector [E = 0, A = 2, C = 1, D = 1, B = 1] to D, B and C. D learns a route towards

B. C and B learn a route towards A.
At this point, all routers have a routing table allowing them to reach all another routers, except router A, which
cannot yet reach router B. A recovers the route towards B once router D sends its updated distance vector [A =
1, B = 2, C = 2, D = 1, E = 1]. This last step is illustrated in figure Routing tables computed by distance vector
after a failure, which shows the routing tables on all routers.
Figure 2.21: Routing tables computed by distance vector after a failure
Consider now that the link between D and E fails. The network is now partitioned into two disjoint parts : (A , D)
and (B, E, C). The routes towards B, C and E expire first on router D. At this time, router D updates its routing
table.
If D sends [D = 0, A = 1, B = , C = , E = ], A learns that B, C and E are unreachable and updates its
routing table.
Unfortunately, if the distance vector sent to A is lost or if A sends its own distance vector ( [A = 0, D = 1, B =
3, C = 3, E = 2] ) at the same time as D sends its distance vector, D updates its routing table to use the
shorter routes advertised by A towards B, C and E. After some time D sends a new distance vector : [D =
0, A = 1, E = 3, C = 4, B = 4]. A updates its routing table and after some time sends its own distance vector
[A = 0, D = 1, B = 5, C = 5, E = 4], etc. This problem is known as the count to infinity problem in networking
literature. Routers A and D exchange distance vectors with increasing costs until these costs reach . This
problem may occur in other scenarios than the one depicted in the above figure. In fact, distance vector routing
may suffer from count to infinity problems as soon as there is a cycle in the network. Cycles are necessary to
have enough redundancy to deal with link and router failures. To mitigate the impact of counting to infinity, some
distance vector protocols consider that 16 = . Unfortunately, this limits the metrics that network operators can
use and the diameter of the networks using distance vectors.
This count to infinity problem occurs because router A advertises to router D a route that it has learned via router
D. A possible solution to avoid this problem could be to change how a router creates its distance vector. Instead
of computing one distance vector and sending it to all its neighbors, a router could create a distance vector that is
specific to each neighbour and only contains the routes that have not been learned via this neighbour. This could
be implemented by the following pseudocode.
Every N seconds:
# one vector for each interface
for l in interfaces:
v=Vector()
for d in R[]:
if (R[d].link != i) :
v=v+Pair(d,R[d.cost])
send(v)
# end for d in R[]
#end for l in interfaces
This technique is called split-horizon. With this technique, the count to infinity problem would not have happened
in the above scenario, as router A would have advertised [A = 0], since it learned all its other routes via router
D. Another variant called split-horizon with poison reverse is also possible. Routers using this variant advertise a

cost of for the destinations that they reach via the router to which they send the distance vector. This can be
implemented by using the pseudo-code below.
Every N seconds:
for l in interfaces:
# one vector for each interface
v=Vector()
for d in R[]:
if (R[d].link != i) :
v=v+Pair(d,R[d.cost])
else:
v=v+Pair(d,infinity);
send(v)
# end for d in R[]
#end for l in interfaces
Unfortunately, split-horizon, is not sufficient to avoid all count to infinity problems with distance vector routing.
Consider the failure of link A-B in the network of four routers below.
Figure 2.22: Count to infinity problem
After having detected the failure, router B sends its distance vectors :
[A = , B = 0, C = , E = 1] to router C
[A = , B = 0, C = 1, E = ] to router E
If, unfortunately, the distance vector sent to router C is lost due to a transmission error or because router C is
overloaded, a new count to infinity problem can occur. If router C sends its distance vector [A = 2, B = 1, C =
0, E = ] to router E, this router installs a route of distance 3 to reach A via C. Router E sends its distance vectors
[A = 3, B = , C = 1, E = 1] to router B and [A = , B = 1, C = , E = 0] to router C. This distance
vector allows B to recover a route of distance 4 to reach A.
Note: Forwarding tables versus routing tables

Routers usually maintain at least two data structures that contain information the reachable destinations. The first
data structure is the routing table. The routing table is a data structure that associates a destination to an outgoing
interface or a nexthop router and a set of additional attributes. Different routing protocols can associate different
attributes for each destination. Distance vector routing protocols will store the cost to reach the destination along
the shortest path. Other routing protocols may store information about the number of hops of the best path, its
lifetime or the number of sub paths. A routing table may store multipath paths towards a given destination and
flag one of them as the best one. The routing table is a software data structure which is updated by (one or more)
routing protocols. The routing table is usually not directly used when forwarding packets. Packet forwarding
relies on a more compact data structure which is the forwarding table. On high-end routers, the forwarding table
is implemented directly in hardware while lower performance routers will use a software implementation. A
forwarding table contains a subset of the information found in the routing table. It only contains the paths that
are used to forward packets and no attributes. A forwarding table will typically associate each destination to an
outgoing interface or nexthop router.

Link state routing
Link state routing is the second family of routing protocols. While distance vector routers use a distributed
algorithm to compute their routing tables, link-state routers exchange messages to allow each router to learn the
entire network topology. Based on this learned topology, each router is then able to compute its routing table by
using a shortest path computation [Dijkstra1959].
For link-state routing, a network is modelled as a directed weighted graph. Each router is a node, and the links
between routers are the edges in the graph. A positive weight is associated to each directed edge and routers use
the shortest path to reach each destination. In practice, different types of weight can be associated to each directed
edge :
unit weight. If all links have a unit weight, shortest path routing prefers the paths with the least number of
intermediate routers.
weight proportional to the propagation delay on the link. If all link weights are configured this way, shortest
path routing uses the paths with the smallest propagation delay.
C
weight = bandwidth where C is a constant larger than the highest link bandwidth in the network. If all link
weights are configured this way, shortest path routing prefers higher bandwidth paths over lower bandwidth
paths
Usually, the same weight is associated to the two directed edges that correspond to a physical link (i.e. R1 R2
and R2 R1). However, nothing in the link state protocols requires this. For example, if the weight is set in
function of the link bandwidth, then an asymmetric ADSL link could have a different weight for the upstream and
downstream directions. Other variants are possible. Some networks use optimisation algorithms to find the best
set of weights to minimize congestion inside the network for a given traffic demand [FRT2002].
When a link-state router boots, it first needs to discover to which routers it is directly connected. For this, each
router sends a HELLO message every N seconds on all of its interfaces. This message contains the routers
address. Each router has a unique address. As its neighbouring routers also send HELLO messages, the router
automatically discovers to which neighbours it is connected. These HELLO messages are only sent to neighbours
who are directly connected to a router, and a router never forwards the HELLO messages that they receive. HELLO
messages are also used to detect link and router failures. A link is considered to have failed if no HELLO message
has been received from the neighbouring router for a period of k N seconds.
Figure 2.23: The exchange of HELLO messages
Once a router has discovered its neighbours, it must reliably distribute its local links to all routers in the network
to allow them to compute their local view of the network topology. For this, each router builds a link-state packet
(LSP) containing the following information :
LSP.Router : identification (address) of the sender of the LSP
LSP.age : age or remaining lifetime of the LSP
LSP.seq : sequence number of the LSP
LSP.Links[] : links advertised in the LSP. Each directed link is represented with the following information
: - LSP.Links[i].Id : identification of the neighbour - LSP.Links[i].cost : cost of the link

These LSPs must be reliably distributed inside the network without using the routers routing table since these
tables can only be computed once the LSPs have been received. The Flooding algorithm is used to efficiently
distribute the LSPs of all routers. Each router that implements flooding maintains a link state database (LSDB)
containing the most recent LSP sent by each router. When a router receives an LSP, it first verifies whether this
LSP is already stored inside its LSDB. If so, the router has already distributed the LSP earlier and it does not need
to forward it. Otherwise, the router forwards the LSP on all links except the link over which the LSP was received.
Flooding can be implemented by using the following pseudo-code.
# links is the set of all links on the router
# Router Rs LSP arrival on link l
if newer(LSP, LSDB(LSP.Router)) :
LSDB.add(LSP)
for i in links :
if i!=l :
send(LSP,i)
else:
# LSP has already been flooded
In this pseudo-code, LSDB(r) returns the most recent LSP originating from router r that is stored in the LSDB.
newer(lsp1,lsp2) returns true if lsp1 is more recent than lsp2. See the note below for a discussion on how newer
can be implemented.
Note: Which is the most recent LSP ?

A router that implements flooding must be able to detect whether a received LSP is newer than the stored LSP.
This requires a comparison between the sequence number of the received LSP and the sequence number of the
LSP stored in the link state database. The ARPANET routing protocol [MRR1979] used a 6 bits sequence number
and implemented the comparison as follows RFC 789
def newer( lsp1, lsp2 ):
return ( ( ( lsp1.seq > lsp2.seq) and ( (lsp1.seq-lsp2.seq)<=32) ) or
( ( lsp1.seq < lsp2.seq) and ( (lsp2.seq-lsp1.seq)> 32) ) )
This comparison takes into account the modulo 26 arithmetic used to increment the sequence numbers. Intuitively,
the comparison divides the circle of all sequence numbers into two halves. Usually, the sequence number of the
received LSP is equal to the sequence number of the stored LSP incremented by one, but sometimes the sequence
numbers of two successive LSPs may differ, e.g. if one router has been disconnected from the network for some
time. The comparison above worked well until October 27, 1980. On this day, the ARPANET crashed completely.
The crash was complex and involved several routers. At one point, LSP 40 and LSP 44 from one of the routers
were stored in the LSDB of some routers in the ARPANET. As LSP 44 was the newest, it should have replaced
LSP 40 on all routers. Unfortunately, one of the ARPANET routers suffered from a memory problem and sequence
number 40 (101000 in binary) was replaced by 8 (001000 in binary) in the buggy router and flooded. Three LSPs
were present in the network and 44 was newer than 40 which is newer than 8, but unfortunately 8 was considered
to be newer than 44... All routers started to exchange these three link state packets for ever and the only solution
to recover from this problem was to shutdown the entire network RFC 789.
Current link state routing protocols usually use 32 bits sequence numbers and include a special mechanism in the
unlikely case that a sequence number reaches the maximum value (using a 32 bits sequence number space takes
136 years if a link state packet is generated every second).
To deal with the memory corruption problem, link state packets contain a checksum. This checksum is computed
by the router that generates the LSP. Each router must verify the checksum when it receives or floods an LSP.
Furthermore, each router must periodically verify the checksums of the LSPs stored in its LSDB.
Flooding is illustrated in the figure below. By exchanging HELLO messages, each router learns its direct neigh-
bours. For example, router E learns that it is directly connected to routers D, B and C. Its first LSP has sequence
number 0 and contains the directed links E->D, E->B and E->C. Router E sends its LSP on all its links and routers
D, B and C insert the LSP in their LSDB and forward it over their other links.
Flooding allows LSPs to be distributed to all routers inside the network without relying on routing tables. In the
example above, the LSP sent by router E is likely to be sent twice on some links in the network. For example,
routers B and C receive Es LSP at almost the same time and forward it over the B-C link. To avoid sending the

Figure 2.24: Flooding : example
same LSP twice on each link, a possible solution is to slightly change the pseudo-code above so that a router waits
for some random time before forwarding a LSP on each link. The drawback of this solution is that the delay to
flood an LSP to all routers in the network increases. In practice, routers immediately flood the LSPs that contain
new information (e.g. addition or removal of a link) and delay the flooding of refresh LSPs (i.e. LSPs that contain
exactly the same information as the previous LSP originating from this router) [FFEB2005].
To ensure that all routers receive all LSPs, even when there are transmissions errors, link state routing protocols
use reliable flooding. With reliable flooding, routers use acknowledgements and if necessary retransmissions
to ensure that all link state packets are successfully transferred to all neighbouring routers. Thanks to reliable
flooding, all routers store in their LSDB the most recent LSP sent by each router in the network. By combining
the received LSPs with its own LSP, each router can compute the entire network topology.
Figure 2.25: Link state databases received by all routers
Note: Static or dynamic link metrics ?

As link state packets are flooded regularly, routers are able to measure the quality (e.g. delay or load) of their
links and adjust the metric of each link according to its current quality. Such dynamic adjustments were included
in the ARPANET routing protocol [MRR1979] . However, experience showed that it was difficult to tune the
dynamic adjustments and ensure that no forwarding loops occur in the network [KZ1989]. Todays link state
routing protocols use metrics that are manually configured on the routers and are only changed by the network
operators or network management tools [FRT2002].
When a link fails, the two routers attached to the link detect the failure by the lack of HELLO messages received
in the last k N seconds. Once a router has detected a local link failure, it generates and floods a new LSP that

no longer contains the failed link and the new LSP replaces the previous LSP in the network. As the two routers
attached to a link do not detect this failure exactly at the same time, some links may be announced in only one
direction. This is illustrated in the figure below. Router E has detected the failures of link E-B and flooded a new
LSP, but router B has not yet detected the failure.
Figure 2.26: The two-way connectivity check
When a link is reported in the LSP of only one of the attached routers, routers consider the link as having failed and
they remove it from the directed graph that they compute from their LSDB. This is called the two-way connectivity
check. This check allows link failures to be flooded quickly as a single LSP is sufficient to announce such bad
news. However, when a link comes up, it can only be used once the two attached routers have sent their LSPs. The
two-way connectivity check also allows for dealing with router failures. When a router fails, all its links fail by
definition. Unfortunately, it does not, of course, send a new LSP to announce its failure. The two-way connectivity
check ensures that the failed router is removed from the graph.
When a router has failed, its LSP must be removed from the LSDB of all routers 4 . This can be done by using the
age field that is included in each LSP. The age field is used to bound the maximum lifetime of a link state packet
in the network. When a router generates a LSP, it sets its lifetime (usually measured in seconds) in the age field.
All routers regularly decrement the age of the LSPs in their LSDB and a LSP is discarded once its age reaches 0.
Thanks to the age field, the LSP from a failed router does not remain in the LSDBs forever.
To compute its forwarding table, each router computes the spanning tree rooted at itself by using Dijkstras shortest
path algorithm [Dijkstra1959]. The forwarding table can be derived automatically from the spanning as shown in
the figure below.
Figure 2.27: Computation of the forwarding table
4 It should be noted that link state routing assumes that all routers in the network have enough memory to store the entire LSDB. The
routers that do not have enough memory to store the entire LSDB cannot participate in link state routing. Some link state routing protocols
allow routers to report that they do not have enough memory and must be removed from the graph by the other routers in the network.

2.3 Applications
The are two important models used to organise a networked application. The first and oldest model is the client-
server model. In this model, a server provides services to clients that exchange information with it. This model is
highly asymmetrical : clients send requests and servers perform actions and return responses. It is illustrated in
the figure below.
Figure 2.28: The client-server model
The client-server model was the first model to be used to develop networked applications. This model comes
naturally from the mainframes and minicomputers that were the only networked computers used until the 1980s.
A minicomputer is a multi-user system that is used by tens or more users at the same time. Each user interacts
with the minicomputer by using a terminal. Those terminals, were mainly a screen, a keyboard and a cable directly
connected to the minicomputer.
There are various types of servers as well as various types of clients. A web server provides information in
response to the query sent by its clients. A print server prints documents sent as queries by the client. An
email server will forward towards their recipient the email messages sent as queries while a music server will
deliver the music requested by the client. From the viewpoint of the application developer, the client and the
server applications directly exchange messages (the horizontal arrows labelled Queries and Responses in the
above figure), but in practice these messages are exchanged thanks to the underlying layers (the vertical arrows in
the above figure). In this chapter, we focus on these horizontal exchanges of messages.
Networked applications do not exchange random messages. In order to ensure that the server is able to understand
the queries sent by a client, and also that the client is able to understand the responses sent by the server, they must
both agree on a set of syntactical and semantic rules. These rules define the format of the messages exchanged as
well as their ordering. This set of rules is called an application-level protocol.
An application-level protocol is similar to a structured conversation between humans. Assume that Alice wants
to know the current time but does not have a watch. If Bob passes close by, the following conversation could take
place :
Alice : Hello
Bob : Hello
Alice : What time is it ?
Bob : 11:55
Alice : Thank you
Bob : Youre welcome
Such a conversation succeeds if both Alice and Bob speak the same language. If Alice meets Tchang who only
speaks Chinese, she wont be able to ask him the current time. A conversation between humans can be more
complex. For example, assume that Bob is a security guard whose duty is to only allow trusted secret agents to
enter a meeting room. If all agents know a secret password, the conversation between Bob and Trudy could be as
follows :
Bob : What is the secret password ?
Trudy : 1234
Bob : This is the correct password, youre welcome
2.3. Applications 51
If Alice wants to enter the meeting room but does not know the password, her conversation could be as follows :
Bob : What is the secret password ?
Alice : 3.1415
Bob : This is not the correct password.
Human conversations can be very formal, e.g. when soldiers communicate with their hierarchy, or informal such
as when friends discuss. Computers that communicate are more akin to soldiers and require well-defined rules to
ensure an successful exchange of information. There are two types of rules that define how information can be
exchanged between computers :
syntactical rules that precisely define the format of the messages that are exchanged. As computers only
process bits, the syntactical rules specify how information is encoded as bit strings
organisation of the information flow. For many applications, the flow of information must be structured and
there are precedence relationships between the different types of information. In the time example above,
Alice must greet Bob before asking for the current time. Alice would not ask for the current time first and
greet Bob afterwards. Such precedence relationships exist in networked applications as well. For example,
a server must receive a username and a valid password before accepting more complex commands from its
clients.
Let us first discuss the syntactical rules. We will later explain how the information flow can be organised by
analysing real networked applications.
Application-layer protocols exchange two types of messages. Some protocols such as those used to support
electronic mail exchange messages expressed as strings or lines of characters. As the transport layer allows hosts
to exchange bytes, they need to agree on a common representation of the characters. The first and simplest method
to encode characters is to use the ASCII table. RFC 20 provides the ASCII table that is used by many protocols
on the Internet. For example, the table defines the following binary representations :
A : 1000011b
0 : 0110000b
z : 1111010b
@ : 1000000b
space : 0100000b
In addition, the ASCII table also defines several non-printable or control characters. These characters were de-
signed to allow an application to control a printer or a terminal. These control characters include CR and LF, that
are used to terminate a line, and the Bell character which causes the terminal to emit a sound.
carriage return (CR) : 0001101b
line feed (LF) : 0001010b
Bell: 0000111b
The ASCII characters are encoded as a seven bits field, but transmitted as an eight-bits byte whose high order bit
is usually set to 0. Bytes are always transmitted starting from the high order or most significant bit.
Most applications exchange strings that are composed of fixed or variable numbers of characters. A common
solution to define the character strings that are acceptable is to define them as a grammar using a Backus-Naur
Form (BNF) such as the Augmented BNF defined in RFC 5234. A BNF is a set of production rules that generate
all valid character strings. For example, consider a networked application that uses two commands, where the
user can supply a username and a password. The BNF for this application could be defined as shown in the figure
below.
The example above defines several terminals and two commands : usercommand and passwordcommand. The
ALPHA terminal contains all letters in upper and lower case. In the ALPHA rule, %x41 corresponds to ASCII
character code 41 in hexadecimal, i.e. capital A. The CR and LF terminals correspond to the carriage return and
linefeed control characters. The CRLF rule concatenates these two terminals to match the standard end of line
termination. The DIGIT terminal contains all digits. The SP terminal corresponds to the white space characters.
The usercommand is composed of two strings separated by white space. In the ABNF rules that define the

Figure 2.29: A simple BNF specification
messages used by Internet applications, the commands are case-insensitive. The rule user corresponds to all
possible cases of the letters that compose the word between brackets, e.g. user, uSeR, USER, usER, ... A username
contains at least one letter and up to 8 letters. User names are case-sensitive as they are not defined as a string
between brackets. The password rule indicates that a password starts with a letter and can contain any number of
letters or digits. The white space and the control characters cannot appear in a password defined by the above rule.
Besides character strings, some applications also need to exchange 16 bits and 32 bits fields such as integers. A
naive solution would have been to send the 16- or 32-bits field as it is encoded in the hosts memory. Unfortunately,
there are different methods to store 16- or 32-bits fields in memory. Some CPUs store the most significant byte
of a 16-bits field in the first address of the field while others store the least significant byte at this location. When
networked applications running on different CPUs exchange 16 bits fields, there are two possibilities to transfer
them over the transport service :
send the most significant byte followed by the least significant byte
send the least significant byte followed by the most significant byte
The first possibility was named big-endian in a note written by Cohen [Cohen1980] while the second was named
little-endian. Vendors of CPUs that used big-endian in memory insisted on using big-endian encoding in net-
worked applications while vendors of CPUs that used little-endian recommended the opposite. Several studies
were written on the relative merits of each type of encoding, but the discussion became almost a religious issue
[Cohen1980]. Eventually, the Internet chose the big-endian encoding, i.e. multi-byte fields are always transmit-
ted by sending the most significant byte first, RFC 791 refers to this encoding as the network-byte order. Most
libraries 5 used to write networked applications contain functions to convert multi-byte fields from memory to the
network byte order and vice versa.
Besides 16 and 32 bit words, some applications need to exchange data structures containing bit fields of various
lengths. For example, a message may be composed of a 16 bits field followed by eight, one bit flags, a 24 bits
field and two 8 bits bytes. Internet protocol specifications will define such a message by using a representation
such as the one below. In this representation, each line corresponds to 32 bits and the vertical lines are used to
delineate fields. The numbers above the lines indicate the bit positions in the 32-bits word, with the high order bit
at position 0.
Figure 2.30: Message format
The message mentioned above will be transmitted starting from the upper 32-bits word in network byte order. The
first field is encoded in 16 bits. It is followed by eight one bit flags (A-H), a 24 bits field whose high order byte is
shown in the first line and the two low order bytes appear in the second line followed by two one byte fields. This
ASCII representation is frequently used when defining binary protocols. We will use it for all the binary protocols
that are discussed in this book.
5 For example, the htonl(3) (resp. ntohl(3)) function the standard C library converts a 32-bits unsigned integer from the byte order
used by the CPU to the network byte order (resp. from the network byte order to the CPU byte order). Similar functions exist in other
programming languages.
2.3. Applications 53
The peer-to-peer model emerged during the last ten years as another possible architecture for networked appli-
cations. In the traditional client-server model, hosts act either as servers or as clients and a server serves a large
number of clients. In the peer-to-peer model, all hosts act as both servers and clients and they play both roles.
The peer-to-peer model has been used to develop various networked applications, ranging from Internet telephony
to file sharing or Internet-wide filesystems. A detailed description of peer-to-peer applications may be found in
[BYL2008]. Surveys of peer-to-peer protocols and applications may be found in [AS2004] and [LCP2005].
2.4 The transport layer
A network is always designed and built to enable applications running on hosts to exchange information. In the
previous chapter, we have explained the principles of the network layer that enables hosts connected to different
types of datalink layers to exchange information through routers. These routers act as relays in the network layer
and ensure the delivery of packets between any pair of hosts attached to the network.
The network layer ensures the delivery of packets on a hop-by-hop basis through intermediate nodes. As such, it
provides a service to the upper layer. In practice, this layer is usually the transport layer that improves the service
provided by the network layer to make it useable by applications.
Figure 2.31: The transport layer
Most networks use a datagram organisation and provide a simple service which is called the connectionless service.
The figure below provides a representation of the connectionless service as a time-sequence diagram. The user
on the left, having address S, issues a Data.request primitive containing Service Data Unit (SDU) M that must be
delivered by the service provider to destination D. The dashed line between the two primitives indicates that the
Data.indication primitive that is delivered to the user on the right corresponds to the Data.request primitive sent
by the user on the left.
Figure 2.32: A simple connectionless service
There are several possible implementations of the connectionless service. Before studying these realisations, it is
useful to discuss the possible characteristics of the connectionless service. A reliable connectionless service is a
service where the service provider guarantees that all SDUs submitted in Data.requests by a user will eventually
be delivered to their destination. Such a service would be very useful for users, but guaranteeing perfect delivery
is difficult in practice. For this reason, network layers usually support an unreliable connectionless service.
An unreliable connectionless service may suffer from various types of problems compared to a reliable connec-
tionless service. First of all, an unreliable connectionless service does not guarantee the delivery of all SDUs.
This can be expressed graphically by using the time-sequence diagram below.
In practice, an unreliable connectionless service will usually deliver a large fraction of the SDUs. However, since
the delivery of SDUs is not guaranteed, the user must be able to recover from the loss of any SDU.

Figure 2.33: An unreliable connectionless service may loose SDUs
A second imperfection that may affect an unreliable connectionless service is that it may duplicate SDUs. Some
packets may be duplicated in a network and be delivered twice to their destination. This is illustrated by the
time-sequence diagram below.
Figure 2.34: An unreliable connectionless service may duplicate SDUs
Finally, some unreliable connectionless service providers may deliver to a destination a different SDU than the
one that was supplied in the Data.request. This is illustrated in the figure below.
Figure 2.35: An unreliable connectionless service may deliver erroneous SDUs
As the transport layer is built on top of the network layer, it is important to know the key features of the net-
work layer service. In this book, we only consider the connectionless network layer service which is the most
widespread. Its main characteristics are :
the connectionless network layer service can only transfer SDUs of limited size
the connectionless network layer service may discard SDUs
the connectionless network layer service may corrupt SDUs
the connectionless network layer service may delay, reorder or even duplicate SDUs
These imperfections of the connectionless network layer service are caused by the operations of the network layer.
This layer is able to deliver packets to their intended destination, but it cannot guarantee this delivery. The main
cause of packet losses and errors are the buffers used on the network nodes. If the buffers of one of these nodes
2.4. The transport layer 55

becomes full, all arriving packets must be discarded. This situation happens frequently in practice. Transmission
errors can also affect packet transmissions on links where reliable transmission techniques are not enabled or
because of errors in the buffers of the network nodes.
2.4.1 Transport layer services
When two applications need to communicate, they need to structure their exchange of information. Structuring
this exchange of information requires solving two different problems. The first problem is how to represent the
information being exchanged knowing that the two applications may be running on hosts that use different oper-
ating systems, different processors and have different conventions to store information. This requires a common
syntax to transfer the information between the two applications. For this chapter, let us assume that this syntax
exists and that the two applications simply need to exchange bytes. We will discuss later how more complex data
can be encoded as sequences of bytes to be exchanged. The second problem is how to organise the interactions
between the application and the underlying network. From the applications viewpoint, the network will appear as
the transport layer service. This transport layer can provide three types of services to the applications :
the connectionless service
the connection oriented service
the request-response service
The connectionless service
The connectionless service that we have described earlier is frequently used by users who need to exchange small
SDUs. It can be easily built on top of the connectionless network layer service that we have described earlier. Users
needing to either send or receive several different and potentially large SDUs, or who need structured exchanges
often prefer the connection-oriented service.
The connection-oriented service
An invocation of the connection-oriented service is divided into three phases. The first phase is the establishment
of a connection. A connection is a temporary association between two users through a service provider. Several
connections may exist at the same time between any pair of users. Once established, the connection is used to
transfer SDUs. Connections usually provide one bidirectional stream supporting the exchange of SDUs between
the two users that are associated through the connection. This stream is used to transfer data during the second
phase of the connection called the data transfer phase. The third phase is the termination of the connection. Once
the users have finished exchanging SDUs, they request to the service provider to terminate the connection. As we
will see later, there are also some cases where the service provider may need to terminate a connection itself.
The establishment of a connection can be modelled by using four primitives : Connect.request, Connect.indication,
Connect.response and Connect.confirm. The Connect.request primitive is used to request the establishment of a
connection. The main parameter of this primitive is the address of the destination user. The service provider
delivers a Connect.indication primitive to inform the destination user of the connection attempt. If it accepts to
establish a connection, it responds with a Connect.response primitive. At this point, the connection is considered to
be established and the destination user can start sending SDUs over the connection. The service provider processes
the Connect.response and will deliver a Connect.confirm to the user who initiated the connection. The delivery

of this primitive terminates the connection establishment phase. At this point, the connection is considered to be
open and both users can send SDUs. A successful connection establishment is illustrated below.
Figure 2.37: Connection establishment
The example above shows a successful connection establishment. However, in practice not all connections are
successfully established. One reason is that the destination user may not agree, for policy or performance reasons,
to establish a connection with the initiating user at this time. In this case, the destination user responds to the
Connect.indication primitive by a Disconnect.request primitive that contains a parameter to indicate why the
connection has been refused. The service provider will then deliver a Disconnect.indication primitive to inform
the initiating user. A second reason is when the service provider is unable to reach the destination user. This
might happen because the destination user is not currently attached to the network or due to congestion. In these
cases, the service provider responds to the Connect.request with a Disconnect.indication primitive whose reason
parameter contains additional information about the failure of the connection.
Figure 2.38: Two types of rejection for a connection establishment attempt
Once the connection has been established, the service provider supplies two data streams to the communicating
users. The first data stream can be used by the initiating user to send SDUs. The second data stream allows
the responding user to send SDUs to the initiating user. The data streams can be organised in different ways. A
first organisation is the message-mode transfer. With the message-mode transfer, the service provider guarantees
that one and only one Data.indication will be delivered to the endpoint of the data stream for each Data.request
primitive issued by the other endpoint. The message-mode transfer is illustrated in the figure below. The main
advantage of the message-transfer mode is that the recipient receives exactly the SDUs that were sent by the other
user. If each SDU contains a command, the receiving user can process each command as soon as it receives a
SDU.
Unfortunately, the message-mode transfer is not widely used on the Internet. On the Internet, the most popular
connection-oriented service transfers SDUs in stream-mode. With the stream-mode, the service provider supplies a
byte stream that links the two communicating users. The sending user sends bytes by using Data.request primitives
that contain sequences of bytes as SDUs. The service provider delivers SDUs containing consecutive bytes to the
receiving user by using Data.indication primitives. The service provider ensures that all the bytes sent at one end
of the stream are delivered correctly in the same order at the other endpoint. However, the service provider does
not attempt to preserve the boundaries of the SDUs. There is no relation enforced by the service provider between
the number of Data.request and the number of Data.indication primitives. The stream-mode is illustrated in the
figure below. In practice, a consequence of the utilisation of the stream-mode is that if the users want to exchange
structured SDUs, they will need to provide the mechanisms that allow the receiving user to separate successive

Figure 2.39: Message-mode transfer in a connection oriented service
SDUs in the byte stream that it receives. Application layer protocols often use specific delimiters such as the end
of line character to delineate SDUs in a bytestream.
Figure 2.40: Stream-mode transfer in a connection oriented service
The third phase of a connection is when it needs to be released. As a connection involves three parties (two users
and one service provider), any of them can request the termination of the connection. Usually, connections are
terminated upon request of one user once the data transfer is finished. However, sometimes the service provider
may be forced to terminate a connection. This can be due to lack of resources inside the service provider or
because one of the users is not reachable anymore through the network. In this case, the service provider will issue
Disconnect.indication primitives to both users. These primitives will contain, as parameter, some information
about the reason for the termination of the connection. Unfortunately, as illustrated in the figure below, when a
service provider is forced to terminate a connection it cannot guarantee that all SDUs sent by each user have been
delivered to the other user. This connection release is said to be abrupt as it can cause losses of data.
Figure 2.41: Abrupt connection release initiated by the service provider
An abrupt connection release can also be triggered by one of the users. If a user needs, for any reason, to terminate
a connection quickly, it can issue a Disconnect.request primitive and to request an abrupt release. The service
provider will process the request, stop the two data streams and deliver the Disconnect.indication primitive to the
remote user as soon as possible. As illustrated in the figure below, this abrupt connection release may cause losses
of SDUs.
To ensure a reliable delivery of the SDUs sent by each user over a connection, we need to consider the two streams
that compose a connection as independent. A user should be able to release the stream that it uses to send SDUs
once it has sent all the SDUs that it planned to send over this connection, but still continue to receive SDUs over

Figure 2.42: Abrupt connection release initiated by a user
the opposite stream. This graceful connection release is usually performed as shown in the figure below. One user
issues a Disconnect.request primitive to its provider once it has issued all its Data.request primitives. The service
provider will wait until all Data.indication primitives have been delivered to the receiving user before issuing the
Disconnnect.indication primitive. This primitive informs the receiving user that it will no longer receive SDUs
over this connection, but it is still able to issue Data.request primitives on the stream in the opposite direction.
Once the user has issued all of its Data.request primitives, it issues a Disconnnect.request primitive to request the
termination of the remaining stream. The service provider will process the request and deliver the corresponding
Disconnect.indication to the other user once it has delivered all the pending Data.indication primitives. At this
point, all data has been delivered, the two streams have been released successfully and the connection is completely
closed.
Figure 2.43: Graceful connection release
Note: Reliability of the connection-oriented service

An important point to note about the connection-oriented service is its reliability. A connection-oriented service
can only guarantee the correct delivery of all SDUs provided that the connection has been released gracefully. This
implies that while the connection is active, there is no guarantee for the actual delivery of the SDUs exchanged as
the connection may need to be released abruptly at any time.
The request-response service
The request-response service is a compromise between the connectionless service and the connection-oriented
service. Many applications need to send a small amount of data and receive a small amount of information back.
This is similar to procedure calls in programming languages. A call to a procedure takes a few arguments and
returns a simple answer. In a network, it is sometimes useful to execute a procedure on a different host and
receive the result of the computation. Executing a procedure on another host is often called Remote Procedure
Call. It is possible to use the connectionless service for this application. However, since this service is usually
unreliable, this would force the application to deal with any type of error that could occur. Using the connection
oriented service is another alternative. This service ensures the reliable delivery of the data, but a connection must
be created before the beginning of the data transfert. This overhead can be important for applications that only
exchange a small amount of data.
The request-response service allows to efficiently exchange small amounts of information in a request and as-
sociate it with the corresponding response. This service can be depicted by using the time-sequence diagram

below.
Host A Service Host B
DATA.req(request)
DATA.ind(request)
DATA.resp(response)
DATA.confirm(response)
Note: Services and layers

In the previous sections, we have described services that are provided by the transport layer. However, it is
important to note that the notion of service is more general than in the transport layer. As explained earlier, the
network layer also provides a service, which in most networks is an unreliable connectionless service. There are
network layers that provide a connection-oriented service. Similarly, the datalink layer also provides services.
Some datalink layers will provide a connectionless service. This will be the case in Local Area Networks for
examples. Other datalink layers, e.g. in public networks, provide a connection oriented service.
2.4.2 The transport layer
The transport layer entity interacts with both a user in the application layer and the network layer. It improves the
network layer service to make it useable by applications. From the applications viewpoint, the main limitations
of the network layer service that its service is unreliable:
the network layer may corrupt data
the network layer may loose data
the network layer may not deliver data in-order
the network layer has an upper bound on maximum length of the data
the network layer may duplicate data
To deal with these issues, the transport layer includes several mechanisms that depend on the service that it
provides. It interacts with both the applications and the underlying network layer.
Figure 2.44: Interactions between the transport layer, its user, and its network layer provider
We have already described in the datalink layers mechanisms to deal with data losses and transmission errors.
These techniques are also used in the transport layer.

Connectionless transport
The simplest service that can be provided in the transport layer is the connectionless transport service. Compared
to the connectionless network layer service, this transport service includes two additional features :
an error detection mechanism that allows to detect corrupted data
a multiplexing technique that enables several applications running on one host to exchange information with
another host
To exchange data, the transport protocol encapsulates the SDU produced by its user inside a segment. The segment
is the unit of transfert of information in the transport layer. Transport layer entities always exchange segments.
When a transport layer entity creates a segment, this segment is encapsulated by the network layer into a packet
which contains the segment as its payload and a network header. The packet is then encapsulated in a frame to be
transmitted in the datalink layer.
A segment also contains control information, usually stored inside a header and the payload that comes from the
application. To detect transmission errors, transport protocols rely on checksums or CRCs like the datalink layer
protocols.
Compared to the connectionless network layer service, the transport layer service allows several applications
running on a host to exchange SDUs with several other applications running on remote hosts. Let us consider two
hosts, e.g. a client and a server. The network layer service allows the client to send information to the server,
but if an application running on the client wants to contact a particular application running on the server, then an
additional addressing mechanism is required other than the network layer address that identifies a host, in order to
differentiate the application running on a host. This additional addressing can be provided by using port numbers.
When a server application is launched on a host, it registers a port number. This port number will be used by the
clients to contact the server process.
The figure below shows a typical usage of port numbers. The client process uses port number 1234 while the
server process uses port number 5678. When the client sends a request, it is identified as originating from port
number 1234 on the client host and destined to port number 5678 on the server host. When the server process
replies to this request, the servers transport layer returns the reply as originating from port 5678 on the server host
and destined to port 1234 on the client host.
Figure 2.45: Utilisation of port numbers
To support the connection-oriented service, the transport layer needs to include several mechanisms to enrich the
connectionless network-layer service. We discuss these mechanisms in the following sections.
Connection establishment
Like the connectionless service, the connection-oriented service allows several applications running on a given
host to exchange data with other hosts. The port numbers described above for the connectionless service are also
used by the connection-oriented service to multiplex several applications. Similarly, connection-oriented protocols
used checksums/CRCs to detect transmission errors and discard segments containing an invalid checksum/CRC.

An important difference between the connectionless service and the connection-oriented one is that the transport
entities in the latter maintain some state during lifetime of the connection. This state is created when a connection
is established and is removed when it is released.
The simplest approach to establish a transport connection would be to define two special control segments : CR and
CA. The CR segment is sent by the transport entity that wishes to initiate a connection. If the remote entity wishes
to accept the connection, it replies by sending a CA segment. The CR and CA segments contain port numbers that
allow to identify the communicating applications. The transport connection is considered to be established once
the CA segment has been received. At that point, data segments can be sent in both directions.
Figure 2.46: Naive transport connection establishment
Unfortunately, this scheme is not sufficient given the unreliable network layer. Since the network layer is imper-
fect, the CR or CA segments can be lost, delayed, or suffer from transmission errors. To deal with these problems,
the control segments must be protected by using a CRC or checksum to detect transmission errors. Furthermore,
since the CA segment acknowledges the reception of the CR segment, the CR segment can be protected by using
a retransmission timer.
Unfortunately, this scheme is not sufficient to ensure the reliability of the transport service. Consider for example
a short-lived transport connection where a single, but important transfer (e.g. money transfer from a bank account)
is sent. Such a short-lived connection starts with a CR segment acknowledged by a CA segment, then the data
segment is sent, acknowledged and the connection terminates. Unfortunately, as the network layer service is
unreliable, delays combined to retransmissions may lead to the situation depicted in the figure below, where a
delayed CR and data segments from a former connection are accepted by the receiving entity as valid segments,
and the corresponding data is delivered to the user. Duplicating SDUs is not acceptable, and the transport protocol
must solve this problem.
Figure 2.47: Duplicate transport connections ?
To avoid these duplicates, transport protocols require the network layer to bound the Maximum Segment Lifetime
(MSL). The organisation of the network must guarantee that no segment remains in the network for longer than
MSL seconds. For example, on todays Internet, MSL is expected to be 2 minutes. To avoid duplicate transport
connections, transport protocol entities must be able to safely distinguish between a duplicate CR segment and
a new CR segment, without forcing each transport entity to remember all the transport connections that it has
established in the past.
A classical solution to avoid remembering the previous transport connections to detect duplicates is to use a clock
inside each transport entity. This transport clock has the following characteristics :
the transport clock is implemented as a k bits counter and its clock cycle is such that 2k cycle >>
M SL. Furthermore, the transport clock counter is incremented every clock cycle and after each connection
establishment. This clock is illustrated in the figure below.

the transport clock must continue to be incremented even if the transport entity stops or reboots
Figure 2.48: Transport clock
It should be noted that transport clocks do not need and usually are not synchronised to the real-time clock.
Precisely synchronising real-time clocks is an interesting problem, but it is outside the scope of this document.
See [Mills2006] for a detailed discussion on synchronizing the real-time clock.
This transport clock can now be combined with an exchange of three segments, called the three way handshake,
to detect duplicates. This three way handshake occurs as follows :
1. The initiating transport entity sends a CR segment. This segment requests the establishment of a
transport connection. It contains a port number (not shown in the figure) and a sequence number
(seq=x in the figure below) whose value is extracted from the transport clock. The transmission
of the CR segment is protected by a retransmission timer.
2. The remote transport entity processes the CR segment and creates state for the connection at-
tempt. At this stage, the remote entity does not yet know whether this is a new connection
attempt or a duplicate segment. It returns a CA segment that contains an acknowledgement
number to confirm the reception of the CR segment (ack=x in the figure below) and a sequence
number (seq=y in the figure below) whose value is extracted from its transport clock. At this
stage, the connection is not yet established.
3. The initiating entity receives the CA segment. The acknowledgement number of this segment
confirms that the remote entity has correctly received the CR segment. The transport connection
is considered to be established by the initiating entity and the numbering of the data segments
starts at sequence number x. Before sending data segments, the initiating entity must acknowl-
edge the received CA segments by sending another CA segment.
4. The remote entity considers the transport connection to be established after having received the
segment that acknowledges its CA segment. The numbering of the data segments sent by the
remote entity starts at sequence number y.
The three way handshake is illustrated in the figure below.
Thanks to the three way handshake, transport entities avoid duplicate transport connections. This is illustrated by
considering the three scenarios below.
The first scenario is when the remote entity receives an old CR segment. It considers this CR segment as a
connection establishment attempt and replies by sending a CA segment. However, the initiating host cannot match
the received CA segment with a previous connection attempt. It sends a control segment (REJECT in the figure
below) to cancel the spurious connection attempt. The remote entity cancels the connection attempt upon reception
of this control segment.
A second scenario is when the initiating entity sends a CR segment that does not reach the remote entity and
receives a duplicate CA segment from a previous connection attempt. This duplicate CA segment cannot contain
a valid acknowledgement for the CR segment as the sequence number of the CR segment was extracted from the
transport clock of the initiating entity. The CA segment is thus rejected and the CR segment is retransmitted upon
expiration of the retransmission timer.
The last scenario is less likely, but it it important to consider it as well. The remote entity receives an old CR
segment. It notes the connection attempt and acknowledges it by sending a CA segment. The initiating entity
does not have a matching connection attempt and replies by sending a REJECT. Unfortunately, this segment never

Figure 2.49: Three-way handshake
Figure 2.50: Three-way handshake : recovery from a duplicate CR
Figure 2.51: Three-way handshake : recovery from a duplicate CA

reaches the remote entity. Instead, the remote entity receives a retransmission of an older CA segment that contains
the same sequence number as the first CR segment. This CA segment cannot be accepted by the remote entity as
a confirmation of the transport connection as its acknowledgement number cannot have the same value as the
sequence number of the first CA segment.
Figure 2.52: Three-way handshake : recovery from duplicates CR and CA
Data transfer
Now that the transport connection has been established, it can be used to transfer data. To ensure a reliable delivery
of the data, the transport protocol will include sliding windows, retransmission timers and go-back-n or selective
repeat. However, we cannot simply reuse the techniques from the datalink because a transport protocol needs to
deal with more types of errors than a reliable protocol in datalink layer. The first difference between the two layers
is the transport layer must face with more variable delays. In the datalink layer, when two hosts are connected
by a link, the transmission delay or the round-trip-time over the link is almost fixed. In a network that can span
the globe, the delays and the round-trip-times can vary significantly on a per packet basis. This variability can
be caused by two factors. First, packets sent through a network do not necessarily follow the same path to reach
their destination. Second, some packets may be queued in the buffers of routers when the load is high and these
queueing delays can lead to increased end-to-end delays. A second difference between the datalink layer and the
transport layer is that a network does not always deliver packets in sequence. This implies that packets may be
reordered by the network. Furthermore, the network may sometimes duplicate packets. The last issue that needs
to be dealt with in the transport layer is the transmission of large SDUs. In the datalink layer, reliable protocols
transmit small frames. Applications could generate SDUs that are much larger than the maximum size of a packet
in the network layer. The transport layer needs to include mechanisms to fragment and reassemble these large
SDUs.
To deal with all these characteristics of the network layer, we need to adapt the techniques that we have introduced
in the datalink layer.
The first point which is common between the two layers is that both use CRCs or checksum to detect transmission
errors. Each segment contains a CRC/checksum which is computed over the entire segment (header and payload)
by the sender and inserted in the header. The receiver recomputes the CRC/checksum for each received segment
and discards all segments with an invalid CRC.
Reliable transport protocols also use sequence numbers and acknowledgement numbers. While reliable protocols
in the datalink layer use one sequence number per frame, reliable transport protocols consider all the data trans-
mitted as a stream of bytes. In these protocols, the sequence number placed in the segment header corresponds to
the position of the first byte of the payload in the bytestream. This sequence number allows to detect losses but
also enables the receiver to reorder the out-of-sequence segments. This is illustrated in the figure below.

Host A Host B
DATA.req(abcde) DATA.ind(abcde)
1:abcd
DATA.req(fghijkl) 5:fghijkl DATA.ind(fghijkl)
Using sequence numbers to count bytes has also one advantage when the transport layer needs to fragment SDUs
in several segments. The figure below shows the fragmentation of a large SDU in two segments. Upon reception
of the segments, the receiver will use the sequence numbers to correctly reorder the data.
Host A Host B
DATA.req(abcdefghijkl)
1:abcde
6:fghijkl DATA.ind(abcdefghijkl)
Compared to reliable protocols in the datalink layer, reliable transport protocols encode their sequence numbers
in more bits. 32 bits and 64 bits sequence numbers are frequent in the transport layer while some datalink layer
protocols encode their sequence numbers in an 8 bits field. This large sequence number space is motivated by two
reasons. First, since the sequence number is incremented for each transmitted byte, a single segment may consume
one or several thousands of sequence numbers. Second, a reliable transport protocol must be able to detect delayed
segments. This can only be done if the number of bytes transmitted during the MSL period is smaller than the
sequence number space. Otherwise, there is a risk of accepting duplicate segments.
Go-back-n and selective repeat can be used in the transport layer as in the datalink layer. Since the network layer
does not guarantee an in-order delivery of the packets, a transport entity should always store the segments that
it receives out-of-sequence. For this reason, most transport protocols will opt for some form of selective repeat
mechanism.
In the datalink layer, the sliding window has usually a fixed size which depends on the amount of buffers allocated
to the datalink layer entity. Such a datalink layer entity usually serves one or a few network layer entities. In
the transport layer, the situation is different. A single transport layer entity serves a large and varying number of
application processes. Each transport layer entity manages a pool of buffers that needs to be shared between all
these processes. Transport entity are usually implemented inside the operating system kernel and shares memory
with other parts of the system. Furthermore, a transport layer entity must support several (possibly hundreds or
thousands) of transport connections at the same time. This implies that the memory which can be used to support
the sending or the receiving buffer of a transport connection may change during the lifetime of the connection 6 .
Thus, a transport protocol must allow the sender and the receiver to adjust their window sizes.
To deal with this issue, transport protocols allow the receiver to advertise the current size of its receiving window
in all the acknowledgements that it sends. The receiving window advertised by the receiver bounds the size of
the sending buffer used by the sender. In practice, the sender maintains two state variables : swin, the size of its
sending window (that may be adjusted by the system) and rwin, the size of the receiving window advertised by the
receiver. At any time, the number of unacknowledged segments cannot be larger than min(swin, rwin) 7 . The
utilisation of dynamic windows is illustrated in the figure below.
The receiver may adjust its advertised receive window based on its current memory consumption, but also to limit
the bandwidth used by the sender. In practice, the receive buffer can also shrink as the application may not able to
process the received data quickly enough. In this case, the receive buffer may be completely full and the advertised
receive window may shrink to 0. When the sender receives an acknowledgement with a receive window set to 0,
it is blocked until it receives an acknowledgement with a positive receive window. Unfortunately, as shown in the
figure below, the loss of this acknowledgement could cause a deadlock as the sender waits for an acknowledgement
while the receiver is waiting for a data segment.
To solve this problem, transport protocols rely on a special timer : the persistence timer. This timer is started
by the sender whenever it receives an acknowledgement advertising a receive window set to 0. When the timer
expires, the sender retransmits an old segment in order to force the receiver to send a new acknowledgement, and
hence send the current receive window size.
6 For a discussion on how the sending buffer can change, see e.g. [SMM1998]
7 Note that if the receive window shrinks, it might happen that the sender has already sent a segment that is not anymore inside its window.
This segment will be discarded by the receiver and the sender will retransmit it later.

Figure 2.53: Dynamic receiving window
Figure 2.54: Risk of deadlock with dynamic windows

To conclude our description of the basic mechanisms found in transport protocols, we still need to discuss the
impact of segments arriving in the wrong order. If two consecutive segments are reordered, the receiver relies on
their sequence numbers to reorder them in its receive buffer. Unfortunately, as transport protocols reuse the same
sequence number for different segments, if a segment is delayed for a prolonged period of time, it might still be
accepted by the receiver. This is illustrated in the figure below where segment D(1,b) is delayed.
Figure 2.55: Ambiguities caused by excessive delays
To deal with this problem, transport protocols combine two solutions. First, they use 32 bits or more to encode
the sequence number in the segment header. This increases the overhead, but also increases the delay between
the transmission of two different segments having the same sequence number. Second, transport protocols require
the network layer to enforce a Maximum Segment Lifetime (MSL). The network layer must ensure that no packet
remains in the network for more than MSL seconds. In the Internet the MSL is assumed 8 to be 2 minutes RFC
793. Note that this limits the maximum bandwidth of a transport protocol. If it uses n bits to encode its sequence
numbers, then it cannot send more than 2n segments every MSL seconds.
Connection release
When we discussed the connection-oriented service, we mentioned that there are two types of connection releases
: abrupt release and graceful release.
The first solution to release a transport connection is to define a new control segment (e.g. the DR segment) and
consider the connection to be released once this segment has been sent or received. This is illustrated in the figure
below.
As the entity that sends the DR segment cannot know whether the other entity has already sent all its data on the
connection, SDUs can be lost during such an abrupt connection release.
The second method to release a transport connection is to release independently the two directions of data transfer.
Once a user of the transport service has sent all its SDUs, it performs a DISCONNECT.req for its direction of data
transfer. The transport entity sends a control segment to request the release of the connection after the delivery of
all previous SDUs to the remote user. This is usually done by placing in the DR the next sequence number and by
delivering the DISCONNECT.ind only after all previous DATA.ind. The remote entity confirms the reception of
the DR segment and the release of the corresponding direction of data transfer by returning an acknowledgement.
8 In reality, the Internet does not strictly enforce this MSL. However, it is reasonable to expect that most packets on the Internet will not
remain in the network during more than 2 minutes. There are a few exceptions to this rule, such as RFC 1149 whose implementation is
described in http://www.blug.linux.no/rfc1149/ but there are few real links supporting RFC 1149 in the Internet.

Figure 2.56: Abrupt connection release
Figure 2.57: Graceful connection release

2.5 Naming and addressing
The network and the transport layers rely on addresses that are encoded as fixed size bit strings. A network layer
uniquely identifies a host. Several transport layer entities can use the service of the same network layer. For
example, a reliable transport protocol and a connectionless transport protocol can coexist on the same host. In
this case, the network layer multiplexes the segments produced by the two protocols. This multiplexing is usually
achieved by placing in the network packet header a field that indicates which transport protocol produced and
should process the segment. Given that there are few different transport protocols, this field does not need to be
long. The port numbers play a similar role in the transport layer since they enable it to multiplex data from several
application processes.
While addresses are natural for the network and transport layer entities, human users prefer to use names when
interacting with servers. Names can be encoded as a character string and a mapping services allows applications
to map a name into the corresponding address. Using names is friendlier for the human users than addresses, but
it also provides a level of indirection which is very useful in various situations. Before looking at these benefits of
names, it is useful to discuss how names are used on the Internet.
In the early days of the Internet, there were only a few number of hosts (mainly minicomputers) connected to the
network. The most popular applications were remote login and file transfer. By 1983, there were already five
hundred hosts attached to the Internet. Each of these hosts were identified by a unique IPv4 address. Forcing
human users to remember the IPv4 addresses of the remote hosts that they want to use was not user-friendly.
Human users prefer to remember names, and use them when needed. Using names as aliases for addresses is a
common technique in Computer Science. It simplifies the development of applications and allows the developer
to ignore the low level details. For example, by using a programming language instead of writing machine code,
a developer can write software without knowing whether the variables that it uses are stored in memory or inside
registers.
Because names are at a higher level than addresses, they allow (both in the example of programming above, and
on the Internet) to treat addresses as mere technical identifiers, which can change at will. Only the names are
stable.
The first solution that allowed applications to use names was the hosts.txt file. This file is similar to the symbol
table found in compiled code. It contains the mapping between the name of each Internet host and its associated IP
address 9 . It was maintained by SRI International that coordinated the Network Information Center (NIC). When
a new host was connected to the network, the system administrator had to register its name and IP address at the
NIC. The NIC updated the hosts.txt file on its server. All Internet hosts regularly retrieved the updated hosts.txt
file from the server maintained by SRI. This file was stored at a well-known location on each Internet host (see
RFC 952) and networked applications could use it to find the IP address corresponding to a name.
A hosts.txt file can be used when there are up to a few hundred hosts on the network. However, it is clearly not
suitable for a network containing thousands or millions of hosts. A key issue in a large network is to define a
suitable naming scheme. The ARPANet initially used a flat naming space, i.e. each host was assigned a unique
name. To limit collisions between names, these names usually contained the name of the institution and a suffix to
identify the host inside the institution (a kind of poor mans hierarchical naming scheme). On the ARPANet few
institutions had several hosts connected to the network.
However, the limitations of a flat naming scheme became clear before the end of the ARPANet and RFC 819
proposed a hierarchical naming scheme. While RFC 819 discussed the possibility of organising the names as a
directed graph, the Internet opted eventually for a tree structure capable of containing all names. In this tree, the
top-level domains are those that are directly attached to the root. The first top-level domain was .arpa 10 . This
top-level name was initially added as a suffix to the names of the hosts attached to the ARPANet and listed in
the hosts.txt file. In 1984, the .gov, .edu, .com, .mil and .org generic top-level domain names were added and
RFC 1032 proposed the utilisation of the two letter ISO-3166 country codes as top-level domain names. Since
ISO-3166 defines a two letter code for each country recognised by the United Nations, this allowed all countries
to automatically have a top-level domain. These domains include .be for Belgium, .fr for France, .us for the USA,
.ie for Ireland or .tv for Tuvalu, a group of small islands in the Pacific and .tm for Turkmenistan. Today, the set
of top-level domain-names is managed by the Internet Corporation for Assigned Names and Numbers (ICANN).
9 The hosts.txt file is not maintained anymore. A historical snapshot retrieved on April 15th, 1984 is available from
http://ftp.univie.ac.at/netinfo/netinfo/hosts.txt
10 See http://www.donelan.com/dnstimeline.html for a time line of DNS related developments.

Recently, ICANN added a dozen of generic top-level domains that are not related to a country and the .cat top-level
domain has been registered for the Catalan language. There are ongoing discussions within ICANN to increase
the number of top-level domains.
Each top-level domain is managed by an organisation that decides how sub-domain names can be registered. Most
top-level domain names use a first-come first served system, and allow anyone to register domain names, but
there are some exceptions. For example, .gov is reserved for the US government, .int is reserved for international
organisations and names in the .ca are mainly reserved for companies or users who are present in Canada.
Figure 2.58: The tree of domain names
RFC 1035 recommended the following BNF for fully qualified domain names, to allow host names with a syntax
which works with all applications (the domain names themselves have a much richer syntax).
Figure 2.59: BNF of the fully qualified host names
This grammar specifies that a host name is an ordered list of labels separated by the dot (.) character. Each label
can contain letters, numbers and the hyphen character (-) 11 . Fully qualified domain names are read from left to
right. The first label is a hostname or a domain name followed by the hierarchy of domains and ending with the
root implicitly at the right. The top-level domain name must be one of the registered TLDs 12 . For example, in the
above figure, www.whitehouse.gov corresponds to a host named www inside the whitehouse domain that belongs
to the gov top-level domain. info.ucl.ac.be corresponds to the info domain inside the ucl domain that is included
in the ac sub-domain of the be top-level domain.
This hierarchical naming scheme is a key component of the Domain Name System (DNS). The DNS is a dis-
tributed database that contains mappings between fully qualified domain names and IP addresses. The DNS uses
the client-server model. The clients are hosts that need to retrieve the mapping for a given name. Each nameserver
stores part of the distributed database and answers the queries sent by clients. There is at least one nameserver that
is responsible for each domain. In the figure below, domains are represented by circles and there are three hosts
inside domain dom (h1, h2 and h3) and three hosts inside domain a.sdom1.dom. As shown in the figure below, a
sub-domain may contain both host names and sub-domains.
A nameserver that is responsible for domain dom can directly answer the following queries :
the IP address of any host residing directly inside domain dom (e.g. h2.dom in the figure above)
the nameserver(s) that are responsible for any direct sub-domain of domain dom (i.e. sdom1.dom and
sdom2.dom in the figure above, but not z.sdom1.dom)
To retrieve the mapping for host h2.dom, a client sends its query to the name server that is responsible for domain
.dom. The name server directly answers the query. To retrieve a mapping for h3.a.sdom1.dom a DNS client first
sends a query to the name server that is responsible for the .dom domain. This nameserver returns the nameserver
that is responsible for the sdom1.dom domain. This nameserver can now be contacted to obtain the nameserver
11 This specification evolved later to support domain names written by using other character sets than us-ASCII RFC 5890. This extension
is important to support languages other than English, but a detailed discussion is outside the scope of this document.
12 The official list of top-level domain names is maintained by IANA at http://data.iana.org/TLD/tlds-alpha-by-domain.txt Additional infor-
mation about these domains may be found at http://en.wikipedia.org/wiki/List_of_Internet_top-level_domains
2.5. Naming and addressing 71

Figure 2.60: A simple tree of domain names
that is responsible for the a.sdom1.dom domain. This nameserver can be contacted to retrieve the mapping for the
h3.a.sdom1.dom name. Thanks to this organisation of the nameservers, it is possible for a DNS client to obtain the
mapping of any host inside the .dom domain or any of its subdomains. To ensure that any DNS client will be able
to resolve any fully qualified domain name, there are special nameservers that are responsible for the root of the
domain name hierarchy. These nameservers are called root nameserver. There are currently about a dozen root
nameservers.
Each root nameserver maintains the list 13 of all the nameservers that are responsible for each of the top-level
domain names and their IP addresses 14 . All root nameservers are synchronised and provide the same answers.
By querying any of the root nameservers, a DNS client can obtain the nameserver that is responsible for any
top-level-domain name. From this nameserver, it is possible to resolve any domain name.
To be able to contact the root nameservers, each DNS client must know their IP addresses. This implies, that
DNS clients must maintain an up-to-date list of the IP addresses of the root nameservers. Without this list, it is
impossible to contact the root nameservers. Forcing all Internet hosts to maintain the most recent version of this list
would be difficult from an operational point of view. To solve this problem, the designers of the DNS introduced
a special type of DNS server : the DNS resolvers. A resolver is a server that provides the name resolution service
for a set of clients. A network usually contains a few resolvers. Each host in these networks is configured to send
all its DNS queries via one of its local resolvers. These queries are called recursive queries as the resolver must
recurse through the hierarchy of nameservers to obtain the answer.
DNS resolvers have several advantages over letting each Internet host query directly nameservers. Firstly, regular
Internet hosts do not need to maintain the up-to-date list of the IP addresses of the root servers. Secondly, regular
Internet hosts do not need to send queries to nameservers all over the Internet. Furthermore, as a DNS resolver
serves a large number of hosts, it can cache the received answers. This allows the resolver to quickly return
answers for popular DNS queries and reduces the load on all DNS servers [JSBM2002].
2.5.1 Benefits of names
Using names instead of addresses inside applications has several important benefits in addition to being more
human friendly. To understand these benefits, let us consider a popular application that provides information
stored on servers. This application involves clients and servers. The server processes provide information upon
requests from client processes running on remote hosts. A first deployment of this application would be to rely
only on addresses. In this case, the server process would be installed on one host and the clients would connect to
this server to retrieve information. Such a deployment has several drawbacks :
if the server process moves to another physical server, all clients must be informed about the new server
address
if there are many concurrent clients, the load of the server will increase without any possibility of adding
another server without changing the server addresses user by the clients
13 A copy of the information maintained by each root nameserver is available at http://www.internic.net/zones/root.zone
14 Until February 2008, the root DNS servers only had IPv4 addresses. IPv6 addresses were added to the root DNS servers slowly to
avoid creating problems as discussed in http://www.icann.org/en/committees/security/sac018.pdf In 2013, several DNS root servers are still
not reachable by using IPv6. The full list is available at http://www.root-servers.org/

Using names solves these problems and provide additional benefits. If clients are configured with the name of the
server, they will query the name service before connecting to the server. The name service will resolve the name
into the corresponding address. If a server process needs to move from one physical server to another, it suffices
to update the name to address mapping of the name service to allow all clients to connect to the new server. The
name service also enables the servers to better sustain be load. Assume a very popular server which is accessed
by millions of user. This service cannot be provided by a single physical server due to performance limitations.
Thanks to the utilisation of names, it is possible to scale this service by mapping a given name to a set of addresses.
When a client queries the name service for the servers name, the name service returns one of the addresses in the
set. Various strategies can be used to select one particular address inside the set of addresses. A first strategy is to
select a random address in the set. A second strategy is to maintain information about the load on the servers and
return the address of the less loaded server. Note that the list of server addresses does not need to remain fixed. It
is possible to add and remove addresses from the list to cope with load fluctuations. Another strategy is to infer
the location of the client from the name request and return the address of the closest server.
Mapping a single name onto a set of addresses allow popular servers to scale dynamically. There are also benefits
in mapping multiple names, possibly a large number of them, onto a single address. Consider the case of informa-
tion servers run by individuals or SMEs. Some of these servers attract only a few clients per day. Using a single
physical server for each of these services would be a waste of resources. A better approach is to use a single server
for a set of services that are all identified by different names. This enables service providers to support a large
number of servers, identifiied by different names, onto a single physical server. If one of these servers becomes
very popular, it will be possible to map its name onto a set of addresses to be able to sustain the load. There are
some deployments where this mapping is done dynamically in function of the load.
Names provide a lot of flexibility compared to addresses. For the network, they play a similar role as variables
in programming languages. No programmer using a high-level programming language would consider using
addresses instead of variables. For the same reasons, all networked applications should depend on names and
avoid dealing with addresses as much as possible.
2.6 Sharing resources
A network is designed to support a potentially large number of users that exchange information with each other.
These users produce and consume information which is exchanged through the network. To support its users, a
network uses several types of resources. It is important to keep in mind the different resources that are shared
inside the network.
The first and more important resource inside a network is the link bandwidth. There are two situations where
link bandwidth needs to be shared between different users. The first situation is when several hosts are attached
to the same physical link. This situation mainly occurs in Local Area Networks (LAN). A LAN is a network
that efficiently interconnects several hosts (usually a few dozens to a few hundreds) in the same room, building
or campus. Consider for a example a network with five hosts. Any of these hosts needs to be able to exchange
information with any of the other five hosts. A first organisation for this LAN is the full-mesh.
Figure 2.61: A Full mesh network
The full-mesh is the most reliable and highest performing network to interconnect these five hosts. However, this
network organisation has two important drawbacks. First, if a network contains n hosts, then n(n1)
2 links are
required. If the network contains more than a few hosts, it becomes impossible to lay down the required physical
2.6. Sharing resources 73

links. Second, if the network contains n hosts, then each host must have n 1 interfaces to terminante n 1 links.
This is beyond the capabilities of most hosts. Furthermore, if a new host is added to the network, new links have
to be laid down and one interface has to be added to each participating host. However, full-mesh has the advantage
of providing the lowest delay between the hosts and the best resiliency against link failures. In practice, full-mesh
networks are rarely used expected when there are few network nodes and resiliency is key.
The second possible physical organisation, which is also used inside computers to connect different extension
cards, is the bus. In a bus network, all hosts are attached to a shared medium, usually a cable through a single
interface. When one host sends an electrical signal on the bus, the signal is received by all hosts attached to the bus.
A drawback of bus-based networks is that if the bus is physically cut, then the network is split into two isolated
networks. For this reason, bus-based networks are sometimes considered to be difficult to operate and maintain,
especially when the cable is long and there are many places where it can break. Such a bus-based topology was
used in early Ethernet networks.
Figure 2.62: A network organized as a Bus
A third organisation of a computer network is a star topology. In such topologies, hosts have a single physical
interface and there is one physical link between each host and the center of the star. The node at the center of
the star can be either a piece of equipment that amplifies an electrical signal, or an active device, such as a piece
of equipment that understands the format of the messages exchanged through the network. Of course, the failure
of the central node implies the failure of the network. However, if one physical link fails (e.g. because the cable
has been cut), then only one node is disconnected from the network. In practice, star-shaped networks are easier
to operate and maintain than bus-shaped networks. Many network administrators also appreciate the fact that
they can control the network from a central point. Administered from a Web interface, or through a console-like
connection, the center of the star is a useful point of control (enabling or disabling devices) and an excellent
observation point (usage statistics).
Figure 2.63: A network organised as a Star
A fourth physical organisation of a network is the ring topology. Like the bus organisation, each host has a single
physical interface connecting it to the ring. Any signal sent by a host on the ring will be received by all hosts
attached to the ring. From a redundancy point of view, a single ring is not the best solution, as the signal only
travels in one direction on the ring; thus if one of the links composing the ring is cut, the entire network fails. In
practice, such rings have been used in local area networks, but are now often replaced by star-shaped networks.
In metropolitan networks, rings are often used to interconnect multiple locations. In this case, two parallel links,
composed of different cables, are often used for redundancy. With such a dual ring, when one ring fails all the
traffic can be quickly switched to the other ring.
A fifth physical organisation of a network is the tree. Such networks are typically used when a large number of
customers must be connected in a very cost-effective manner. Cable TV networks are often organised as trees.

Figure 2.64: A network organised as a ring
Figure 2.65: A network organized as a Tree
2.6.1 Sharing bandwidth
In all these networks, except the full-mesh, the link bandwidth is shared among all connected hosts. Various
algorithms have been proposed and are used to efficiently share the access to this resource. We explain several of
them in the Medium Access Control section below.
Note: Fairness in computer networks

Sharing resources is important to ensure that the network efficiently serves its user. In practice, there are many
ways to share resources. Some resource sharing schemes consider that some users are more important than others
and should obtain more resources. For example, on the highways, police cars and ambulances have priority to
use the highways. In some cities, traffic lanes are reserved for buses to promote public services, ... In computer
networks, the same problem arise. Given that resources are limited, the network needs to enable users to efficiently
share them. Before designing an efficient resource sharing scheme, one needs to first formalize its objectives. In
computer networks, the most popular objective for resource sharing schemes is that they must be fair. In a
simple situation, for example two hosts using a shared 2 Mbps link, the sharing scheme should allocate the same
bandwidth to each user, in this case 1 Mbps. However, in a large networks, simply dividing the available resources
by the number of users is not sufficient. Consider the network shown in the figure below where A1 sends data to
A2, B1 to B2, ... In this network, how should we divide the bandwidth among the different flows ? A first approach
would be to allocate the same bandwidth to each flow. In this case, each flow would obtain 5 Mbps and the link
between R2 and R3 would not be fully loaded. Another approach would be to allocate 10 Mbps to A1-A2, 20
Mbps to C1-C2 and nothing to B1-B2. This is clearly unfair.

A1
R1 10 Mbps
B1
R2
A2 20 Mbps
R3
C1 B2
C2
In large networks, fairness is always a compromise. The most widely used definition of fairness is the max-min
fairness. A bandwidth allocation in a network is said to be max-min fair if it is such that it is impossible to
allocate more bandwidth to one of the flows without reducing the bandwidth of a flow that already has a smaller
allocation than the flow that we want to increase. If the network is completely known, it is possible to derive a
max-min fair allocation as follows. Initially, all flows have a null bandwidth and they are placed in the candidate
set. The bandwidth allocation of all flows in the candidate set is increased until one link becomes congested. At
this point, the flows that use the congested link have reached their maximum allocation. They are removed from
the candidate set and the process continues until the candidate set becomes empty.
In the above network, the allocation of all flows would grow until A1-A2 and B1-B2 reach 5 Mbps. At this point,
link R1-R2 becomes congested and these two flows have reached their maximum. The allocation for flow C1-C2
can increase until reaching 15 Mbps. At this point, link R2-R3 is congested. To increase the bandwidth allocated
to C1-C2, one would need to reduce the allocation to flow B1-B2. Similarly, the only way to increase the allocation
to flow B1-B2 would require a decrease of the allocation to A1-A2.
2.6.2 Network congestion
Sharing bandwidth among the hosts directly attached to a link is not the only sharing problem that occurs in
computer networks. To understand the general problem, let us consider a very simple network which contains
only point-to-point links. This network contains three hosts and two network nodes. All links inside the network
have the same capacity. For example, let us assume that all links have a bandwidth of 1000 bits per second and
that the hosts send packets containing exactly one thousand bits.

R1 R2
C
In the network above, consider the case where host A is transmitting packets to destination C. A can send one
packet per second and its packets will be delivered to C. Now, let us explore what happens when host B also starts
to transmit a packet. Node R1 will receive two packets that must be forwarded to R2. Unfortunately, due to the
limited bandwidth on the R1-R2 link, only one of these two packets can be transmitted. The outcome of the second
packet will depend on the available buffers on R1. If R1 has one available buffer, it could store the packet that
has not been transmitted on the R1-R2 link until the link becomes available. If R1 does not have available buffers,
then the packet needs to be discarded.
Besides the link bandwidth, the buffers on the network nodes are the second type of resource that needs to be
shared inside the network. The node buffers play an important role in the operation of the network because that
can be used to absorb transient traffic peaks. Consider again the example above. Assume that one average host
A and host B send a group of three packets every ten seconds. Their combined transmission rate (0.6 packets
per second) is, on average, lower than the network capacity (1 packet per second). However, if they both start
to transmit at the same time, node R1 will have to absorb a burst of packets. This burst of packets is a small
network congestion. We will say that P a network is congested, when the sum of the traffic demand from the hosts
is larger than the network capacity demand > capacity. This network congestion problem is one of the most
difficult resource sharing problem in computer networks. Congestion occurs in almost all networks. Minimizing
the amount of congestion is a key objective for many network operators. In most cases, they will have to accept
transient congestion, i.e. congestion lasting a few seconds or perhaps minutes, but will want to prevent congestion
that lasts days or months. For this, they can rely on a wide range of solutions. We briefly present some of these in
the paragraphs below.
If R1 has enough buffers, it will be able to absorb the load without having to discard packets. The packets sent by
hosts A and B will reach their final destination C, but will experience a longer delay than when they are transmitting
alone. The amount of buffering on the network node is the first paper that a network operator can tune to control
congestion inside his network. Given the decreasing cost of memory, one could be tempted to put as many buffers
15
as possible on the network nodes. Let us consider this case in the network above and assume that R1 has infinite
buffers. Assume now that hosts A and B try to transmit a file that corresponds to one thousand packets each.
Both are using a reliable protocol that relies on go-back-n to recover from transmission errors. The transmission
starts and packets start to accumulate in R1s buffers. These presence of these packets in the buffers increases the
delay between the transmission of a packet by A and the return of the corresponding acknowledgement. Given the
increasing delay, host A (and B as well) will consider that some of the packets that it sent have been lost. These
packets will be retransmitted and will enter the buffers of R1. The occupancy of the buffers of R1 will continue
to increase and the delays as well. This will cause new retransmissions, ... In the end, several copies of the same
packet will be transmitted over the R1-R2, but only one file will be delivered (very slowly) to the destination.
This is known as the congestion collapse problem RFC 896. Congestion collapse is the nightmare for network
operators. When it happens, the network carries packets without delivering useful data to the end users.
15 There are still some vendors that try to put as many buffers as possible on their network nodes. A recent example is the buffer bloat
problem that plagues some low-end Internet routers [GN2011].

Note: Congestion collapse on the Internet

Congestion collapse is unfortunately not only an academic experience. Van Jacobson reports in [Jacobson1988]
one of these events that affected him while he was working at the Lawrence Berkeley Laboratory (LBL). LBL was
two network nodes away from the University of California in Berkeley. At that time, the link between the two
sites had a bandwidth of 32 Kbps, but some hosts were already attached to 10 Mbps LANs. In October 1986, the
data throughput from LBL to UC Berkeley ... dropped from 32 Kbps to 40 bps. We were fascinated by this sudden
factor-of-thousand drop in bandwidth and embarked on an investigation of why things had gotten so bad. This
work lead to the development of various congestion control techniques that have allowed the Internet to continue
to grow without experiencing widespread congestion collapse events.
Besides bandwidth and memory, a third resource that needs to be shared inside a network is the (packet) processing
capacity. To forward a packet, a network node needs bandwidth on the outgoing link, but it also needs to analyze
the packet header to perform a lookup inside its forwarding table. Performing these lookup operations require
resources such as CPU cycles or memory accesses. Network nodes are usually designed to be able to sustain a
given packet processing rate, measured in packets per second.
Note: Packets per second versus bits per second

The performance of network nodes can be characterized by two key metrics :
the nodes capacity measured in bits per second
the nodes lookup performance measured in packets per second
The nodes capacity in bits per second mainly depends on the physical interfaces that it uses and also on the
capacity of the internal interconnection (bus, crossbar switch, ...) between the different interfaces inside the node.
Many vendors, in particular for low-end devices will use the sum of the bandwidth of the nodes interfaces as the
node capacity in bits per second. Measurements do not always match this maximum theoretical capacity. A well
designed network node will usually have a capacity in bits per second larger than the sum of its link capacities.
Such nodes will usually reach this maximum capacity when forwarding large packets.
When a network node forwards small packets, its performance is usually limited by the number of lookup op-
erations that it can perform every second. This lookup performance is measured in packets per second. The
performance may depend on the length of the forwarded packets. The key performance factor is the number of
minimal size packets that are forwarded by the node every second. This rate can lead to a capacity in bits per
second which is much lower than the sum of the bandwidth of the nodes links.
Let us now try to present a broad overview of the congestion problem in networks. We will assume that the
network is composed of dedicated links having a fixed bandwidth 16 . A network contains hosts that generate
and receive packets and nodes P that forward packets. Assuming that each host is connected via a single link to the
network, the largest demand is AccessLinks. In practice, this largest demand is never reached and the network
will be engineered to sustain a much lower traffic demand. The difference between the worst-case traffic demand
and the sustainable traffic demand can be large, up to several orders of magnitude. Fortunately, the hosts are not
completely dumb and they can adapt their traffic demand to the current state of the network and the available
bandwidth. For this, the hosts need to sense the current level of congestion and adjust their own traffic demand
based on the estimated congestion. Network nodes can react in different ways to network congestion and hosts
can sense the level of congestion in different ways.
Let us first explore which mechanisms can be used inside a network to control congestion and how these mecha-
nisms can influence the behavior of the end hosts.
As explained earlier, one of the first manifestation of congestion on network nodes is the saturation of the network
links that leads to a growth in the occupancy of the buffers of the node. This growth of the buffer occupancy implies
that some packets will spend more time in the buffer and thus in the network. If hosts measure the network delays
(e.g. by measuring the round-trip-time between the transmission of a packet and the return of the corresponding
acknowledgement) they could start to sense congestion. On low bandwidth links, a growth in the buffer occupancy
can lead to an increase of the delays which can be easily measured by the end hosts. On high bandwidth links, a
16 Some networking technologies allow to adjust dynamically the bandwidth of links. For example, some devices can reduce their bandwidth
to preserve energy. We ignore these technologies in this basic course and assume that all links used inside the network have a fixed bandwidth.

few packets inside the buffer will cause a small variation in the delay which may not necessarily be larger that the
natural fluctuations of the delay measurements.
If the buffers occupancy continues to grow, it will overflow and packets will need to be discarded. Discarding
packets during congestion is the second possible reaction of a network node to congestion. Before looking at
how a node can discard packets, it is interesting to discuss qualitatively the impact of the buffer occupancy on the
reliable delivery of data through a network. This is illustrated by the figure below, adapted from [Jain1990].
Figure 2.66: Network congestion
When the network load is low, buffer occupancy and link utilizations are low. The buffers on the network nodes
are mainly used to absorb very short bursts of packets, but on average the traffic demand is lower than the network
capacity. If the demand increases, the average buffer occupancy will increase as well. Measurements have shown
that the total throughput increases as well. If the buffer occupancy is zero or very low, transmission opportunities
on network links can be missed. This is not the case when the buffer occupancy is small but non zero. However, if
the buffer occupancy continues to increase, the buffer becomes overloaded and the throughput does not increase
anymore. When the buffer occupancy is close to the maximum, the throughput may decrease. This drop in
throughput can be caused by excessive retransmissions of reliable protocols that incorrectly assume that previously
sent packets have been lost while they are still waiting in the buffer. The network delay on the other hand increases
with the buffer occupancy. In practice, a good operating point for a network buffer is a low occupancy to achieve
high link utilization and also low delay for interactive applications.
Discarding packets is one of the signals that the network nodes can use to inform the hosts of the current level of
congestion. Buffers on network nodes are usually used as FIFO queues to preserve packet ordering. Several packet
discard mechanisms have been proposed for network nodes. These techniques basically answer two different
questions :
What triggers a packet to be discarded ? What are the conditions that lead a network node to decide to
discard a packet. The simplest answer to this question is : When the buffer is full. Although this is a good
congestion indication, it is probably not the best one from a performance viewpoint. An alternative is to
discard packets when the buffer occupancy grows too much. In this case, it is likely that the buffer will
become full shortly. Since packet discarding is an information that allows hosts to adapt their transmission
rate, discarding packets early could allow hosts to react earlier and thus prevent congestion from happening.
Which packet(s) should be discarded ? Once the network node has decided to discard packets, it needs to
actually discard real packets.
By combining different answers to these questions, network researchers have developed different packet discard
mechanisms.
tail drop is the simplest packet discard technique. When a buffer is full, the arriving packet is discarded. Tail
drop can be easily implemented. This is, by far, the most widely used packet discard mechanism. However,
it suffers from two important drawbacks. First, since tail drop discards packets only when the buffer is full,
buffers tend to be congested and realtime applications may suffer from the increased delays. Second, tail
drop is blind when it discards a packet. It may discard a packet from a low bandwidth interactive flow while
most of the buffer is used by large file transfers.
drop from front is an alternative packet discard technique. Instead of removing the arriving packet, it re-
moves the packet that was at the head of the queue. Discarding this packet instead of the arriving one can

have two advantages. First, it already stayed a long time in the buffer. Second, hosts should be able to detect
the loss (and thus the congestion) earlier.
probabilistic drop. Various random drop techniques have been proposed. Compared to the previous tech-
niques. A frequently cited technique is Random Early Discard (RED) [FJ1993]. RED measures the average
buffer occupancy and probabilistically discards packets when this average occupancy is too high. Com-
pared to tail drop and drop from front, an advantage of RED is that thanks to the probabilistic drops, packets
should be discarded from different flows in proportion of their bandwidth.
Discarding packets is a frequent reaction to network congestion. Unfortunately, discarding packets is not optimal
since a packet which is discarded on a network node has already consumed resources on the upstream nodes.
There are other ways for the network to inform the end hosts of the current congestion level. A first solution is to
mark the packets when a node is congested. Several networking technologies have relied on this kind of packet
marking.
In datagram networks, Forward Explicit Congestion Notification (FECN) can be used. One field of the packet
header, typically one bit, is used to indicate congestion. When a host sends a packet, the congestion bit is reset.
If the packet passes through a congested node, the congestion bit is set. The destination can then determine the
current congestion level by measuring the fraction of the packets that it received with the congestion bit set. It may
then return this information to the sending host to allow it to adapt its retransmission rate. Compared to packet
discarding, the main advantage of FECN is that hosts can detect congestion explicitly without having to rely on
packet losses.
In virtual circuit networks, packet marking can be improved if the return packets follow the reverse path of the
forward packets. It this case, a network node can detect congestion on the forward path (e.g. due to the size of its
buffer), but mark the packets on the return path. Marking the return packets (e.g. the acknowledgements used by
reliable protocols) provides a faster feedback to the sending hosts compared to FECN. This technique is usually
called Backward Explicit Congestion Notification (BECN).
If the packet header does not contain any bit in the header to represent the current congestion level, an alternative
is to allow the network nodes to send a control packet to the source to indicate the current congestion level. Some
networking technologies use such control packets to explicitly regulate the transmission rate of sources. However,
their usage is mainly restricted to small networks. In large networks, network nodes usually avoid using such
control packets. These controlled packets are even considered to be dangerous in some networks. First, using
them increases the network load when the network is congested. Second, while network nodes are optimized to
forward packets, they are usually pretty slow at creating new packets.
Dropping and marking packets is not the only possible reaction of a router that becomes congested. A router
could also selectively delay packets belonging to some flows. There are different algorithms that can be used by a
router to delay packets. If the objective of the router is to fairly distribute to bandwidth of an output link among
competing flows, one possibility is to organize the buffers of the router as a set of queues. For simplicity, let us
assume that the router is capable of supporting a fixed number of concurrent flows, say N. One of the queues of the
router is associated to each flow and when a packet arrives, it is placed at the tail of the corresponding queue. All
the queues are controlled by a scheduler. A scheduler is an algorithm that is run each time there is an opportunity
to transmit a packet on the outgoing link. Various schedulers have been proposed in the scientific literature and
some are used in real routers.
Figure 2.67: A round-robin scheduler

A very simple scheduler is the round-robin scheduler. This scheduler serves all the queues in a round-robin
fashion. If all flows send packets of the same size, then the round-robin scheduler allocates the bandwidth fairly
among the different flows. Otherwise, it favors flows that are using larger packets. Extensions to the round-
robin scheduler have been proposed to provide a fair distribution of the bandwidth with variable-length packets
[SV1995] but these are outside the scope of this chapter.
# N queues
# state variable : next_queue
next_queue=0
while (true) :
if isEmpty(buffer) :
wait
# wait for next packet in buffer
if !isEmpty(queue[next_queue]) :
# Send packet at head of next_queue
p=remove_packet(queue[next_queue])
send(p)
next_queue=(next_queue+1)%N
# end while
2.6.3 Distributing the load across the network
Delays, packet discards, packet markings and control packets are the main types of information that the network
can exchange with the end hosts. Discarding packets is the main action that a network node can perform if the
congestion is too severe. Besides tackling congestion at each node, it is also possible to change the divert some
traffic flows from heavily loaded links to reduce congestion. Early routing algorithms [MRR1980] have used delay
measurements to detect congestion between network nodes and update the link weights dynamically. By reflect-
ing the delay perceived by applications in the link weights used for the shortest paths computation, these routing
algorithms managed to dynamically change the forwarding paths in reaction to congestion. However, deployment
experience showed that these dynamic routing algorithms could cause oscillations and did not necessarily lower
congestion. Deployed datagram networks rarely use dynamic routing algorithms, except in some wireless net-
works. In datagram networks, the state of the art reaction to long term congestion, i.e. congestion lasting hours,
days or more, is to measure the traffic demand and then select the link weights [FRT2002] that allow to minimize
the maximum link loads. If the congestion lasts longer, changing the weights is not sufficient anymore and the
network needs to be upgraded with few or faster links. However, in Wide Area Networks, adding new links can
take months.
In virtual circuit networks, another way to manage or prevent congestion is to limit the number of circuits that
use the network at any time. This technique is usually called connection admission control. When a host requests
the creation of a new circuit in the network, it specifies the destination and in some networking technologies the
required bandwidth. With this information, the network can check whether there are enough resources available
to reach this particular destination. If yes, the circuit is established. However, the requested is denied and the
host will have to defer the creation of its virtual circuit. Connection admission control schemes are widely used in
the telephone networks. In these networks, a busy tone corresponds to an unavailable destination or a congested
network.
In datagram networks, this technique cannot be easily used since the basic assumption of such a network is that a
host can send any packet towards any destination at any time. A host does not need to request the authorization of
the network to send packets towards a particular destination.
Based on the feedback received from the network, the hosts can adjust their transmission rate. We discuss in
section Congestion control some techniques that allow hosts to react to congestion.
Another way to share the network resources is to distribute the load across multiple links. Many techniques have
been designed to spread the load over the network. As an illustration, let us briefly consider how load can be
shared when accessing some content. Consider a large and popular file such as the image of a Linux distribution
or the upgrade of a commercial operating system that will be downloaded by many users. There are many ways
to distribute this large file. A naive solution is to place one copy of the file on a server and allow all users to
download this file from the server. If the file is popular and millions of users want to download it, the server will
quickly become overloaded. There are two classes of solutions that can be used to serve a large number of users.

A first approach is to store the file on servers whose name is known by the clients. Before retrieving the file, each
client will query the name service to obtain the address of the server. If the file is available from many servers,
the name service can provide different addresses to different clients. This will automatically spread the load since
different clients will download the file from different servers. Most large content providers use such a solution to
distribute large files or videos.
There is another solution that allows to spread the load among many sources without relying on the name service.
The popular bittorent service is an example of this approach. With this solution, each file is divided in blocks of a
fixed size. To retrieve a file, a client needs to retrieve all the blocks that compose the file. However, nothing forces
the client to retrieve all the blocks in sequence and from the same server. Each file is associated with metadata
that indicates for each block a list of addresses of hosts that store this block. To retrieve a complete file, a client
first downloads the metadata. Then, it tries to retrieve each block from one of the hosts that store the block. In
practice, implementations often try to download several blocks in parallel. Once one block has been successfully
downloaded, the next block can be requested. If a host is slow to provide one block or becomes unavailable,
the client can contact another host listed in the metadata. Most deployments of bittorrent allow the clients to
participate to the distribution of blocks. Once a client has downloaded one block, it contacts the server which
stores the metadata to indicate that it can also provide this block. With this scheme, when a file is popular, its
blocks are downloaded by many hosts that automatically participate in the distribution of the blocks. Thus, the
number of servers that are capable of providing blocks from a popular file automatically increases with the files
popularity.
Now that we have provided a broad overview of the techniques that can be used to spread the load and allocate
resources in the network, let us analyze two techniques in more details : Medium Access Control and Congestion
control.
2.6.4 Medium Access Control algorithms
The common problem among Local Area Networks is how to efficiently share the available bandwidth. If two
devices send a frame at the same time, the two electrical, optical or radio signals that correspond to these frames
will appear at the same time on the transmission medium and a receiver will not be able to decode either frame.
Such simultaneous transmissions are called collisions. A collision may involve frames transmitted by two or more
devices attached to the Local Area Network. Collisions are the main cause of errors in wired Local Area Networks.
All Local Area Network technologies rely on a Medium Access Control algorithm to regulate the transmissions to
either minimise or avoid collisions. There are two broad families of Medium Access Control algorithms :
1. Deterministic or pessimistic MAC algorithms. These algorithms assume that collisions are a very severe
problem and that they must be completely avoided. These algorithms ensure that at any time, at most one
device is allowed to send a frame on the LAN. This is usually achieved by using a distributed protocol which
elects one device that is allowed to transmit at each time. A deterministic MAC algorithm ensures that no
collision will happen, but there is some overhead in regulating the transmission of all the devices attached
to the LAN.
2. Stochastic or optimistic MAC algorithms. These algorithms assume that collisions are part of the normal
operation of a Local Area Network. They aim to minimise the number of collisions, but they do not try to
avoid all collisions. Stochastic algorithms are usually easier to implement than deterministic ones.
We first discuss a simple deterministic MAC algorithm and then we describe several important optimistic algo-
rithms, before coming back to a distributed and deterministic MAC algorithm.
Static allocation methods
A first solution to share the available resources among all the devices attached to one Local Area Network is to
define, a priori, the distribution of the transmission resources among the different devices. If N devices need to
share the transmission capacities of a LAN operating at b Mbps, each device could be allocated a bandwidth of Nb
Mbps.
Limited resources need to be shared in other environments than Local Area Networks. Since the first radio trans-
missions by Marconi more than one century ago, many applications that exchange information through radio
signals have been developed. Each radio signal is an electromagnetic wave whose power is centered around a

given frequency. The radio spectrum corresponds to frequencies ranging between roughly 3 KHz and 300 GHz.
Frequency allocation plans negotiated among governments reserve most frequency ranges for specific applications
such as broadcast radio, broadcast television, mobile communications, aeronautical radio navigation, amateur ra-
dio, satellite, etc. Each frequency range is then subdivided into channels and each channel can be reserved for a
given application, e.g. a radio broadcaster in a given region.
Frequency Division Multiplexing (FDM) is a static allocation scheme in which a frequency is allocated to each
device attached to the shared medium. As each device uses a different transmission frequency, collisions cannot
occur. In optical networks, a variant of FDM called Wavelength Division Multiplexing (WDM) can be used. An
optical fiber can transport light at different wavelengths without interference. With WDM, a different wavelength
is allocated to each of the devices that share the same optical fiber.
Time Division Multiplexing (TDM) is a static bandwidth allocation method that was initially defined for the tele-
phone network. In the fixed telephone network, a voice conversation is usually transmitted as a 64 Kbps signal.
Thus, a telephone conservation generates 8 KBytes per second or one byte every 125 microseconds. Telephone
conversations often need to be multiplexed together on a single line. For example, in Europe, thirty 64 Kbps voice
signals are multiplexed over a single 2 Mbps (E1) line. This is done by using Time Division Multiplexing (TDM).
TDM divides the transmission opportunities into slots. In the telephone network, a slot corresponds to 125 mi-
croseconds. A position inside each slot is reserved for each voice signal. The figure below illustrates TDM on a
link that is used to carry four voice conversations. The vertical lines represent the slot boundaries and the letters
the different voice conversations. One byte from each voice conversation is sent during each 125 microseconds
slot. The byte corresponding to a given conversation is always sent at the same position in each slot.
Figure 2.68: Time-division multiplexing
TDM as shown above can be completely static, i.e. the same conversations always share the link, or dynamic. In
the latter case, the two endpoints of the link must exchange messages specifying which conversation uses which
byte inside each slot. Thanks to these signalling messages, it is possible to dynamically add and remove voice
conversations from a given link.
TDM and FDM are widely used in telephone networks to support fixed bandwidth conversations. Using them
in Local Area Networks that support computers would probably be inefficient. Computers usually do not send
information at a fixed rate. Instead, they often have an on-off behaviour. During the on period, the computer tries
to send at the highest possible rate, e.g. to transfer a file. During the off period, which is often much longer than
the on period, the computer does not transmit any packet. Using a static allocation scheme for computers attached
to a LAN would lead to huge inefficiencies, as they would only be able to transmit at N1 of the total bandwidth
during their on period, despite the fact that the other computers are in their off period and thus do not need to
transmit any information. The dynamic MAC algorithms discussed in the remainder of this chapter aim solve this
problem.
ALOHA
In the 1960s, computers were mainly mainframes with a few dozen terminals attached to them. These terminals
were usually in the same building as the mainframe and were directly connected to it. In some cases, the terminals
were installed in remote locations and connected through a modem attached to a dial-up line. The university
of Hawaii chose a different organisation. Instead of using telephone lines to connect the distant terminals, they
developed the first packet radio technology [Abramson1970]. Until then, computer networks were built on top of
either the telephone network or physical cables. ALOHANet showed that it was possible to use radio signals to
interconnect computers.
The first version of ALOHANet, described in [Abramson1970], operated as follows: First, the terminals and the
mainframe exchanged fixed-length frames composed of 704 bits. Each frame contained 80 8-bit characters, some
control bits and parity information to detect transmission errors. Two channels in the 400 MHz range were reserved
for the operation of ALOHANet. The first channel was used by the mainframe to send frames to all terminals.

The second channel was shared among all terminals to send frames to the mainframe. As all terminals share the
same transmission channel, there is a risk of collision. To deal with this problem as well as transmission errors,
the mainframe verified the parity bits of the received frame and sent an acknowledgement on its channel for each
correctly received frame. The terminals on the other hand had to retransmit the unacknowledged frames. As for
TCP, retransmitting these frames immediately upon expiration of a fixed timeout is not a good approach as several
terminals may retransmit their frames at the same time leading to a network collapse. A better approach, but still
far from perfect, is for each terminal to wait a random amount of time after the expiration of its retransmission
timeout. This avoids synchronisation among multiple retransmitting terminals.
The pseudo-code below shows the operation of an ALOHANet terminal. We use this python syntax for all Medium
Access Control algorithms described in this chapter. The algorithm is applied to each new frame that needs to be
transmitted. It attempts to transmit a frame at most max times (while loop). Each transmission attempt is performed
as follows: First, the frame is sent. Each frame is protected by a timeout. Then, the terminal waits for either a
valid acknowledgement frame or the expiration of its timeout. If the terminal receives an acknowledgement, the
frame has been delivered correctly and the algorithm terminates. Otherwise, the terminal waits for a random time
and attempts to retransmit the frame.
# ALOHA
N=1
while N<= max :
send(frame)
wait(ack_on_return_channel or timeout)
if (ack_on_return_channel):
break # transmission was successful
else:
# timeout
wait(random_time)
N=N+1
else:
# Too many transmission attempts
[Abramson1970] analysed the performance of ALOHANet under particular assumptions and found that ALO-
HANet worked well when the channel was lightly loaded. In this case, the frames are rarely retransmitted and the
channel traffic, i.e. the total number of (correct and retransmitted) frames transmitted per unit of time is close to
the channel utilization, i.e. the number of correctly transmitted frames per unit of time. Unfortunately, the analysis
1
also reveals that the channel utilization reaches its maximum at 2e = 0.186 times the channel bandwidth. At
higher utilization, ALOHANet becomes unstable and the network collapses due to collided retransmissions.
Note: Amateur packet radio

Packet radio technologies have evolved in various directions since the first experiments performed at the University
of Hawaii. The Amateur packet radio service developed by amateur radio operators is one of the descendants
ALOHANet. Many amateur radio operators are very interested in new technologies and they often spend countless
hours developing new antennas or transceivers. When the first personal computers appeared, several amateur radio
operators designed radio modems and their own datalink layer protocols [KPD1985] [BNT1997]. This network
grew and it was possible to connect to servers in several European countries by only using packet radio relays.
Some amateur radio operators also developed TCP/IP protocol stacks that were used over the packet radio service.
Some parts of the amateur packet radio network are connected to the global Internet and use the 44.0.0.0/8 prefix.
Many improvements to ALOHANet have been proposed since the publication of [Abramson1970], and this tech-
nique, or some of its variants, are still found in wireless networks today. The slotted technique proposed in
[Roberts1975] is important because it shows that a simple modification can significantly improve channel utiliza-
tion. Instead of allowing all terminals to transmit at any time, [Roberts1975] proposed to divide time into slots
and allow terminals to transmit only at the beginning of each slot. Each slot corresponds to the time required to
transmit one fixed size frame. In practice, these slots can be imposed by a single clock that is received by all
terminals. In ALOHANet, it could have been located on the central mainframe. The analysis in [Roberts1975]
reveals that this simple modification improves the channel utilization by a factor of two.

Carrier Sense Multiple Access
ALOHA and slotted ALOHA can easily be implemented, but unfortunately, they can only be used in networks that
are very lightly loaded. Designing a network for a very low utilisation is possible, but it clearly increases the cost
of the network. To overcome the problems of ALOHA, many Medium Access Control mechanisms have been
proposed which improve channel utilization. Carrier Sense Multiple Access (CSMA) is a significant improvement
compared to ALOHA. CSMA requires all nodes to listen to the transmission channel to verify that it is free before
transmitting a frame [KT1975]. When a node senses the channel to be busy, it defers its transmission until the
channel becomes free again. The pseudo-code below provides a more detailed description of the operation of
CSMA.
# persistent CSMA
N=1
while N<= max :
wait(channel_becomes_free)
send(frame)
wait(ack or timeout)
if ack :
else :
# timeout
N=N+1
# end of while loop
The above pseudo-code is often called persistent CSMA [KT1975] as the terminal will continuously listen to the
channel and transmit its frame as soon as the channel becomes free. Another important variant of CSMA is the
non-persistent CSMA [KT1975]. The main difference between persistent and non-persistent CSMA described
in the pseudo-code below is that a non-persistent CSMA node does not continuously listen to the channel to
determine when it becomes free. When a non-persistent CSMA terminal senses the transmission channel to be
busy, it waits for a random time before sensing the channel again. This improves channel utilization compared to
persistent CSMA. With persistent CSMA, when two terminals sense the channel to be busy, they will both transmit
(and thus cause a collision) as soon as the channel becomes free. With non-persistent CSMA, this synchronisation
does not occur, as the terminals wait a random time after having sensed the transmission channel. However, the
higher channel utilization achieved by non-persistent CSMA comes at the expense of a slightly higher waiting
time in the terminals when the network is lightly loaded.
# Non persistent CSMA
N=1
while N<= max :
listen(channel)
if free(channel):
send(frame)
if received(ack) :
else :
# timeout
N=N+1
else:
wait(random_time)
# end of while loop
[KT1975] analyzes in detail the performance of several CSMA variants. Under some assumptions about the trans-
mission channel and the traffic, the analysis compares ALOHA, slotted ALOHA, persistent and non-persistent
CSMA. Under these assumptions, ALOHA achieves a channel utilization of only 18.4% of the channel capacity.
Slotted ALOHA is able to use 36.6% of this capacity. Persistent CSMA improves the utilization by reaching
52.9% of the capacity while non-persistent CSMA achieves 81.5% of the channel capacity.

Carrier Sense Multiple Access with Collision Detection
CSMA improves channel utilization compared to ALOHA. However, the performance can still be improved,
especially in wired networks. Consider the situation of two terminals that are connected to the same cable. This
cable could, for example, be a coaxial cable as in the early days of Ethernet [Metcalfe1976]. It could also be built
with twisted pairs. Before extending CSMA, it is useful to understand more intuitively, how frames are transmitted
in such a network and how collisions can occur. The figure below illustrates the physical transmission of a frame
on such a cable. To transmit its frame, host A must send an electrical signal on the shared medium. The first step
is thus to begin the transmission of the electrical signal. This is point (1) in the figure below. This electrical signal
will travel along the cable. Although electrical signals travel fast, we know that information cannot travel faster
than the speed of light (i.e. 300.000 kilometers/second). On a coaxial cable, an electrical signal is slightly slower
than the speed of light and 200.000 kilometers per second is a reasonable estimation. This implies that if the cable
has a length of one kilometer, the electrical signal will need 5 microseconds to travel from one end of the cable to
the other. The ends of coaxial cables are equipped with termination points that ensure that the electrical signal is
not reflected back to its source. This is illustrated at point (3) in the figure, where the electrical signal has reached
the left endpoint and host B. At this point, B starts to receive the frame being transmitted by A. Notice that there is
a delay between the transmission of a bit on host A and its reception by host B. If there were other hosts attached
to the cable, they would receive the first bit of the frame at slightly different times. As we will see later, this timing
difference is a key problem for MAC algorithms. At point (4), the electrical signal has reached both ends of the
cable and occupies it completely. Host A continues to transmit the electrical signal until the end of the frame. As
shown at point (5), when the sending host stops its transmission, the electrical signal corresponding to the end of
the frame leaves the coaxial cable. The channel becomes empty again once the entire electrical signal has been
removed from the cable.
Figure 2.69: Frame transmission on a shared bus
Now that we have looked at how a frame is actually transmitted as an electrical signal on a shared bus, it is
interesting to look in more detail at what happens when two hosts transmit a frame at almost the same time. This
is illustrated in the figure below, where hosts A and B start their transmission at the same time (point (1)). At this
time, if host C senses the channel, it will consider it to be free. This will not last a long time and at point (2) the
electrical signals from both host A and host B reach host C. The combined electrical signal (shown graphically as
the superposition of the two curves in the figure) cannot be decoded by host C. Host C detects a collision, as it
receives a signal that it cannot decode. Since host C cannot decode the frames, it cannot determine which hosts
are sending the colliding frames. Note that host A (and host B) will detect the collision after host C (point (3) in
the figure below).
As shown above, hosts detect collisions when they receive an electrical signal that they cannot decode. In a wired

Figure 2.70: Frame collision on a shared bus
network, a host is able to detect such a collision both while it is listening (e.g. like host C in the figure above) and
also while it is sending its own frame. When a host transmits a frame, it can compare the electrical signal that it
transmits with the electrical signal that it senses on the wire. At points (1) and (2) in the figure above, host A senses
only its own signal. At point (3), it senses an electrical signal that differs from its own signal and can thus detects
the collision. At this point, its frame is corrupted and it can stop its transmission. The ability to detect collisions
while transmitting is the starting point for the Carrier Sense Multiple Access with Collision Detection (CSMA/CD)
Medium Access Control algorithm, which is used in Ethernet networks [Metcalfe1976] [IEEE802.3] . When an
Ethernet host detects a collision while it is transmitting, it immediately stops its transmission. Compared with
pure CSMA, CSMA/CD is an important improvement since when collisions occur, they only last until colliding
hosts have detected it and stopped their transmission. In practice, when a host detects a collision, it sends a special
jamming signal on the cable to ensure that all hosts have detected the collision.
To better understand these collisions, it is useful to analyse what would be the worst collision on a shared bus
network. Let us consider a wire with two hosts attached at both ends, as shown in the figure below. Host A
starts to transmit its frame and its electrical signal is propagated on the cable. Its propagation time depends on the
physical length of the cable and the speed of the electrical signal. Let us use to represent this propagation delay
in seconds. Slightly less than seconds after the beginning of the transmission of As frame, B decides to start
transmitting its own frame. After seconds, B senses As frame, detects the collision and stops transmitting. The
beginning of Bs frame travels on the cable until it reaches host A. Host A can thus detect the collision at time
+ 2 . An important point to note is that a collision can only occur during the first 2 seconds of
its transmission. If a collision did not occur during this period, it cannot occur afterwards since the transmission
channel is busy after seconds and CSMA/CD hosts sense the transmission channel before transmitting their
frame.
Furthermore, on the wired networks where CSMA/CD is used, collisions are almost the only cause of transmission
errors that affect frames. Transmission errors that only affect a few bits inside a frame seldom occur in these wired
networks. For this reason, the designers of CSMA/CD chose to completely remove the acknowledgement frames
in the datalink layer. When a host transmits a frame, it verifies whether its transmission has been affected by a
collision. If not, given the negligible Bit Error Ratio of the underlying network, it assumes that the frame was
received correctly by its destination. Otherwise the frame is retransmitted after some delay.
Removing acknowledgements is an interesting optimisation as it reduces the number of frames that are exchanged
on the network and the number of frames that need to be processed by the hosts. However, to use this optimisation,
we must ensure that all hosts will be able to detect all the collisions that affect their frames. The problem is
important for short frames. Let us consider two hosts, A and B, that are sending a small frame to host C as
illustrated in the figure below. If the frames sent by A and B are very short, the situation illustrated below may
occur. Hosts A and B send their frame and stop transmitting (point (1)). When the two short frames arrive at the
location of host C, they collide and host C cannot decode them (point (2)). The two frames are absorbed by the

Figure 2.71: The worst collision on a shared bus
ends of the wire. Neither host A nor host B have detected the collision. They both consider their frame to have
been received correctly by its destination.
Figure 2.72: The short-frame collision problem
To solve this problem, networks using CSMA/CD require hosts to transmit for at least 2 seconds. Since
the network transmission speed is fixed for a given network technology, this implies that a technology that uses
CSMA/CD enforces a minimum frame size. In the most popular CSMA/CD technology, Ethernet, 2 is called
the slot time 17 .
The last innovation introduced by CSMA/CD is the computation of the retransmission timeout. As for ALOHA,
this timeout cannot be fixed, otherwise hosts could become synchronised and always retransmit at the same time.
Setting such a timeout is always a compromise between the network access delay and the amount of collisions. A
short timeout would lead to a low network access delay but with a higher risk of collisions. On the other hand,
a long timeout would cause a long network access delay but a lower risk of collisions. The binary exponential
back-off algorithm was introduced in CSMA/CD networks to solve this problem.
To understand binary exponential back-off, let us consider a collision caused by exactly two hosts. Once it has
detected the collision, a host can either retransmit its frame immediately or defer its transmission for some time.
If each colliding host flips a coin to decide whether to retransmit immediately or to defer its retransmission, four
cases are possible :
17 This name should not be confused with the duration of a transmission slot in slotted ALOHA. In CSMA/CD networks, the slot time is
the time during which a collision can occur at the beginning of the transmission of a frame. In slotted ALOHA, the duration of a slot is the
transmission time of an entire fixed-size frame.

1. Both hosts retransmit immediately and a new collision occurs

2. The first host retransmits immediately and the second defers its retransmission
3. The second host retransmits immediately and the first defers its retransmission
4. Both hosts defer their retransmission and a new collision occurs
In the second and third cases, both hosts have flipped different coins. The delay chosen by the host that defers
its retransmission should be long enough to ensure that its retransmission will not collide with the immediate
retransmission of the other host. However the delay should not be longer than the time necessary to avoid the
collision, because if both hosts decide to defer their transmission, the network will be idle during this delay. The
slot time is the optimal delay since it is the shortest delay that ensures that the first host will be able to retransmit
its frame completely without any collision.
If two hosts are competing, the algorithm above will avoid a second collision 50% of the time. However, if the
network is heavily loaded, several hosts may be competing at the same time. In this case, the hosts should be able
to automatically adapt their retransmission delay. The binary exponential back-off performs this adaptation based
on the number of collisions that have affected a frame. After the first collision, the host flips a coin and waits
0 or 1 slot time. After the second collision, it generates a random number and waits 0, 1, 2 or 3 slot times, etc.
The duration of the waiting time is doubled after each collision. The complete pseudo-code for the CSMA/CD
algorithm is shown in the figure below.
# CSMA/CD pseudo-code
N=1
while N<= max :
wait(channel_becomes_free)
send(frame)
wait_until (end_of_frame) or (collision)
if collision detected:
stop transmitting
send(jamming)
k = min (10, N)
r = random(0, 2**k - 1)
wait(r*slotTime)
N=N+1
else :
wait(inter-frame_delay)
break
# end of while loop
The inter-frame delay used in this pseudo-code is a short delay corresponding to the time required by a network
adapter to switch from transmit to receive mode. It is also used to prevent a host from sending a continuous
stream of frames without leaving any transmission opportunities for other hosts on the network. This contributes
to the fairness of CSMA/CD. Despite this delay, there are still conditions where CSMA/CD is not completely fair
[RY1994]. Consider for example a network with two hosts : a server sending long frames and a client sending
acknowledgments. Measurements reported in [RY1994] have shown that there are situations where the client
could suffer from repeated collisions that lead it to wait for long periods of time due to the exponential back-off
algorithm.
Carrier Sense Multiple Access with Collision Avoidance
The Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) Medium Access Control algorithm was
designed for the popular WiFi wireless network technology [IEEE802.11]. CSMA/CA also senses the transmis-
sion channel before transmitting a frame. Furthermore, CSMA/CA tries to avoid collisions by carefully tuning the
timers used by CSMA/CA devices.
CSMA/CA uses acknowledgements like CSMA. Each frame contains a sequence number and a CRC. The CRC
is used to detect transmission errors while the sequence number is used to avoid frame duplication. When a
device receives a correct frame, it returns a special acknowledgement frame to the sender. CSMA/CA introduces
a small delay, named Short Inter Frame Spacing (SIFS), between the reception of a frame and the transmission of

the acknowledgement frame. This delay corresponds to the time that is required to switch the radio of a device
between the reception and transmission modes.
Compared to CSMA, CSMA/CA defines more precisely when a device is allowed to send a frame. First,
CSMA/CA defines two delays : DIFS and EIFS. To send a frame, a device must first wait until the channel
has been idle for at least the Distributed Coordination Function Inter Frame Space (DIFS) if the previous frame
was received correctly. However, if the previously received frame was corrupted, this indicates that there are
collisions and the device must sense the channel idle for at least the Extended Inter Frame Space (EIFS), with
SIF S < DIF S < EIF S. The exact values for SIFS, DIFS and EIFS depend on the underlying physical layer
[IEEE802.11].
The figure below shows the basic operation of CSMA/CA devices. Before transmitting, host A verifies that the
channel is empty for a long enough period. Then, its sends its data frame. After checking the validity of the
received frame, the recipient sends an acknowledgement frame after a short SIFS delay. Host C, which does not
participate in the frame exchange, senses the channel to be busy at the beginning of the data frame. Host C can
use this information to determine how long the channel will be busy for. Note that as SIF S < DIF S < EIF S,
even a device that would start to sense the channel immediately after the last bit of the data frame could not decide
to transmit its own frame during the transmission of the acknowledgement frame.
Figure 2.73: Operation of a CSMA/CA device
The main difficulty with CSMA/CA is when two or more devices transmit at the same time and cause collisions.
This is illustrated in the figure below, assuming a fixed timeout after the transmission of a data frame. With
CSMA/CA, the timeout after the transmission of a data frame is very small, since it corresponds to the SIFS plus
the time required to transmit the acknowledgement frame.
To deal with this problem, CSMA/CA relies on a backoff timer. This backoff timer is a random delay that is
chosen by each device in a range that depends on the number of retransmissions for the current frame. The
range grows exponentially with the retransmissions as in CSMA/CD. The minimum range for the backoff timer
is [0, 7 slotT ime] where the slotTime is a parameter that depends on the underlying physical layer. Compared
to CSMA/CDs exponential backoff, there are two important differences to notice. First, the initial range for
the backoff timer is seven times larger. This is because it is impossible in CSMA/CA to detect collisions as
they happen. With CSMA/CA, a collision may affect the entire frame while with CSMA/CD it can only affect
the beginning of the frame. Second, a CSMA/CA device must regularly sense the transmission channel during
its back off timer. If the channel becomes busy (i.e. because another device is transmitting), then the back off
timer must be frozen until the channel becomes free again. Once the channel becomes free, the back off timer
is restarted. This is in contrast with CSMA/CD where the back off is recomputed after each collision. This is
illustrated in the figure below. Host A chooses a smaller backoff than host C. When C senses the channel to be
busy, it freezes its backoff timer and only restarts it once the channel is free again.

Figure 2.74: Collisions with CSMA/CA
Figure 2.75: Detailed example with CSMA/CA

The pseudo-code below summarizes the operation of a CSMA/CA device. The values of the SIFS, DIFS, EIFS
and slotTime depend on the underlying physical layer technology [IEEE802.11]
# CSMA/CA simplified pseudo-code
N=1
while N<= max :
waitUntil(free(channel))
if correct(last_frame) :
wait(channel_free_during_t >=DIFS)
else:
wait(channel_free_during_t >=EIFS)
back-off_time = int(random[0,min(255,7*(2^(N-1)))])*slotTime
wait(channel free during backoff_time)
# backoff timer is frozen while channel is sensed to be busy
send(frame)
if received(ack)
# frame received correctly
break
else:
# retransmission required
N=N+1
# end of while loop
Another problem faced by wireless networks is often called the hidden station problem. In a wireless network,
radio signals are not always propagated same way in all directions. For example, two devices separated by a wall
may not be able to receive each others signal while they could both be receiving the signal produced by a third
host. This is illustrated in the figure below, but it can happen in other environments. For example, two devices that
are on different sides of a hill may not be able to receive each others signal while they are both able to receive the
signal sent by a station at the top of the hill. Furthermore, the radio propagation conditions may change with time.
For example, a truck may temporarily block the communication between two nearby devices.
Figure 2.76: The hidden station problem
To avoid collisions in these situations, CSMA/CA allows devices to reserve the transmission channel for some
time. This is done by using two control frames : Request To Send (RTS) and Clear To Send (CTS). Both are very
short frames to minimize the risk of collisions. To reserve the transmission channel, a device sends a RTS frame
to the intended recipient of the data frame. The RTS frame contains the duration of the requested reservation. The
recipient replies, after a SIFS delay, with a CTS frame which also contains the duration of the reservation. As the
duration of the reservation has been sent in both RTS and CTS, all hosts that could collide with either the sender
or the reception of the data frame are informed of the reservation. They can compute the total duration of the
transmission and defer their access to the transmission channel until then. This is illustrated in the figure below
where host A reserves the transmission channel to send a data frame to host B. Host C notices the reservation and
defers its transmission.
The utilization of the reservations with CSMA/CA is an optimisation that is useful when collisions are frequent.

Figure 2.77: Reservations with CSMA/CA
If there are few collisions, the time required to transmit the RTS and CTS frames can become significant and in
particular when short frames are exchanged. Some devices only turn on RTS/CTS after transmission errors.
Deterministic Medium Access Control algorithms
During the 1970s and 1980s, there were huge debates in the networking community about the best suited Medium
Access Control algorithms for Local Area Networks. The optimistic algorithms that we have described until
now were relatively easy to implement when they were designed. From a performance perspective, mathematical
models and simulations showed the ability of these optimistic techniques to sustain load. However, none of the
optimistic techniques are able to guarantee that a frame will be delivered within a given delay bound and some
applications require predictable transmission delays. The deterministic MAC algorithms were considered by a
fraction of the networking community as the best solution to fulfill the needs of Local Area Networks.
Both the proponents of the deterministic and the opportunistic techniques lobbied to develop standards for Local
Area networks that would incorporate their solution. Instead of trying to find an impossible compromise between
these diverging views, the IEEE 802 committee that was chartered to develop Local Area Network standards
chose to work in parallel on three different LAN technologies and created three working groups. The IEEE
802.3 working group became responsible for CSMA/CD. The proponents of deterministic MAC algorithms agreed
on the basic principle of exchanging special frames called tokens between devices to regulate the access to the
transmission medium. However, they did not agree on the most suitable physical layout for the network. IBM
argued in favor of Ring-shaped networks while the manufacturing industry, led by General Motors, argued in
favor of a bus-shaped network. This led to the creation of the IEEE 802.4 working group to standardise Token Bus
networks and the IEEE 802.5 working group to standardise Token Ring networks. Although these techniques are
not widely used anymore today, the principles behind a token-based protocol are still important.
The IEEE 802.5 Token Ring technology is defined in [IEEE802.5]. We use Token Ring as an example to explain
the principles of the token-based MAC algorithms in ring-shaped networks. Other ring-shaped networks include
the almost defunct FDDI [Ross1989] or the more recent Resilient Pack Ring [DYGU2004] . A good survey of the
token ring networks may be found in [Bux1989] .
A Token Ring network is composed of a set of stations that are attached to a unidirectional ring. The basic principle
of the Token Ring MAC algorithm is that two types of frames travel on the ring : tokens and data frames. When the
Token Ring starts, one of the stations sends the token. The token is a small frame that represents the authorization
to transmit data frames on the ring. To transmit a data frame on the ring, a station must first capture the token by
removing it from the ring. As only one station can capture the token at a time, the station that owns the token can
safely transmit a data frame on the ring without risking collisions. After having transmitted its frame, the station
must remove it from the ring and resend the token so that other stations can transmit their own frames.
While the basic principles of the Token Ring are simple, there are several subtle implementation details that add
complexity to Token Ring networks. To understand these details let us analyse the operation of a Token Ring

Figure 2.78: A Token Ring network
interface on a station. A Token Ring interface serves three different purposes. Like other LAN interfaces, it must
be able to send and receive frames. In addition, a Token Ring interface is part of the ring, and as such, it must be
able to forward the electrical signal that passes on the ring even when its station is powered off.
When powered-on, Token Ring interfaces operate in two different modes : listen and transmit. When operating
in listen mode, a Token Ring interface receives an electrical signal from its upstream neighbour on the ring,
introduces a delay equal to the transmission time of one bit on the ring and regenerates the signal before sending
it to its downstream neighbour on the ring.
The first problem faced by a Token Ring network is that as the token represents the authorization to transmit, it
must continuously travel on the ring when no data frame is being transmitted. Let us assume that a token has been
produced and sent on the ring by one station. In Token Ring networks, the token is a 24 bits frame whose structure
is shown below.
Figure 2.79: 802.5 token format
The token is composed of three fields. First, the Starting Delimiter is the marker that indicates the beginning of a
frame. The first Token Ring networks used Manchester coding and the Starting Delimiter contained both symbols
representing 0 and symbols that do not represent bits. The last field is the Ending Delimiter which marks the end
of the token. The Access Control field is present in all frames, and contains several flags. The most important is
the Token bit that is set in token frames and reset in other frames.
Let us consider the five station network depicted in figure A Token Ring network above and assume that station S1
sends a token. If we neglect the propagation delay on the inter-station links, as each station introduces a one bit
delay, the first bit of the frame would return to S1 while it sends the fifth bit of the token. If station S1 is powered
off at that time, only the first five bits of the token will travel on the ring. To avoid this problem, there is a special
station called the Monitor on each Token Ring. To ensure that the token can travel forever on the ring, this Monitor
inserts a delay that is equal to at least 24 bit transmission times. If station S3 was the Monitor in figure A Token
Ring network, S1 would have been able to transmit the entire token before receiving the first bit of the token from
its upstream neighbor.
Now that we have explained how the token can be forwarded on the ring, let us analyse how a station can capture
a token to transmit a data frame. For this, we need some information about the format of the data frames. An
802.5 data frame begins with the Starting Delimiter followed by the Access Control field whose Token bit is reset,

a Frame Control field that allows for the definition of several types of frames, destination and source address, a
payload, a CRC, the Ending Delimiter and a Frame Status field. The format of the Token Ring data frames is
illustrated below.
Figure 2.80: 802.5 data frame format
To capture a token, a station must operate in Listen mode. In this mode, the station receives bits from its upstream
neighbour. If the bits correspond to a data frame, they must be forwarded to the downstream neighbour. If they
correspond to a token, the station can capture it and transmit its data frame. Both the data frame and the token
are encoded as a bit string beginning with the Starting Delimiter followed by the Access Control field. When the
station receives the first bit of a Starting Delimiter, it cannot know whether this is a data frame or a token and
must forward the entire delimiter to its downstream neighbour. It is only when it receives the fourth bit of the
Access Control field (i.e. the Token bit) that the station knows whether the frame is a data frame or a token. If
the Token bit is reset, it indicates a data frame and the remaining bits of the data frame must be forwarded to the
downstream station. Otherwise (Token bit is set), this is a token and the station can capture it by resetting the
bit that is currently in its buffer. Thanks to this modification, the beginning of the token is now the beginning of
a data frame and the station can switch to Transmit mode and send its data frame starting at the fifth bit of the
Access Control field. Thus, the one-bit delay introduced by each Token Ring station plays a key role in enabling
the stations to efficiently capture the token.
After having transmitted its data frame, the station must remain in Transmit mode until it has received the last bit
of its own data frame. This ensures that the bits sent by a station do not remain in the network forever. A data
frame sent by a station in a Token Ring network passes in front of all stations attached to the network. Each station
can detect the data frame and analyse the destination address to possibly capture the frame.
The text above describes the basic operation of a Token Ring network when all stations work correctly. Unfor-
tunately, a real Token Ring network must be able to handle various types of anomalies and this increases the
complexity of Token Ring stations. We briefly list the problems and outline their solutions below. A detailed
description of the operation of Token Ring stations may be found in [IEEE802.5]. The first problem is when all
the stations attached to the network start. One of them must bootstrap the network by sending the first token.
For this, all stations implement a distributed election mechanism that is used to select the Monitor. Any station
can become a Monitor. The Monitor manages the Token Ring network and ensures that it operates correctly. Its
first role is to introduce a delay of 24 bit transmission times to ensure that the token can travel smoothly on the
ring. Second, the Monitor sends the first token on the ring. It must also verify that the token passes regularly.
According to the Token Ring standard [IEEE802.5], a station cannot retain the token to transmit data frames for a
duration longer than the Token Holding Time (THT) (slightly less than 10 milliseconds). On a network containing
N stations, the Monitor must receive the token at least every N T HT seconds. If the Monitor does not receive
a token during such a period, it cuts the ring for some time and then reinitialises the ring and sends a token.
Several other anomalies may occur in a Token Ring network. For example, a station could capture a token and be
powered off before having resent the token. Another station could have captured the token, sent its data frame and
be powered off before receiving all of its data frame. In this case, the bit string corresponding to the end of a frame
would remain in the ring without being removed by its sender. Several techniques are defined in [IEEE802.5] to
allow the Monitor to handle all these problems. If unfortunately, the Monitor fails, another station will be elected
to become the new Monitor.

2.6.5 Congestion control
Most networks contain links having different bandwidth. Some hosts can use low bandwidth wireless networks.
Some servers are attached via 10 Gbps interfaces and inter-router links may vary from a few tens of kilobits per
second up to hundred Gbps. Despite these huge differences in performance, any host should be able to efficiently
exchange segments with a high-end server.
To understand this problem better, let us consider the scenario shown in the figure below, where a server (A)
attached to a 10 Mbps link needs to reliably transfer segments to another computer (C) through a path that contains
a 2 Mbps link.
Figure 2.81: Reliable transport with heterogeneous links
In this network, the segments sent by the server reach router R1. R1 forwards the segments towards router R2.
Router R1 can potentially receive segments at 10 Mbps, but it can only forward them at 2 Mbps to router R2 and
then to host C. Router R1 includes buffers that allow it to store the packets that cannot immediately be forwarded
to their destination. To understand the operation of a reliable transport protocol in this environment, let us consider
a simplified model of this network where host A is attached to a 10 Mbps link to a queue that represents the buffers
of router R1. This queue is emptied at a rate of 2 Mbps.
Figure 2.82: Self clocking
Let us consider that host A uses a window of three segments. It thus sends three back-to-back segments at 10
Mbps and then waits for an acknowledgement. Host A stops sending segments when its window is full. These
segments reach the buffers of router R2. The first segment stored in this buffer is sent by router R2 at a rate of 2
Mbps to the destination host. Upon reception of this segment, the destination sends an acknowledgement. This
acknowledgement allows host A to transmit a new segment. This segment is stored in the buffers of router R2 while
it is transmitting the second segment that was sent by host A... Thus, after the transmission of the first window
of segments, the reliable transport protocol sends one data segment after the reception of each acknowledgement
returned by the destination. In practice, the acknowledgements sent by the destination serve as a kind of clock
that allows the sending host to adapt its transmission rate to the rate at which segments are received by the

destination. This self-clocking is the first mechanism that allows a window-based reliable transport protocol to
adapt to heterogeneous networks [Jacobson1988]. It depends on the availability of buffers to store the segments
that have been sent by the sender but have not yet been transmitted to the destination.
However, transport protocols are not only used in this environment. In the global Internet, a large number of hosts
send segments to a large number of receivers. For example, let us consider the network depicted below which is
similar to the one discussed in [Jacobson1988] and RFC 896. In this network, we assume that the buffers of the
router are infinite to ensure that no packet is lost.
Figure 2.83: The congestion collapse problem
If many senders are attached to the left part of the network above, they all send a window full of segments. These
segments are stored in the buffers of the router before being transmitted towards their destination. If there are
many senders on the left part of the network, the occupancy of the buffers quickly grows. A consequence of
the buffer occupancy is that the round-trip-time, measured by the transport protocol, between the sender and the
receiver increases. Consider a network where 10,000 bits segments are sent. When the buffer is empty, such a
segment requires 1 millisecond to be transmitted on the 10 Mbps link and 5 milliseconds to be the transmitted
on the 2 Mbps link. Thus, the measured round-trip-time measured is roughly 6 milliseconds if we ignore the
propagation delay on the links. If the buffer contains 100 segments, the round-trip-time becomes 1 + 100 5 + 5
milliseconds as new segments are only transmitted on the 2 Mbps link once all previous segments have been
transmitted. Unfortunately, if the reliable transport protocol uses a retransmission timer and performs go-back-n
to recover from transmission errors it will retransmit a full window of segments. This increases the occupancy of
the buffer and the delay through the buffer... Furthermore, the buffer may store and send on the low bandwidth
links several retransmissions of the same segment. This problem is called congestion collapse. It occurred several
times during the late 1980s on the Internet [Jacobson1988].
The congestion collapse is a problem that all heterogeneous networks face. Different mechanisms have been
proposed in the scientific literature to avoid or control network congestion. Some of them have been implemented
and deployed in real networks. To understand this problem in more detail, let us first consider a simple network
with two hosts attached to a high bandwidth link that are sending segments to destination C attached to a low
bandwidth link as depicted below.
Figure 2.84: The congestion problem
To avoid congestion collapse, the hosts must regulate their transmission rate 18 by using a congestion control
mechanism. Such a mechanism can be implemented in the transport layer or in the network layer. In TCP/IP
networks, it is implemented in the transport layer, but other technologies such as Asynchronous Transfer Mode
(ATM) or Frame Relay include congestion control mechanisms in lower layers.
18 In this section, we focus on congestion control mechanisms that regulate the transmission rate of the hosts. Other types of mecha-
nisms have been proposed in the literature. For example, credit-based flow-control has been proposed to avoid congestion in ATM networks
[KR1995]. With a credit-based mechanism, hosts can only send packets once they have received credits from the routers and the credits depend
on the occupancy of the routers buffers.

Let us first consider the simple problem of a set of i hosts that share a single bottleneck link as shown in the
example above. In this network, the congestion control scheme must achieve the following objectives [CJ1989] :
1. The congestion control scheme must avoid congestion. In practice, this means that the bottle-
neck link cannot be overloaded. If ri (t) is the transmission rate allocated to host i at time t and
R the bandwidth
P of the bottleneck link, then the congestion control scheme should ensure that,
on average, t ri (t) R.
2. The congestion control scheme must be efficient. The bottleneck link is usually both a shared
and an expensive resource. Usually, bottleneck links are wide area links that are much more
expensive to upgrade than the local area networks. The congestion control scheme should ensure
thatPsuch links are efficiently used. Mathematically, the control scheme should ensure that
t ri (t) R.
3. The congestion control scheme should be fair. Most congestion schemes aim at achieving max-
min fairness. An allocation of transmission rates to sources is said to be max-min fair if :
no link in the network is congested
the rate allocated to source j cannot be increased without decreasing the rate allocated to a
source i whose allocation is smaller than the rate allocated to source j [Leboudec2008] .
Depending on the network, a max-min fair allocation may not always exist. In practice, max-min fairness is an
ideal objective that cannot necessarily be achieved. When there is a single bottleneck link as in the example above,
max-min fairness implies that each source should be allocated the same transmission rate.
To visualise the different rate allocations, it is useful to consider the graph shown below. In this graph, we plot
on the x-axis (resp. y-axis) the rate allocated to host B (resp. A). A point in the graph (rB , rA ) corresponds to a
possible allocation of the transmission rates. Since there is a 2 Mbps bottleneck link in this network, the graph
can be divided into two regions. The lower left part of the graph contains all allocations (rB , rA ) such that the
bottleneck link is not congested (rA + rB < 2). The right border of this region is the efficiency line, i.e. the set
of allocations that completely utilise the bottleneck link (rA + rB = 2). Finally, the fairness line is the set of fair
allocations.
Figure 2.85: Possible allocated transmission rates
As shown in the graph above, a rate allocation may be fair but not efficient (e.g. rA = 0.7, rB = 0.7), fair and
efficient ( e.g. rA = 1, rB = 1) or efficient but not fair (e.g. rA = 1.5, rB = 0.5). Ideally, the allocation should be
both fair and efficient. Unfortunately, maintaining such an allocation with fluctuations in the number of flows that
use the network is a challenging problem. Furthermore, there might be several thousands flows that pass through
the same link 19 .
To deal with these fluctuations in demand, which result in fluctuations in the available bandwidth, computer
networks use a congestion control scheme. This congestion control scheme should achieve the three objectives
19 For example, the measurements performed in the Sprint network in 2004 reported more than 10k active TCP connections on a link, see
https://research.sprintlabs.com/packstat/packetoverview.php. More recent information about backbone links may be obtained from caida s
realtime measurements, see e.g. http://www.caida.org/data/realtime/passive/

listed above. Some congestion control schemes rely on a close cooperation between the endhosts and the routers,
while others are mainly implemented on the endhosts with limited support from the routers.
A congestion control scheme can be modelled as an algorithm that adapts the transmission rate (ri (t)) of host i
based on the feedback received from the network. Different types of feedbacks are possible. The simplest scheme
is a binary feedback [CJ1989] [Jacobson1988] where the hosts simply learn whether the network is congested or
not. Some congestion control schemes allow the network to regularly send an allocated transmission rate in Mbps
to each host [BF1995].
Let us focus on the binary feedback scheme which is the most widely used today. Intuitively, the congestion
control scheme should decrease the transmission rate of a host when congestion has been detected in the network,
in order to avoid congestion collapse. Furthermore, the hosts should increase their transmission rate when the
network is not congested. Otherwise, the hosts would not be able to efficiently utilise the network. The rate
allocated to each host fluctuates with time, depending on the feedback received from the network. The figure
below illustrates the evolution of the transmission rates allocated to two hosts in our simple network. Initially, two
hosts have a low allocation, but this is not efficient. The allocations increase until the network becomes congested.
At this point, the hosts decrease their transmission rate to avoid congestion collapse. If the congestion control
scheme works well, after some time the allocations should become both fair and efficient.
Figure 2.86: Evolution of the transmission rates
Various types of rate adaption algorithms are possible. Dah Ming Chiu and Raj Jain have analysed, in [CJ1989],
different types of algorithms that can be used by a source to adapt its transmission rate to the feedback received
from the network. Intuitively, such a rate adaptation algorithm increases the transmission rate when the network
is not congested (ensure that the network is efficiently used) and decrease the transmission rate when the network
is congested (to avoid congestion collapse).
The simplest form of feedback that the network can send to a source is a binary feedback (the network is congested
or not congested). In this case, a linear rate adaptation algorithm can be expressed as :
rate(t + 1) = C + C rate(t) when the network is congested
rate(t + 1) = N + N rate(t) when the network is not congested
With a linear adaption algorithm, C , N , C and N are constants. The analysis of [CJ1989] shows that to
be fair and efficient, such a binary rate adaption mechanism must rely on Additive Increase and Multiplicative
Decrease. When the network is not congested, the hosts should slowly increase their transmission rate (N =
1 and N > 0). When the network is congested, the hosts must multiplicatively decrease their transmission rate
(C < 1 and C = 0). Such an AIMD rate adaptation algorithm can be implemented by the pseudo-code below.
# Additive Increase Multiplicative Decrease
if congestion :
rate=rate*betaC # multiplicative decrease, betaC<1
else
rate=rate+alphaN # additive increase, v0>0
Note: Which binary feedback ?

Two types of binary feedback are possible in computer networks. A first solution is to rely on implicit feedback.
This is the solution chosen for TCP. TCPs congestion control scheme [Jacobson1988] does not require any coop-
eration from the router. It only assumes that they use buffers and that they discard packets when there is congestion.
TCP uses the segment losses as an indication of congestion. When there are no losses, the network is assumed to
be not congested. This implies that congestion is the main cause of packet losses. This is true in wired networks,
but unfortunately not always true in wireless networks. Another solution is to rely on explicit feedback. This
is the solution proposed in the DECBit congestion control scheme [RJ1995] and used in Frame Relay and ATM
networks. This explicit feedback can be implemented in two ways. A first solution would be to define a special
message that could be sent by routers to hosts when they are congested. Unfortunately, generating such messages
may increase the amount of congestion in the network. Such a congestion indication packet is thus discouraged
RFC 1812. A better approach is to allow the intermediate routers to indicate, in the packets that they forward,
their current congestion status. Binary feedback can be encoded by using one bit in the packet header. With such a
scheme, congested routers set a special bit in the packets that they forward while non-congested routers leave this
bit unmodified. The destination host returns the congestion status of the network in the acknowledgements that it
sends. Details about such a solution in IP networks may be found in RFC 3168. Unfortunately, as of this writing,
this solution is still not deployed despite its potential benefits.
Congestion control in a window-based transport protocol
AIMD controls congestion by adjusting the transmission rate of the sources in reaction to the current congestion
level. If the network is not congested, the transmission rate increases. If congestion is detected, the transmission
rate is multiplicatively decreased. In practice, directly adjusting the transmission rate can be difficult since it
requires the utilisation of fine grained timers. In reliable transport protocols, an alternative is to dynamically
adjust the sending window. This is the solution chosen for protocols like TCP and SCTP that will be described in
more details later. To understand how window-based protocols can adjust their transmission rate, let us consider
the very simple scenario of a reliable transport protocol that uses go-back-n. Consider the very simple scenario
shown in the figure below.

A B
R1
500 kbps
R2
The links between the hosts and the routers have a bandwidth of 1 Mbps while the link between the two routers
has a bandwidth of 500 Kbps. There is no significant propagation delay in this network. For simplicity, assume
that hosts A and B send 1000 bits packets. The transmission of such a packet on a host-router (resp. router-router
) link requires 1 msec (resp. 2 msec). If there is no traffic in the network, round-trip-time measured by host A
is slightly larger than 4 msec. Let us observe the flow of packets with different window sizes to understand the
relationship between sending window and transmission rate.
Consider first a window of one segment. This segment takes 4 msec to reach host D. The destination replies with
an acknowledgement and the next segment can be transmitted. With such a sending window, the transmission rate
is roughly 250 segments per second of 250 Kbps.

+-----+----------+----------+----------+
|Time | A-R1 | R1-R2 | R2-D |
+=====+==========+==========+==========+
|t0 | data(0) | | |
+-----+----------+----------+ |
|t0+1 | | | |
+-----+ | data(0) | |
|t0+2 | | | |
+-----+ +----------+----------+
|t0+3 | | | data(0) |
+-----+----------+ +----------+
|t0+4 | data(1) | | |
+-----+----------+----------+ |
|t0+5 | | | |
+-----+ | data(1) | |
|t0+6 | | | |
+-----+ +----------+----------+
|t0+7 | | | data(1) |
+-----+----------+ +----------+
|t0+8 | data(2) | |
+-----+----------+----------------------
Consider now a window of two segments. Host A can send two segments within 2 msec on its 1 Mbps link. If the
first segment is sent at time t0 , it reaches host D at t0 + 4. Host D replies with an acknowledgement that opens the
sending window on host A and enables it to transmit a new segment. In the meantime, the second segment was
buffered by router R1. It reaches host D at t0 + 6 and an acknowledgement is returned. With a window of two
segments, host A transmits at roughly 500 Kbps, i.e. the transmission rate of the bottleneck link.
+-----+----------+----------+----------+
|Time | A-R1 | R1-R2 | R2-D |
+=====+==========+==========+==========+
|t0 | data(0) | | |
+-----+----------+----------+ |
|t0+1 | data(1) | | |
+-----+----------+ data(0) | |
|t0+2 | | | |
+-----+ +----------+----------+
|t0+3 | | | data(0) |
+-----+----------+ data(1) +----------+
|t0+4 | data(2) | | |
+-----+----------+----------+----------+
|t0+5 | | | data(1) |
+-----+----------+ data(2) +----------+
|t0+6 | data(3) | | |
+-----+----------+----------+----------+
Our last example is a window of four segments. These segments are sent at t0 , t0 + 1, t0 + 2 and t0 + 3. The first
segment reaches host D at t0 + 4. Host D replies to this segment by sending an acknowledgement that enables host
A to transmit its fifth segment. This segment reaches router R1 at t0 + 5. At that time, router R1 is transmitting
the third segment to router R2 and the fourth segment is still in its buffers. At time t0 + 6, host D receives the
second segment and returns the corresponding acknowledgement. This acknowledgement enables host A to send
its sixth segment. This segment reaches router R1 at roughly t0 + 7. At that time, the router starts to transmit the
fourth segment to router R2. Since link R1-R2 can only sustain 500 Kbps, packets will accumulate in the buffers
of R1. On average, there will be two packets waiting in the buffers of R1. The presence of these two packets
will induce an increase of the round-trip-time as measured by the transport protocol. While the first segment was
acknowledged within 4 msec, the fifth segment (data(4)) that was transmitted at time t0 + 4 is only acknowledged
at time t0 + 11. On average, the sender transmits at 500 Kbps, but the utilisation of a large window induces a
longer delay through the network.
+-----+----------+----------+----------+
|Time | A-R1 | R1-R2 | R2-D |
+=====+==========+==========+==========+
|t0 | data(0) | | |

+-----+----------+----------+ |
|t0+1 | data(1) | | |
+-----+----------+ data(0) | |
|t0+2 | data(2) | | |
+-----+----------+----------+----------+
|t0+3 | data(3) | | data(0) |
+-----+----------+ data(1) +----------+
|t0+4 | data(4) | | |
+-----+----------+----------+----------+
|t0+5 | | | data(1) |
+-----+----------+ data(2) +----------+
|t0+6 | data(5) | | |
+-----+----------+----------+----------+
|t0+7 | | | data(2) |
+-----+----------+ data(3) +----------+
|t0+8 | data(6) | | |
+-----+----------+----------+----------+
|t0+9 | | | data(3) |
+-----+----------+ data(4) +----------+
|t0+10| data(7) | | |
+-----+----------+----------+----------+
|t0+11| | | data(4) |
+-----+----------+ data(5) +----------+
|t0+12| data(8) | | |
+-----+----------+----------+----------+
From the above example, we can adjust the transmission rate by adjusting the sending window of a reliable
transport protocol. A reliable transport protocol cannot send data faster than window
rtt where window is current
sending window. To control the transmission rate, we introduce a congestion window. This congestion window
limits the sending window. A any time, the sending window is restricted to min(swin, cwin), where swin is the
sending window and cwin the current congestion window. Of course, the window is further constrained by the
receive window advertised by the remote peer. With the utilization of a congestion window, a simple reliable
transport protocol that uses fixed size segments could implement AIMD as follows.
For the Additive Increase part our simple protocol would simply increase its congestion window by one segment
every round-trip-time. The Multiplicative Decrease part of AIMD could be implemented by halving the congestion
window when congestion is detected. For simplicity, we assume that congestion is detected thanks to a binary
feedback and that no segments are lost. We will discuss in more details how losses affect a real transport protocol
like TCP.
A congestion control scheme for our simple transport protocol could be implemented as follows.
# Initialisation
cwin = 1 # congestion window measured in segments
# Ack arrival
if newack : # new ack, no congestion
# increase cwin by one every rtt
cwin = cwin+ (1/cwin)
else:
# no increase
Congestion detected:
cwnd=cwin/2 # only once per rtt
In the above pseudocode, cwin contains the congestion window stored as a real in segments. This congestion
window is updated upon the arrival of each acknowledgment and when congestion is detected. For simplicity, we
assume that cwin is stored as a floating point number but only full segments can be transmitted.
As an illustration, let us consider the network scenario above and assume that the router implements the DECBit
binary feedback scheme [RJ1995]. This scheme uses a form of Forward Explicit Congestion Notification and a
router marks the congestion bit in arriving packets when its buffer contains one or more packets. In the figure
below, we use a * to indicate a marked packet.

+-----+----------+----------+----------+
|Time | A-R1 | R1-R2 | R2-D |
+-----+==========+==========+==========+
|t0 | data(0) | | |
+-----+----------+----------+ |
|t0+1 | | | |
+-----+ | data(0) | |
|t0+2 | | | |
+-----+ +----------+----------+
|t0+3 | | | data(0) |
+-----+----------+ +----------+
|t0+4 | data(1) | | |
+-----+----------+----------+ |
|t0+5 | data(2) | | |
+-----+----------+ data(1) | |
|t0+6 | | | |
+-----+ +----------+----------+
|t0+7 | | | data(1) |
+-----+----------+ data(2) +----------+
|t0+8 | data(3) | | |
+-----+----------+----------+----------+
|t0+9 | | | data(2) |
+-----+----------+ data(3) +----------+
|t0+10| data(4) | | |
+-----+----------+----------+----------+
|t0+11| data(5) | | data(3) |
+-----+----------+ data(4) +----------+
|t0+12| data(6) | | |
+-----+----------+----------+----------+
|t0+13| | | data(4) |
+-----+----------+ data(5) +----------+
|t0+14| data(7) | | |
+-----+----------+----------+----------+
|t0+15| | | data(5) |
+-----+----------+ data*(6) +----------+
|t0+16| data(8) | | |
+-----+----------+----------+----------+
|t0+17| data(9) | | data*(6) |
+-----+----------+ data*(7) +----------+
|t0+18| | | |
+-----+ |----------+----------
|t0+19| | | data*(7) |
+-----+ | data*(8) +----------+
|t0+20| | | |
+-----+ |----------+----------+
|t0+21| | | data*(8) |
+-----+----------+ data*(9) +----------+
|t0+22| data(10) | | |
+-----+----------+----------+----------+
When the connection starts, its congestion window is set to one segment. Segment data(0) is sent at acknowl-
edgment at roughly t0 + 4. The congestion window is increased by one segment and data(1) and data(2) are
transmitted at time t0 + 4 and t0 + 5. The corresponding acknowledgements are received at times t0 + 8 and
t0 + 10. Upon reception of this last acknowledgement, the congestion window reaches 3 and segments can be sent
(data(4) and data(5)). When segment data(6) reaches router R1, its buffers already contain data(5). The packet
containing data(6) is thus marked to inform the sender of the congestion. Note that the sender will only notice
the congestion once it receives the corresponding acknowledgement at t0 + 18. In the meantime, the congestion
window continues to increase. At t0 + 16, upon reception of the acknowledgement for data(5), it reaches 4. When
congestion is detected, the congestion window is decreased down to 2. This explains the idle time between the
reception of the acknowledgement for data*(6) and the transmission of data(10).

2.7 The reference models
Given the growing complexity of computer networks, during the 1970s network researchers proposed various
reference models to facilitate the description of network protocols and services. Of these, the Open Systems
Interconnection (OSI) model [Zimmermann80] was probably the most influential. It served as the basis for the
standardisation work performed within the ISO to develop global computer network standards. The reference
model that we use in this book can be considered as a simplified version of the OSI reference model 20 .
2.7.1 The five layers reference model
Our reference model is divided into five layers, as shown in the figure below.
Figure 2.87: The five layers of the reference model
2.7.2 The Physical layer
Starting from the bottom, the first layer is the Physical layer. Two communicating devices are linked through a
physical medium. This physical medium is used to transfer an electrical or optical signal between two directly
connected devices.
An important point to note about the Physical layer is the service that it provides. This service is usually an
unreliable connection-oriented service that allows the users of the Physical layer to exchange bits. The unit of
information transfer in the Physical layer is the bit. The Physical layer service is unreliable because :
the Physical layer may change, e.g. due to electromagnetic interferences, the value of a bit being transmitted
the Physical layer may deliver more bits to the receiver than the bits sent by the sender
the Physical layer may deliver fewer bits to the receiver than the bits sent by the sender
Figure 2.88: The Physical layer
20 An interesting historical discussion of the OSI-TCP/IP debate may be found in [Russel06]
2.7. The reference models 105

2.7.3 The Datalink layer
The Datalink layer builds on the service provided by the underlying physical layer. The Datalink layer allows
two hosts that are directly connected through the physical layer to exchange information. The unit of information
exchanged between two entities in the Datalink layer is a frame. A frame is a finite sequence of bits. Some
Datalink layers use variable-length frames while others only use fixed-length frames. Some Datalink layers
provide a connection-oriented service while others provide a connectionless service. Some Datalink layers provide
reliable delivery while others do not guarantee the correct delivery of the information.
An important point to note about the Datalink layer is that although the figure below indicates that two entities
of the Datalink layer exchange frames directly, in reality this is slightly different. When the Datalink layer entity
on the left needs to transmit a frame, it issues as many Data.request primitives to the underlying physical layer
as there are bits in the frame. The physical layer will then convert the sequence of bits in an electromagnetic
or optical signal that will be sent over the physical medium. The physical layer on the right hand side of the
figure will decode the received signal, recover the bits and issue the corresponding Data.indication primitives to
its Datalink layer entity. If there are no transmission errors, this entity will receive the frame sent earlier.
Figure 2.89: The Datalink layer
2.7.4 The Network layer
The Datalink layer allows directly connected hosts to exchange information, but it is often necessary to exchange
information between hosts that are not attached to the same physical medium. This is the task of the network
layer. The network layer is built above the datalink layer. Network layer entities exchange packets. A packet is
a finite sequence of bytes that is transported by the datalink layer inside one or more frames. A packet usually
contains information about its origin and its destination, and usually passes through several intermediate devices
called routers on its way from its origin to its destination.
Figure 2.90: The network layer
2.7.5 The Transport layer
Most realisations of the network layer, including the internet, do not provide a reliable service. However, many
applications need to exchange information reliably and so using the network layer service directly would be
very difficult for them. Ensuring the reliable delivery of the data produced by applications is the task of the
transport layer. Transport layer entities exchange segments. A segment is a finite sequence of bytes that are
transported inside one or more packets. A transport layer entity issues segments (or sometimes part of segments)
as Data.request to the underlying network layer entity.
There are different types of transport layers. The most widely used transport layers on the Internet are TCP
,that provides a reliable connection-oriented bytestream transport service, and UDP ,that provides an unreliable
connection-less transport service.

2.7.6 The Application layer
The upper layer of our architecture is the Application layer. This layer includes all the mechanisms and data
structures that are necessary for the applications. We will use Application Data Unit (ADU) or the generic Service
Data Unit (SDU) term to indicate the data exchanged between two entities of the Application layer.
Figure 2.92: The Application layer
In the remaining chapters of this text, we will often refer to the information exchanged between entities located in
different layers. To avoid any confusion, we will stick to the terminology defined earlier, i.e. :
physical layer entities exchange bits
datalink layer entities exchange frames
network layer entities exchange packets
transport layer entities exchange segments
application layer entities exchange SDUs
2.7.7 Reference models
Two reference models have been successful in the networking community : the OSI reference model and the
TCP/IP reference model. We discuss them briefly in this section.
The TCP/IP reference model
In contrast with OSI, the TCP/IP community did not spend a lot of effort defining a detailed reference model; in
fact, the goals of the Internet architecture were only documented after TCP/IP had been deployed [Clark88]. RFC
1122 , which defines the requirements for Internet hosts, mentions four different layers. Starting from the top,
these are :
the Application layer
the Transport layer
the Internet layer which is equivalent to the network layer of our reference model
the Link layer which combines the functionalities of the physical and datalink layers of our five-layer refer-
ence model
Besides this difference in the lower layers, the TCP/IP reference model is very close to the five layers that we use
throughout this document.
The OSI reference model
Compared to the five layers reference model explained above, the OSI reference model defined in [X200] is
divided in seven layers. The four lower layers are similar to the four lower layers described above. The OSI
reference model refined the application layer by dividing it in three layers :
the Session layer. The Session layer contains the protocols and mechanisms that are necessary to organize
and to synchronize the dialogue and to manage the data exchange of presentation layer entities. While one
of the main functions of the transport layer is to cope with the unreliability of the network layer, the sessions
layer objective is to hide the possible failures of transport-level connections to the upper layer higher. For
this, the Session Layer provides services that allow to establish a session-connection, to support orderly data
exchange (including mechanisms that allow to recover from the abrupt release of an underlying transport
connection), and to release the connection in an orderly manner.
2.7. The reference models 107

the Presentation layer was designed to cope with the different ways of representing information on com-
puters. There are many differences in the way computer store information. Some computers store integers
as 32 bits field, others use 64 bits field and the same problem arises with floating point number. For textual
information, this is even more complex with the many different character codes that have been used 21 . The
situation is even more complex when considering the exchange of structured information such as database
records. To solve this problem, the Presentation layer contains provides for a common representation of the
data transferred. The ASN.1 notation was designed for the Presentation layer and is still used today by some
protocols.
the Application layer that contains the mechanisms that do not fit in neither the Presentation nor the Session
layer. The OSI Application layer was itself further divided in several generic service elements.
Figure 2.93: The seven layers of the OSI reference model
21 There is now a rough consensus for the greater use of the Unicode character format. Unicode can represent more than 100,000 different
characters from the known written languages on Earth. Maybe one day, all computers will only use Unicode to represent all their stored
characters and Unicode could become the standard format to exchange characters, but we are not yet at this stage today.

CHAPTER 3
Part 2: Protocols
3.1 The application layer
Networked applications rely on the transport service. As explained earlier, there are two main types of transport
services :
the connectionless service
the connection-oriented or byte-stream service
The connectionless service allows applications to easily exchange messages or Service Data Units. On the Internet,
this service is provided by the UDP protocol that will be explained in the next chapter. The connectionless transport
service on the Internet is unreliable, but is able to detect transmission errors. This implies that an application will
not receive data that has been corrupted due to transmission errors.
The connectionless transport service allows networked application to exchange messages. Several networked
applications may be running at the same time on a single host. Each of these applications must be able to exchange
SDUs with remote applications. To enable these exchanges of SDUs, each networked application running on a
host is identified by the following information :
the host on which the application is running
the port number on which the application listens for SDUs
On the Internet, the port number is an integer and the host is identified by its network address. There are two types
of Internet Addresses :
IP version 4 addresses that are 32 bits wide
IP version 6 addresses that are 128 bits wide
IPv4 addresses are usually represented by using a dotted decimal representation where each decimal number
corresponds to one byte of the address, e.g. 203.0.113.56. IPv6 addresses are usually represented as a set of
hexadecimal numbers separated by semicolons, e.g. 2001:db8:3080:2:217:f2ff:fed6:65c0. Today, most Internet
hosts have one IPv4 address. A small fraction of them also have an IPv6 address. In the future, we can expect that
more and more hosts will have IPv6 addresses and that some of them will not have an IPv4 address anymore. A
host that only has an IPv4 address cannot communicate with a host having only an IPv6 address. The figure below
illustrates two that are using the datagram service provided by UDP on hosts that are using IPv4 addresses.
Note: Textual representation of IPv6 addresses
109
Figure 3.1: The connectionless or datagram service
It is sometimes necessary to write IPv6 addresses in text format, e.g. when manually configuring addresses or for
documentation purposes. The preferred format for writing IPv6 addresses is x:x:x:x:x:x:x:x, where the x s are
hexadecimal digits representing the eight 16-bit parts of the address. Here are a few examples of IPv6 addresses :
abcd:ef01:2345:6789:abcd:ef01:2345:6789
2001:db8:0:0:8:800:200c:417a
fe80:0:0:0:219:e3ff:fed7:1204
IPv6 addresses often contain a long sequence of bits set to 0. In this case, a compact notation has been defined.
With this notation, :: is used to indicate one or more groups of 16 bits blocks containing only bits set to 0. For
example,
2001:db8:0:0:8:800:200c:417a is represented as 2001:db8::8:800:200c:417a
ff01:0:0:0:0:0:0:101 is represented as ff01::101
0:0:0:0:0:0:0:1 is represented as ::1
0:0:0:0:0:0:0:0 is represented as ::
The second transport service is the connection-oriented service. On the Internet, this service is often called the
byte-stream service as it creates a reliable byte stream between the two applications that are linked by a transport
connection. Like the datagram service, the networked applications that use the byte-stream service are identified
by the host on which they run and a port number. These hosts can be identified by an address or a name. The
figure below illustrates two applications that are using the byte-stream service provided by the TCP protocol on
IPv6 hosts. The byte stream service provided by TCP is reliable and bidirectional.
3.2 The Domain Name System
We have already explained the main principles that underlie the utilisation of names on the Internet and their
mapping to addresses. The last component of the Domain Name System is the DNS protocol. The DNS protocol
runs above both the datagram service and the bytestream services. In practice, the datagram service is used when
short queries and responses are exchanged, and the bytestream service is used when longer responses are expected.
In this section, we will only discuss the utilisation of the DNS protocol above the datagram service. This is the
most frequent utilisation of the DNS.
DNS messages are composed of five parts that are named sections in RFC 1035. The first three sections are
mandatory and the last two sections are optional. The first section of a DNS message is its Header. It contains
information about the type of message and the content of the other sections. The second section contains the
Question sent to the name server or resolver. The third section contains the Answer to the Question. When a client
sends a DNS query, the Answer section is empty. The fourth section, named Authority, contains information about
the servers that can provide an authoritative answer if required. The last section contains additional information
that is supplied by the resolver or server but was not requested in the question.
110 Chapter 3. Part 2: Protocols

Figure 3.2: The connection-oriented or byte-stream service
The header of DNS messages is composed of 12 bytes and its structure is shown in the figure below.
Figure 3.3: DNS header
The ID (identifier) is a 16-bits random value chosen by the client. When a client sends a question to a DNS server,
it remembers the question and its identifier. When a server returns an answer, it returns in the ID field the identifier
chosen by the client. Thanks to this identifier, the client can match the received answer with the question that it
sent.
The QR flag is set to 0 in DNS queries and 1 in DNS answers. The Opcode is used to specify the type of query.
For instance, a standard query is when a client sends a name and the server returns the corresponding data and an
update request is when the client sends a name and new data and the server then updates its database.
The AA bit is set when the server that sent the response has authority for the domain name found in the question
section. In the original DNS deployments, two types of servers were considered : authoritative servers and non-
authoritative servers. The authoritative servers are managed by the system administrators responsible for a given
domain. They always store the most recent information about a domain. Non-authoritative servers are servers or
3.2. The Domain Name System 111

resolvers that store DNS information about external domains without being managed by the owners of a domain.
They may thus provide answers that are out of date. From a security point of view, the authoritative bit is not an
absolute indication about the validity of an answer. Securing the Domain Name System is a complex problem that
was only addressed satisfactorily recently by the utilisation of cryptographic signatures in the DNSSEC extensions
to DNS described in RFC 4033. However, these extensions are outside the scope of this chapter.
The RD (recursion desired) bit is set by a client when it sends a query to a resolver. Such a query is said to be
recursive because the resolver will recurse through the DNS hierarchy to retrieve the answer on behalf of the client.
In the past, all resolvers were configured to perform recursive queries on behalf of any Internet host. However,
this exposes the resolvers to several security risks. The simplest one is that the resolver could become overloaded
by having too many recursive queries to process. As of this writing, most resolvers 1 only allow recursive queries
from clients belonging to their company or network and discard all other recursive queries. The RA bit indicates
whether the server supports recursion. The RCODE is used to distinguish between different types of errors. See
RFC 1035 for additional details. The last four fields indicate the size of the Question, Answer, Authority and
Additional sections of the DNS message.
The last four sections of the DNS message contain Resource Records (RR). All RRs have the same top level format
Figure 3.4: DNS Resource Records
In a Resource Record (RR), the Name indicates the name of the node to which this resource record pertains. The
two bytes Type field indicate the type of resource record. The Class field was used to support the utilisation of the
DNS in other environments than the Internet.
The TTL field indicates the lifetime of the Resource Record in seconds. This field is set by the server that returns
an answer and indicates for how long a client or a resolver can store the Resource Record inside its cache. A long
TTL indicates a stable RR. Some companies use short TTL values for mobile hosts and also for popular servers.
For example, a web hosting company that wants to spread the load over a pool of hundred servers can configure
its nameservers to return different answers to different clients. If each answer has a small TTL, the clients will be
forced to send DNS queries regularly. The nameserver will reply to these queries by supplying the address of the
less loaded server.
1 Some DNS resolvers allow any host to send queries. Google operates a public DNS resolver at addresses 2001:4860:4860::8888 and
2001:4860:4860::8844

The RDLength field is the length of the RData field that contains the information of the type specified in the Type
field.
Several types of DNS RR are used in practice. The A type is used to encode the IPv4 address that corresponds to
the specified name. The AAAA type is used to encode the IPv6 address that corresponds to the specified name. A
NS record contains the name of the DNS server that is responsible for a given domain. For example, a query for
the AAAA record associated to the www.ietf.org name returns the following answer.
Figure 3.5: Query for the AAAA record of www.ietf.org
This answer contains several pieces of information. First, the name www.ietf.org is associated to IP address
2001:1890:123a::1:1e. Second, the ietf.org domain is managed by six different nameservers. Five of these
nameservers are reachable via IPv4 and IPv6.
CNAME (or canonical names) are used to define aliases. For example www.example.com could be a CNAME for
pc12.example.com that is the actual name of the server on which the web server for www.example.com runs.
Note: Reverse DNS

The DNS is mainly used to find the address that corresponds to a given name. However, it is sometimes use-
ful to obtain the name that corresponds to an IP address. This done by using the PTR (pointer) RR. The RData
part of a PTR RR contains the name while the Name part of the RR contains the IP address encoded in the
in-addr.arpa domain. IPv4 addresses are encoded in the in-addr.arpa by reversing the four digits that com-
pose the dotted decimal representation of the address. For example, consider IPv4 address 192.0.2.11. The
hostname associated to this address can be found by requesting the PTR RR that corresponds to 11.2.0.192.in-
addr.arpa. A similar solution is used to support IPv6 addresses RFC 3596, but slightly more complex given
the length of the IPv6 addresses. For example, consider IPv6 address 2001:1890:123a::1:1e. To obtain
the name that corresponds to this address, we need first to convert it in a reverse dotted decimal notation :
e.1.0.0.1.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.a.3.2.1.0.9.8.1.1.0.0.2. In this notation, each character between dots cor-
responds to one nibble, i.e. four bits. The low-order byte (e) appears first and the high order (2) last. To ob-
tain the name that corresponds to this address, one needs to append the ip6.arpa domain name and query for
e.1.0.0.1.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.a.3.2.1.0.9.8.1.1.0.0.2.ip6.arpa. In practice, tools and libraries do the con-
version automatically and the user does not need to worry about it.
3.2. The Domain Name System 113

An important point to note regarding the Domain Name System is its extensibility. Thanks to the Type and
RDLength fields, the format of the Resource Records can easily be extended. Furthermore, a DNS implementation
that receives a new Resource Record that it does not understand can ignore the record while still being able to
process the other parts of the message. This allows, for example, a DNS server that only supports IPv6 to ignore
the IPv4 addresses listed in the DNS reply for www.ietf.org while still being able to correctly parse the Resource
Records that it understands. This extensibility allowed the Domain Name System to evolve over the years while
still preserving the backward compatibility with already deployed DNS implementations.
3.3 Electronic mail
Electronic mail, or email, is a very popular application in computer networks such as the Internet. Email appeared
in the early 1970s and allows users to exchange text based messages. Initially, it was mainly used to exchange
short messages, but over the years its usage has grown. It is now not only used to exchange small, but also long
messages that can be composed of several parts as we will see later.
Before looking at the details of Internet email, let us consider a simple scenario illustrated in the figure below,
where Alice sends an email to Bob. Alice prepares her email by using an email clients and sends it to her email
server. Alices email server extracts Bobs address from the email and delivers the message to Bobs server. Bob
retrieves Alices message on his server and reads it by using his favourite email client or through his webmail
interface.
Figure 3.6: Simplified architecture of the Internet email
The email system that we consider in this book is composed of four components :
a message format, that defines how valid email messages are encoded
protocols, that allow hosts and servers to exchange email messages
client software, that allows users to easily create and read email messages
software, that allows servers to efficiently exchange email messages
We will first discuss the format of email messages followed by the protocols that are used on todays Internet to
exchange and retrieve emails. Other email systems have been developed in the past [Bush1993] [Genilloud1990]
[GC2000], but today most email solutions have migrated to the Internet email. Information about the software
that is used to compose and deliver emails may be found on wikipedia among others, for both email clients and
email servers. More detailed information about the full Internet Mail Architecture may be found in RFC 5598.
Email messages, like postal mail, are composed of two parts :
a header that plays the same role as the letterhead in regular mail. It contains metadata about the message.
the body that contains the message itself.
Email messages are entirely composed of lines of ASCII characters. Each line can contain up to 998 characters
and is terminated by the CR and LF control characters RFC 5322. The lines that compose the header appear
before the message body. An empty line, containing only the CR and LF characters, marks the end of the header.
The email header contains several lines that all begin with a keyword followed by a colon and additional informa-
tion. The format of email messages and the different types of header lines are defined in RFC 5322. Two of these
header lines are mandatory and must appear in all email messages :

Figure 3.7: The structure of email messages
The sender address. This header line starts with From:. This contains the (optional) name of the sender
followed by its email address between < and >. Email addresses are always composed of a username
followed by the @ sign and a domain name.
The date. This header line starts with Date:. RFC 5322 precisely defines the format used to encode a date.
Other header lines appear in most email messages. The Subject: header line allows the sender to indicate the topic
discussed in the email. Three types of header lines can be used to specify the recipients of a message :
2
the To: header line contains the email addresses of the primary recipients of the message . Several
addresses can be separated by using commas.
the cc: header line is used by the sender to provide a list of email addresses that must receive a carbon copy
of the message. Several addresses can be listed in this header line, separated by commas. All recipients of
the email message receive the To: and cc: header lines.
the bcc: header line is used by the sender to provide a list of comma separated email addresses that must
receive a blind carbon copy of the message. The bcc: header line is not delivered to the recipients of the
email message.
A simple email message containing the From:, To:, Subject: and Date: header lines and two lines of body is shown
below.
From: Bob Smith <[email protected]>
To: Alice Doe <[email protected]>, Alice Smith <[email protected]>
Subject: Hello
Date: Mon, 8 Mar 2010 19:55:06 -0600
This is the "Hello world" of email messages.

This is the second line of the body
Note the empty line after the Date: header line; this empty line contains only the CR and LF characters, and marks
the boundary between the header and the body of the message.
Several other optional header lines are defined in RFC 5322 and elsewhere 3 . Furthermore, many email clients
and servers define their own header lines starting from X-. Several of the optional header lines defined in RFC
5322 are worth being discussed here :
the Message-Id: header line is used to associate a unique identifier to each email. Email identifiers are
usually structured like string@domain where string is a unique character string or sequence number chosen
by the sender of the email and domain the domain name of the sender. Since domain names are unique,
a host can generate globally unique message identifiers concatenating a locally unique identifier with its
domain name.
the In-reply-to: is used when a message was created in reply to a previous message. In this case, the end of
the In-reply-to: line contains the identifier of the original message.
2It could be surprising that the To: is not mandatory inside an email message. While most email messages will contain this header line an
email that does not contain a To: header line and that relies on the bcc: to specify the recipient is valid as well.
3 The list of all standard email header lines may be found at http://www.iana.org/assignments/message-headers/message-header-index.html
3.3. Electronic mail 115

the Received: header line is used when an email message is processed by several servers before reaching its
destination. Each intermediate email server adds a Received: header line. These header lines are useful to
debug problems in delivering email messages.
The figure below shows the header lines of one email message. The message originated at a host named
wira.firstpr.com.au and was received by smtp3.sgsi.ucl.ac.be. The Received: lines have been wrapped for read-
ability.
Received: from smtp3.sgsi.ucl.ac.be (Unknown [10.1.5.3])
by mmp.sipr-dc.ucl.ac.be
(Sun Java(tm) System Messaging Server 7u3-15.01 64bit (built Feb 12 2010))
with ESMTP id <[email protected]>; Mon,
08 Mar 2010 11:37:17 +0100 (CET)
Received: from mail.ietf.org (mail.ietf.org [64.170.98.32])
by smtp3.sgsi.ucl.ac.be (Postfix) with ESMTP id B92351C60D7; Mon,
08 Mar 2010 11:36:51 +0100 (CET)
Received: from [127.0.0.1] (localhost [127.0.0.1]) by core3.amsl.com (Postfix)
with ESMTP id F066A3A68B9; Mon, 08 Mar 2010 02:36:38 -0800 (PST)
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix)
with ESMTP id A1E6C3A681B for <[email protected]>; Mon,
08 Mar 2010 02:36:37 -0800 (PST)
Received: from mail.ietf.org ([64.170.98.32])
by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
with ESMTP id erw8ih2v8VQa for <[email protected]>; Mon,
08 Mar 2010 02:36:36 -0800 (PST)
Received: from gair.firstpr.com.au (gair.firstpr.com.au [150.101.162.123])
by core3.amsl.com (Postfix) with ESMTP id 03E893A67ED for <[email protected]>; Mon,
08 Mar 2010 02:36:35 -0800 (PST)
Received: from [10.0.0.6] (wira.firstpr.com.au [10.0.0.6])
by gair.firstpr.com.au (Postfix) with ESMTP id D0A49175B63; Mon,
08 Mar 2010 21:36:37 +1100 (EST)
Date: Mon, 08 Mar 2010 21:36:38 +1100
From: Robin Whittle <[email protected]>
Subject: Re: [rrg] Recommendation and what happens next
In-reply-to: <C7B9C21A.4FAB%[email protected]>
To: RRG <[email protected]>
Message-id: <[email protected]>
Message content removed
Initially, email was used to exchange small messages of ASCII text between computer scientists. However, with
the growth of the Internet, supporting only ASCII text became a severe limitation for two reasons. First of all,
non-English speakers wanted to write emails in their native language that often required more characters than
those of the ASCII character table. Second, many users wanted to send other content than just ASCII text by
email such as binary files, images or sound.
To solve this problem, the IETF developed the Multipurpose Internet Mail Extensions (MIME). These extensions
were carefully designed to allow Internet email to carry non-ASCII characters and binary files without breaking
the email servers that were deployed at that time. This requirement for backward compatibility forced the MIME
designers to develop extensions to the existing email message format RFC 822 instead of defining a completely
new format that would have been better suited to support the new types of emails.
RFC 2045 defines three new types of header lines to support MIME :
The MIME-Version: header indicates the version of the MIME specification that was used to encode the
email message. The current version of MIME is 1.0. Other versions of MIME may be defined in the future.
Thanks to this header line, the software that processes email messages will be able to adapt to the MIME
version used to encode the message. Messages that do not contain this header are supposed to be formatted
according to the original RFC 822 specification.
The Content-Type: header line indicates the type of data that is carried inside the message (see below)
The Content-Transfer-Encoding: header line is used to specify how the message has been encoded. When
MIME was designed, some email servers were only able to process messages containing characters encoded

using the 7 bits ASCII character set. MIME allows the utilisation of other character encodings.
Inside the email header, the Content-Type: header line indicates how the MIME email message is structured. RFC
2046 defines the utilisation of this header line. The two most common structures for MIME messages are :
Content-Type: multipart/mixed. This header line indicates that the MIME message contains several inde-
pendent parts. For example, such a message may contain a part in plain text and a binary file.
Content-Type: multipart/alternative. This header line indicates that the MIME message contains several
representations of the same information. For example, a multipart/alternative message may contain both a
plain text and an HTML version of the same text.
To support these two types of MIME messages, the recipient of a message must be able to extract the different
parts from the message. In RFC 822, an empty line was used to separate the header lines from the body. Using an
empty line to separate the different parts of an email body would be difficult as the body of email messages often
contains one or more empty lines. Another possible option would be to define a special line, e.g. *-LAST_LINE-*
to mark the boundary between two parts of a MIME message. Unfortunately, this is not possible as some emails
may contain this string in their body (e.g. emails sent to students to explain the format of MIME messages). To
solve this problem, the Content-Type: header line contains a second parameter that specifies the string that has
been used by the sender of the MIME message to delineate the different parts. In practice, this string is often
chosen randomly by the mail client.
The email message below, copied from RFC 2046 shows a MIME message containing two parts that are both in
plain text and encoded using the ASCII character set. The string simple boundary is defined in the Content-Type:
header as the marker for the boundary between two successive parts. Another example of MIME messages may
be found in RFC 2046.
Date: Mon, 20 Sep 1999 16:33:16 +0200
From: Nathaniel Borenstein <[email protected]>
To: Ned Freed <[email protected]>
Subject: Test
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="simple boundary"
preamble, to be ignored
--simple boundary
Content-Type: text/plain; charset=us-ascii
First part
--simple boundary
Content-Type: text/plain; charset=us-ascii
Second part
--simple boundary
The Content-Type: header can also be used inside a MIME part. In this case, it indicates the type of data placed
in this part. Each data type is specified as a type followed by a subtype. A detailed description may be found in
RFC 2046. Some of the most popular Content-Type: header lines are :
text. The message part contains information in textual format. There are several subtypes : text/plain for
regular ASCII text, text/html defined in RFC 2854 for documents in HTML format or the text/enriched
format defined in RFC 1896. The Content-Type: header line may contain a second parameter that specifies
the character set used to encode the text. charset=us-ascii is the standard ASCII character table. Other
frequent character sets include charset=UTF8 or charset=iso-8859-1. The list of standard character sets is
maintained by IANA
image. The message part contains a binary representation of an image. The subtype indicates the format of
the image such as gif, jpg or png.
audio. The message part contains an audio clip. The subtype indicates the format of the audio clip like wav
or mp3

video. The message part contains a video clip. The subtype indicates the format of the video clip like avi or
mp4
application. The message part contains binary information that was produced by the particular application
listed as the subtype. Email clients use the subtype to launch the application that is able to decode the
received binary information.
Note: From ASCII to Unicode

The first computers used different techniques to represent characters in memory and on disk. During the 1960s,
computers began to exchange information via tape or telephone lines. Unfortunately, each vendor had its own
proprietary character set and exchanging data between computers from different vendors was often difficult. The
7 bits ASCII character table RFC 20 set was adopted by several vendors and by many Internet protocols. However,
ASCII became a problem with the internationalisation of the Internet and the desire of more and more users to use
character sets that support their own written language. A first attempt at solving this problem was the definition
of the ISO-8859 character sets by ISO. This family of standards specified various character sets that allowed the
representation of many European written languages by using 8 bits characters. Unfortunately, an 8-bits character
set is not sufficient to support some widely used languages, such as those used in Asian countries. Fortunately, at
the end of the 1980s, several computer scientists proposed to develop a standard that supports all written languages
used on Earth today. The Unicode standard [Unicode] has now been adopted by most computer and software
vendors. For example, Java uses Unicode natively to manipulate characters, Python can handle both ASCII and
Unicode characters. Internet applications are slowly moving towards complete support for the Unicode character
sets, but moving from ASCII to Unicode is an important change that can have a huge impact on current deployed
implementations. See for example, the work to completely internationalise email RFC 4952 and domain names
RFC 5890.
The last MIME header line is Content-Transfer-Encoding:. This header line is used after the Content-Type: header
line, within a message part, and specifies how the message part has been encoded. The default encoding is to use
7 bits ASCII. The most frequent encodings are quoted-printable and Base64. Both support encoding a sequence
of bytes into a set of ASCII lines that can be safely transmitted by email servers. quoted-printable is defined in
RFC 2045. We briefly describe base64 which is defined in RFC 2045 and RFC 4648.
Base64 divides the sequence of bytes to be encoded into groups of three bytes (with the last group possibly being
partially filled). Each group of three bytes is then divided into four six-bit fields and each six bit field is encoded
as a character from the table below.
Value Encoding Value Encoding Value Encoding Value Encoding
0 A 17 R 34 i 51 z
1 B 18 S 35 j 52 0
2 C 19 T 36 k 53 1
3 D 20 U 37 l 54 2
4 E 21 V 38 m 55 3
5 F 22 W 39 n 56 4
6 G 23 X 40 o 57 5
7 H 24 Y 41 p 58 6
8 I 25 Z 42 q 59 7
9 J 26 a 43 r 60 8
10 K 27 b 44 s 61 9
11 L 28 c 45 t 62 +
12 M 29 d 46 u 63 /
13 N 30 e 47 v
14 O 31 f 48 w
15 P 32 g 49 x
16 Q 33 h 50 y
The example below, from RFC 4648, illustrates the Base64 encoding.

Input data 0x14fb9c03d97e

8-bit 00010100 11111011 10011100 00000011 11011001 01111110
6-bit 000101 001111 101110 011100 000000 111101 100101 111110
Decimal 5 15 46 28 0 61 37 62
Encoding FPucA9l+
The last point to be discussed about base64 is what happens when the length of the sequence of bytes to be
encoded is not a multiple of three. In this case, the last group of bytes may contain one or two bytes instead of
three. Base64 reserves the = character as a padding character. This character is used once when the last group
contains two bytes and twice when it contains one byte as illustrated by the two examples below.
Input data 0x14
8-bit 00010100
6-bit 000101 000000
Decimal 50
Encoding FA==
Input data 0x14b9
8-bit 00010100 11111011
6-bit 000101 001111 101100
Decimal 5 15 44
Encoding FPs=
Now that we have explained the format of the email messages, we can discuss how these messages can be ex-
changed through the Internet. The figure below illustrates the protocols that are used when Alice sends an email
message to Bob. Alice prepares her email with an email client or on a webmail interface. To send her email to
Bob, Alices client will use the Simple Mail Transfer Protocol (SMTP) to deliver her message to her SMTP server.
Alices email client is configured with the name of the default SMTP server for her domain. There is usually at
least one SMTP server per domain. To deliver the message, Alices SMTP server must find the SMTP server that
contains Bobs mailbox. This can be done by using the Mail eXchange (MX) records of the DNS. A set of MX
records can be associated to each domain. Each MX record contains a numerical preference and the fully qualified
domain name of a SMTP server that is able to deliver email messages destined to all valid email addresses of this
domain. The DNS can return several MX records for a given domain. In this case, the server with the lowest
numerical preference is used first RFC 2821. If this server is not reachable, the second most preferred server is
used etc. Bobs SMTP server will store the message sent by Alice until Bob retrieves it using a webmail interface
or protocols such as the Post Office Protocol (POP) or the Internet Message Access Protocol (IMAP).
Figure 3.8: Email delivery protocols
3.3.1 The Simple Mail Transfer Protocol
The Simple Mail Transfer Protocol (SMTP) defined in RFC 5321 is a client-server protocol. The SMTP specifi-
cation distinguishes between five types of processes involved in the delivery of email messages. Email messages
are composed on a Mail User Agent (MUA). The MUA is usually either an email client or a webmail. The MUA
sends the email message to a Mail Submission Agent (MSA). The MSA processes the received email and forwards
it to the Mail Transmission Agent (MTA). The MTA is responsible for the transmission of the email, directly or
via intermediate MTAs to the MTA of the destination domain. This destination MTA will then forward the mes-
sage to the Mail Delivery Agent (MDA) where it will be accessed by the recipients MUA. SMTP is used for the

interactions between MUA and MSA 4 , MSA-MTA and MTA-MTA.

SMTP is a text-based protocol like many other application-layer protocols on the Internet. It relies on the byte-
stream service. Servers listen on port 25. Clients send commands that are each composed of one line of ASCII
text terminated by CR+LF. Servers reply by sending ASCII lines that contain a three digit numerical error/success
code and optional comments.
The SMTP protocol, like most text-based protocols, is specified as a BNF. The full BNF is defined in RFC 5321.
The main SMTP commands are defined by the BNF rules shown in the figure below.
Figure 3.9: BNF specification of the SMTP commands
In this BNF, atext corresponds to printable ASCII characters. This BNF rule is defined in RFC 5322. The five
main commands are EHLO, MAIL FROM:, RCPT TO:, DATA and QUIT 5 . Postmaster is the alias of the system
administrator who is responsible for a given domain or SMTP server. All domains must have a Postmaster alias.
The SMTP responses are defined by the BNF shown in the figure below.
Figure 3.10: BNF specification of the SMTP responses
SMTP servers use structured reply codes containing three digits and an optional comment. The first digit of
the reply code indicates whether the command was successful or not. A reply code of 2xy indicates that the
command has been accepted. A reply code of 3xy indicates that the command has been accepted, but additional
information from the client is expected. A reply code of 4xy indicates a transient negative reply. This means that
for some reason, which is indicated by either the other digits or the comment, the command cannot be processed
immediately, but there is some hope that the problem will only be transient. This is basically telling the client to
try the same command again later. In contrast, a reply code of 5xy indicates a permanent failure or error. In this
case, it is useless for the client to retry the same command later. Other application layer protocols such as FTP
RFC 959 or HTTP RFC 2616 use a similar structure for their reply codes. Additional details about the other reply
codes may be found in RFC 5321.
Examples of SMTP reply codes include the following :
500 Syntax error, command unrecognized
501 Syntax error in parameters or arguments
502 Command not implemented
503 Bad sequence of commands
220 <domain> Service ready
221 <domain> Service closing transmission channel
421 <domain> Service not available, closing transmission channel
250 Requested mail action okay, completed
450 Requested mail action not taken: mailbox unavailable
4 During the last years, many Internet Service Providers, campus and enterprise networks have deployed SMTP extensions RFC 4954 on
their MSAs. These extensions force the MUAs to be authenticated before the MSA accepts an email message from the MUA.
5 The first versions of SMTP used HELO as the first command sent by a client to a SMTP server. When SMTP was extended to support
newer features such as 8 bits characters, it was necessary to allow a server to recognise whether it was interacting with a client that supported
the extensions or not. EHLO became mandatory with the publication of RFC 2821.

452 Requested action not taken: insufficient system storage

550 Requested action not taken: mailbox unavailable
354 Start mail input; end with <CRLF>.<CRLF>
The first four reply codes correspond to errors in the commands sent by the client. The fourth reply code would
be sent by the server when the client sends commands in an incorrect order (e.g. the client tries to send an email
before providing the destination address of the message). Reply code 220 is used by the server as the first message
when it agrees to interact with the client. Reply code 221 is sent by the server before closing the underlying
transport connection. Reply code 421 is returned when there is a problem (e.g. lack of memory/disk resources)
that prevents the server from accepting the transport connection. Reply code 250 is the standard positive reply that
indicates the success of the previous command. Reply codes 450 and 452 indicate that the destination mailbox
is temporarily unavailable, for various reasons, while reply code 550 indicates that the mailbox does not exist or
cannot be used for policy reasons. Reply code 354 indicates that the client can start transmitting its email message.
The transfer of an email message is performed in three phases. During the first phase, the client opens a transport
connection with the server. Once the connection has been established, the client and the server exchange greetings
messages (EHLO command). Most servers insist on receiving valid greeting messages and some of them drop the
underlying transport connection if they do not receive a valid greeting. Once the greetings have been exchanged,
the email transfer phase can start. During this phase, the client transfers one or more email messages by indicating
the email address of the sender (MAIL FROM: command), the email address of the recipient (RCPT TO: command)
followed by the headers and the body of the email message (DATA command). Once the client has finished sending
all its queued email messages to the SMTP server, it terminates the SMTP association (QUIT command).
A successful transfer of an email message is shown below
S: 220 smtp.example.com ESMTP MTA information
C: EHLO mta.example.org
S: 250 Hello mta.example.org, glad to meet you
C: MAIL FROM:<[email protected]>
S: 250 Ok
C: RCPT TO:<[email protected]>
S: 250 Ok
C: DATA
S: 354 End data with <CR><LF>.<CR><LF>
C: From: "Alice Doe" <[email protected]>
C: To: Bob Smith <[email protected]>
C: Date: Mon, 9 Mar 2010 18:22:32 +0100
C: Subject: Hello
C:
C: Hello Bob
C: This is a small message containing 4 lines of text.
C: Best regards,
C: Alice
C: .
S: 250 Ok: queued as 12345
C: QUIT
S: 221 Bye
In the example above, the MTA running on mta.example.org opens a TCP connection to the SMTP server on host
smtp.example.com. The lines prefixed with S: (resp. C:) are the responses sent by the server (resp. the commands
sent by the client). The server sends its greetings as soon as the TCP connection has been established. The client
then sends the EHLO command with its fully qualified domain name. The server replies with reply-code 250 and
sends its greetings. The SMTP association can now be used to exchange an email.
To send an email, the client must first provide the address of the recipient with RCPT TO:. Then it uses the MAIL
FROM: with the address of the sender. Both the recipient and the sender are accepted by the server. The client
can now issue the DATA command to start the transfer of the email message. After having received the 354 reply
code, the client sends the headers and the body of its email message. The client indicates the end of the message
by sending a line containing only the . (dot) character 6 . The server confirms that the email message has been
6 This implies that a valid email message cannot contain a line with one dot followed by CR and LF. If a user types such a line in an email,
his email client will automatically add a space character before or after the dot when sending the message over SMTP.

queued for delivery or transmission with a reply code of 250. The client issues the QUIT command to close the
session and the server confirms with reply-code 221, before closing the TCP connection.
Note: Open SMTP relays and spam

Since its creation in 1971, email has been a very useful tool that is used by many users to exchange lots of
information. In the early days, all SMTP servers were open and anyone could use them to forward emails towards
their final destination. Unfortunately, over the years, some unscrupulous users have found ways to use email for
marketing purposes or to send malware. The first documented abuse of email for marketing purposes occurred in
1978 when a marketer who worked for a computer vendor sent a marketing email to many ARPANET users. At
that time, the ARPANET could only be used for research purposes and this was an abuse of the acceptable use
policy. Unfortunately, given the extremely low cost of sending emails, the problem of unsolicited emails has not
stopped. Unsolicited emails are now called spam and a study carried out by ENISA in 2009 reveals that 95% of
email was spam and this number seems to continue to grow. This places a burden on the email infrastructure of
Internet Service Providers and large companies that need to process many useless messages.
Given the amount of spam messages, SMTP servers are no longer open RFC 5068. Several extensions to SMTP
have been developed in recent years to deal with this problem. For example, the SMTP authentication scheme
defined in RFC 4954 can be used by an SMTP server to authenticate a client. Several techniques have also been
proposed to allow SMTP servers to authenticate the messages sent by their users RFC 4870 RFC 4871 .
3.3.2 The Post Office Protocol
When the first versions of SMTP were designed, the Internet was composed of minicomputers that were used by
an entire university department or research lab. These minicomputers were used by many users at the same time.
Email was mainly used to send messages from a user on a given host to another user on a remote host. At that
time, SMTP was the only protocol involved in the delivery of the emails as all hosts attached to the network were
running an SMTP server. On such hosts, an email destined to local users was delivered by placing the email in a
special directory or file owned by the user. However, the introduction of personal computers in the 1980s, changed
this environment. Initially, users of these personal computers used applications such as telnet to open a remote
session on the local minicomputer to read their email. This was not user-friendly. A better solution appeared
with the development of user friendly email client applications on personal computers. Several protocols were
designed to allow these client applications to retrieve the email messages destined to a user from his/her server.
Two of these protocols became popular and are still used today. The Post Office Protocol (POP), defined in RFC
1939, is the simplest one. It allows a client to download all the messages destined to a given user from his/her
email server. We describe POP briefly in this section. The second protocol is the Internet Message Access Protocol
(IMAP), defined in RFC 3501. IMAP is more powerful, but also more complex than POP. IMAP was designed to
allow client applications to efficiently access in real-time to messages stored in various folders on servers. IMAP
assumes that all the messages of a given user are stored on a server and provides the functions that are necessary
to search, download, delete or filter messages.
POP is another example of a simple line-based protocol. POP runs above the bytestream service. A POP server
usually listens to port 110. A POP session is composed of three parts : an authorisation phase during which
the server verifies the clients credential, a transaction phase during which the client downloads messages and an
update phase that concludes the session. The client sends commands and the server replies are prefixed by +OK
to indicate a successful command or by -ERR to indicate errors.
When a client opens a transport connection with the POP server, the latter sends as banner an ASCII-line starting
with +OK. The POP session is at that time in the authorisation phase. In this phase, the client can send its
username (resp. password) with the USER (resp. PASS) command. The server replies with +OK if the username
(resp. password) is valid and -ERR otherwise.
Once the username and password have been validated, the POP session enters in the transaction phase. In this
phase, the client can issue several commands. The STAT command is used to retrieve the status of the server.
Upon reception of this command, the server replies with a line that contains +OK followed by the number of
messages in the mailbox and the total size of the mailbox in bytes. The RETR command, followed by a space and
an integer, is used to retrieve the nth message of the mailbox. The DELE command is used to mark for deletion
the nth message of the mailbox.

Once the client has retrieved and possibly deleted the emails contained in the mailbox, it must issue the QUIT
command. This command terminates the POP session and allows the server to delete all the messages that have
been marked for deletion by using the DELE command.
The figure below provides a simple POP session. All lines prefixed with C: (resp. S:) are sent by the client (resp.
server).
S: +OK POP3 server ready
C: USER alice
S: +OK
C PASS 12345pass
S: +OK alices maildrop has 2 messages (620 octets)
C: STAT
S: +OK 2 620
C: LIST
S: +OK 2 messages (620 octets)
S: 1 120
S: 2 500
S: .
C: RETR 1
S: +OK 120 octets
S: <the POP3 server sends message 1>
S: .
C: DELE 1
S: +OK message 1 deleted
C: QUIT
S: +OK POP3 server signing off (1 message left)
In this example, a POP client contacts a POP server on behalf of the user named alice. Note that in this example,
Alices password is sent in clear by the client. This implies that if someone is able to capture the packets sent by
Alice, he will know Alices password 7 . Then Alices client issues the STAT command to know the number of
messages that are stored in her mailbox. It then retrieves and deletes the first message of the mailbox.
3.4 The HyperText Transfer Protocol
In the early days of the Internet was mainly used for remote terminal access with telnet, email and file transfer.
The default file transfer protocol, FTP, defined in RFC 959 was widely used and FTP clients and servers are still
included in most operating systems.
Many FTP clients offer a user interface similar to a Unix shell and allow the client to browse the file system on
the server and to send and retrieve files. FTP servers can be configured in two modes :
authenticated : in this mode, the ftp server only accepts users with a valid user name and password. Once
authenticated, they can access the files and directories according to their permissions
anonymous : in this mode, clients supply the anonymous userid and their email address as password. These
clients are granted access to a special zone of the file system that only contains public files.
ftp was very popular in the 1990s and early 2000s, but today it has mostly been superseded by more recent
protocols. Authenticated access to files is mainly done by using the Secure Shell (ssh) protocol defined in RFC
4251 and supported by clients such as scp or sftp. Nowadays, anonymous access is mainly provided by web
protocols.
In the late 1980s, high energy physicists working at CERN had to efficiently exchange documents about their
ongoing and planned experiments. Tim Berners-Lee evaluated several of the documents sharing techniques that
were available at that time [B1989]. As none of the existing solutions met CERNs requirements, they chose to
develop a completely new document sharing system. This system was initially called the mesh, but was quickly
renamed the world wide web. The starting point for the world wide web are hypertext documents. An hypertext
document is a document that contains references (hyperlinks) to other documents that the reader can immediately
access. Hypertext was not invented for the world wide web. The idea of hypertext documents was proposed in
7 RFC 1939 defines the APOP authentication scheme that is not vulnerable to such attacks.
3.4. The HyperText Transfer Protocol 123

1945 [Bush1945] and the first experiments were done during the 1960s [Nelson1965] [Myers1998] . Compared to
the hypertext documents that were used in the late 1980s, the main innovation introduced by the world wide web
was to allow hyperlinks to reference documents stored on remote machines.
Figure 3.11: World-wide web clients and servers
A document sharing system such as the world wide web is composed of three important parts.
1. A standardised addressing scheme that allows unambiguous identification of documents
2. A standard document format : the HyperText Markup Language
3. A standardised protocol that facilitates efficient retrieval of documents stored on a server
Note: Open standards and open implementations

Open standards have, and are still playing a key role in the success of the world wide web as we know it to-
day. Without open standards, the world wide web would never have reached its current size. In addition to open
standards, another important factor for the success of the web was the availability of open and efficient imple-
mentations of these standards. When CERN started to work on the web, their objective was to build a running
system that could be used by physicists. They developed open-source implementations of the first web servers and
web clients. These open-source implementations were powerful and could be used as is, by institutions willing to
share information on the web. They were also extended by other developers who contributed to new features. For
example, NCSA added support for images in their Mosaic browser that was eventually used to create Netscape
Communications.
The first components of the world wide web are the Uniform Resource Identifiers (URI), defined in RFC 3986. A
URI is a character string that unambiguously identifies a resource on the world wide web. Here is a subset of the
BNF for URIs
URI = scheme ":" "//" authority path [ "?" query ] [ "#" fragment ]
scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
authority = [ userinfo "@" ] host [ ":" port ]
query = *( pchar / "/" / "?" )
fragment = *( pchar / "/" / "?" )
pchar = unreserved / pct-encoded / sub-delims / ":" / "@"
query = *( pchar / "/" / "?" )
fragment = *( pchar / "/" / "?" )
pct-encoded = "%" HEXDIG HEXDIG
unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~"
reserved = gen-delims / sub-delims

gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@"

sub-delims = "!" / "$" / "&" / "" / "(" / ")" / "*" / "+" / "," / ";" / "="
The first component of a URI is its scheme. A scheme can be seen as a selector, indicating the meaning of the
fields after it. In practice, the scheme often identifies the application-layer protocol that must be used by the client
to retrieve the document, but it is not always the case. Some schemes do not imply a protocol at all and some do
not indicate a retrievable document 8 . The most frequent scheme is http that will be described later. A URI scheme
can be defined for almost any application layer protocol 9 . The characters : and // follow the scheme of any URI.
The second part of the URI is the authority. With retrievable URI, this includes the DNS name or the IP address
of the server where the document can be retrieved using the protocol specified via the scheme. This name can
be preceded by some information about the user (e.g. a user name) who is requesting the information. Earlier
definitions of the URI allowed the specification of a user name and a password before the @ character (RFC
1738), but this is now deprecated as placing a password inside a URI is insecure. The host name can be followed
by the semicolon character and a port number. A default port number is defined for some protocols and the port
number should only be included in the URI if a non-default port number is used (for other protocols, techniques
like service DNS records are used).
The third part of the URI is the path to the document. This path is structured as filenames on a Unix host (but
it does not imply that the files are indeed stored this way on the server). If the path is not specified, the server
will return a default document. The last two optional parts of the URI are used to provide a query and indicate a
specific part (e.g. a section in an article) of the requested document. Sample URIs are shown below.
http://tools.ietf.org/html/rfc3986.html
mailto:[email protected]?subject=current-issue
http://docs.python.org/library/basehttpserver.html?highlight=http#BaseHTTPServer.BaseHTTPRequestHa
telnet://[2001:db8:3080:3::2]:80/
ftp://cnn.example.com&[email protected]/top_story.htm
The first URI corresponds to a document named rfc3986.html that is stored on the server named tools.ietf.org and
can be accessed by using the http protocol on its default port. The second URI corresponds to an email message,
with subject current-issue, that will be sent to user infobot in domain example.com. The mailto: URI scheme is
defined in RFC 6068. The third URI references the portion BaseHTTPServer.BaseHTTPRequestHandler of the
document basehttpserver.html that is stored in the library directory on server docs.python.org. This document can
be retrieved by using the http protocol. The query highlight=http is associated to this URI. The fourth example is a
server that operates the telnet protocol, uses IPv6 address 2001:db8:3080:3::2 and is reachable on port 80. The last
URI is somewhat special. Most users will assume that it corresponds to a document stored on the cnn.example.com
server. However, to parse this URI, it is important to remember that the @ character is used to separate the user
name from the host name in the authorisation part of a URI. This implies that the URI points to a document named
top_story.htm on host having IPv4 address 10.0.0.1. The document will be retrieved by using the ftp protocol with
the user name set to cnn.example.com&story=breaking_news.
The second component of the word wide web is the HyperText Markup Language (HTML). HTML defines the
format of the documents that are exchanged on the web. The first version of HTML was derived from the Standard
Generalized Markup Language (SGML) that was standardised in 1986 by ISO. SGML was designed to allow
large project documents in industries such as government, law or aerospace to be shared efficiently in a machine-
readable manner. These industries require documents to remain readable and editable for tens of years and insisted
on a standardised format supported by multiple vendors. Today, SGML is no longer widely used beyond specific
applications, but its descendants including HTML and XML are now widespread.
A markup language is a structured way of adding annotations about the formatting of the document within the
document itself. Example markup languages include troff, which is used to write the Unix man pages or Latex.
HTML uses markers to annotate text and a document is composed of HTML elements. Each element is usually
composed of three items: a start tag that potentially includes some specific attributes, some text (often including
other elements), and an end tag. A HTML tag is a keyword enclosed in angle brackets. The generic form of a
HTML element is
8 An example of a non-retrievable URI is urn:isbn:0-380-81593-1 which is an unique identifier for a book, through the urn scheme
(see RFC 3187). Of course, any URI can be made retrievable via a dedicated server or a new protocol but this one has no explicit proto-
col. Same thing for the scheme tag (see RFC 4151), often used in Web syndication (see RFC 4287 about the Atom syndication format).
Even when the scheme is retrievable (for instance with http), it is often used only as an identifier, not as a way to get a resource. See
http://norman.walsh.name/2006/07/25/namesAndAddresses for a good explanation.
9 The list of standard URI schemes is maintained by IANA at http://www.iana.org/assignments/uri-schemes.html

<tag>Some text to be displayed</tag>
More complex HTML elements can also include optional attributes in the start tag
<tag attribute1="value1" attribute2="value2">some text to be displayed</tag>
The HTML document shown below is composed of two parts : a header, delineated by the <head> and </head>
markers, and a body (between the <body> and </body> markers). In the example below, the header only contains
a title, but other types of information can be included in the header. The body contains an image, some text and a
list with three hyperlinks. The image is included in the web page by indicating its URI between brackets inside the
<img src=...> marker. The image can, of course, reside on any server and the client will automatically download
it when rendering the web page. The <h1>...</h1> marker is used to specify the first level of headings. The <ul>
marker indicates an unnumbered list while the <li> marker indicates a list item. The <a href=URI>text</a>
indicates a hyperlink. The text will be underlined in the rendered web page and the client will fetch the specified
URI when the user clicks on the link.
Figure 3.12: A simple HTML page
Additional details about the various extensions to HTML may be found in the official specifications maintained
by W3C.
The third component of the world wide web is the HyperText Transfert Protocol (HTTP). HTTP is a text-based
protocol, in which the client sends a request and the server returns a response. HTTP runs above the bytestream
service and HTTP servers listen by default on port 80. The design of HTTP has largely been inspired by the
Internet email protocols. Each HTTP request contains three parts :
a method , that indicates the type of request, a URI, and the version of the HTTP protocol used by the client
a header , that is used by the client to specify optional parameters for the request. An empty line is used to
mark the end of the header
an optional MIME document attached to the request
The response sent by the server also contains three parts :
a status line , that indicates whether the request was successful or not
a header , that contains additional information about the response. The response header ends with an empty
line.
a MIME document
Several types of method can be used in HTTP requests. The three most important ones are :
the GET method is the most popular one. It is used to retrieve a document from a server. The
GET method is encoded as GET followed by the path of the URI of the requested document and
the version of HTTP used by the client. For example, to retrieve the http://www.w3.org/MarkUp/
URI, a client must open a TCP on port 80 with host www.w3.org and send a HTTP request
containing the following line :

Figure 3.13: HTTP requests and responses
GET /MarkUp/ HTTP/1.0
the HEAD method is a variant of the GET method that allows the retrieval of the header lines
for a given URI without retrieving the entire document. It can be used by a client to verify if a
document exists, for instance.
the POST method can be used by a client to send a document to a server. The sent document is
attached to the HTTP request as a MIME document.
HTTP clients and servers can include many different HTTP headers in HTTP requests and responses. Each HTTP
header is encoded as a single ASCII-line terminated by CR and LF. Several of these headers are briefly described
below. A detailed discussion of all standard headers may be found in RFC 1945. The MIME headers can appear
in both HTTP requests and HTTP responses.
the Content-Length: header is the MIME header that indicates the length of the MIME document in bytes.
the Content-Type: header is the MIME header that indicates the type of the attached MIME document.
HTML pages use the text/html type.
the Content-Encoding: header indicates how the MIME document has been encoded. For example, this
header would be set to x-gzip for a document compressed using the gzip software.
RFC 1945 and RFC 2616 define headers that are specific to HTTP responses. These server headers include :
the Server: header indicates the version of the web server that has generated the HTTP response. Some
servers provide information about their software release and optional modules that they use. For security
reasons, some system administrators disable these headers to avoid revealing too much information about
their server to potential attackers.
the Date: header indicates when the HTTP response has been produced by the server.
the Last-Modified: header indicates the date and time of the last modification of the document attached to
the HTTP response.
Similarly, the following header lines can only appear inside HTTP requests sent by a client :
the User-Agent: header provides information about the client that has generated the HTTP request. Some
servers analyse this header line and return different headers and sometimes different documents for different
user agents.
the If-Modified-Since: header is followed by a date. It enables clients to cache in memory or on disk the
recent or most frequently used documents. When a client needs to request a URI from a server, it first checks

whether the document is already in its cache. If it is, the client sends a HTTP request with the If-Modified-
Since: header indicating the date of the cached document. The server will only return the document attached
to the HTTP response if it is newer than the version stored in the clients cache.
the Referrer: header is followed by a URI. It indicates the URI of the document that the client visited before
sending this HTTP request. Thanks to this header, the server can know the URI of the document containing
the hyperlink followed by the client, if any. This information is very useful to measure the impact of
advertisements containing hyperlinks placed on websites.
the Host: header contains the fully qualified domain name of the URI being requested.
Note: The importance of the Host: header line

The first version of HTTP did not include the Host: header line. This was a severe limitation for web host-
ing companies. For example consider a web hosting company that wants to serve both web.example.com and
www.example.net on the same physical server. Both web sites contain a /index.html document. When a client
sends a request for either http://web.example.com/index.html or http://www.example.net/index.html, the HTTP 1.0
request contains the following line :
GET /index.html HTTP/1.0
By parsing this line, a server cannot determine which index.html file is requested. Thanks to the
Host: header line, the server knows whether the request is for http://web.example.com/index.html or
http://www.dummy.net/index.html. Without the Host: header, this is impossible. The Host: header line allowed
web hosting companies to develop their business by supporting a large number of independent web servers on the
same physical server.
The status line of the HTTP response begins with the version of HTTP used by the server (usually HTTP/1.0
defined in RFC 1945 or HTTP/1.1 defined in RFC 2616) followed by a three digit status code and additional
information in English. HTTP status codes have a similar structure as the reply codes used by SMTP.
All status codes starting with digit 2 indicate a valid response. 200 Ok indicates that the HTTP request was
successfully processed by the server and that the response is valid.
All status codes starting with digit 3 indicate that the requested document is no longer available on the
server. 301 Moved Permanently indicates that the requested document is no longer available on this server.
A Location: header containing the new URI of the requested document is inserted in the HTTP response.
304 Not Modified is used in response to an HTTP request containing the If-Modified-Since: header. This
status line is used by the server if the document stored on the server is not more recent than the date indicated
in the If-Modified-Since: header.
All status codes starting with digit 4 indicate that the server has detected an error in the HTTP request sent
by the client. 400 Bad Request indicates a syntax error in the HTTP request. 404 Not Found indicates that
the requested document does not exist on the server.
All status codes starting with digit 5 indicate an error on the server. 500 Internal Server Error indicates that
the server could not process the request due to an error on the server itself.
In both the HTTP request and the HTTP response, the MIME document refers to a representation of the document
with the MIME headers indicating the type of document and its size.
As an illustration of HTTP/1.0, the transcript below shows a HTTP request for http://www.ietf.org and the corre-
sponding HTTP response. The HTTP request was sent using the curl command line tool. The User-Agent: header
line contains more information about this client software. There is no MIME document attached to this HTTP
request, and it ends with a blank line.
GET / HTTP/1.0
User-Agent: curl/7.19.4 (universal-apple-darwin10.0) libcurl/7.19.4 OpenSSL/0.9.8l zlib/1.2.3
Host: www.ietf.org
The HTTP response indicates the version of the server software used with the modules included. The Last-
Modified: header indicates that the requested document was modified about one week before the request. A

HTML document (not shown) is attached to the response. Note the blank line between the header of the HTTP
response and the attached MIME document. The Server: header line has been truncated in this output.
HTTP/1.1 200 OK
Date: Mon, 15 Mar 2010 13:40:38 GMT
Server: Apache/2.2.4 (Linux/SUSE) mod_ssl/2.2.4 OpenSSL/0.9.8e (truncated)
Last-Modified: Tue, 09 Mar 2010 21:26:53 GMT
Content-Length: 17019
Content-Type: text/html
<!DOCTYPE HTML PUBLIC .../HTML>
HTTP was initially designed to share self-contained text documents. For this reason, and to ease the implemen-
tation of clients and servers, the designers of HTTP chose to open a TCP connection for each HTTP request.
This implies that a client must open one TCP connection for each URI that it wants to retrieve from a server as
illustrated on the figure below. For a web page containing only text documents this was a reasonable design choice
as the client usually remains idle while the (human) user is reading the retrieved document.
Figure 3.14: HTTP 1.0 and the underlying TCP connection
However, as the web evolved to support richer documents containing images, opening a TCP connection for each
URI became a performance problem [Mogul1995]. Indeed, besides its HTML part, a web page may include
dozens of images or more. Forcing the client to open a TCP connection for each component of a web page
has two important drawbacks. First, the client and the server must exchange packets to open and close a TCP
connection as we will see later. This increases the network overhead and the total delay of completely retrieving
all the components of a web page. Second, a large number of established TCP connections may be a performance
bottleneck on servers.
This problem was solved by extending HTTP to support persistent TCP connections RFC 2616. A persistent
connection is a TCP connection over which a client may send several HTTP requests. This is illustrated in the
figure below.
To allow the clients and servers to control the utilisation of these persistent TCP connections, HTTP 1.1 RFC
2616 defines several new HTTP headers :
The Connection: header is used with the Keep-Alive argument by the client to indicate that it expects the
underlying TCP connection to be persistent. When this header is used with the Close argument, it indicates
that the entity that sent it will close the underlying TCP connection at the end of the HTTP response.
The Keep-Alive: header is used by the server to inform the client about how it agrees to use the persistent
connection. A typical Keep-Alive: contains two parameters : the maximum number of requests that the
server agrees to serve on the underlying TCP connection and the timeout (in seconds) after which the server
will close an idle connection
The example below shows the operation of HTTP/1.1 over a persistent TCP connection to retrieve three URIs
stored on the same server. Once the connection has been established, the client sends its first request with the
Connection: keep-alive header to request a persistent connection.

Figure 3.15: HTTP 1.1 persistent connections
GET / HTTP/1.1
Host: www.kame.net
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_6_2; en-us)
Connection: keep-alive
The server replies with the Connection: Keep-Alive header and indicates that it accepts a maximum of 100 HTTP
requests over this connection and that it will close the connection if it remains idle for 15 seconds.
HTTP/1.1 200 OK
Date: Fri, 19 Mar 2010 09:23:37 GMT
Server: Apache/2.0.63 (FreeBSD) PHP/5.2.12 with Suhosin-Patch
Keep-Alive: timeout=15, max=100
Connection: Keep-Alive
Content-Type: text/html
<html>... </html>
The client sends a second request for the style sheet of the retrieved web page.
GET /style.css HTTP/1.1
Host: www.kame.net
Referer: http://www.kame.net/
The server replies with the requested style sheet and maintains the persistent connection. Note that the server only
accepts 99 remaining HTTP requests over this persistent connection.
HTTP/1.1 200 OK
Date: Fri, 19 Mar 2010 09:23:37 GMT
Last-Modified: Mon, 10 Apr 2006 05:06:39 GMT
Content-Type: text/css
...

Then the client automatically requests the web servers icon 10 , that could be displayed by the browser. This server
does not contain such URI and thus replies with a 404 HTTP status. However, the underlying TCP connection is
not closed immediately.
GET /favicon.ico HTTP/1.1
Host: www.kame.net
Referer: http://www.kame.net/
HTTP/1.1 404 Not Found

Date: Fri, 19 Mar 2010 09:23:40 GMT
Content-Length: 318
Content-Type: text/html; charset=iso-8859-1
<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN"> ...
As illustrated above, a client can send several HTTP requests over the same persistent TCP connection. However,
it is important to note that all of these HTTP requests are considered to be independent by the server. Each HTTP
request must be self-contained. This implies that each request must include all the header lines that are required
by the server to understand the request. The independence of these requests is one of the important design choices
of HTTP. As a consequence of this design choice, when a server processes a HTTP request, it doesnt use any
other information than what is contained in the request itself. This explains why the client adds its User-Agent:
header in all of the HTTP requests it sends over the persistent TCP connection.
However, in practice, some servers want to provide content tuned for each user. For example, some servers
can provide information in several languages or other servers want to provide advertisements that are targeted to
different types of users. To do this, servers need to maintain some information about the preferences of each user
and use this information to produce content matching the users preferences. HTTP contains several mechanisms
that enable to solve this problem. We discuss three of them below.
A first solution is to force the users to be authenticated. This was the solution used by FTP to control the files that
each user could access. Initially, user names and passwords could be included inside URIs RFC 1738. However,
placing passwords in the clear in a potentially publicly visible URI is completely insecure and this usage has now
been deprecated RFC 3986. HTTP supports several extension headers RFC 2617 that can be used by a server
to request the authentication of the client by providing his/her credentials. However, user names and passwords
have not been popular on web servers as they force human users to remember one user name and one password
per server. Remembering a password is acceptable when a user needs to access protected content, but users will
not accept the need for a user name and password only to receive targeted advertisements from the web sites that
they visit.
A second solution to allow servers to tune that content to the needs and capabilities of the user is to rely on
the different types of Accept-* HTTP headers. For example, the Accept-Language: can be used by the client to
indicate its preferred languages. Unfortunately, in practice this header is usually set based on the default language
of the browser and it is not possible for a user to indicate the language it prefers to use by selecting options on
each visited web server.
The third, and widely adopted, solution are HTTP cookies. HTTP cookies were initially developed as a private
extension by Netscape. They are now part of the standard RFC 6265. In a nutshell, a cookie is a short string that
is chosen by a server to represent a given client. Two HTTP headers are used : Cookie: and Set-Cookie:. When a
server receives an HTTP request from a new client (i.e. an HTTP request that does not contain the Cookie: header),
it generates a cookie for the client and includes it in the Set-Cookie: header of the returned HTTP response. The
Set-Cookie: header contains several additional parameters including the domain names for which the cookie is
valid. The client stores all received cookies on disk and every time it sends a HTTP request, it verifies whether
it already knows a cookie for this domain. If so, it attaches the Cookie: header to the HTTP request. This is
illustrated in the figure below with HTTP 1.1, but cookies also work with HTTP 1.0.
10 Favorite icons are small icons that are used to represent web servers in the toolbar of Internet browsers. Microsoft added this feature
in their browsers without taking into account the W3C standards. See http://www.w3.org/2005/10/howto-favicon for a discussion on how to
cleanly support such favorite icons.

Figure 3.16: HTTP cookies
Note: Privacy issues with HTTP cookies

The HTTP cookies introduced by Netscape are key for large e-commerce websites. However, they have also
raised many discussions concerning their potential misuses. Consider ad.com, a company that delivers lots of
advertisements on web sites. A web site that wishes to include ad.coms advertisements next to its content will
add links to ad.com inside its HTML pages. If ad.com is used by many web sites, ad.com could be able to track the
interests of all the users that visit its client websites and use this information to provide targeted advertisements.
Privacy advocates have even sued online advertisement companies to force them to comply with the privacy
regulations. More recent related technologies also raise privacy concerns
3.5 Remote Procedure Calls
In the previous sections, we have described several protocols that enable humans to exchange messages and access
to remote documents. This is not the only usage of computer networks and in many situations applications use
the network to exchange information with other applications. When an application needs to perform a large
computation on a host, it can sometimes be useful to request computations from other hosts. Many distributed
systems have been built by distributing applications on different hosts and using Remote Procedure Calls as a
basic building block.
In traditional programming languages, procedure calls allow programmers to better structure their code. Each
procedure is identified by a name, a return type and a set of parameters. When a procedure is called, the current
flow of program execution is diverted to execute the procedure. This procedure uses the provided parameters to
perform its computation and returns one or more values. This programming model was designed with a single
host in mind. In a nutshell, most programming languages support it as follows :
1. The caller places the values of the parameters at a location (register, stack, ...) where the callee can access
them
2. The caller transfers the control of execution to the callees procedure
3. The callee accesses the parameters and performs the requested computation
4. The callee places the return value(s) at a location (register, stack, ...) where the caller can access them
5. The callee returns the control of execution to the callers
This model was developed with a single host in mind. How should it be modified if the caller and the callee are
different hosts connected through a network ? Since the two hosts can be different, the two main problems are the

fact they do not share the same memory and that they do not necessarily use the same representation for numbers,
characters, ... Let us examine how the five steps identified above can be supported through a network.
The first problem to be solved is how to transfer the information from the caller to the callee. This problem is not
simple and includes two sub-problems. The first subproblem is the encoding of the information. How to encode
the values of the parameters so that they can be transferred correctly through the network ? The second problem is
how to reach the callee through the network ? The callee is identified by a procedure name, but to use the transport
service, we need to convert this name into an address and a port number.
3.5.1 Encoding data
The encoding problem exists in a wide range of applications. In the previous sections, we have described how
character-based encodings are used by email and http. Although standard encoding techniques such as ASN.1
[Dubuisson2000] have been defined to cover most application needs, many applications have defined their specific
encoding. Remote Procedure Call are no exception to this rule. The three most popular encoding methods are
probably XDR RFC 1832 used by ONC-RPC RFC 1831, XML, used by XML-RPC and JSON RFC 4627.
The eXternal Data Representation (XDR) Standard, defined in RFC 1832 is an early specification that describes
how information exchanged during Remote Procedure Calls should be encoded before being transmitted through
a network. Since the transport service allows to transfer a block of bytes (with the connectionless service) or a
stream of bytes (by using the connection-oriented service), XDR maps each datatype onto a sequence of bytes.
The caller encodes each data in the appropriate sequence and the callee decodes the received information. Here
are a few examples extracted from RFC 1832 to illustrate how this encoding/decoding can be performed.
For basic data types, RFC 1832 simply maps their representation into a sequence of bytes. For example a 32 bits
integer is transmitted as follows (with the most significant byte first, which corresponds to big-endian encoding).
XDR also supports 64 bits integers and booleans. The booleans are mapped onto integers (0 for false and 1 for
true). For the floating point numbers, the encoding defined in the IEEE standard is used.
In this representation, the first bit (S) is the sign (0 represents positive). The next 11 bits represent the exponent of
the number (E), in base 2, and the remaining 52 bits are the fractional part of the number (F). The floating point
number that corresponds to this representation is (1)S 2E1023 1.F . XDR also allows to encode complex
data types. A first example is the string of bytes. A string of bytes is composed of two parts : a length (encoded
as an integer) and a sequence of bytes. For performance reasons, the encoding of a string is aligned to 32 bits
boundaries. This implies that some padding bytes may be inserted during the encoding operation is the length of
the string is not a multiple of 4. The structure of the string is shown below (source RFC 1832).
In some situations, it is necessary to encode fixed or variable length arrays. XDR RFC 1832 supports such
arrays. For example, the encoding below corresponds to a variable length array containing n elements. The
3.5. Remote Procedure Calls 133

encoded representation starts with an integer that contains the number of elements and follows with all elements
in sequence. It is also possible to encode a fixed-length array. In this case, the first integer is missing.
XDR also supports the definition of unions, structures, ... Additional details are provided in RFC 1832.
A second popular method to encode data is the JavaScript Object Notation (JSON). This syntax was initially
defined to allow applications written in JavaScript to exchange data, but it has now wider usages. JSON RFC
4627 is a text-based representation. The simplest data type is the integer. It is represented as a sequence of
digits in ASCII. Strings can also be encoding by using JSON. A JSON string always starts and ends with a quote
character () as in the C language. As in the C language, some characters (like or \) must be escaped if they
appear in a string. RFC 4627 describes this in details. Booleans are also supported by using the strings false and
true. Like XDR, JSON supports more complex data types. A structure or object is defined as a comma separated
list of elements enclosed in curly brackets. RFC 4627 provides the following example as an illustration.
{
"Image": {
"Width": 800,
"Height": 600,
"Title": "View from 15th Floor",
"Thumbnail": {
"Url": "http://www.example.com/image/481989943",
"Height": 125,
"Width": 100
},
"ID": 1234
}
}
This object has one field named Image. It has five attributes. The first one, Width, is an integer set to 800. The
third one is a string. The fourth attribute, Thumbnail is also an object composed of three different attributes, one
string and two integers. JSON can also be used to encode arrays or lists. In this case, square brackets are used as
delimiters. The snippet below shows an array which contains the prime integers that are smaller than ten.
{
"Primes" : [ 2, 3, 5, 7 ]
}
Compared with XDR, the main advantage of JSON is that the transfer syntax is easily readable by a human.
However, this comes at the expense of a less compact encoding. Some data encoded in JSON will usually take
more space than when it is encoded with XDR. More compact encoding schemes have been defined, see e.g.
[BH2013] and the references therein.
3.5.2 Reaching the callee
The second subproblem is how to reach the callee. A simple solution to this problem is to make sure that the
callee listens on a specific port on the remote machine and then exchange information with this server process.
This is the solution chosen for JSON-RPC [JSON-RPC2]. JSON-RPC can be used over the connectionless or the
connection-oriented transport. A JSON-RPC request contains the following information :
jsonrpc: a string indicating the version of the protocol used. This is important to allow the protocol to evolve
in the future.
method: a string that contains the name of the procedure which is invoked
params: a structure that contains the values of the parameters that are passed to the method
id: an identifier chosen by the caller

The JSON-RPC is encoded as a JSON object. For example, the example below shows an invokation of a method
called sum with 1 and 3 as parameters.
{"jsonrpc": "2.0", "method": "sum", "params": [1, 3], "id": 1}
Upon reception of this JSON structure, the callee parses the object, locates the corresponding method and passes
the parameters. This method returns a response which is also encoded as a JSON structure. This response contains
the following information :
jsonrpc: a string indicating the version of the protocol used to encode the response
id: the same identifier as the identifier chosen by the caller
result: if the request succeeded, this member contains the result of the request (in our example, value 4).
error: if the method called does not exist or its execution causes an error, the result element will be replaced
by an error element which contains the following members :
code: a number that indicates the type of error. Several error codes are defined in [JSON-RPC2]. For
example, -32700 indicates an error in parsing the request, -32602 indicates invalid parameters and
-32601 indicates that the method could not be found on the server. Other error codes are listed in
[JSON-RPC2].
message: a string (limited to one sentence) that provides a short description of the error.
data: an optional field that provides additional information about the error.
Coming back to our example with the call for the sum procedure, it would return the following JSON structure.
{ "jsonrpc": "2.0", "result": 4, "id": 1}
If the sum method is not implemented on the server, it would reply with the following response.
{ "jsonrpc": "2.0", "error": {"code": -32601, "message": "Method not found"}, "id": "1"}
The id field, which is present in the request and the response plays the same role as the identifier field in the
DNS message. It allows the caller to match the response with the request that it sent. This id is very important
when JSON-RPC is used over the connectionless service which is unreliable. If a request is sent, it may need to
be retransmitted and it is possible that a callee will receive twice the same request (e.g. if the response for the
first request was lost). In the DNS, when a request is lost, it can be retransmitted without causing any difficulty.
However with remote procedure calls in general, losses can cause some problems. Consider a method which is
used to deposit money on a bank account. If the request is lost, it will be retransmitted and the deposit will be
eventually performed. However, if the response is lost, the caller will also retransmit its request. This request will
be received by the callee that will deposit the money again. To prevent this problem from affecting the application,
either the programmer must ensure that the remote procedures that it calls can be safely called multiple times or the
application must verify whether the request has been transmitted earlier. In most deployments, the programmers
use remote methods that can be safely called multiple times without breaking the application logic.
ONC-RPC uses a more complex method to allow a caller to reach the callee. On a host, server processes can run
on different ports and given the limited number of port values (216 per host on the Internet), it is impossible to
reserve one port number for each method. The solution used in ONC-RPC RFC 1831 is to use a special method
which is called the portmapper RFC 1833. The portmapper is a kind of directory that runs on a server that
hosts methods. The portmapper runs on a standard port (111 for ONC-RPC RFC 1833). A server process that
implements a method registers its method on the local portmapper. When a caller needs to call a method on a
remote server, it first contacts the portmapper to obtain the port number of the server process which implements
the method. The response from the portmapper allows it to directly contact the server process which implements
the method.
3.6 Internet transport protocols
3.6. Internet transport protocols 135

Transport protocols rely on the service provided by the network layer. On the Internet, the network layer provides
a connectionless service. The network layer identifies each (interface of a) host by using an IP address. It enables
hosts to transmit packets that contain up to 64 KBytes of payload to any destination reachable through the network.
The network layer does not guarantee the delivery of information, cannot detect transmission errors and does not
preserve sequence integrity.
Several transport protocols have been designed to provide a richer service to the applications. The two most
widely deployed transport protocols on the Internet are the User Datagram Protocol (UDP) and the Transmission
Control Protocol (TCP). A third important transport protocol, the Stream Control Transmission Protocol (SCTP)
RFC 4960 appeared in the early 2000s. It is currently used by some particular applications such as signaling
in Voice over IP networks. We also describe SCTP in this section to present a different design than TCP. The
Real Time Transport Protocol (RTP), defined in RFC 3550 is another important protocol that is used by many
multimedia applications. It includes functions that belong to the transport layer, but also functions that are related
to the encoding of the information. Due to space limitations, we do not discuss it in details in this section.
3.7 The User Datagram Protocol
The User Datagram Protocol (UDP) is defined in RFC 768. It provides an unreliable connectionless transport
service on top of the unreliable network layer connectionless service. The main characteristics of the UDP service
are :
the UDP service cannot deliver SDUs that are larger than 65467 bytes 11
the UDP service does not guarantee the delivery of SDUs (losses and desequencing can occur)
the UDP service will not deliver a corrupted SDU to the destination
Compared to the connectionless network layer service, the main advantage of the UDP service is that it allows
several applications running on a host to exchange SDUs with several other applications running on remote hosts.
Let us consider two hosts, e.g. a client and a server. The network layer service allows the client to send information
to the server, but if an application running on the client wants to contact a particular application running on the
server, then an additional addressing mechanism is required other than the IP address that identifies a host, in order
to differentiate the application running on a host. This additional addressing is provided by port numbers. When
a server using UDP is enabled on a host, this server registers a port number. This port number will be used by the
clients to contact the server process via UDP.
The figure below shows a typical usage of the UDP port numbers. The client process uses port number 1234 while
the server process uses port number 5678. When the client sends a request, it is identified as originating from port
number 1234 on the client host and destined to port number 5678 on the server host. When the server process
replies to this request, the servers UDP implementation will send the reply as originating from port 5678 on the
server host and destined to port 1234 on the client host.
Figure 3.17: Usage of the UDP port numbers
11 This limitation is due to the fact that the network layer cannot transport packets that are larger than 64 KBytes. As UDP does not include
any segmentation/reassembly mechanism, it cannot split a SDU before sending it. The UDP header consumes 8 bytes and the IPv6 header 60.
With IPv4, the IPv4 header only consumes 20 bytes and thus the maximum UDP payload size is 65507 bytes.

UDP uses a single segment format shown in the figure below.
Figure 3.18: UDP Header Format
The UDP header contains four fields :

a 16 bits source port
a 16 bits destination port
a 16 bits length field
a 16 bits checksum
As the port numbers are encoded as a 16 bits field, there can be up to only 65535 different server processes that are
bound to a different UDP port at the same time on a given server. In practice, this limit is never reached. However,
it is worth noticing that most implementations divide the range of allowed UDP port numbers into three different
ranges :
the privileged port numbers (1 < port < 1024 )
the ephemeral port numbers ( officially 12 49152 <= port <= 65535 )
the registered port numbers (officially 1024 <= port < 49152)
In most Unix variants, only processes having system administrator privileges can be bound to port numbers smaller
than 1024. Well-known servers such as DNS, NTP or RPC use privileged port numbers. When a client needs to
use UDP, it usually does not require a specific port number. In this case, the UDP implementation will allocate
the first available port number in the ephemeral range. The range of registered port numbers should be used by
servers. In theory, developers of network servers should register their port number officially through IANA, but
few developers do this.
Note: Computation of the UDP checksum

The checksum of the UDP segment is computed over :
a pseudo header RFC 2460 containing the source address, the destination address, the packet length encoded
as a 32 bits number and a 32 bits bit field containing the three most significant bytes set to 0 and the low
order byte set to 17
the entire UDP segment, including its header
This pseudo-header allows the receiver to detect errors affecting the source or destination addresses placed in
the IP layer below. This is a violation of the layering principle that dates from the time when UDP and IP were
elements of a single protocol. It should be noted that if the checksum algorithm computes value 0x0000, then
value 0xffff is transmitted. A UDP segment whose checksum is set to 0x0000 is a segment for which the
transmitter did not compute a checksum upon transmission. Some NFS servers chose to disable UDP checksums
for performance reasons when running over IPv4, but this caused problems that were difficult to diagnose. Over
IPv6, the UDP checksum cannot be disabled. A detailed discussion of the implementation of the Internet checksum
may be found in RFC 1071
Several types of applications rely on UDP. As a rule of thumb, UDP is used for applications where delay must
be minimised or losses can be recovered by the application itself. A first class of the UDP-based applications are
applications where the client sends a short request and expects a quick and short answer. The DNS is an example of
a UDP application that is often used in the wide area. However, in local area networks, many distributed systems
12 A discussion of the ephemeral port ranges used by different TCP/UDP implementations may be found in
http://www.ncftp.com/ncftpd/doc/misc/ephemeral_ports.html
3.7. The User Datagram Protocol 137

rely on Remote Procedure Call (RPC) that is often used on top of UDP. In Unix environments, the Network File
System (NFS) is built on top of RPC and runs frequently on top of UDP. A second class of UDP-based applications
are the interactive computer games that need to frequently exchange small messages, such as the players location
or their recent actions. Many of these games use UDP to minimise the delay and can recover from losses. A
third class of applications are multimedia applications such as interactive Voice over IP or interactive Video over
IP. These interactive applications expect a delay shorter than about 200 milliseconds between the sender and the
receiver and can recover from losses directly inside the application.
3.8 The Transmission Control Protocol
The Transmission Control Protocol (TCP) was initially defined in RFC 793. Several parts of the protocol have
been improved since the publication of the original protocol specification 13 . However, the basics of the protocol
remain and an implementation that only supports RFC 793 should inter-operate with todays implementation.
TCP provides a reliable bytestream, connection-oriented transport service on top of the unreliable connectionless
network service provided by IP. TCP is used by a large number of applications, including :
Email (SMTP, POP, IMAP)
World wide web ( HTTP, ...)
Most file transfer protocols ( ftp, peer-to-peer file sharing applications , ...)
remote computer access : telnet, ssh, X11, VNC, ...
non-interactive multimedia applications : flash
On the global Internet, most of the applications used in the wide area rely on TCP. Many studies 14 have reported
that TCP was responsible for more than 90% of the data exchanged in the global Internet.
To provide this service, TCP relies on a simple segment format that is shown in the figure below. Each TCP
segment contains a header described below and, optionally, a payload. The default length of the TCP header is
twenty bytes, but some TCP headers contain options.
Figure 3.19: TCP header format
A TCP header contains the following fields :

Source and destination ports. The source and destination ports play an important role in TCP, as they
allow the identification of the connection to which a TCP segment belongs. When a client opens a TCP
connection, it typically selects an ephemeral TCP port number as its source port and contacts the server by
using the servers port number. All the segments that are sent by the client on this connection have the same
source and destination ports. The server sends segments that contain as source (resp. destination) port, the
destination (resp. source) port of the segments sent by the client (see figure Utilization of the TCP source
and destination ports). A TCP connection is always identified by four pieces of information :
13 A detailed presentation of all standardisation documents concerning TCP may be found in RFC 4614
14 Several researchers have analysed the utilisation of TCP and UDP in the global Internet. Most of these stud-
ies have been performed by collecting all the packets transmitted over a given link during a period of a few hours
or days and then analysing their headers to infer the transport protocol used, the type of application, ... Recent
studies include http://www.caida.org/research/traffic-analysis/tcpudpratio/, https://research.sprintlabs.com/packstat/packetoverview.php or
http://www.nanog.org/meetings/nanog43/presentations/Labovitz_internetstats_N43.pdf

the address of the client

the address of the server
the port chosen by the client
the port chosen by the server
the sequence number (32 bits), acknowledgement number (32 bits) and window (16 bits) fields are used
to provide a reliable data transfer, using a window-based protocol. In a TCP bytestream, each byte of the
stream consumes one sequence number. Their utilisation will be described in more detail in section TCP
reliable data transfer
the Urgent pointer is used to indicate that some data should be considered as urgent in a TCP bytestream.
However, it is rarely used in practice and will not be described here. Additional details about the utilisation
of this pointer may be found in RFC 793, RFC 1122 or [Stevens1994]
the flags field contains a set of bit flags that indicate how a segment should be interpreted by the TCP entity
receiving it :
the SYN flag is used during connection establishment
the FIN flag is used during connection release
the RST is used in case of problems or when an invalid segment has been received
when the ACK flag is set, it indicates that the acknowledgment field contains a valid number. Other-
wise, the content of the acknowledgment field must be ignored by the receiver
the URG flag is used together with the Urgent pointer
the PSH flag is used as a notification from the sender to indicate to the receiver that it should pass all
the data it has received to the receiving process. However, in practice TCP implementations do not
allow TCP users to indicate when the PSH flag should be set and thus there are few real utilizations of
this flag.
the checksum field contains the value of the Internet checksum computed over the entire TCP segment and
a pseudo-header as with UDP
the Reserved field was initially reserved for future utilization. It is now used by RFC 3168.
the TCP Header Length (THL) or Data Offset field is a four bits field that indicates the size of the TCP
header in 32 bit words. The maximum size of the TCP header is thus 64 bytes.
the Optional header extension is used to add optional information to the TCP header. Thanks to this header
extension, it is possible to add new fields to the TCP header that were not planned in the original specifi-
cation. This allowed TCP to evolve since the early eighties. The details of the TCP header extension are
explained in sections TCP connection establishment and TCP reliable data transfer.
Figure 3.20: Utilization of the TCP source and destination ports
3.8. The Transmission Control Protocol 139

The rest of this section is organised as follows. We first explain the establishment and the release of a TCP
connection, then we discuss the mechanisms that are used by TCP to provide a reliable bytestream service. We
end the section with a discussion of network congestion and explain the mechanisms that TCP uses to avoid
congestion collapse.
3.8.1 TCP connection establishment
A TCP connection is established by using a three-way handshake. The connection establishment phase uses the
sequence number, the acknowledgment number and the SYN flag. When a TCP connection is established, the two
communicating hosts negotiate the initial sequence number to be used in both directions of the connection. For
this, each TCP entity maintains a 32 bits counter, which is supposed to be incremented by one at least every 4
microseconds and after each connection establishment 15 . When a client host wants to open a TCP connection
with a server host, it creates a TCP segment with :
the SYN flag set
the sequence number set to the current value of the 32 bits counter of the client hosts TCP entity
Upon reception of this segment (which is often called a SYN segment), the server host replies with a segment
containing :
the SYN flag set
the sequence number set to the current value of the 32 bits counter of the server hosts TCP entity
the ACK flag set
the acknowledgment number set to the sequence number of the received SYN segment incremented by 1
(mod 232 ). When a TCP entity sends a segment having x+1 as acknowledgment number, this indicates that
it has received all data up to and including sequence number x and that it is expecting data having sequence
number x+1. As the SYN flag was set in a segment having sequence number x, this implies that setting the
SYN flag in a segment consumes one sequence number.
This segment is often called a SYN+ACK segment. The acknowledgment confirms to the client that the server has
correctly received the SYN segment. The sequence number of the SYN+ACK segment is used by the server host to
verify that the client has received the segment. Upon reception of the SYN+ACK segment, the client host replies
with a segment containing :
the ACK flag set
the acknowledgment number set to the sequence number of the received SYN+ACK segment incremented
by 1 (mod 232 )
At this point, the TCP connection is open and both the client and the server are allowed to send TCP segments
containing data. This is illustrated in the figure below.
Figure 3.21: Establishment of a TCP connection

15 This 32 bits counter was specified in RFC 793. A 32 bits counter that is incremented every 4 microseconds wraps in about 4.5 hours.
This period is much larger than the Maximum Segment Lifetime that is fixed at 2 minutes in the Internet (RFC 791, RFC 1122).

In the figure above, the connection is considered to be established by the client once it has received the SYN+ACK
segment, while the server considers the connection to be established upon reception of the ACK segment. The first
data segment sent by the client (server) has its sequence number set to x+1 (resp. y+1).
Note: Computing TCPs initial sequence number

In the original TCP specification RFC 793, each TCP entity maintained a clock to compute the initial sequence
number (ISN) placed in the SYN and SYN+ACK segments. This made the ISN predictable and caused a security
issue. The typical security problem was the following. Consider a server that trusts a host based on its IP address
and allows the system administrator to login from this host without giving a password 16 . Consider now an attacker
who knows this particular configuration and is able to send IP packets having the clients address as source. He
can send fake TCP segments to the server, but does not receive the servers answers. If he can predict the ISN that
is chosen by the server, he can send a fake SYN segment and shortly after the fake ACK segment confirming the
reception of the SYN+ACK segment sent by the server. Once the TCP connection is open, he can use it to send
any command to the server. To counter this attack, current TCP implementations add randomness to the ISN. One
of the solutions, proposed in RFC 1948 is to compute the ISN as
ISN = M + H(localhost, localport, remotehost, remoteport, secret).
where M is the current value of the TCP clock and H is a cryptographic hash function. localhost and remotehost
(resp. localport and remoteport ) are the IP addresses (port numbers) of the local and remote host and secret is a
random number only known by the server. This method allows the server to use different ISNs for different clients
at the same time. Measurements performed with the first implementations of this technique showed that it was
difficult to implement it correctly, but todays TCP implementation now generate good ISNs.
A server could, of course, refuse to open a TCP connection upon reception of a SYN segment. This refusal may be
due to various reasons. There may be no server process that is listening on the destination port of the SYN segment.
The server could always refuse connection establishments from this particular client (e.g. due to security reasons)
or the server may not have enough resources to accept a new TCP connection at that time. In this case, the server
would reply with a TCP segment having its RST flag set and containing the sequence number of the received SYN
segment as its acknowledgment number. This is illustrated in the figure below. We discuss the other utilizations
of the TCP RST flag later (see TCP connection release).
Figure 3.22: TCP connection establishment rejected by peer
TCP connection establishment can be described as the four state Finite State Machine shown below. In this FSM,
!X (resp. ?Y) indicates the transmission of segment X (resp. reception of segment Y) during the corresponding
transition. Init is the initial state.
A client host starts in the Init state. It then sends a SYN segment and enters the SYN Sent state where it waits
for a SYN+ACK segment. Then, it replies with an ACK segment and enters the Established state where data can
be exchanged. On the other hand, a server host starts in the Init state. When a server process starts to listen to
a destination port, the underlying TCP entity creates a TCP control block and a queue to process incoming SYN
segments. Upon reception of a SYN segment, the servers TCP entity replies with a SYN+ACK and enters the SYN
16 On many departmental networks containing Unix workstations, it was common to allow users on one of the hosts to use rlogin RFC 1258
to run commands on any of the workstations of the network without giving any password. In this case, the remote workstation authenticated
the client host based on its IP address. This was a bad practice from a security viewpoint.

Figure 3.23: TCP FSM for connection establishment
RCVD state. It remains in this state until it receives an ACK segment that acknowledges its SYN+ACK segment,
with this it then enters the Established state.
Apart from these two paths in the TCP connection establishment FSM, there is a third path that corresponds to the
case when both the client and the server send a SYN segment to open a TCP connection 17 . In this case, the client
and the server send a SYN segment and enter the SYN Sent state. Upon reception of the SYN segment sent by the
other host, they reply by sending a SYN+ACK segment and enter the SYN RCVD state. The SYN+ACK that arrives
from the other host allows it to transition to the Established state. The figure below illustrates such a simultaneous
establishment of a TCP connection.
Figure 3.24: Simultaneous establishment of a TCP connection
17 Of course, such a simultaneous TCP establishment can only occur if the source port chosen by the client is equal to the destination
port chosen by the server. This may happen when a host can serve both as a client as a server or in peer-to-peer applications when the
communicating hosts do not use ephemeral port numbers.

Denial of Service attacks

When a TCP entity opens a TCP connection, it creates a Transmission Control Block (TCB). The TCB
contains the entire state that is maintained by the TCP entity for each TCP connection. During connection
establishment, the TCB contains the local IP address, the remote IP address, the local port number, the
remote port number, the current local sequence number, the last sequence number received from the remote
entity. Until the mid 1990s, TCP implementations had a limit on the number of TCP connections that could
be in the SYN RCVD state at a given time. Many implementations set this limit to about 100 TCBs. This limit
was considered sufficient even for heavily load http servers given the small delay between the reception of a
SYN segment and the reception of the ACK segment that terminates the establishment of the TCP connection.
When the limit of 100 TCBs in the SYN Rcvd state is reached, the TCP entity discards all received TCP SYN
segments that do not correspond to an existing TCB.
This limit of 100 TCBs in the SYN Rcvd state was chosen to protect the TCP entity from the risk of overload-
ing its memory with too many TCBs in the SYN Rcvd state. However, it was also the reason for a new type of
Denial of Service (DoS) attack RFC 4987. A DoS attack is defined as an attack where an attacker can render
a resource unavailable in the network. For example, an attacker may cause a DoS attack on a 2 Mbps link
used by a company by sending more than 2 Mbps of packets through this link. In this case, the DoS attack
was more subtle. As a TCP entity discards all received SYN segments as soon as it has 100 TCBs in the SYN
Rcvd state, an attacker simply had to send a few 100 SYN segments every second to a server and never reply
to the received SYN+ACK segments. To avoid being caught, attackers were of course sending these SYN
segments with a different address than their own IP address a . On most TCP implementations, once a TCB
entered the SYN Rcvd state, it remained in this state for several seconds, waiting for a retransmission of the
initial SYN segment. This attack was later called a SYN flood attack and the servers of the ISP named panix
were among the first to be affected by this attack.
To avoid the SYN flood attacks, recent TCP implementations no longer enter the SYN Rcvd state upon recep-
tion of a SYN segment. Instead, they reply directly with a SYN+ACK segment and wait until the reception
of a valid ACK. This implementation trick is only possible if the TCP implementation is able to verify that
the received ACK segment acknowledges the SYN+ACK segment sent earlier without storing the initial se-
quence number of this SYN+ACK segment in a TCB. The solution to solve this problem, which is known as
SYN cookies is to compute the 32 bits of the ISN as follows :
the high order bits contain the low order bits of a counter that is incremented slowly
the low order bits contain a hash value computed over the local and remote IP addresses and ports and
a random secret only known to the server
The advantage of the SYN cookies is that by using them, the server does not need to create a TCB upon
reception of the SYN segment and can still check the returned ACK segment by recomputing the SYN cookie.
The main disadvantage is that they are not fully compatible with the TCP options. This is why they are not
enabled by default on a typical system.
a Sending a packet with a different source IP address than the address allocated to the host is called sending a spoofed packet.
Retransmitting the first SYN segment

As IP provides an unreliable connectionless service, the SYN and SYN+ACK segments sent to open a TCP
connection could be lost. Current TCP implementations start a retransmission timer when they send the first
SYN segment. This timer is often set to three seconds for the first retransmission and then doubles after each
retransmission RFC 2988. TCP implementations also enforce a maximum number of retransmissions for
the initial SYN segment.
As explained earlier, TCP segments may contain an optional header extension. In the SYN and SYN+ACK seg-
ments, these options are used to negotiate some parameters and the utilisation of extensions to the basic TCP
specification.
The first parameter which is negotiated during the establishment of a TCP connection is the Maximum Segment
Size (MSS). The MSS is the size of the largest segment that a TCP entity is able to process. According to RFC
879, all TCP implementations must be able to receive TCP segments containing 536 bytes of payload. However,
most TCP implementations are able to process larger segments. Such TCP implementations use the TCP MSS
Option in the SYN/SYN+ACK segment to indicate the largest segment they are able to process. The MSS value
indicates the maximum size of the payload of the TCP segments. The client (resp. server) stores in its TCB the

MSS value announced by the server (resp. the client).

Another utilisation of TCP options during connection establishment is to enable TCP extensions. For example,
consider RFC 1323 (which is discussed in TCP reliable data transfer). RFC 1323 defines TCP extensions to
support timestamps and larger windows. If the client supports RFC 1323, it adds a RFC 1323 option to its SYN
segment. If the server understands this RFC 1323 option and wishes to use it, it replies with an RFC 1323
option in the SYN+ACK segment and the extension defined in RFC 1323 is used throughout the TCP connection.
Otherwise, if the servers SYN+ACK does not contain the RFC 1323 option, the client is not allowed to use this
extension and the corresponding TCP header options throughout the TCP connection. TCPs option mechanism
is flexible and it allows the extension of TCP while maintaining compatibility with older implementations.
The TCP options are encoded by using a Type Length Value format where :
the first byte indicates the type of the option.
the second byte indicates the total length of the option (including the first two bytes) in bytes
the last bytes are specific for each type of option
RFC 793 defines the Maximum Segment Size (MSS) TCP option that must be understood by all TCP implemen-
tations. This option (type 2) has a length of 4 bytes and contains a 16 bits word that indicates the MSS supported
by the sender of the SYN segment. The MSS option can only be used in TCP segments having the SYN flag set.
RFC 793 also defines two special options that must be supported by all TCP implementations. The first option
is End of option. It is encoded as a single byte having value 0x00 and can be used to ensure that the TCP header
extension ends on a 32 bits boundary. The No-Operation option, encoded as a single byte having value 0x01, can
be used when the TCP header extension contains several TCP options that should be aligned on 32 bit boundaries.
All other options 18 are encoded by using the TLV format.
Note: The robustness principle

The handling of the TCP options by TCP implementations is one of the many applications of the robustness
principle which is usually attributed to Jon Postel and is often quoted as Be liberal in what you accept, and
conservative in what you send RFC 1122
Concerning the TCP options, the robustness principle implies that a TCP implementation should be able to accept
TCP options that it does not understand, in particular in received SYN segments, and that it should be able to parse
any received segment without crashing, even if the segment contains an unknown TCP option. Furthermore, a
server should not send in the SYN+ACK segment or later, options that have not been proposed by the client in the
SYN segment.
3.8.2 TCP reliable data transfer
The original TCP data transfer mechanisms were defined in RFC 793. Based on the experience of using TCP
on the growing global Internet, this part of the TCP specification has been updated and improved several times,
always while preserving the backward compatibility with older TCP implementations. In this section, we review
the main data transfer mechanisms used by TCP.
TCP is a window-based transport protocol that provides a bi-directional byte stream service. This has several
implications on the fields of the TCP header and the mechanisms used by TCP. The three fields of the TCP header
are :
sequence number. TCP uses a 32 bits sequence number. The sequence number placed in the header of a
TCP segment containing data is the sequence number of the first byte of the payload of the TCP segment.
acknowledgement number. TCP uses cumulative positive acknowledgements. Each TCP segment contains
the sequence number of the next byte that the sender of the acknowledgement expects to receive from the
remote host. In theory, the acknowledgement number is only valid if the ACK flag of the TCP header is set.
In practice almost all 19 TCP segments have their ACK flag set.
18 The full list of all TCP options may be found at http://www.iana.org/assignments/tcp-parameters/
19 In practice, only the SYN segment do not have their ACK flag set.

window. a TCP receiver uses this 16 bits field to indicate the current size of its receive window expressed
in bytes.
Note: The Transmission Control Block

For each established TCP connection, a TCP implementation must maintain a Transmission Control Block (TCB).
A TCB contains all the information required to send and receive segments on this connection RFC 793. This
includes 20 :
the local IP address
the remote IP address
the local TCP port number
the remote TCP port number
the current state of the TCP FSM
the maximum segment size (MSS)
snd.nxt : the sequence number of the next byte in the byte stream (the first byte of a new data segment that
you send uses this sequence number)
snd.una : the earliest sequence number that has been sent but has not yet been acknowledged
snd.wnd : the current size of the sending window (in bytes)
rcv.nxt : the sequence number of the next byte that is expected to be received from the remote host
rcv.wnd : the current size of the receive window advertised by the remote host
sending buffer : a buffer used to store all unacknowledged data
receiving buffer : a buffer to store all data received from the remote host that has not yet been delivered
to the user. Data may be stored in the receiving buffer because either it was not received in sequence or
because the user is too slow to process it
The original TCP specification can be categorised as a transport protocol that provides a byte stream service and
uses go-back-n.
To send new data on an established connection, a TCP entity performs the following operations on the correspond-
ing TCB. It first checks that the sending buffer does not contain more data than the receive window advertised by
the remote host (rcv.wnd). If the window is not full, up to MSS bytes of data are placed in the payload of a TCP
segment. The sequence number of this segment is the sequence number of the first byte of the payload. It is set to
the first available sequence number : snd.nxt and snd.nxt is incremented by the length of the payload of the TCP
segment. The acknowledgement number of this segment is set to the current value of rcv.nxt and the window field
of the TCP segment is computed based on the current occupancy of the receiving buffer. The data is kept in the
sending buffer in case it needs to be retransmitted later.
When a TCP segment with the ACK flag set is received, the following operations are performed. rcv.wnd is set
to the value of the window field of the received segment. The acknowledgement number is compared to snd.una.
The newly acknowledged data is removed from the sending buffer and snd.una is updated. If the TCP segment
contained data, the sequence number is compared to rcv.nxt. If they are equal, the segment was received in
sequence and the data can be delivered to the user and rcv.nxt is updated. The contents of the receiving buffer is
checked to see whether other data already present in this buffer can be delivered in sequence to the user. If so,
rcv.nxt is updated again. Otherwise, the segments payload is placed in the receiving buffer.
Segment transmission strategies
In a transport protocol such as TCP that offers a bytestream, a practical issue that was left as an implementation
choice in RFC 793 is to decide when a new TCP segment containing data must be sent. There are two simple and
20 A complete TCP implementation contains additional information in its TCB, notably to support the urgent pointer. However, this part of
TCP is not discussed in this book. Refer to RFC 793 and RFC 2140 for more details about the TCB.

extreme implementation choices. The first implementation choice is to send a TCP segment as soon as the user
has requested the transmission of some data. This allows TCP to provide a low delay service. However, if the
user is sending data one byte at a time, TCP would place each user byte in a segment containing 20 bytes of TCP
header 21 . This is a huge overhead that is not acceptable in wide area networks. A second simple solution would
be to only transmit a new TCP segment once the user has produced MSS bytes of data. This solution reduces the
overhead, but at the cost of a potentially very high delay.
An elegant solution to this problem was proposed by John Nagle in RFC 896. John Nagle observed that the
overhead caused by the TCP header was a problem in wide area connections, but less in local area connections
where the available bandwidth is usually higher. He proposed the following rules to decide to send a new data
segment when a new data has been produced by the user or a new ack segment has been received
if rcv.wnd>= MSS and len(data) >= MSS :
send one MSS-sized segment
else
if there are unacknowledged data:
place data in buffer until acknowledgement has been received
else
send one TCP segment containing all buffered data
The first rule ensures that a TCP connection used for bulk data transfer always sends full TCP segments. The
second rule sends one partially filled TCP segment every round-trip-time.
This algorithm, called the Nagle algorithm, takes a few lines of code in all TCP implementations. These lines of
code have a huge impact on the packets that are exchanged in TCP/IP networks. Researchers have analysed the
distribution of the packet sizes by capturing and analysing all the packets passing through a given link. These
studies have shown several important results :
in TCP/IP networks, a large fraction of the packets are TCP segments that contain only an acknowledgement.
These packets usually account for 40-50% of the packets passing through the studied link
in TCP/IP networks, most of the bytes are exchanged in long packets, usually packets containing about 1440
bytes of payload which is the default MSS for hosts attached to an Ethernet network, the most popular type
of LAN
Recent measurements indicate that these packet size distributions are still valid in todays Internet, although the
packet distribution tends to become bimodal with small packets corresponding to TCP pure acks and large 1440-
bytes packets carrying most of the user data [SMASU2012].
3.8.3 TCP windows
From a performance point of view, one of the main limitations of the original TCP specification is the 16 bits
window field in the TCP header. As this field indicates the current size of the receive window in bytes, it limits the
TCP receive window at 65535 bytes. This limitation was not a severe problem when TCP was designed since at
that time high-speed wide area networks offered a maximum bandwidth of 56 kbps. However, in todays network,
this limitation is not acceptable anymore. The table below provides the rough 22 maximum throughput that can be
achieved by a TCP connection with a 64 KBytes window in function of the connections round-trip-time
RTT Maximum Throughput
1 msec 524 Mbps
10 msec 52.4 Mbps
100 msec 5.24 Mbps
500 msec 1.05 Mbps
To solve this problem, a backward compatible extension that allows TCP to use larger receive windows was
proposed in RFC 1323. Today, most TCP implementations support this option. The basic idea is that instead of
storing snd.wnd and rcv.wnd as 16 bits integers in the TCB, they should be stored as 32 bits integers. As the TCP
segment header only contains 16 bits to place the window field, it is impossible to copy the value of snd.wnd in
21 This TCP segment is then placed in an IP header. We describe IPv6 in the next chapter. The minimum size of the IPv6 (resp. IPv4)
header is 40 bytes (resp. 20 bytes).

22 A precise estimation of the maximum bandwidth that can be achieved by a TCP connection should take into account the overhead of the
TCP and IP headers as well.

each sent TCP segment. Instead the header contains snd.wnd >> S where S is the scaling factor ( 0 S 14)
negotiated during connection establishment. The client adds its proposed scaling factor as a TCP option in the
SYN segment. If the server supports RFC 1323, it places in the SYN+ACK segment the scaling factor that it uses
when advertising its own receive window. The local and remote scaling factors are included in the TCB. If the
server does not support RFC 1323, it ignores the received option and no scaling is applied.
By using the window scaling extensions defined in RFC 1323, TCP implementations can use a receive buffer
of up to 1 GByte. With such a receive buffer, the maximum throughput that can be achieved by a single TCP
connection becomes :
RTT Maximum Throughput
1 msec 8590 Gbps
10 msec 859 Gbps
100 msec 86 Gbps
500 msec 17 Gbps
These throughputs are acceptable in todays networks. However, there are already servers having 10 Gbps in-
terfaces... Early TCP implementations had fixed receiving and sending buffers 23 . Todays high performance
implementations are able to automatically adjust the size of the sending and receiving buffer to better support high
bandwidth flows [SMM1998]
3.8.4 TCPs retransmission timeout
In a go-back-n transport protocol such as TCP, the retransmission timeout must be correctly set in order to achieve
good performance. If the retransmission timeout expires too early, then bandwidth is wasted by retransmitting
segments that have already been correctly received; whereas if the retransmission timeout expires too late, then
bandwidth is wasted because the sender is idle waiting for the expiration of its retransmission timeout.
A good setting of the retransmission timeout clearly depends on an accurate estimation of the round-trip-time of
each TCP connection. The round-trip-time differs between TCP connections, but may also change during the
lifetime of a single connection. For example, the figure below shows the evolution of the round-trip-time between
two hosts during a period of 45 seconds.
Figure 3.25: Evolution of the round-trip-time between two hosts
The easiest solution to measure the round-trip-time on a TCP connection is to measure the delay between the
transmission of a data segment and the reception of a corresponding acknowledgement 24 . As illustrated in the
figure below, this measurement works well when there are no segment losses.
23 See http://fasterdata.es.net/tuning.html for more information on how to tune a TCP implementation
24 In theory, a TCP implementation could store the timestamp of each data segment transmitted and compute a new estimate for the round-
trip-time upon reception of the corresponding acknowledgement. However, using such frequent measurements introduces a lot of noise in
practice and many implementations still measure the round-trip-time once per round-trip-time by recording the transmission time of one
segment at a time RFC 2988

Figure 3.26: How to measure the round-trip-time ?
However, when a data segment is lost, as illustrated in the bottom part of the figure, the measurement is ambiguous
as the sender cannot determine whether the received acknowledgement was triggered by the first transmission of
segment 123 or its retransmission. Using incorrect round-trip-time estimations could lead to incorrect values of
the retransmission timeout. For this reason, Phil Karn and Craig Partridge proposed, in [KP91], to ignore the
round-trip-time measurements performed during retransmissions.
To avoid this ambiguity in the estimation of the round-trip-time when segments are retransmitted, recent TCP
implementations rely on the timestamp option defined in RFC 1323. This option allows a TCP sender to place
two 32 bit timestamps in each TCP segment that it sends. The first timestamp, TS Value (TSval) is chosen by the
sender of the segment. It could for example be the current value of its real-time clock 25 . The second value, TS
Echo Reply (TSecr), is the last TSval that was received from the remote host and stored in the TCB. The figure
below shows how the utilization of this timestamp option allows for the disambiguation of the round-trip-time
measurement when there are retransmissions.
Figure 3.27: Disambiguating round-trip-time measurements with the RFC 1323 timestamp option
Once the round-trip-time measurements have been collected for a given TCP connection, the TCP entity must
compute the retransmission timeout. As the round-trip-time measurements may change during the lifetime of a
connection, the retransmission timeout may also change. At the beginning of a connection 26 , the TCP entity that
sends a SYN segment does not know the round-trip-time to reach the remote host and the initial retransmission
timeout is usually set to 3 seconds RFC 2988.
The original TCP specification proposed in RFC 793 to include two additional variables in the TCB :
srtt : the smoothed round-trip-time computed as srrt = ( srtt) + ((1 ) rtt) where rtt is the
round-trip-time measured according to the above procedure and a smoothing factor (e.g. 0.8 or 0.9)
rto : the retransmission timeout is computed as rto = min(60, max(1, srtt)) where is used to take
25 Some security experts have raised concerns that using the real-time clock to set the TSval in the timestamp option can leak information
such as the systems up-time. Solutions proposed to solve this problem may be found in [CNPI09]
26 As a TCP client often establishes several parallel or successive connections with the same server, RFC 2140 has proposed to reuse for
a new connection some information that was collected in the TCB of a previous connection, such as the measured rtt. However, this solution
has not been widely implemented.

into account the delay variance (value : 1.3 to 2.0). The 60 and 1 constants are used to ensure that the rto is
not larger than one minute nor smaller than 1 second.
However, in practice, this computation for the retransmission timeout did not work well. The main problem was
that the computed rto did not correctly take into account the variations in the measured round-trip-time. Van Ja-
cobson proposed in his seminal paper [Jacobson1988] an improved algorithm to compute the rto and implemented
it in the BSD Unix distribution. This algorithm is now part of the TCP standard RFC 2988.
Jacobsons algorithm uses two state variables, srtt the smoothed rtt and rttvar the estimation of the variance of
the rtt and two parameters : and . When a TCP connection starts, the first rto is set to 3 seconds. When a first
estimation of the rtt is available, the srtt, rttvar and rto are computed as follows :
srtt=rtt
rttvar=rtt/2
rto=srtt+4*rttvar
Then, when other rtt measurements are collected, srtt and rttvar are updated as follows :
rttvar = (1 ) rttvar + |srtt rtt|
srtt = (1 ) srtt + rtt
rto = srtt + 4 rttvar
The proposed values for the parameters are = 81 and = 41 . This allows a TCP implementation, implemented
in the kernel, to perform the rtt computation by using shift operations instead of the more costly floating point
operations [Jacobson1988]. The figure below illustrates the computation of the rto upon rtt changes.
Figure 3.28: Example computation of the rto
3.8.5 Advanced retransmission strategies
The default go-back-n retransmission strategy was defined in RFC 793. When the retransmission timer expires,
TCP retransmits the first unacknowledged segment (i.e. the one having sequence number snd.una). After each
expiration of the retransmission timeout, RFC 2988 recommends to double the value of the retransmission time-
out. This is called an exponential backoff. This doubling of the retransmission timeout after a retransmission
was included in TCP to deal with issues such as network/receiver overload and incorrect initial estimations of the
retransmission timeout. If the same segment is retransmitted several times, the retransmission timeout is doubled
after every retransmission until it reaches a configured maximum. RFC 2988 suggests a maximum retransmission
timeout of at least 60 seconds. Once the retransmission timeout reaches this configured maximum, the remote
host is considered to be unreachable and the TCP connection is closed.
This retransmission strategy has been refined based on the experience of using TCP on the Internet. The first
refinement was a clarification of the strategy used to send acknowledgements. As TCP uses piggybacking, the
easiest and less costly method to send acknowledgements is to place them in the data segments sent in the other

direction. However, few application layer protocols exchange data in both directions at the same time and thus this
method rarely works. For an application that is sending data segments in one direction only, the remote TCP entity
returns empty TCP segments whose only useful information is their acknowledgement number. This may cause
a large overhead in wide area network if a pure ACK segment is sent in response to each received data segment.
Most TCP implementations use a delayed acknowledgement strategy. This strategy ensures that piggybacking is
used whenever possible, otherwise pure ACK segments are sent for every second received data segments when
there are no losses. When there are losses or reordering, ACK segments are more important for the sender and
they are sent immediately RFC 813 RFC 1122. This strategy relies on a new timer with a short delay (e.g. 50
milliseconds) and one additional flag in the TCB. It can be implemented as follows :
reception of a data segment:
if pkt.seq==rcv.nxt: # segment received in sequence
if delayedack :
send pure ack segment
cancel acktimer
delayedack=False
else:
delayedack=True
start acktimer
else: # out of sequence segment
if delayedack:
delayedack=False
cancel acktimer
transmission of a data segment: # piggyback ack

if delayedack:
delayedack=False
cancel acktimer
acktimer expiration:
delayedack=False
Due to this delayed acknowledgement strategy, during a bulk transfer, a TCP implementation usually acknowl-
edges every second TCP segment received.
The default go-back-n retransmission strategy used by TCP has the advantage of being simple to implement, in
particular on the receiver side, but when there are losses, a go-back-n strategy provides a lower performance than
a selective repeat strategy. The TCP developers have designed several extensions to TCP to allow it to use a
selective repeat strategy while maintaining backward compatibility with older TCP implementations. These TCP
extensions assume that the receiver is able to buffer the segments that it receives out-of-sequence.
The first extension that was proposed is the fast retransmit heuristic. This extension can be implemented on TCP
senders and thus does not require any change to the protocol. It only assumes that the TCP receiver is able to
buffer out-of-sequence segments.
From a performance point of view, one issue with TCPs retransmission timeout is that when there are isolated
segment losses, the TCP sender often remains idle waiting for the expiration of its retransmission timeouts. Such
isolated losses are frequent in the global Internet [Paxson99]. A heuristic to deal with isolated losses without
waiting for the expiration of the retransmission timeout has been included in many TCP implementations since
the early 1990s. To understand this heuristic, let us consider the figure below that shows the segments exchanged
over a TCP connection when an isolated segment is lost.
As shown above, when an isolated segment is lost the sender receives several duplicate acknowledgements since
the TCP receiver immediately sends a pure acknowledgement when it receives an out-of-sequence segment. A
duplicate acknowledgement is an acknowledgement that contains the same acknowledgement number as a previous
segment. A single duplicate acknowledgement does not necessarily imply that a segment was lost, as a simple
reordering of the segments may cause duplicate acknowledgements as well. Measurements [Paxson99] have
shown that segment reordering is frequent in the Internet. Based on these observations, the fast retransmit heuristic
has been included in most TCP implementations. It can be implemented as follows :

Figure 3.29: Detecting isolated segment losses
ack arrival:
if tcp.ack==snd.una: # duplicate acknowledgement
dupacks++
if dupacks==3:
retransmit segment(snd.una)
else:
dupacks=0
# process acknowledgement
This heuristic requires an additional variable in the TCB (dupacks). Most implementations set the default number
of duplicate acknowledgements that trigger a retransmission to 3. It is now part of the standard TCP specification
RFC 2581. The fast retransmit heuristic improves the TCP performance provided that isolated segments are lost
and the current window is large enough to allow the sender to send three duplicate acknowledgements.
The figure below illustrates the operation of the fast retransmit heuristic.
Figure 3.30: TCP fast retransmit heuristics
When losses are not isolated or when the windows are small, the performance of the fast retransmit heuristic
decreases. In such environments, it is necessary to allow a TCP sender to use a selective repeat strategy instead
of the default go-back-n strategy. Implementing selective-repeat requires a change to the TCP protocol as the
receiver needs to be able to inform the sender of the out-of-order segments that it has already received. This can
be done by using the Selective Acknowledgements (SACK) option defined in RFC 2018. This TCP option is

negotiated during the establishment of a TCP connection. If both TCP hosts support the option, SACK blocks can
be attached by the receiver to the segments that it sends. SACK blocks allow a TCP receiver to indicate the blocks
of data that it has received correctly but out of sequence. The figure below illustrates the utilisation of the SACK
blocks.
Figure 3.31: TCP selective acknowledgements
An SACK option contains one or more blocks. A block corresponds to all the sequence numbers between the left
edge and the right edge of the block. The two edges of the block are encoded as 32 bit numbers (the same size as
the TCP sequence number) in an SACK option. As the SACK option contains one byte to encode its type and one
byte for its length, a SACK option containing b blocks is encoded as a sequence of 2 + 8 b bytes. In practice,
the size of the SACK option can be problematic as the optional TCP header extension cannot be longer than 44
bytes. As the SACK option is usually combined with the RFC 1323 timestamp extension, this implies that a TCP
segment cannot usually contain more than three SACK blocks. This limitation implies that a TCP receiver cannot
always place in the SACK option that it sends, information about all the received blocks.
To deal with the limited size of the SACK option, a TCP receiver currently having more than 3 blocks inside its
receiving buffer must select the blocks to place in the SACK option. A good heuristic is to put in the SACK option
the blocks that have most recently changed, as the sender is likely to be already aware of the older blocks.
When a sender receives an SACK option indicating a new block and thus a new possible segment loss, it usually
does not retransmit the missing segments immediately. To deal with reordering, a TCP sender can use a heuristic
similar to fast retransmit by retransmitting a gap only once it has received three SACK options indicating this gap.
It should be noted that the SACK option does not supersede the acknowledgement number of the TCP header. A
TCP sender can only remove data from its sending buffer once they have been acknowledged by TCPs cumulative
acknowledgements. This design was chosen for two reasons. First, it allows the receiver to discard parts of its
receiving buffer when it is running out of memory without loosing data. Second, as the SACK option is not
transmitted reliably, the cumulative acknowledgements are still required to deal with losses of ACK segments
carrying only SACK information. Thus, the SACK option only serves as a hint to allow the sender to optimise its
retransmissions.
3.8.6 TCP connection release
TCP, like most connection-oriented transport protocols, supports two types of connection releases :
graceful connection release, where each TCP user can release its own direction of data transfer after having
transmitted all data
abrupt connection release, where either one user closes both directions of data transfer or one TCP entity
is forced to close the connection (e.g. because the remote host does not reply anymore or due to lack of
resources)

The abrupt connection release mechanism is very simple and relies on a single segment having the RST bit set. A
TCP segment containing the RST bit can be sent for the following reasons :
a non-SYN segment was received for a non-existing TCP connection RFC 793
by extension, some implementations respond with an RST segment to a segment that is received on an
existing connection but with an invalid header RFC 3360. This causes the corresponding connection to be
closed and has caused security attacks RFC 4953
by extension, some implementations send an RST segment when they need to close an existing TCP con-
nection (e.g. because there are not enough resources to support this connection or because the remote host
is considered to be unreachable). Measurements have shown that this usage of TCP RST is widespread
[AW05]
When an RST segment is sent by a TCP entity, it should contain the current value of the sequence number for the
connection (or 0 if it does not belong to any existing connection) and the acknowledgement number should be set
to the next expected in-sequence sequence number on this connection.
Note: TCP RST wars

TCP implementers should ensure that two TCP entities never enter a TCP RST war where host A is sending a RST
segment in response to a previous RST segment that was sent by host B in response to a TCP RST segment sent by
host A ... To avoid such an infinite exchange of RST segments that do not carry data, a TCP entity is never allowed
to send a RST segment in response to another RST segment.
The normal way of terminating a TCP connection is by using the graceful TCP connection release. This mecha-
nism uses the FIN flag of the TCP header and allows each host to release its own direction of data transfer. As for
the SYN flag, the utilisation of the FIN flag in the TCP header consumes one sequence number. The figure FSM
for TCP connection release shows the part of the TCP FSM used when a TCP connection is released.
Figure 3.32: FSM for TCP connection release
Starting from the Established state, there are two main paths through this FSM.
The first path is when the host receives a segment with sequence number x and the FIN flag set. The utilisation of
the FIN flag indicates that the byte before sequence number x was the last byte of the byte stream sent by the remote
host. Once all of the data has been delivered to the user, the TCP entity sends an ACK segment whose ack field is
set to (x+1) (mod 232 ) to acknowledge the FIN segment. The FIN segment is subject to the same retransmission
mechanisms as a normal TCP segment. In particular, its transmission is protected by the retransmission timer. At
this point, the TCP connection enters the CLOSE_WAIT state. In this state, the host can still send data to the
remote host. Once all its data have been sent, it sends a FIN segment and enter the LAST_ACK state. In this state,
the TCP entity waits for the acknowledgement of its FIN segment. It may still retransmit unacknowledged data
segments e.g. if the retransmission timer expires. Upon reception of the acknowledgement for the FIN segment,
the TCP connection is completely closed and its TCB can be discarded.

The second path is when the host has transmitted all data. Assume that the last transmitted sequence number is
z. Then, the host sends a FIN segment with sequence number (z + 1) (mod 232 ) and enters the FIN_WAIT1
state. It this state, it can retransmit unacknowledged segments but cannot send new data segments. It waits for an
acknowledgement of its FIN segment (i.e. sequence number (z + 1) (mod 232 )), but may receive a FIN segment
sent by the remote host. In the first case, the TCP connection enters the FIN_WAIT2 state. In this state, new data
segments from the remote host are still accepted until the reception of the FIN segment. The acknowledgement
for this FIN segment is sent once all data received before the FIN segment have been delivered to the user and
the connection enters the TIME_WAIT state. In the second case, a FIN segment is received and the connection
enters the Closing state once all data received from the remote host have been delivered to the user. In this state,
no new data segments can be sent and the host waits for an acknowledgement of its FIN segment before entering
the TIME_WAIT state.
The TIME_WAIT state is different from the other states of the TCP FSM. A TCP entity enters this state after
having sent the last ACK segment on a TCP connection. This segment indicates to the remote host that all the
data that it has sent have been correctly received and that it can safely release the TCP connection and discard
the corresponding TCB. After having sent the last ACK segment, a TCP connection enters the TIME_WAIT and
remains in this state for 2 M SL seconds. During this period, the TCB of the connection is maintained. This
ensures that the TCP entity that sent the last ACK maintains enough state to be able to retransmit this segment
if this ACK segment is lost and the remote host retransmits its last FIN segment or another one. The delay of
2 M SL seconds ensures that any duplicate segments on the connection would be handled correctly without
causing the transmission of an RST segment. Without the TIME_WAIT state and the 2 M SL seconds delay, the
connection release would not be graceful when the last ACK segment is lost.
Note: TIME_WAIT on busy TCP servers

The 2 M SL seconds delay in the TIME_WAIT state is an important operational problem on servers having
thousands of simultaneously opened TCP connections [FTY99]. Consider for example a busy web server that
processes 10.000 TCP connections every second. If each of these connections remain in the TIME_WAIT state
for 4 minutes, this implies that the server would have to maintain more than 2 million TCBs at any time. For this
reason, some TCP implementations prefer to perform an abrupt connection release by sending a RST segment to
close the connection [AW05] and immediately discard the corresponding TCB. However, if the RST segment is
lost, the remote host continues to maintain a TCB for a connection no longer exists. This optimisation reduces the
number of TCBs maintained by the host sending the RST segment but at the potential cost of increased processing
on the remote host when the RST segment is lost.
3.9 The Stream Control Transmission Protocol
The Stream Control Transmission Protocol (SCTP) RFC 4960 was defined in the late 1990s, early 2000s as
an alternative to the Transmission Control Protocol. The initial design of SCTP was motivated by the need to
efficiently support signaling protocols that are used in Voice over IP networks. These signaling protocols allow
to create, control and terminate voice calls. They have different requirements than regular applications like email,
http that are well served by TCPs bytestream service.
One of the first motivations for SCTP was the need to efficiently support multihomed hosts, i.e. hosts equipped
with two or more network interfaces. The Internet architecture and TCP in particular were not designed to handle
efficiently such hosts. On the Internet, when a host is multihomed, it needs to use several IP addresses, one per
interface. Consider for example a smartphone connected to both WiFi and 3G. The smartphone uses one IP address
on its WiFi interface and a different one on its 3G interface. When it establishes a TCP connection through its
WiFi interface, this connection is bound to the IP address of the WiFi interface and the segments corresponding
to this connection must always be transmitted through the WiFi interface. If the WiFi interface is not anymore
connected to the network (e.g. because the smartphone user moved), the TCP connection stops and need to be
explicitly reestablished by the application over the 3G interface. SCTP was designed to support seamless failover
from one interface to another during the lifetime of a connection. This is a major change compared to TCP 27 .
27 Recently, the IETF approved the Multipath TCP extension RFC 6824 that allows TCP to efficiently support multihomed hosts. A detailed
presentation of Multipath TCP is outside the scope of this document, but may be found in [RIB2013] and on http://www.multipath-tcp.org

A second motivation for designing SCTP was to provide a different service than TCPs bytestream to the applica-
tions. A first service brought by SCTP is the ability exchange messages instead of only a stream of bytes. This
is a major modification which has many benefits for applications. Unfortunately, there are many deployed appli-
cations that have been designed under the assumption of the bytestream service. Rewriting them to benefit from
a message-mode service will require a lot of effort. It seems unlikely as of this writing to expect old applications
to be rewritten to fully support SCTP and use it. However, some new applications are considering using SCTP
instead of TCP. Voice over IP signaling protocols are a frequently cited example. The Real-Time Communica-
tion in Web-browsers working group is also considering the utilization of SCTP for some specific data channels
[JLT2013]. From a service viewpoint, a second advantage of SCTP compared to TCP is its ability to support
several simultaneous streams. Consider a web application that needs to retrieve five objects from a remote server.
With TCP, one possibility is to open one TCP connection for each object, send a request over each connection and
retrieve one object per connection. This is the solution used by HTTP/1.0 as explained earlier. The drawback of
this approach is that the application needs to maintain several concurrent TCP connections. Another solution is
possible with HTTP/1.1 [NGB+1997] . With HTTP/1.1, the client can use pipelining to send several HTTP Re-
quests without waiting for the answer of each request. The server replies to these requests in sequence, one after
the other. If the server replies to the requests in the sequence, this may lead to head-of-line blocking problems.
Consider that the objects different sizes. The first object is a large 10 MBytes image while the other objects are
small javascript files. In this case, delivering the objects in sequence will cause a very long delay for the javascript
files since they will only be transmitted once the large image has been sent.
With SCTP, head-of-line blocking can be mitigated. SCTP can open a single connection and divide it in five logical
streams so that the five objects are sent in parallel over the single connection. SCTP controls the transmission of
the segments over the connection and ensures that the data is delivered efficiently to the application. In the example
above, the small javascript files could be delivered as independent messages before the large image.
Another extension to SCTP RFC 3758 supports partially-reliable delivery. With this extension, an SCTP sender
can be instructed to expire data based on one of several events, such as a timeout, the sender can signal the SCTP
receiver to move on without waiting for the expired data. This partially reliable service could be useful to provide
timed delivery for example. With this service, there is an upper limit on the time required to deliver a message to
the receiver. If the transport layer cannot deliver the data within the specified delay, the data is discarded by the
sender without causing any stall in the stream.
3.9.1 SCTP segments
SCTP entities exchange segments. In contrast with TCP that uses a simple segment format with a limited space
for the options, the designers of SCTP have learned from the experience of using and extending TCP during
almost two decades. An SCTP segment is always composed of a fixed size common header followed by a variable
number of chunks. The common header is 12 bytes long and contains four fields. The first two fields and the
Source and Destination ports that allow to identify the SCTP connection. The Verification tag is a field that is
set during connection establishment and placed in all segments exchanged during a connection to validate the
received segments. The last field of the common header is a 32bits CRC. This CRC is computed over the entire
segment (common header and all chunks). It is computed by the sender and verified by the receiver. Note that
although this field is named Checksum RFC 4960 it is computed by using the CRC-32 algorithm that has much
stronger error detection capabilities than the Internet checksum algorithm used by TCP [SGP98].
Figure 3.33: The SCTP segment format
3.9. The Stream Control Transmission Protocol 155

The SCTP chunks play a key role in the extensibility of SCTP. In TCP, the extensibility of the protocol is provided
by the utilisation of options that allow to extend the TCP header. However, even with options the TCP header
cannot be longer than 64 bytes. This severely restricts our ability to significantly extend TCP [RIB2013]. In
SCTP, a segment, which must be transmitted inside a single network packet, like a TCP segment, can contain a
variable number of chunks and each chunk has a variable length. The payload that contains the data provided by
the user is itself a chunk. The SCTP chunks are a good example of a protocol format that can be easily extended.
Each chunk is encoded as four fields shown in the figure below.
Figure 3.34: The SCTP chunk format
The first byte indicates the chunk type. 15 chunk types are defined in RFC 4960 and new ones can be easily
added. The low-order 16 bits of the first word contain the length of the chunk in bytes. The presence of the length
field ensures that any SCTP implementation will be able to correctly parse any received SCTP segment, even if it
contains unknown or new chunks. To further ease the processing of unknown chunks, RFC 4960 uses the first two
bits of the chunk type to specify how an SCTP implementation should react when receiving an unknown chunk.
If the two high-order bits of the type of the unknown are set to 00, then the entire SCTP segment containing
the chunk should be discarded. It is expected that all SCTP implementations are capable of recognizing and
processing these chunks. If the first two bits of the chunk type are set to 01 the SCTP segment must be discarded
and an error reported to the sender. If the two high order bits of the type are set to 10 (resp. 11), the chunk must
be ignored, but the processing of the other chunks in the SCTP segment continues (resp. and an error is reported).
The second byte contains flags that are used for some chunks.
3.9.2 Connection establishment
The SCTP protocol was designed shortly after the first Denial of Service attacks against the three-way handshake
used by TCP. These attacks have heavily influenced the connection establishment mechanism chosen for SCTP.
An SCTP connection is established by using a four-way handshake.
The SCTP connection establishment uses several chunks to specify the values of some parameters that are ex-
changed. The SCTP four-way handshake uses four segments as shown in the figure below.
INIT, Itag=1234
INIT-ACK,cookie,ITag=5678
COOKIE-ECHO, cookie, Vtag=5678
COOKIE-ACK,VTag=1234
The first segment contains the INIT chunk. To establish an SCTP connection with a server, the client first creates
some local state for this connection. The most important parameter of the INIT chunk is the Initiation tag. This
value is a random number that is used to identify the connection on the client host for its entire lifetime. This
Initiation tag is placed as the Verification tag in all segments sent by the server. This is an important change
compared to TCP where only the source and destination ports are used to identify a given connection. The INIT
chunk may also contain the other addresses owned by the client. The server responds by sending an INIT-ACK
chunk. This chunk also contains an Initiation tag chosen by the server and a copy of the Initiation tag chosen by
the client. The INIT and INIT-ACK chunks also contain an initial sequence number. A key difference between
TCPs three-way handshake and SCTPs four-way handshake is that an SCTP server does not create any state
when receiving an INIT chunk. For this, the server places inside the INIT-ACK reply a State cookie chunk.

This State cookie is an opaque block of data that contains information computed from the INIT and INIT-ACK
chunks that the server would have had stored locally, some lifetime information and a signature. The format of
the State cookie is flexible and the server could in theory place almost any information inside this chunk. The
only requirement is that the State cookie must be echoed back by the client to confirm the establishment of the
connection. Upon reception of the COOKIE-ECHO chunk, the server verifies the signature of the State cookie.
The client may provide some user data and an initial sequence number inside the COOKIE-ECHO chunk. The
server then responds with a COOKIE-ACK chunk that acknowledges the COOKIE-ECHO chunk. The SCTP
connection between the client and the server is now established. This four-way handshake is both more secure
and more flexible than the three-way handshake used by TCP. The detailed formats of the INIT, INIT-ACK,
COOKIE-ECHO and COOKIE-ACK chunks may be found in RFC 4960.
3.9.3 Reliable data transfert
SCTP provides a slightly different service model RFC 3286. Once an SCTP connection has been established, the
communicating hosts can access two or more message streams. A message stream is a stream of variable length
messages. Each message is composed of an integer number of bytes. The connection-oriented service provided
by SCTP preserves the message boundaries. It is interesting to analyze how SCTP provides the message-mode
service and contrast SCTP with TCP. Data is exchanged by using data chunks. The format of these chunks is
Figure 3.35: The SCTP DATA chunk
An SCTP DATA chunk contains several fields as shown in the figure above. The detailed description of this chunk
may be found in RFC 4960. For simplicity, we focus on an SCTP connection that supports a single stream. SCTP
uses the Transmission Sequence Number (TSN) to sequence the data chunks that are sent. The TSN is also used to
reorder the received DATA chunks and detect lost chunks. This TSN is encoded as a 32 bits field, as the sequence
number by the TCP. However, the TSN is only incremented by one for each data chunk. This implies that the TSN
space does not wrap as quickly as the TCP sequence number. When a small message needs to be sent, the SCTP
entity creates a new data chunk with the next available TSN and places the data inside the chunk. A single SCTP
segment may contain several data chunks, e.g. when small messages are transmitted. Each message is identified
by its TSN and within a stream all messages are delivered in sequence. If the message to be transmitted is larger
than the underlying network packet, SCTP needs to fragment the message in several chunks that are placed in
subsequent segments. The packing of the message in successive segments must still enable the receiver to detect
the message boundaries. This is achieved by using the B and E bits of the second high-order byte of the data
chunk. The B (Begin) bit is set when the first byte of the User data field of the data chunk is the first byte of the
message. The E (End) bit is set when the last byte of the User data field of the data chunk is the last byte of the
message. A small message is always a sent as chunk whose B and E bits are set to 1. A message which is larger
than one network packet will be fragmented in several chunks. Consider for example a message that needs to be
divided in three chunks sent in three different SCTP segments. The first chunk will have its B bit set to 1 and its E
bit set to 0 and a TSN (say x). The second chunk will have both its B and E bits set to 0 and its TSN will be x+1.
The third, and last, chunk will have its B bit set to 0, its E bit set to 1 and its TSN will be x+2. All the chunks
that correspond to a given message must have successive TSNs. The B and E bits allow the receiver to recover the
message from the received data chunks.
The data chunks are only one part of the reliable data transfert. To reliably transfer data, a transport protocol
must also use acknowledgements, retransmissions and flow-control. In SCTP, all these mechanisms rely on the
Selective Acknowledgements (Sack) chunk whose format is shown in the figure below.
3.9. The Stream Control Transmission Protocol 157

Figure 3.36: The SCTP Sack chunk
This chunk is sent by a sender when it needs to send feedback about the reception of data chunks or its buffer
space to the remote sender. The Cumulative TSN ack contains the TSN of the last data chunk that was received
in sequence. This cumulative indicates which TSN has been reliably received by the receiver. The evolution of
this field shows the progress of the reliable transmission. This is the first feedback provided by SCTP. Note that
in SCTP the acknowledgements are at the chunk level and not at the byte level in contrast with TCP. While SCTP
transfers messages divided in chunks, buffer space is still measured in bytes and not in variable-length messages
or chunks. The Advertised Receiver Window Credit field of the Sack chunk provides the current receive window
of the receiver. This window is measured in bytes and its left edge is the last byte of the last in-sequence data
chunk.
The Sack chunk also provides information about the received out-of-sequence chunks (if any). The Sack chunk
contains gap blocks that are in principle similar to the TCP Sack option. However, there are some differences
between TCP and SCTP. The Sack option used by TCP has a limited size. This implies that if there are many gaps
that need to be reported, a TCP receiver must decide which gaps to include in the SACK option. The SCTP Sack
chunk is only limited by the network packet length, which is not a problem in practice. A second difference is
that SCTP can also provide feedback about the reception of duplicate chunks. If several copies of the same data
chunk have been received, this probably indicates a bad heuristic on the sender. The last part of the Sack chunk
provides the list of duplicate TSN received to enable a sender to tune its retransmission mechanism based on this
information. Some details on a possible use of this field may be found in RFC 3708. The last difference with
the TCP SACK option is that the gaps are encoded as deltas relative to the Cumulative TSN ack. These deltas are
encoded as 16 bits integers and allow to reduce the length of the chunk.
3.9.4 Connection release
SCTP uses a different approach to terminante connections. When an application requests a shutdown of a con-
nection, SCTP performs a three-way handshake. This handshake uses the SHUTDOWN, SHUTDOWN-ACK and
SHUTDOWN-COMPLETE chunks. The SHUTDOWN chunk is sent once all outgoing data has been acknowledged.
It contains the last cumulative sequence number. Upon reception of a SHUTDOWN chunk, an SCTP entity in-
forms its application that it cannot accept anymore data over this connection. It then ensures that all outstanding
data have been delivered correctly. At that point, it sends a SHUTDOWN-ACK to confirm the reception of the
SHUTDOWN segment. The three-way handshake completes with the transmission of the SHUTDOWN-COMPLETE
chunk RFC 4960.

SHUTDOWN(TSN=last)
SHUTDOWN-ACK
SHUTDOWN-COMPLETE
Note that in contrast with TCPs four-way handshake, the utilisation of a three-way handshake to close an SCTP
connection implies that the client (resp. server) may close the connection when the application at the other end
has still some data to transmit. Upon reception of the SHUTDOWN chunk, an SCTP entity must stop accepting new
data from the application, but it still needs to retransmit the unacknowledged data chunks (the SHUTDOWN chunk
may be placed in the same segment as a Sack chunk that indicates gaps in the received chunks).
SCTP also provides the equivalent to TCPs RST segment. The ABORT chunk can be used to refuse a connection,
react to the reception of an invalid segment or immediately close a connection (e.g. due to lack of resources).
3.10 Congestion control
In an internetwork, i.e. a networking composed of different types of networks, such as the Internet, congestion
control could be implemented either in the network layer or the transport layer. The congestion problem was
clearly identified in the later 1980s and the researchers who developed techniques to solve the problem opted for
a solution in the transport layer. Adding congestion control to the transport layer makes sense since this layer
provides a reliable data transfert and avoiding congestion is a factor in this reliable delivery. The transport layer
already deals with heterogeneous networks thanks to its self-clocking property that we have already described.
In this section, we explain how congestion control has been added to TCP (and SCTP whose congestion control
scheme is very close to TCPs congestion control) and how this mechanism could be improved in the future.
The TCP congestion control scheme was initially proposed by Van Jacobson in [Jacobson1988]. The current
specification may be found in RFC 5681. TCP relies on Additive Increase and Multiplicative Decrease (AIMD).
To implement AIMD, a TCP host must be able to control its transmission rate. A first approach would be to use
timers and adjust their expiration times in function of the rate imposed by AIMD. Unfortunately, maintaining such
timers for a large number of TCP connections can be difficult. Instead, Van Jacobson noted that the rate of TCP
congestion can be artificially controlled by constraining its sending window. A TCP connection cannot send data
faster than window
rtt where window is the maximum between the hosts sending window and the window advertised
by the receiver.
TCPs congestion control scheme is based on a congestion window. The current value of the congestion window
(cwnd) is stored in the TCB of each TCP connection and the window that can be used by the sender is constrained
by min(cwnd, rwin, swin) where swin is the current sending window and rwin the last received receive win-
dow. The Additive Increase part of the TCP congestion control increments the congestion window by MSS bytes
every round-trip-time. In the TCP literature, this phase is often called the congestion avoidance phase. The Mul-
tiplicative Decrease part of the TCP congestion control divides the current value of the congestion window once
congestion has been detected.
When a TCP connection begins, the sending host does not know whether the part of the network that it uses
to reach the destination is congested or not. To avoid causing too much congestion, it must start with a small
congestion window. [Jacobson1988] recommends an initial window of MSS bytes. As the additive increase part
of the TCP congestion control scheme increments the congestion window by MSS bytes every round-trip-time,
the TCP connection may have to wait many round-trip-times before being able to efficiently use the available
bandwidth. This is especially important in environments where the bandwidth rtt product is high. To avoid
waiting too many round-trip-times before reaching a congestion window that is large enough to efficiently utilise
the network, the TCP congestion control scheme includes the slow-start algorithm. The objective of the TCP
slow-start phase is to quickly reach an acceptable value for the cwnd. During slow-start, the congestion window is
doubled every round-trip-time. The slow-start algorithm uses an additional variable in the TCB : ssthresh (slow-
start threshold). The ssthresh is an estimation of the last value of the cwnd that did not cause congestion. It is
initialised at the sending window and is updated after each congestion event.
3.10. Congestion control 159

A key question that must be answered by any congestion control scheme is how congestion is detected. The
first implementations of the TCP congestion control scheme opted for a simple and pragmatic approach : packet
losses indicate congestion. If the network is congested, router buffers are full and packets are discarded. In
wired networks, packet losses are mainly caused by congestion. In wireless networks, packets can be lost due to
transmission errors and for other reasons that are independent of congestion. TCP already detects segment losses
to ensure a reliable delivery. The TCP congestion control scheme distinguishes between two types of congestion :
mild congestion. TCP considers that the network is lightly congested if it receives three duplicate acknowl-
edgements and performs a fast retransmit. If the fast retransmit is successful, this implies that only one
segment has been lost. In this case, TCP performs multiplicative decrease and the congestion window is
divided by 2. The slow-start threshold is set to the new value of the congestion window.
severe congestion. TCP considers that the network is severely congested when its retransmission timer
expires. In this case, TCP retransmits the first segment, sets the slow-start threshold to 50% of the congestion
window. The congestion window is reset to its initial value and TCP performs a slow-start.
The figure below illustrates the evolution of the congestion window when there is severe congestion. At the
beginning of the connection, the sender performs slow-start until the first segments are lost and the retransmission
timer expires. At this time, the ssthresh is set to half of the current congestion window and the congestion window
is reset at one segment. The lost segments are retransmitted as the sender again performs slow-start until the
congestion window reaches the sshtresh. It then switches to congestion avoidance and the congestion window
increases linearly until segments are lost and the retransmission timer expires ...
Figure 3.37: Evaluation of the TCP congestion window with severe congestion
The figure below illustrates the evolution of the congestion window when the network is lightly congested and
all lost segments can be retransmitted using fast retransmit. The sender begins with a slow-start. A segment is
lost but successfully retransmitted by a fast retransmit. The congestion window is divided by 2 and the sender
immediately enters congestion avoidance as this was a mild congestion.
Figure 3.38: Evaluation of the TCP congestion window when the network is lightly congested
Most TCP implementations update the congestion window when they receive an acknowledgement. If we assume
that the receiver acknowledges each received segment and the sender only sends MSS sized segments, the TCP

congestion control scheme can be implemented using the simplified pseudo-code 28 below.
# Initialization
cwnd = MSS # congestion window in bytes
ssthresh= swin # in bytes
# Ack arrival
if tcp.ack > snd.una : # new ack, no congestion
if cwnd < ssthresh :
# slow-start : increase quickly cwnd
# double cwnd every rtt
cwnd = cwnd + MSS
else:
# congestion avoidance : increase slowly cwnd
# increase cwnd by one mss every rtt
cwnd = cwnd+ mss*(mss/cwnd)
else: # duplicate or old ack
if tcp.ack==snd.una: # duplicate acknowledgement
dupacks++
if dupacks==3:
retransmitsegment(snd.una)
ssthresh=max(cwnd/2,2*MSS)
cwnd=ssthresh
else: # ack for old segment, ignored
dupacks=0
Expiration of the retransmission timer:

send(snd.una) # retransmit first lost segment
sshtresh=max(cwnd/2,2*MSS)
cwnd=MSS
Furthermore when a TCP connection has been idle for more than its current retransmission timer, it should reset its
congestion window to the congestion window size that it uses when the connection begins, as it no longer knows
the current congestion state of the network.
Note: Initial congestion window

The original TCP congestion control mechanism proposed in [Jacobson1988] recommended that each TCP con-
nection should begin by setting cwnd = M SS. However, in todays higher bandwidth networks, using such a
small initial congestion window severely affects the performance for short TCP connections, such as those used
by web servers.In 2002, RFC 3390 allowed an initial congestion window of about 4 KBytes, which corresponds
to 3 segments in many environments. Recently, researchers from google proposed to further increase the initial
window up to 15 KBytes [DRC+2010]. The measurements that they collected show that this increase would not
significantly increase congestion but would significantly reduce the latency of short HTTP responses. Unsur-
prisingly, the chosen initial window corresponds to the average size of an HTTP response from a search engine.
This proposed modification has been adopted as an experimental modification in RFC 6928 and popular TCP
implementations support it.
3.10.1 Controlling congestion without losing data
In todays Internet, congestion is controlled by regularly sending packets at a higher rate than the network capacity.
These packets fill the buffers of the routers and are eventually discarded. But shortly after, TCP senders retransmit
packets containing exactly the same data. This is potentially a waste of ressources since these successive retrans-
missions consume resources upstream of the router that discards the packets. Packet losses are not the only signal
to detect congestion inside the network. An alternative is to allow to routers to explicitly indicate their current
level of congestion when forwarding packets. This approach was proposed in the late 1980s [RJ1995] and used
28 In this pseudo-code, we assume that TCP uses unlimited sequence and acknowledgement numbers. Furthermore, we do not detail how
the cwnd is adjusted after the retransmission of the lost segment by fast retransmit. Additional details may be found in RFC 5681.

in some networks. Unfortunately, it took almost a decade before the Internet community agreed to consider this
approach. In the mean time, a large number of TCP implementations and routers were deployed on the Internet.
As explained earlier, Explicit Congestion Notification RFC 3168, improves the detection of congestion by al-
lowing routers to explicitly mark packets when they are lightly congested. In theory, a single bit in the packet
header [RJ1995] is sufficient to support this congestion control scheme. When a host receives a marked packet, it
returns the congestion information to the source that adapts its transmission rate accordingly. Although the idea is
relatively simple, deploying it on the entire Internet has proven to be challenging [KNT2013]. It is interesting to
analyze the different factors that have hindered the deployment of this technique.
The first difficulty in adding Explicit Congestion Notification (ECN) in TCP/IP network was to modify the format
of the network packet and transport segment headers to carry the required information. In the network layer, one
bit was required to allow the routers to mark the packets they forward during congestion periods. In the IP network
layer, this bit is called the Congestion Experienced (CE) bit and is part of the packet header. However, using a
single bit to mark packets is not sufficient. Consider a simple scenario with two sources, one congested router
and one destination. Assume that the first sender and the destination support ECN, but not the second sender. If
the router is congested it will mark packets from both senders. The first sender will react to the packet markings
by reducing its transmission rate. However since the second sender does not support ECN, it will not react to the
markings. Furthermore, this sender could continue to increase its transmission rate, which would lead to more
packets being marked and the first source would decrease again its transmission rate, ... In the end, the sources
that implement ECN are penalized compared to the sources that do not implement it. This unfairness issue is a
major hurdle to widely deploy ECN on the public Internet 29 . The solution proposed in RFC 3168 to deal with
this problem is to use a second bit in the network packet header. This bit, called the ECN-capable transport (ECT)
bit, indicates whether the packet contains a segment produced by a transport protocol that supports ECN or not.
Transport protocols that support ECN set the ECT bit in all packets. When a router is congested, it first verifies
whether the ECT bit is set. In this case, the CE bit of the packet is set to indicate congestion. Otherwise, the packet
is discarded. This improves the deployability of ECN 30 .
The second difficulty is how to allow the receiver to inform the sender of the reception of network packets marked
with the CE bit. In reliable transport protocols like TCP and SCTP, the acknowledgements can be used to provide
this feedback. For TCP, two options were possible : change some bits in the TCP segment header or define a new
TCP option to carry this information. The designers of ECN opted for reusing spare bits in the TCP header. More
precisely, two TCP flags have been added in the TCP header to support ECN. The ECN-Echo (ECE) is set in the
acknowledgements when the CE was set in packets received on the forward path.
Figure 3.39: The TCP flags
The third difficulty is to allow an ECN-capable sender to detect whether the remote host also supports ECN. This
is a classical negotiation of extensions to a transport protocol. In TCP, this could have been solved by defining a
new TCP option used during the three-way handshake. To avoid wasting space in the TCP options, the designers
of ECN opted in RFC 3168 for using the ECN-Echo and CWR bits in the TCP header to perform this negotiation.
In the end, the result is the same with fewer bits exchanged. SCTP defines in [STD2013] the ECN Support
parameter which can be included in the INIT and INIT-ACK chunks to negotiate the utilization of ECN. The
solution adopted for SCTP is cleaner than the solution adopted for TCP.
Thanks to the ECT, CE and ECE, routers can mark packets during congestion and receivers can return the conges-
tion information back to the TCP senders. However, these three bits are not sufficient to allow a server to reliably
send the ECE bit to a TCP sender. TCP acknowledgements are not sent reliably. A TCP acknowledgement al-
ways contains the next expected sequence number. Since TCP acknowledgements are cumulative, the loss of one
acknowledgement is recovered by the correct reception of a subsequent acknowledgement.
29 In enterprise networks or datacenters, the situation is different since a single company typically controls all the sources and all the routers.
In such networks it is possible to ensure that all hosts and routers have been upgraded before turning on ECN on the routers.
30 With the ECT bit, the deployment issue with ECN is solved provided that all sources cooperate. If some sources do not support ECN but
still set the ECT bit in the packets that they sent, they will have an unfair advantage over the sources that correctly react to packet markings.
Several solutions have been proposed to deal with this problem RFC 3540, but they are outside the scope of this book.

If TCP acknowledgements are overloaded to carry the ECE bit, the situation is different. Consider the example
shown in the figure below. A client sends packets to a server through a router. In the example below, the first packet
is marked. The server returns an acknowledgement with the ECE bit set. Unfortunately, this acknowledgement
is lost and never reaches the client. Shortly after, the server sends a data segment that also carries a cumulative
acknowledgement. This acknowledgement confirms the reception of the data to the client, but it did not receive
the congestion information through the ECE bit.
client router server
data[seq=1,ECT=1,CE=0]
ack=2,ECE=1
ack=2,ECE=1
data[seq=x,ack=2,ECE=0,ECT=1,CE=0]
To solve this problem, RFC 3168 uses an additional bit in the TCP header : the Congestion Window Reduced
(CWR) bit.
client router server
ack=2,ECE=1
ack=2,ECE=1
data[seq=1,ECT=1,CE=0,CWR=1]
data[seq=1,ECT=1,CE=1,CWR=1]
The CWR bit of the TCP header provides some form of acknowledgement for the ECE bit. When a TCP receiver
detects a packet marked with the CE bit, it sets the ECE bit in all segments that it returns to the sender. Upon
reception of an acknowledgement with the ECE bit set, the sender reduces its congestion window to reflect a mild
congestion and sets the CWR bit. This bit remains set as long as the segments received contained the ECE bit set.
A sender should only react once per round-trip-time to marked packets.
SCTP uses a different approach to inform the sender once congestion has been detected. Instead of using one bit
to carry the congestion notification from the receiver to the sender, SCTP defines an entire ECN Echo chunk for
this. This chunk contains the lowest TSN that was received in a packet with the CE bit set and the number of
marked packets received. The SCTP CWR chunk allows to acknowledge the reception of an ECN Echo chunk. It
echoes the lowest TSN placed in the ECN Echo chunk.
The last point that needs to be discussed about Explicit Congestion Notification is the algorithm that is used by

routers to detect congestion. On a router, congestion manifests itself by the number of packets that are stored
inside the router buffers. As explained earlier, we need to distinguish between two types of routers :
routers that have a single FIFO queue
routers that have several queues served by a round-robin scheduler
Routers that use a single queue measure their buffer occupancy as the number of bytes of packets stored in the
queue 31 . A first method to detect congestion is to measure the instantaneous buffer occupancy and consider
the router to be congested as soon as this occupancy is above a threshold. Typical values of the threshold could
be 40% of the total buffer. Measuring the instantaneous buffer occupancy is simple since it only requires one
counter. However, this value is fragile from a control viewpoint since it changes frequently. A better solution is to
measure the average buffer occupancy and consider the router to be congested when this average occupancy is too
high. Random Early Detection (RED) [FJ1993] is an algorithm that was designed to support Explicit Congestion
Notification. In addition to measuring the average buffer occupancy, it also uses probabilistic marking. When
the router is congested, the arriving packets are marked with a probability that increases with the average buffer
occupancy. The main advantage of using probabilistic marking instead of marking all arriving packets is that flows
will be marked in proportion of the number of packets that they transmit. If the router marks 10% of the arriving
packets when congested, then a large flow that sends hundred packets per second will be marked 10 times while a
flow that only sends one packet per second will not be marked. This probabilistic marking allows to mark packets
in proportion of their usage of the network ressources.
If the router uses several queues served by a scheduler, the situation is different. If a large and a small flow are
competing for bandwidth, the scheduler will already favor the small flow that is not using its fair share of the
bandwidth. The queue for the small flow will be almost empty while the queue for the large flow will build up.
On routers using such schedulers, a good way of marking the packets is to set a threshold on the occupancy of
each queue and mark the packets that arrive in a particular queue as soon as its occupancy is above the configured
threshold.
3.10.2 Modeling TCP congestion control
Thanks to its congestion control scheme, TCP adapts its transmission rate to the losses that occur in the net-
work. Intuitively, the TCP transmission rate decreases when the percentage of losses increases. Researchers have
proposed detailed models that allow the prediction of the throughput of a TCP connection when losses occur
[MSMO1997] . To have some intuition about the factors that affect the performance of TCP, let us consider a
very simple model. Its assumptions are not completely realistic, but it gives us good intuition without requiring
complex mathematics.
This model considers a hypothetical TCP connection that suffers from equally spaced segment losses. If p is the
segment loss ratio, then the TCP connection successfully transfers p1 1 segments and the next segment is lost.
If we ignore the slow-start at the beginning of the connection, TCP in this environment is always in congestion
avoidance as there are only isolated losses that can be recovered by using fast retransmit. The evolution of the
congestion window is thus as shown in the figure below. Note the that x-axis of this figure represents time measured
in units of one round-trip-time, which is supposed to be constant in the model, and the y-axis represents the size
of the congestion window measured in MSS-sized segments.
Figure 3.40: Evolution of the congestion window with regular losses

31 The buffers of a router can be implemented as variable or fixed-length slots. If the router uses variable length slots to store the queued
packets, then the occupancy is usually measured in bytes. Some routers have use fixed-length slots with each slot large enough to store a
maximum-length packet. In this case, the buffer occupancy is measured in packets.

As the losses are equally spaced, the congestion window always starts at some value ( W 2 ), and is incremented by
one MSS every round-trip-time until it reaches twice this value (W). At this point, a segment is retransmitted and
the cycle starts again. If the congestion window is measured in MSS-sized segments, a cycle lasts W 2 round-trip-
times. The bandwidth of the TCP connection is the number of bytes that have been transmitted during a given
period of time. During a cycle, the number of segments that are sent on the TCP connection is equal to the area of
the yellow trapeze in the figure. Its area is thus :
3W 2
area = ( W 2
2 ) +
1
2 (W 2
2 ) = 8
However, given the regular losses that we consider, the q number of segments that are sent between two losses (i.e.
1 8
during a cycle) is by definition equal to p . Thus, W = 3p = kp . The throughput (in bytes per second) of the
TCP connection is equal to the number of segments transmitted divided by the duration of the cycle :
3W 2
q
T hroughput = areaM time
SS
= W
8
rtt
or, after having eliminated W, T hroughput = 3 M SS
2 rtt p

2
More detailed models and the analysis of simulations have shown that a first order model of the TCP throughput
when losses occur was T hroughput kM SS
rtt p . This is an important result which shows that :

TCP connections with a small round-trip-time can achieve a higher throughput than TCP connections having
a longer round-trip-time when losses occur. This implies that the TCP congestion control scheme is not
completely fair since it favors the connections that have the shorter round-trip-time
TCP connections that use a large MSS can achieve a higher throughput that the TCP connections that use
a shorter MSS. This creates another source of unfairness between TCP connections. However, it should be
noted that today most hosts are using almost the same MSS, roughly 1460 bytes.
In general, the maximum throughput that can be achieved by a TCP connection depends on its maximum window
size and the round-trip-time if there are no losses. If there are losses, it depends on the MSS, the round-trip-time
and the loss ratio.
T hroughput < min( window kMSS
rtt , rtt p )
Note: The TCP congestion control zoo

The first TCP congestion control scheme was proposed by Van Jacobson in [Jacobson1988]. In addition to writing
the scientific paper, Van Jacobson also implemented the slow-start and congestion avoidance schemes in release
4.3 Tahoe of the BSD Unix distributed by the University of Berkeley. Later, he improved the congestion control
by adding the fast retransmit and the fast recovery mechanisms in the Reno release of 4.3 BSD Unix. Since
then, many researchers have proposed, simulated and implemented modifications to the TCP congestion control
scheme. Some of these modifications are still used today, e.g. :
NewReno (RFC 3782), which was proposed as an improvement of the fast recovery mechanism in the Reno
implementation
TCP Vegas, which uses changes in the round-trip-time to estimate congestion in order to avoid it [BOP1994]
CUBIC, which was designed for high bandwidth links and is the default congestion control scheme in the
Linux 2.6.19 kernel [HRX2008]
Compound TCP, which was designed for high bandwidth links is the default congestion control scheme in
several Microsoft operating systems [STBT2009]
A search of the scientific literature (RFC 6077) will probably reveal more than 100 different variants of the
TCP congestion control scheme. Most of them have only been evaluated by simulations. However, the TCP
implementation in the recent Linux kernels supports several congestion control schemes and new ones can be
easily added. We can expect that new TCP congestion control schemes will always continue to appear.

3.11 The network layer
The main objective of the network layer is to allow endsystems, connected to different networks, to exchange
information through intermediate systems called router. The unit of information in the network layer is called a
packet.
Figure 3.41: The network layer in the reference model
Before explaining the network layer in detail, it is useful to begin by analysing the service provided by the datalink
layer. There are many variants of the datalink layer. Some provide a connection-oriented service while others
provide a connectionless service. In this section, we focus on connectionless datalink layer services as they are
the most widely used. Using a connection-oriented datalink layer causes some problems that are beyond the scope
of this chapter. See RFC 3819 for a discussion on this topic.
Figure 3.42: The point-to-point datalink layer
There are three main types of datalink layers. The simplest datalink layer is when there are only two communi-
cating systems that are directly connected through the physical layer. Such a datalink layer is used when there is
a point-to-point link between the two communicating systems. The two systems can be endsystems or routers.
PPP, defined in RFC 1661, is an example of such a point-to-point datalink layer. Datalink layers exchange frames
and a datalink frame sent by a datalink layer entity on the left is transmitted through the physical layer, so that
it can reach the datalink layer entity on the right. Point-to-point datalink layers can either provide an unreliable
service (frames can be corrupted or lost) or a reliable service (in this case, the datalink layer includes retransmis-
sion mechanisms similar to the ones used in the transport layer). The unreliable service is frequently used above
physical layers (e.g. optical fiber, twisted pairs) having a low bit error ratio while reliability mechanisms are often
used in wireless networks to recover locally from transmission errors.
The second type of datalink layer is the one used in Local Area Networks (LAN). Conceptually, a LAN is a set of
communicating devices such that any two devices can directly exchange frames through the datalink layer. Both
endsystems and routers can be connected to a LAN. Some LANs only connect a few devices, but there are LANs
that can connect hundreds or even thousands of devices.
In the next chapter, we describe the organisation and the operation of Local Area Networks. An important dif-
ference between the point-to-point datalink layers and the datalink layers used in LANs is that in a LAN, each
communicating device is identified by a unique datalink layer address. This address is usually embedded in the
hardware of the device and different types of LANs use different types of datalink layer addresses. Most LANs
use 48-bits long addresses that are usually called MAC addresses. A communicating device attached to a LAN
can send a datalink frame to any other communicating device that is attached to the same LAN. Most LANs also
support special broadcast and multicast datalink layer addresses. A frame sent to the broadcast address of the

Figure 3.43: A local area network
LAN is delivered to all communicating devices that are attached to the LAN. The multicast addresses are used to
identify groups of communicating devices. When a frame is sent towards a multicast datalink layer address, it is
delivered by the LAN to all communicating devices that belong to the corresponding group.
The third type of datalink layers are used in Non-Broadcast Multi-Access (NBMA) networks. These networks are
used to interconnect devices like a LAN. All devices attached to an NBMA network are identified by a unique
datalink layer address. However, and this is the main difference between an NBMA network and a traditional
LAN, the NBMA service only supports unicast. The datalink layer service provided by an NBMA network
supports neither broadcast nor multicast.
Unfortunately no datalink layer is able to send frames of unlimited side. Each datalink layer is characterised by a
maximum frame size. There are more than a dozen different datalink layers and unfortunately most of them use a
different maximum frame size. The network layer must cope with the heterogeneity of the datalink layer.
3.11.1 IP version 6
In the late 1980s and early 1990s the growth of the Internet was causing several operational problems on routers.
Many of these routers had a single CPU and up to 1 MByte of RAM to store their operating system, packet buffers
and routing tables. Given the rate of allocation of IPv4 prefixes to companies and universities willing to join the
Internet, the routing tables where growing very quickly and some feared that all IPv4 prefixes would quickly be
allocated. In 1987, a study cited in RFC 1752, estimated that there would be 100,000 networks in the near future.
In August 1990, estimates indicated that the class B space would be exhausted by March 1994. Two types of
solution were developed to solve this problem. The first short term solution was the introduction of Classless Inter
Domain Routing (CIDR). A second short term solution was the Network Address Translation (NAT) mechanism,
defined in RFC 1631. NAT allowed multiple hosts to share a single public IPv4 address.
However, in parallel with these short-term solutions, which have allowed the IPv4 Internet to continue to be usable
until now, the Internet Engineering Task Force started to work on developing a replacement for IPv4. This work
started with an open call for proposals, outlined in RFC 1550. Several groups responded to this call with proposals
for a next generation Internet Protocol (IPng) :
TUBA proposed in RFC 1347 and RFC 1561
PIP proposed in RFC 1621
SIPP proposed in RFC 1710
The IETF decided to pursue the development of IPng based on the SIPP proposal. As IP version 5 was already
used by the experimental ST-2 protocol defined in RFC 1819, the successor of IP version 4 is IP version 6. The
initial IP version 6 defined in RFC 1752 was designed based on the following assumptions :
IPv6 addresses are encoded as a 128 bits field
The IPv6 header has a simple format that can easily be parsed by hardware devices
A host should be able to configure its IPv6 address automatically
3.11. The network layer 167

Security must be part of IPv6
Note: The IPng address size

When the work on IPng started, it was clear that 32 bits was too small to encode an IPng address and all proposals
used longer addresses. However, there were many discussions about the most suitable address length. A first
approach, proposed by SIP in RFC 1710, was to use 64 bit addresses. A 64 bits address space was 4 billion times
larger than the IPv4 address space and, furthermore, from an implementation perspective, 64 bit CPUs were being
considered and 64 bit addresses would naturally fit inside their registers. Another approach was to use an existing
address format. This was the TUBA proposal (RFC 1347) that reuses the ISO CLNP 20 bytes addresses. The
20 bytes addresses provided room for growth, but using ISO CLNP was not favored by the IETF partially due to
political reasons, despite the fact that mature CLNP implementations were already available. 128 bits appeared to
be a reasonable compromise at that time.
IPv6 addressing architecture
The experience of IPv4 revealed that the scalability of a network layer protocol heavily depends
on its addressing architecture. The designers of IPv6 spent a lot of effort defining its address-
ing architecture RFC 3513. All IPv6 addresses are 128 bits wide. This implies that there are
340, 282, 366, 920, 938, 463, 463, 374, 607, 431, 768, 211, 456(3.4 1038 ) different IPv6 addresses. As the sur-
face of the Earth is about 510,072,000 km2 , this implies that there are about 6.67 1023 IPv6 addresses per
square meter on Earth. Compared to IPv4, which offers only 8 addresses per square kilometer, this is a significant
improvement on paper.
IPv6 supports unicast, multicast and anycast addresses. An IPv6 unicast address is used to identify one datalink-
layer interface on a host. If a host has several datalink layer interfaces (e.g. an Ethernet interface and a WiFi
interface), then it needs several IPv6 addresses. In general, an IPv6 unicast address is structured as shown in the
figure below.
Note: Textual representation of IPv6 addresses

It is sometimes necessary to write IPv6 addresses in text format, e.g. when manually configuring addresses or
for documentation purposes. The preferred format for writing IPv6 addresses is x:x:x:x:x:x:x:x, where the
x s are hexadecimal digits representing the eight 16-bit parts of the address. Here are a few examples of IPv6
addresses :
abcd:ef01:2345:6789:abcd:ef01:2345:6789
2001:db8:0:0:8:800:200c:417a
fe80:0:0:0:219:e3ff:fed7:1204
IPv6 addresses often contain a long sequence of bits set to 0. In this case, a compact notation has been defined.
With this notation, :: is used to indicate one or more groups of 16 bits blocks containing only bits set to 0. For
example,
2001:db8:0:0:8:800:200c:417a is represented as 2001:db8::8:800:200c:417a
ff01:0:0:0:0:0:0:101 is represented as ff01::101
0:0:0:0:0:0:0:1 is represented as ::1
0:0:0:0:0:0:0:0 is represented as ::
An IPv6 prefix can be represented as address/length, where length is the length of the prefix in bits. For example,
the three notations below correspond to the same IPv6 prefix :
2001:0db8:0000:cd30:0000:0000:0000:0000/60
2001:0db8::cd30:0:0:0:0/60
2001:0db8:0:cd30::/60

Figure 3.44: Structure of IPv6 unicast addresses
An IPv6 unicast address is composed of three parts :

1. A global routing prefix that is assigned to the Internet Service Provider that owns this block of addresses
2. A subnet identifier that identifies a customer of the ISP
3. An interface identifier that identifies a particular interface on an endsystem
The subnet identifier plays a key role in the scalability of network layer addressing architecture. An important
point to be defined in a network layer protocol is the allocation of the network layer addresses. A naive allocation
scheme would be to provide an address to each host when the host is attached to the Internet on a first come
first served basis. With this solution, a host in Belgium could have address 2001:db8::1 while another host
located in Africa would use address 2001:db8::2. Unfortunately, this would force all routers on the Internet
to maintain one route towards each host. In the network layer, scalability is often a function of the number of
routes stored on the router. A network will usually work better if its routers store fewer routes and network
administrators usually try to minimize the number of routes that are known by their routers. For this, they often
divide their network prefix in smaller subblocks. For example, consider a company with three campuses, a large
one and two smaller ones. The network administrator would probably divide his block of addresses as follows :
the bottom half is used for the large campus
the top half is divided in two smaller blocks, one for each small campus
Inside each campus, the same division can be done, for example on a per building basis, starting from the buildings
that host the largest number of nodes, e.g. the company datacenter. In each building, the same division can be done
on a per floor basis, ... The advantage of such a hierarchical allocation of the addresses is that the routers in the
large campus only need one route to reach a router in the smaller campus. The routers in the large campus would
know more routes about the buildings in their campus, but they do not need to know the details of the organisation
of each smaller campus.
To preserve the scalability of the routing system, it is important to minimize the number of routes that are stored on
each router. A router cannot store and maintain one route for each of the almost 1 billion hosts that are connected
to todays Internet. Routers should only maintain routes towards blocks of addresses and not towards individual
hosts. For this, hosts are grouped in subnets based on their location in the network. A typical subnet groups
all the hosts that are part of the same enterprise. An enterprise network is usually composed of several LANs
interconnected by routers. A small block of addresses from the Enterprises block is usually assigned to each
LAN.
In todays deployments, interface identifiers are always 64 bits wide. This implies that while there are 2128
different IPv6 addresses, they must be grouped in 264 subnets. This could appear as a waste of resources, however
using 64 bits for the host identifier allows IPv6 addresses to be auto-configured and also provides some benefits
from a security point of view, as explained in section ICMPv6
In practice, there are several types of IPv6 unicast address. Most of the IPv6 unicast addresses are allocated in
blocks under the responsibility of IANA. The current IPv6 allocations are part of the 2000::/3 address block.
Regional Internet Registries (RIR) such as RIPE in Europe, ARIN in North-America or AfriNIC in Africa have
each received a block of IPv6 addresses that they sub-allocate to Internet Service Providers in their region. The
ISPs then sub-allocate addresses to their customers.

When considering the allocation of IPv6 addresses, two types of address allocations are often distinguished. The
RIRs allocate provider-independent (PI) addresses. PI addresses are usually allocated to Internet Service Providers
and large companies that are connected to at least two different ISPs [CSP2009]. Once a PI address block has
been allocated to a company, this company can use its address block with the provider of its choice and change
its provider at will. Internet Service Providers allocate provider-aggregatable (PA) address blocks from their own
PI address block to their customers. A company that is connected to only one ISP should only use PA addresses.
The drawback of PA addresses is that when a company using a PA address block changes its provider, it needs to
change all the addresses that it uses. This can be a nightmare from an operational perspective and many companies
are lobbying to obtain PI address blocks even if they are small and connected to a single provider. The typical size
of the IPv6 address blocks are :
/32 for an Internet Service Provider
/48 for a single company
/56 for small user sites
/64 for a single user (e.g. a home user connected via ADSL)
/128 in the rare case when it is known that no more than one endhost will be attached
There is one difficulty with the utilisation of these IPv6 prefixes. Consider Belnet, the Belgian research ISP
that has been allocated the 2001:6a8::/32 prefix. Universities are connected to Belnet. UCL uses prefix
2001:6a8:3080::/48 while the University of Liege uses 2001:6a8:2d80::/48. A commercial ISP
uses prefix 2a02:2788::/32. Both Belnet and the commercial ISP are connected to the global Internet.
Belnet
2001:6a8::/32 ISP1
2a02:2788::/32
ULg UCL
alpha.com
2001:6a8:2d80::/48 2001:6a8:3080::/48
The Belnet network advertises prefix 2001:6a8::/32 that includes the prefixes from both UCL and ULg.
These two subnetworks can be easily reached from any internet connected host. After a few years, UCL decides
to increase the redundancy of its Internet connectivity and buys transit service from ISP1. A direct link between
UCL and the commercial ISP appears on the network and UCL expects to receive packets from both Belnet and
the commercial ISP.
Now, consider how a router inside alpha.com would reach a host in the UCL network. This router has two
routes towards 2001:6a8:3080::1. The first one, for prefix 2001:6a8:3080::/48 is via the direct link
between the commercial ISP and UCL. The second one, for prefix 2001:6a8::/32 is via the Internet and
Belnet. Since RFC 1519 when a router knows several routes towards the same destination address, it must
forward packets along the route having the longest prefix length. In the case of 2001:6a8:3080::1, this is
the route 2001:6a8:3080::/48 that is used to forward the packet. This forwarding rule is called the longest
prefix match or the more specific match. All IP routers implement this forwarding rule.
To understand the longest prefix match forwarding, consider the IPv6 routing below.
Destination Gateway
::/0 fe80::dead:beef
::1 ::1
2a02:2788:2c4:16f::/64 eth0
2001:6a8:3080::/48 fe80::bad:cafe
2001:6a8:2d80::/48 fe80::bad:bad
2001:6a8::/32 fe80::aaaa:bbbb
With the longest match rule, the route ::/0 plays a particular role. As this route has a prefix length of 0 bits, it
matches all destination addresses. This route is often called the default route.

a packet with destination 2a02:2788:2c4:16f::1 received by router R is destined to a host on interface

eth0 .
a packet with destination 2001:6a8:3080::1234 matches three routes : ::/0, 2001:6a8::/32
and 2001:6a8:3080::/48. The packet is forwarded via gateway fe80::bad:cafe
a packet with destination 2001:1890:123a::1:1e matches one route : ::/0. The packet is forwarded
via fe80::dead:beef
a packet with destination 2001:6a8:3880:40::2 matches two routes : 2001:6a8::/32 and ::/0. The
packet is forwarded via fe80::aaaa:bbbb
The longest prefix match can be implemented by using different data structures. One possibility is to use a trie.
Details on how to implement efficient packet forwarding algorithms may be found in [Varghese2005].
For the companies that want to use IPv6 without being connected to the IPv6 Internet, RFC 4193 defines the
Unique Local Unicast (ULA) addresses (fc00::/7). These ULA addresses play a similar role as the private
IPv4 addresses defined in RFC 1918. However, the size of the fc00::/7 address block allows ULA to be much
more flexible than private IPv4 addresses.
Furthermore, the IETF has reserved some IPv6 addresses for a special usage. The two most important ones are :
0:0:0:0:0:0:0:1 (::1 in compact form) is the IPv6 loopback address. This is the address of a logical
interface that is always up and running on IPv6 enabled hosts.
0:0:0:0:0:0:0:0 (:: in compact form) is the unspecified IPv6 address. This is the IPv6 address that
a host can use as source address when trying to acquire an official address.
The last type of unicast IPv6 addresses are the Link Local Unicast addresses. These addresses are part of the
fe80::/10 address block and are defined in RFC 4291. Each host can compute its own link local address by
concatenating the fe80::/64 prefix with the 64 bits identifier of its interface. Link local addresses can be used
when hosts that are attached to the same link (or local area network) need to exchange packets. They are used
notably for address discovery and auto-configuration purposes. Their usage is restricted to each link and a router
cannot forward a packet whose source or destination address is a link local address. Link local addresses have also
been defined for IPv4 RFC 3927. However, the IPv4 link local addresses are only used when a host cannot obtain
a regular IPv4 address, e.g. on an isolated LAN.
Figure 3.45: IPv6 link local address structure
Note: All IPv6 hosts have several addresses

An important consequence of the IPv6 unicast addressing architecture and the utilisation of link-local addresses is
that each IPv6 host has several IPv6 addresses. This implies that all IPv6 stacks must be able to handle multiple
IPv6 addresses.
The addresses described above are unicast addresses. These addresses are used to identify (interfaces on) hosts
and routers. They can appear as source and destination addresses in the IPv6 packets. When a host sends a packet
towards a unicast address, this packet is delivered by the network to its final destination. There are situations,
such as when delivering video or television signal to a large number of receivers, where it is useful to have a
network that can efficiently deliver the same packet to a large number of receivers. This is the multicast service. A
multicast service can be provided in a LAN. In this case, a multicast address identifies a set of receivers and each
frame sent towards this address is delivered to all receivers in the group. Multicast can also be used in a network
containing routers and hosts. In this case, a multicast address identifies also a group of receivers and the network
delivers efficiently each multicast packet to all members of the group. Consider for example the network below.

R1
B
R2 R3
R4
C

Assume that B and D are part of a multicast group. If A sends a multicast packet towards this group, then R1 will
replicate the packet to forward it to R2 and R3. R2 would forward the packet towards B. R3 would forward the
packet towards R4 that would deliver it to D.
Finally, RFC 4291 defines the structure of the IPv6 multicast addresses 32 . This structure is depicted in the figure
below
Figure 3.46: IPv6 multicast address structure
The low order 112 bits of an IPv6 multicast address are the groups identifier. The high order bits are used as a
marker to distinguish multicast addresses from unicast addresses. Notably, the 4 bits flag field indicates whether
the address is temporary or permanent. Finally, the scope field indicates the boundaries of the forwarding of
packets destined to a particular address. A link-local scope indicates that a router should not forward a packet
destined to such a multicast address. An organisation local-scope indicates that a packet sent to such a multicast
destination address should not leave the organisation. Finally the global scope is intended for multicast groups
spanning the global Internet.
Among these addresses, some are well known. For example, all endsystem automatically belong to the ff02::1
multicast group while all routers automatically belong to the ff02::2 multicast group. A detailed discussion of
IPv6 multicast is outside the scope of this chapter.
IPv6 packet format
The IPv6 packet format was heavily inspired by the packet format proposed for the SIPP protocol in RFC 1710.
The standard IPv6 header defined in RFC 2460 occupies 40 bytes and contains 8 different fields, as shown in the
figure below.
Figure 3.47: The IP version 6 header (RFC 2460)
Apart from the source and destination addresses, the IPv6 header contains the following fields :
32 The full list of allocated IPv6 multicast addresses is available at http://www.iana.org/assignments/ipv6-multicast-addresses

version : a 4 bits field set to 6 and intended to allow IP to evolve in the future if needed
Traffic class : this 8 bits field allows to indicate the type of service expected by this packet and contains the
CE and ECT flags that are used by Explicit Congestion Notification
Flow label : this field was initially intended to be used to tag packets belonging to the same flow. A recent
document, RFC 6437 describes some possible usages of this field, but it is too early to tell whether it will
be really used.
Payload length : this is the size of the packet payload in bytes. As the length is encoded as a 16 bits field,
an IPv6 packet can contain up to 65535 bytes of payload.
Hop Limit : this 8 bits field indicates the number of routers that can forward the packet. It is decremented
by one by each router and prevents packets from looping forever inside the network.
Next Header : this 8 bits field indicates the type 33 of header that follows the IPv6 header. It can be a
transport layer header (e.g. 6 for TCP or 17 for UDP) or an IPv6 option.
It is interesting to note that there is no checksum inside the IPv6 header. This is mainly because all datalink layers
and transport protocols include a checksum or a CRC to protect their frames/segments against transmission errors.
Adding a checksum in the IPv6 header would have forced each router to recompute the checksum of all packets,
with limited benefit in detecting errors. In practice, an IP checksum allows for catching errors that occur inside
routers (e.g. due to memory corruption) before the packet reaches its destination. However, this benefit was found
to be too small given the reliability of current memories and the cost of computing the checksum on each router
34
.
When a host receives an IPv6 packet, it needs to determine which transport protocol (UDP, TCP, SCTP, ...) needs
to handle the payload of the packet. This is the first role of the Next header field. The IANA which manages the
allocation of Internet ressources and protocol parameters, maintains an official list of transport protocols 2 . The
following protocol numbers are reserved :
TCP uses Next Header number 6
UDP uses Next Header number 17
SCTP uses Next Header number 132
For example, an IPv6 packet that contains an SCTP segment would appear as shown in the figure below. However,
Figure 3.48: An IPv6 packet containing an SCTP segment

33The IANA maintains the list of all allocated Next Header types at http://www.iana.org/assignments/protocol-numbers/
34 When IPv4 was designed, the situation was different. The IPv4 header includes a checksum that only covers the network header. This
checksum is computed by the source and updated by all intermediate routers that decrement the TTL, which is the IPv4 equivalent of the
HopLimit used by IPv6.

the Next header has broader usages than simply indicating the transport protocol which is responsible for the
packet payload. An IPv6 packet can contain a chain of headers and the last one indicates the transport protocol
that is responsible for the packet payload. Supporting a chain of headers is a clever design from an extensibility
viewpoint. As we will seen, this chain of headers has several usages.
RFC 2460 defines several types of IPv6 extension headers that could be added to an IPv6 packet :
Hop-by-Hop Options header. This option is processed by routers and endhosts.
Destination Options header. This option is processed only by endhosts.
Routing header. This option is processed by some nodes.
Fragment header. This option is processed only by endhosts.
Authentication header. This option is processed only by endhosts.
Encapsulating Security Payload. This option is processed only by endhosts.
The last two headers are used to add security above IPv6 and implement IPSec. They are described in RFC 2402
and RFC 2406 and are outside the scope of this document.
The Hop-by-Hop Options header was designed to allow IPv6 to be easily extended. In theory, this option could
be used to define new fields that were not foreseen when IPv6 was designed. It is intended to be processed by
both routers and endhosts. Deploying an extension to a network protocol can be difficult in practice since some
nodes already support the extensions while others still use the old version and do not understand the extension.
To deal with this issue, the IPv6 designers opted for a Type-Length-Value encoding of these IPv6 options. The
Hop-by-Hop Options header is encoded as shown below.
Figure 3.49: The IPv6 Hop-by-Hop Options header
In this optional header, the Next Header field is used to support the chain of headers. It indicates the type of the
next header in the chain. IPv6 headers have different lengths. The Hdr Ext Len field indicates the total length of
the option header in bytes. The Opt. Type field indicates the type of option. These types are encoded such that
their high order bits specify how the header needs to be handled by nodes that do not recognize it. The following
values are defined for the two high order bits :
00 : if a node does not recognize this header, it can be safely skipped and the processing continues with the
subsequent header
01 : if a node does not recognize this header, the packet must be discarded
10 (resp. 11) : if a node does not recognize this header, it must return a control packet (ICMP, see later)
back to the source (resp. except if the destination was a multicast address)
This encoding allows the designers of protocol extensions to specify whether the option must be supported by all
nodes on a path or not. Still, deploying such an extension can be difficult in practice.
Two hop-by-hop options have been defined. RFC 2675 specifies the jumbogram that enables IPv6 to support
packets containing a payload larger than 65535 bytes. These jumbo packets have their payload length set to 0 and
the jumbogram option contains the packet length as a 32 bits field. Such packets can only be sent from a source
to a destination if all the routers on the path support this option. However, as of this writing it does not seem that
the jumbogram option has been implemented. The router alert option defined in RFC 2711 is the second example
of a hop-by-hop option. The packets that contain this option should be processed in a special way by intermediate
routers. This option is used for IP packets that carry Resource Reservation Protocol (RSVP) messages, but this is
outside the scope of this book.
The Destinations Option header uses the same format as the Hop-by-Hop Options header. It has some usages, e.g.
to support mobile nodes RFC 6275, but these are outside the scope of this document.

The Fragment Options header is more important. An important problem in the network layer is the ability to
handle heterogeneous datalink layers. Most datalink layer technologies can only transmit and receive frames
that are shorter than a given maximum frame size. Unfortunately, all datalink layer technologies use different
maximum frames sizes.
Each datalink layer has its own characteristics and as indicated earlier, each datalink layer is characterised by
a maximum frame size. From IPs point of view, a datalink layer interface is characterised by its Maximum
Transmission Unit (MTU). The MTU of an interface is the largest packet (including header) that it can send. The
table below provides some common MTU sizes.
Datalink layer MTU
Ethernet 1500 bytes
WiFi 2272 bytes
ATM (AAL5) 9180 bytes
802.15.4 102 or 81 bytes
Token Ring 4464 bytes
FDDI 4352 bytes
Although IPv6 can send 64 KBytes long packets, few datalink layer technologies that are used today are able to
send a 64 KBytes packet inside a frame. Furthermore, as illustrated in the figure below, another problem is that a
host may send a packet that would be too large for one of the datalink layers used by the intermediate routers.
Figure 3.50: The need for fragmentation and reassembly
To solve these problems, IPv6 includes a packet fragmentation and reassembly mechanism. In IPv4, fragmenta-
tion was performed by both the endhosts and the intermediate routers. However, experience with IPv4 has shown
that fragmenting packets in routers was costly [KM1995]. For this reason, the developers of IPv6 have decided
that routers would not fragment packets anymore. In IPv6, fragmentation is only performed by the source host. If
a source has to send a packet which is larger than the MTU of the outgoing interface, the packet needs to be frag-
mented before being transmitted. In IPv6, each packet fragment is an IPv6 packet that includes the Fragmentation
header. This header is included by the source in each packet fragment. The receiver uses them to reassemble the
received fragments.
Figure 3.51: IPv6 fragmentation header
If a router receives a packet that is too long to be forwarded, the packet is dropped and the router returns an
ICMPv6 message to inform the sender of the problem. The sender can then either fragment the packet or perform
Path MTU discovery. In IPv6, packet fragmentation is performed only by the source by using IPv6 options.
In IPv6, fragmentation is performed exclusively by the source host and relies on the fragmentation header. This
64 bits header is composed of six fields :
a Next Header field that indicates the type of the header that follows the fragmentation header
a reserved field set to 0.
the Fragment Offset is a 13-bit unsigned integer that contains the offset, in 8 bytes units, of the data following
this header, relative to the start of the original packet.
the More flag, which is set to 0 in the last fragment of a packet and to 1 in all other fragments.

the 32 bits Identification field indicates to which original packet a fragment belongs. When a host sends
fragmented packets, it should ensure that it does not reuse the same identification field for packets sent to
the same destination during a period of MSL seconds. This is easier with the 32 bits identification used in
the IPv6 fragmentation header, than with the 16 bits identification field of the IPv4 header.
Some IPv6 implementations send the fragments of a packet in increasing fragment offset order, starting from the
first fragment. Others send the fragments in reverse order, starting from the last fragment. The latter solution can
be advantageous for the host that needs to reassemble the fragments, as it can easily allocate the buffer required to
reassemble all fragments of the packet upon reception of the last fragment. When a host receives the first fragment
of an IPv6 packet, it cannot know a priori the length of the entire IPv6 packet.
The figure below provides an example of a fragmented IPv6 packet containing a UDP segment. The Next Header
type reserved for the IPv6 fragmentation option is 44.
Figure 3.52: IPv6 fragmentation example
The following pseudo-code details the IPv6 fragmentation, assuming that the packet does not contain options.
#mtu : maximum size of the packet (including header) of outgoing link
if p.len < mtu :
send(p)
# packet is too large
maxpayload=8*int((mtu-40)/8) # must be n times 8 bytes
# packet must be fragmented
payload=p[IP].payload
pos=0
id=globalCounter;
globalCounter++;
while len(payload) > 0 :
if len(payload) > maxpayload :
toSend=IP(dest=p.dest,src=p.src,
hoplimit=p.hoplimit, id,
frag=p.frag+(pos/8), m=True,
len=mtu, nextheader=p.nextheader)/payload[0:maxpayload]
pos=pos+maxpayload
payload=payload[maxpayload+1:]
else
toSend=IP(dest=p.dest,src=p.src,
hoplimit=p.hoplimit, id,
frag=p.frag+(pos/8), m=False,
len=len(payload), nextheader=p.nextheader)/payload
forward(toSend)

In the above pseudocode, we maintain a single 32 bits counter that is incremented for each packet that needs
to be fragmented. Other implementations to compute the packet identification are possible. RFC 2460 only
requires that two fragmented packets that are sent within the MSL between the same pair of hosts have different
identifications.
The fragments of an IPv6 packet may arrive at the destination in any order, as each fragment is forwarded inde-
pendently in the network and may follow different paths. Furthermore, some fragments may be lost and never
reach the destination.
The reassembly algorithm used by the destination host is roughly as follows. First, the destination can verify
whether a received IPv6 packet is a fragment or not by checking whether it contains a fragment header. If so,
all fragments with the some identification must be reassembled together. The reassembly algorithm relies on
the Identification field of the received fragments to associate a fragment with the corresponding packet being
reassembled. Furthermore, the Fragment Offset field indicates the position of the fragment payload in the original
unfragmented packet. Finally, the packet with the M flag reset allows the destination to determine the total length
of the original unfragmented packet.
Note that the reassembly algorithm must deal with the unreliability of the IP network. This implies that a fragment
may be duplicated or a fragment may never reach the destination. The destination can easily detect fragment
duplication thanks to the Fragment Offset. To deal with fragment losses, the reassembly algorithm must bound the
time during which the fragments of a packet are stored in its buffer while the packet is being reassembled. This
can be implemented by starting a timer when the first fragment of a packet is received. If the packet has not been
reassembled upon expiration of the timer, all fragments are discarded and the packet is considered to be lost.
Note: Header compression on low bandwidth links

Given the size of the IPv6 header, it can cause huge overhead on low bandwidth links, especially when small
packets are exchanged such as for Voice over IP applications. In such environments, several techniques can be
used to reduce the overhead. A first solution is to use data compression in the datalink layer to compress all the
information exchanged [Thomborson1992]. These techniques are similar to the data compression algorithms used
in tools such as compress(1) or gzip(1) RFC 1951. They compress streams of bits without taking advantage
of the fact that these streams contain IP packets with a known structure. A second solution is to compress the IP
and TCP header. These header compression techniques, such as the one defined in RFC 5795 take advantage of
the redundancy found in successive packets from the same flow to significantly reduce the size of the protocol
headers. Another solution is to define a compressed encoding of the IPv6 header that matches the capabilities of
the underlying datalink layer RFC 4944.
The last type of IPv6 header extension is the Routingheader. The type 0 routing header defined in RFC 2460
is an example of an IPv6 option that must be processed by some routers. This option is encoded as shown below.
The type 0 routing option was intended to allow a host to indicate a loose source route that should be followed by
a packet by specifying the addresses of some of the routers that must forward this packet. Unfortunately, further
work with this routing header, including an entertaining demonstration with scapy [BE2007] , revealed severe
security problems with this routing header. For this reason, loose source routing with the type 0 routing header
has been removed from the IPv6 specification RFC 5095.
3.11.2 ICMP version 6
It is sometimes necessary for intermediate routers or the destination host to inform the sender of the packet of
a problem that occurred while processing a packet. In the TCP/IP protocol suite, this reporting is done by the
Internet Control Message Protocol (ICMP). ICMPv6 is defined in RFC 4443. It is used both to report problems
that occurred while processing an IPv6 packet, but also when distributing addresses.
ICMPv6 messages are carried inside IPv6 packets (the Next Header field for ICMPv6 is 58). Each ICMP message
contains a 32 bits header with an 8 bits type field, a code field and a 16 bits checksum computed over the entire
ICMPv6 message. The message body contains a copy of the IPv6 packet in error.
ICMPv6 specifies two classes of messages : error messages that indicate a problem in handling a packet and
informational messages. Four types of error messages are defined in RFC 4443 :

Figure 3.53: The Type 0 routing header (RFC 2460)
Figure 3.54: ICMP version 6 packet format

1 [Destination Unreachable. Such an ICMPv6 message is sent when the destination address of a packet
is unreachable. The code field of the ICMP header contains additional information about the type of
unreachability. The following codes are specified in RFC 4443]
0 : No route to destination. This indicates that the router that sent the ICMPv6 message did not
have a route towards the packets destination
1 : Communication with destination administratively prohibited. This indicates that a firewall has
refused to forward the packet towards its final destination.
2 : Beyond scope of source address. This message can be sent if the source is using link-local
addresses to reach a global unicast address outside its subnet.
3 : Address unreachable. This message indicates that the packet reached the subnet of the desti-
nation, but the host that owns this destination address cannot be reached.
4 : Port unreachable. This message indicates that the IPv6 packet was received by the destination,
but there was no application listening to the specified port.
2 : Packet Too Big. The router that was to send the ICMPv6 message received an IPv6 packet that is larger
than the MTU of the outgoing link. The ICMPv6 message contains the MTU of this link in bytes. This
allows the sending host to implement Path MTU discovery RFC 1981
3 : Time Exceeded. This error message can be sent either by a router or by a host. A router would set code
to 0 to report the reception of a packet whose Hop Limit reached 0. A host would set code to 1 to report that
it was unable to reassemble received IPv6 fragments.
4 : Parameter Problem. This ICMPv6 message is used to report either the reception of an IPv6 packet with
an erroneous header field (type 0) or an unknown Next Header or IP option (types 1 and 2). In this case, the
message body contains the erroneous IPv6 packet and the first 32 bits of the message body contain a pointer
to the error.
The Destination Unreachable ICMP error message is returned when a packet cannot be forwarded to its final
destination. The first four ICMPv6 error messages (type 1, codes 0-3) are generated by routers while endhosts
may return code 4 when there is no application bound to the corresponding port number.
The Packet Too Big ICMP messages enable the source endhost to discover the MTU size that it can safely use to
reach a given destination. To understand its operation, consider the (academic) scenario shown in the figure below.
In this figure, the labels on each link represent the maximum packet size supported by this link.

1500
R1
1400
R2
1300
R3
1500

If A sends a 1500 bytes packet, R1 will return an ICMPv6 error message indicating a maximum packet length of
1400 bytes. A would then fragment the packet before retransmitting it. The small fragment would go through, but
the large fragment will be refused by R2 that would return an ICMPv6 error message. A can refragment the packet
and send it to the final destination as two fragments.
In practice, an IPv6 implementation does not store the transmitted packets to be able to retransmit them if needed.
However, since TCP (and SCTP) buffer the segments that they transmit, a similar approach can be used in transport
protocols to detect the maximum MTU on a path towards a given destination. This technique is called PathMTU
Discovery RFC 1981.
When a TCP segment is transported in an IP packet that is fragmented in the network, the loss of a single fragment
forces TCP to retransmit the entire segment (and thus all the fragments). If TCP was able to send only packets
that do not require fragmentation in the network, it could retransmit only the information that was lost in the
network. In addition, IP reassembly causes several challenges at high speed as discussed in RFC 4963. Using IP
fragmentation to allow UDP applications to exchange large messages raises several security issues [KPS2003].
ICMPv6 is used by TCP implementations to discover the largest MTU size that is allowed to reach a destination
host without causing network fragmentation. A TCP implementation parses the Packets Too Big ICMP mes-
sages that it receives. These ICMP messages contain the MTU of the routers outgoing link in their Data field.
Upon reception of such an ICMP message, the source TCP implementation adjusts its Maximum Segment Size
(MSS) so that the packets containing the segments that it sends can be forwarded by this router without requiring
fragmentation.
Two types of informational ICMPv6 messages are defined in RFC 4443 : echo request and echo reply, which are
used to test the reachability of a destination by using ping6(8). Each host is supposed 35 to reply with an ICMP
Echo reply message when its receives an ICMP Echo request message. A sample usage of ping6(8) is shown
below.
#ping6 www.ietf.org
PING6(56=40+8+8 bytes) 2001:6a8:3080:2:3403:bbf4:edae:afc3 --> 2001:1890:123a::1:1e
16 bytes from 2001:1890:123a::1:1e, icmp_seq=0 hlim=49 time=156.905 ms
^C
--- www.ietf.org ping6 statistics ---
8 packets transmitted, 8 packets received, 0.0% packet loss
round-trip min/avg/max/std-dev = 155.325/155.810/156.905/0.447 ms
Another very useful debugging tool is traceroute6(8). The traceroute man page describes this tool as print
the route packets take to network host. traceroute uses the Time exceeded ICMP messages to discover the inter-
mediate routers on the path towards a destination. The principle behind traceroute is very simple. When a router
receives an IP packet whose Hop Limit is set to 1 it is forced to return to the sending host a Time exceeded ICMP
message containing the header and the first bytes of the discarded packet. To discover all routers on a network
path, a simple solution is to first send a packet whose Hop Limit is set to 1, then a packet whose Hop Limit is set
to 2, etc. A sample traceroute6 output is shown below.
#traceroute6 www.ietf.org
traceroute6 to www.ietf.org (2001:1890:1112:1::20) from 2001:6a8:3080:2:217:f2ff:fed6:65c0, 30 hop
1 2001:6a8:3080:2::1 13.821 ms 0.301 ms 0.324 ms
2 2001:6a8:3000:8000::1 0.651 ms 0.51 ms 0.495 ms
3 10ge.cr2.bruvil.belnet.net 3.402 ms 3.34 ms 3.33 ms
4 10ge.cr2.brueve.belnet.net 3.668 ms 10ge.cr2.brueve.belnet.net 3.988 ms 10ge.cr2.brueve.beln
5 belnet.rt1.ams.nl.geant2.net 10.598 ms 7.214 ms 10.082 ms
6 so-7-0-0.rt2.cop.dk.geant2.net 20.19 ms 20.002 ms 20.064 ms
7 kbn-ipv6-b1.ipv6.telia.net 21.078 ms 20.868 ms 20.864 ms
35Until a few years ago, all hosts replied to Echo request ICMP messages. However, due to the security problems that have affected TCP/IP
implementations, many of these implementations can now be configured to disable answering Echo request ICMP messages.

8 s-ipv6-b1-link.ipv6.telia.net 31.312 ms 31.113 ms 31.411 ms

9 s-ipv6-b1-link.ipv6.telia.net 61.986 ms 61.988 ms 61.994 ms
10 2001:1890:61:8909::1 121.716 ms 121.779 ms 121.177 ms
11 2001:1890:61:9117::2 203.709 ms 203.305 ms 203.07 ms
12 mail.ietf.org 204.172 ms 203.755 ms 203.748 ms
Note: Rate limitation of ICMP messages

High-end hardware based routers use special purpose chips on their interfaces to forward IPv6 packets at line
rate. These chips are optimised to process correct IP packets. They are not able to create ICMP messages at line
rate. When such a chip receives an IP packet that triggers an ICMP message, it interrupts the main CPU of the
router and the software running on this CPU processes the packet. This CPU is much slower than the hardware
acceleration found on the interfaces [Gill2004]. It would be overloaded if it had to process IP packets at line rate
and generate one ICMP message for each received packet. To protect this CPU, high-end routers limit the rate at
which the hardware can interrupt the main CPU and thus the rate at which ICMP messages can be generated. This
implies that not all erroneous IP packets cause the transmission of an ICMP message. The risk of overloading the
main CPU of the router is also the reason why using hop-by-hop IPv6 options, including the router alter option is
discouraged 36 .
3.12 The IPv6 subnet
Until now, we have focussed our discussion on the utilisation of IPv6 on point-to-point links. Although there
are point-to-point links in the Internet, mainly between routers and sometimes for endhosts, most of the endhosts
are attached to datalink layer networks such as Ethernet LANs or WiFi networks. These datalink layer networks
play an important role in todays Internet and have heavily influenced the design of the operation of IPv6. To
understand IPv6 and ICMPv6 completely, we first need to correctly understand the key principles behind these
datalink layer technologies.
As explained earlier, devices attached to a Local Area Network can directly exchange frames among themselves.
For this, each datalink layer interface on a device (endhost, router, ...) attached to such a network is identified
by a MAC address. Each datalink layer interface includes a unique hardwired MAC address. MAC addresses
are allocated to manufacturers in blocks and interface is numbered with a unique address. Thanks to the global
unicity of the MAC addresses, the datalink layer service can assume that two hosts attached to a LAN have
different addresses. Most LANs provide an unreliable connectionless service and a datalink layer frame has a
header containing :
the source MAC address
the destination MAC address
some multiplexing information to indicate the network layer protocol that is responsible for the payload of
the frame
LANs also provide a broadcast and a multicast service. The broadcast service enables a device to send a single
frame to all the devices attached to the same LAN. This is done by reserving a special broadcast MAC address
(typically all bits of the address are set to one). To broadcast a frame, a device simply needs to send a frame whose
destination is the broadcast address. All devices attached to the datalink network will receive the frame.
The broadcast service allows to easily reach all devices attached to a datalink layer network. It has been widely
used to support IP version 4. A drawback of using the broadcast service to support a network layer protocol is that
a broadcast frame that contains a network layer packet is always delivered to all devices attached to the datalink
network, even if some of these devices do not support the network layer protocol. The multicast service is a useful
alternative to the broadcast service. To understand its operation, it is important to understand how a datalink layer
36 For a discussion of the issues with the router alert IP option, see http://tools.ietf.org/html/draft-rahman-rtg-router-alert-dangerous-00 or
http://tools.ietf.org/html/draft-rahman-rtg-router-alert-considerations-03
3.12. The IPv6 subnet 183

interface operates. In shared media LANs, all devices are attached to the same physical medium and all frames are
delivered to all devices. When such a frame is received by a datalink layer interface, it compares the destination
address with the MAC address of the device. If the two addresses match, or the destination address is the broadcast
address, the frame is destined to the device and its payload is delivered to the network layer protocol. The multicast
service exploits this principle. A multicast address is a logical address. To receive frames destined to a multicast
address in a shared media LAN, a device captures all frames having this multicast address as their destination. All
IPv6 nodes are capable of capturing datalink layer frames destined to different multicast addresses.
3.12.1 Interactions between IPv6 and the datalink layer
IPv6 hosts and routers frequently interact with the datalink layer service. To understand the main interactions, it
is useful to analyze all the packets that are exchanged when a simple network containing a few hosts and routers
is built. Let us first start with a LAN containing two hosts 37 .
A B
MAC : 0023:4567:89ab MAC : 0034:5678:9abc
lan
Hosts A and B are attached to the same datalink layer network. They can thus exchange frames by using the MAC
addresses shown in the figure above. To be able to use IPv6 to exchange packets, they need to have an IPv6 address.
One possibility would be to manually configure an IPv6 address on each host. However, IPv6 provides a better so-
lution thanks to the link-local IPv6 addresses. A link-local IPv6 address is an address that is composed by concate-
nating the fe80:://64 prefix with the MAC address of the device. In the example above, host A would use IPv6
link-local address fe80::0223:45FF:FE67:89ab and host B fe80::0234:5678:9aFF:FEbc:dede.
With these two IPv6 addresses, the hosts can exchange IPv6 packets.
Note: Converting MAC addresses in host identifiers

Appendix A of RFC 4291 provides the algorithm used to convert a 48 bits MAC address into a 64 bits host
identifier. This algorithm builds upon the structure of the MAC addresses. A MAC address is represented as
MAC addresses are allocated in blocks of 220 . When a company registers for a block of MAC addresses, it receives
an identifier. company identifier is then used to populated the c bits of the MAC addresses. The company can
allocate all addresses in starting with this prefix and mangages the m bits as it wishes.
37 For simplicity, you assume that each datalink layer interface is assigned a 64 bits MAC address. As we will see later, todays datalink
layer technologies mainly use 48 bits MAC addresses, but the smaller addresses can easily be converted into 64 bits addresses.

Figure 3.55: A MAC address
Figure 3.56: A MAC address converted into a 64 bits host identifier
Inside a MAC address, the two bits indicated as 0 and g in the figure above play a special role. The first bit
indicates whether the address is universal or local. The g bit indicates whether this is a multicast address or a
unicast address. The MAC address can be converted into a 64 bits host identifier by flipping the value of the 0
bit and inserting FFFE, i.e. 1111111111111110 in binary, in the middle of the address as shown in the figure
below. The c, m and g bits of the MAC address are not modified.
The next step is to connect the LAN to the Internet. For this, a router is attached to the LAN.
router
A B 0045:6789:abcd
MAC : 0023:4567:89ab MAC : 0034:5678:9abc
lan
Assume that the LAN containing the two hosts and the router is assigned prefix 2001:db8:1234:5678/64.
A first solution to configure the IPv6 addresses in this network is to assign them manually. A possible assignment
is :
2001:db8:1234:5678::1 is assigned to router
2001:db8:1234:5678::AA is assigned to hostA
2001:db8:1234:5678::BB is assigned to hostB
To be able to exchange IPv6 packets with hostB, hostA needs to know the MAC address of the interface of
hostB on the LAN. This is the address resolution problem. In IPv6, this problem is solved by using the Neighbor
Discovery Protocol (NDP). NDP is specified in RFC 4861. This protocol is part of ICMPv6 and uses the multicast
datalink layer service.
NDP allows a host to discover the MAC address used by any other host attached to the same LAN. NDP operates in
two steps. First, the querier sends a multicast ICMPv6 Neighbor Solicitation message that contains as parameter
the queried IPv6 address. This multicast ICMPv6 NS is placed inside a multicast frame 38 . The queried node
receives the frame, parses it and replies with a unicast ICMPv6 Neighbor Advertisement that provides its own
IPv6 and MAC addresses. Upon reception of the Neighbor Advertisement message, the querier stores the mapping
38 RFC 4291 and RFC 4861 explain in more details how the IPv6 multicast address is determined from the target IPv6 unicast address.
These details are outside the scope of this book, but may matter if you try to understand a packet trace.

between the IPv6 and the MAC address inside its NDP table. This table is a data structure that maintains a cache
of the recently received Neighbor Advertisement. Thanks to this cache, a host only needs to send a Neighbor
Sollicitation message for the first packet that it sends to a given host. After this initial packet, the NDP table can
provide the mapping between the destination IPv6 address and the corresponding MAC address.
router hostA hostB
NS : Who has 2001:db8:1234:5678::BB
NA : 1234:5678:9abc:dede
The NS message can also be used to verify the reachability of a host in the local subnet. For this usage, NS
messages can be sent in unicast since other nodes on the subnet do not need to process the message.
When an entry in the NDP table times out on a host, it may either be deleted or the host may try to revalidate it by
sending the NS message again.
This is not the only usage of the Neighbor Solicitation and Neighbor Advertisement messages. They are also
used to detect the utilization of duplicate addresses. In the network above, consider what happens when a
new host is connected to the LAN. If this host is configured by mistake with the same address as hostA (i.e.
2001:db8:1234:5678::AA), problems could occur. Indeed, if two hosts have the same IPv6 address on the
LAN, but different MAC addresses, it will be difficult to correctly reach them. IPv6 anticipated this problem and
includes a Duplicate Address Detection Algorithm (DAD). When an IPv6 address 39 is configured on a host, by
any means, the host must verify the uniqueness of this address on the LAN. For this, it multicasts an ICMPv6
Neighbor Solicitation that queries the network for its newly configured address. The IPv6 source address of this
NS is set to :: (i.e. the reserved unassigned address) if the host does not already have an IPv6 address on this
subnet). If the NS does not receive any answer, the new address is considered to be unique and can safely be
used. Otherwise, the new address is refused and an error message should be returned to the system administra-
tor or a new IPv6 address should be generated. The Duplicate Address Detection Algorithm can prevent various
operational problems that are often difficult to debug.
Few users manually configure the IPv6 addresses on their hosts. They prefer to rely on protocols that can auto-
matically configure their IPv6 addresses. IPv6 supports two such protocols : DHCPv6 and the Stateless Address
Autoconfiguration (SLAAC).
The Stateless Address Autoconfiguration (SLAAC) mechanism defined in RFC 4862 enables hosts to automat-
ically configure their addresses without maintaining any state. When a host boots, it derives its identifier from
its datalink layer address 40 as explained earlier and concatenates this 64 bits identifier to the FE80::/64 prefix
to obtain its link-local IPv6 address. It then multicasts a Neighbour Solicitation with its link-local address as a
target to verify whether another host is using the same link-local address on this subnet. If it receives a Neighbour
Advertisement indicating that the link-local address is used by another host, it generates another 64 bits identifier
and sends again a Neighbour Solicitation. If there is no answer, the host considers its link-local address to be
valid. This address will be used as the source address for all NDP messages sent on the subnet.
To automatically configure its global IPv6 address, the host must know the globally routable IPv6 prefix that is
used on the local subnet. IPv6 routers regularly multicast ICMPv6 Router Advertisement messages that indicate
the IPv6 prefix assigned to the subnet. The Router Advertisement message contains several interesting fields.
This message is sent from the link-local address of the router on the subnet. Its destination is the IPv6 multicast
address that targets all IPv6 enabled hosts (i.e. ff02::1). The Cur Hop Limit field, if different from zero, allows
to specify the default Hop Limit that hosts should use when sending IPv6 from this subnet. 64 is a frequently used
value. The M and O bits are used to indicate that some information can be obtained from DHCPv6. The Router
Lifetime parameter provides the expected lifetime (in seconds) of the sending router acting as a default router.
This lifetime allows to plan the replacement of a router by another one in the same subnet. The Reachable Time
and the Retrans Timer parameter are used to configure the utilisation of the NDP protocol on the hosts attached to
the subnet.
39 The DAD algorithm is also used with link-local addresses.
40 Using a datalink layer address to derive a 64 bits identifier for each host raises privacy concerns as the host will always use the same
identifier. Attackers could use this to track hosts on the Internet. An extension to the Stateless Address Configuration mechanism that does not
raise privacy concerns is defined in RFC 4941. These privacy extensions allow a host to generate its 64 bits identifier randomly every time it
attaches to a subnet. It then becomes impossible for an attacker to use the 64-bits identifier to track a host.

Figure 3.57: Format of the ICMPv6 Router Advertisement message
Several options can be included in the Router Advertisement message. The simplest one is the MTU option that
indicates the MTU to be used within the subnet. Thanks to this option, it is possible to ensure that all devices
attached to the same subnet use the same MTU. Otherwise, operational problems could occur. The Prefix option
is more important. It provides information about the prefix(es) that is (are) advertised by the router on the subnet.
Figure 3.58: The Prefix information option
The key information placed in this option are the prefix and its length. This allows the hosts attached to the
subnet to automatically configure their own IPv6 address. The Valid and Preferred Lifetimes provide information
about the expected lifetime of the prefixes. Associating some time validity to prefixes is a good practice from an
operational viewpoint. There are some situations where the prefix assigned to a subnet needs to change without
impacting the hosts attached to the subnet. This is often called the IPv6 renumbering problem in the literature
RFC 7010. A very simple scenario is the following. An SME subscribes to one ISP. Its router is attached to
another router of this ISP and advertises a prefix assigned by the ISP. The SME is composed of a single subnet
and all its hosts rely on stateless address configuration. After a few years, the SME decides to change of network
provider. It connects its router to the second ISP and receives a different prefix from this ISP. At this point, two
prefixes are advertised on the SMEs subnet. The old prefix can be advertised with a short lifetime to ensure that
hosts will stop using it while the new one is advertised with a longer lifetime. After sometime, the router stops
advertising the old prefix and the hosts stop using it. The old prefix can now be returned back to the first ISP. In
larger networks, renumbering an IPv6 remains a difficult operational problem [LeB2009].
Upon reception of this message, the host can derive its global IPv6 address by concatenating its 64 bits identifier
with the received prefix. It concludes the SLAAC by sending a Neighbour Solicitation message targeted at its
global IPv6 address to ensure that no other host is not using the same IPv6 address.
Note: Router Advertisements and Hop Limits

ICMPv6 Router Advertisements messages are regularly sent by routers. They are destined to all devices attached
to the local subnet and no router should ever forward them to another subnet. Still, these messages are sent inside
IPv6 packets whose Hop Limit is always set to 255. Given that the packet should not the forwarded outside of
the local subnet, the reader could expect instead a Hop Limit set to 1. Using a Hop Limit set to 255 provides one
important benefit from a security viewpoint and this hack has been adapted in several Internet protocols. When a
host receives a Router Advertisement message, it expects that this message has been generated by a router attached
to the same subnet. Using a Hop Limit of 255 provides a simple check for this. If the message was generated by
an attacker outside the subnet, it would reach the subnet with a decremented Hop Limit. Checking that the Hop

Limit is set to 255 is a simple 41 verification that the packet was generated on this particular subnet. RFC 5082
provides other examples of protocols that use this hack and discuss its limitations.
Routers regularly send Router Advertisement messages. These messages are triggered by a timer that is often set
at approximately 30 seconds. Usually, hosts wait for the arrival of a Router Advertisement message to configure
their address. This implies that hosts could sometimes need to wait 30 seconds before being able to configure their
address. If this delay is too long, a host can also send a Router Solicitation message. This message is sent towards
the multicast address that corresponds to all IPv6 routers (i.e. FF01::2) and the default router will reply.
The last point that needs to be explained about ICMPv6 is the Redirect message. This message is used when there
is more than one router on a subnet as shown in the figure below.
router1 router2
A B 0045:6789:abcd 0012:3456:7878
MAC : 0023:4567:89ab MAC : 0034:5678:9abc
lan
In this network, router1 is the default router for all hosts. The second router, router2 provides connectivity
to a specific IPv6 subnet, e.g. 2001:db8:abcd::/48. These two routers attached to the same subnet can be
used in different ways. First, it is possible to manually configure the routing tables on all hosts to add a route
towards 2001:db8:abcd::/48 via router2. Unfortunately, forcing such manual configuration boils down
all the benefits of using address auto-configuration in IPv6. The second approach is to automatically configure
a default route via router1 on all hosts. With such route, when a host needs to send a packet to any address
within 2001:db8:abcd::/48, it will send it to router1. router1 would consult its routing table and find
that the packet needs to be sent again on the subnet to reach router2. This is a waste of time. A better approach
would be to enable the hosts to automatically learn the new route. This is possible thanks to the ICMPv6 Redirect
message. When router1 receives a packet that needs to be forwarded back on the same interface, it replies
with a Redirect message that indicates that the packet should have been sent via router2. Upon reception of a
Redirect message, the host updates it forwarding table to include a new transient entry for the destination reported
in the message. A timeout is usually associated with this transient entry to automatically delete it after some time.
An alternative is the Dynamic Host Configuration Protocol (DHCP) defined in RFC 2131 and RFC 3315. DHCP
allows a host to automatically retrieve its assigned IPv6 address, but relies on server. A DHCP server is associated
to each subnet 42 . Each DHCP server manages a pool of IPv6 addresses assigned to the subnet. When a host is
first attached to the subnet, it sends a DHCP request message in a UDP segment (the DHCP server listens on port
67). As the host knows neither its IPv6 address nor the IPv6 address of the DHCP server, this UDP segment is sent
inside a multicast packet target at the DHCP servers. The DHCP request may contain various options such as the
name of the host, its datalink layer address, etc. The server captures the DHCP request and selects an unassigned
address in its address pool. It then sends the assigned IPv6 address in a DHCP reply message which contains the
datalink layer address of the host and additional information such as the subnet mask, the address of the default
router or the address of the DNS resolver. The DHCP reply also specifies the lifetime of the address allocation.
This forces the host to renew its address allocation once it expires. Thanks to the limited lease time, IP addresses
are automatically returned to the pool of addresses when hosts are powered off.
Both SLAAC and DHCPv6 can be extended to provide additional information beyond the IPv6 prefix/address. For
example, RFC 6106 defines options for the ICMPv6 ND message that can carry the IPv6 address of the recursive
DNS resolver and a list of default domain search suffixes. It is also possible to combine SLAAC with DHCPv6.
41 Using a Hop Limit of 255 prevents one family of attacks against ICMPv6, but other attacks still remain possible. A detailed discussion
of the security issues with IPv6 is outside the scope of this book. It is possible to secure NDP by using the Cryptographically Generated IPv6
Addresses (CGA) defined in RFC 3972. The Secure Neighbour Discovery Protocol is defined in RFC 3971. A detailed discussion of the
security of IPv6 may be found in [HV2008].
42 In practice, there is usually one DHCP server per group of subnets and the routers capture on each subnet the DHCP messages and
forward them to the DHCP server.

RFC 3736 defines a stateless variant of DHCPv6 that can be used to distribute DNS information while SLAAC is
used to distribute the prefixes.
Warning: This is an unpolished draft of the second edition of this ebook. If you find any error or have
suggestions to improve the text, please create an issue via https://github.com/obonaventure/cnp3/issues/new
3.13 Routing in IP networks
In a large IP network such as the global Internet, routers need to exchange routing information. The Internet is an
interconnection of networks, often called domains, that are under different responsibilities. As of this writing, the
Internet is composed on more than 40,000 different domains and this number is still growing 43 . A domain can
be a small enterprise that manages a few routers in a single building, a larger enterprise with a hundred routers
at multiple locations, or a large Internet Service Provider managing thousands of routers. Two classes of routing
protocols are used to allow these domains to efficiently exchange routing information.
Figure 3.59: Organisation of a small Internet
The first class of routing protocols are the intradomain routing protocols (sometimes also called the interior gate-
way protocols or IGP). An intradomain routing protocol is used by all routers inside a domain to exchange routing
information about the destinations that are reachable inside the domain. There are several intradomain routing
protocols. Some domains use RIP, which is a distance vector protocol. Other domains use link-state routing pro-
tocols such as OSPF or IS-IS. Finally, some domains use static routing or proprietary protocols such as IGRP or
EIGRP.
These intradomain routing protocols usually have two objectives. First, they distribute routing information that
corresponds to the shortest path between two routers in the domain. Second, they should allow the routers to
quickly recover from link and router failures.
The second class of routing protocols are the interdomain routing protocols (sometimes also called the exterior
gateway protocols or EGP). The objective of an interdomain routing protocol is to distribute routing information
between domains. For scalability reasons, an interdomain routing protocol must distribute aggregated routing
information and considers each domain as a black box.
A very important difference between intradomain and interdomain routing are the routing policies that are used
by each domain. Inside a single domain, all routers are considered equal, and when several routes are available
to reach a given destination prefix, the best route is selected based on technical criteria such as the route with the
shortest delay, the route with the minimum number of hops or the route with the highest bandwidth.
When we consider the interconnection of domains that are managed by different organisations, this is no longer
true. Each domain implements its own routing policy. A routing policy is composed of three elements : an import
filter that specifies which routes can be accepted by a domain, an export filter that specifies which routes can be
advertised by a domain and a ranking algorithm that selects the best route when a domain knows several routes
43 See http://bgp.potaroo.net/index-as.html for reports on the evolution of the number of Autonomous Systems over time.
3.13. Routing in IP networks 189

towards the same destination prefix. As we will see later, another important difference is that the objective of the
interdomain routing protocol is to find the cheapest route towards each destination. There is only one interdomain
routing protocol : BGP.
3.14 Intradomain routing
In this section, we briefly describe the key features of the two main intradomain unicast routing protocols : RIP
and OSPF. The basic principles of distance vector and link-state routing have been presented earlier.
3.14.1 RIP
The Routing Information Protocol (RIP) is the simplest routing protocol that was standardised for the TCP/IP
protocol suite. RIP is defined in RFC 2453. Additional information about RIP may be found in [Malkin1999]
RIP routers periodically exchange RIP messages. The format of these messages is shown below. A RIP message
is sent inside a UDP segment whose destination port is set to 521. A RIP message contains several fields. The
Cmd field indicates whether the RIP message is a request or a response. When a router boots, its routing table is
empty and it cannot forward any packet. To speedup the discovery of the network, it can send a request message to
the RIP IPv6 multicast address, FF02::9. All RIP routers listen to this multicast address and any router attached
to the subnet will reply by sending its own routing table as a sequence of RIP messages. In steady state, routers
multicast one of more RIP response messages every 30 seconds. These messages contain the distance vectors that
summarize the routers routing table. The current version of RIP is version 2 defined in RFC 2453 for IPv4 and
RFC 2080 for IPv6.
Figure 3.60: The RIP message format
Each RIP message contains a set of route entries. Each route entry is encoded as a 20 bytes field whose format is
shown below. RIP was initially designed to be suitable for different network layer protocols. Some implementa-
tions of RIP were used in XNS or IPX networks RFC 2453. The format of the route entries used by RFC 2080
is shown below. Plen is the length of the subnet identifier in bits and the metric is encoded as one byte. The
maximum metric supported by RIP is 15.
Figure 3.61: Format of the RIP IPv6 route entries
Note: A note on timers

The first RIP implementations sent their distance vector exactly every 30 seconds. This worked well in most
networks, but some researchers noticed that routers were sometimes overloaded because they were processing
too many distance vectors at the same time [FJ1994]. They collected packet traces in these networks and found
that after some time the routers timers became synchronised, i.e. almost all routers were sending their distance
vectors at almost the same time. This synchronisation of the transmission times of the distance vectors caused
an overload on the routers CPU but also increased the convergence time of the protocol in some cases. This
was mainly due to the fact that all routers set their timers to the same expiration time after having processed the
received distance vectors. Sally Floyd and Van Jacobson proposed in [FJ1994] a simple solution to solve this
synchronisation problem. Instead of advertising their distance vector exactly after 30 seconds, a router should
send its next distance vector after a delay chosen randomly in the [15,45] interval RFC 2080. This randomisation
of the delays prevents the synchronisation that occurs with a fixed delay and is now a recommended practice for
protocol designers.
3.14.2 OSPF
Link-state routing protocols are used in IP networks. Open Shortest Path First (OSPF), defined in RFC 2328, is the
link state routing protocol that has been standardised by the IETF. The last version of OSPF, which supports IPv6,
is defined in RFC 5340. OSPF is frequently used in enterprise networks and in some ISP networks. However,
ISP networks often use the IS-IS link-state routing protocol [ISO10589] , which was developed for the ISO CLNP
protocol but was adapted to be used in IP RFC 1195 networks before the finalisation of the standardisation
of OSPF. A detailed analysis of ISIS and OSPF may be found in [BMO2006] and [Perlman2000]. Additional
information about OSPF may be found in [Moy1998].
Compared to the basics of link-state routing protocols that we discussed in section Link state routing, there are
some particularities of OSPF that are worth discussing. First, in a large network, flooding the information about
all routers and links to thousands of routers or more may be costly as each router needs to store all the information
about the entire network. A better approach would be to introduce hierarchical routing. Hierarchical routing
divides the network into regions. All the routers inside a region have detailed information about the topology of
the region but only learn aggregated information about the topology of the other regions and their interconnections.
OSPF supports a restricted variant of hierarchical routing. In OSPFs terminology, a region is called an area.
OSPF imposes restrictions on how a network can be divided into areas. An area is a set of routers and links that
are grouped together. Usually, the topology of an area is chosen so that a packet sent by one router inside the area
can reach any other router in the area without leaving the area 44 . An OSPF area contains two types of routers
RFC 2328:
Internal router : A router whose directly connected networks belong to the area
Area border routers : A router that is attached to several areas.
For example, the network shown in the figure below has been divided into three areas : area 1, containing routers
R1, R3, R4, R5 and RA, area 2 containing R7, R8, R9, R10, RB and RC. OSPF areas are identified by a 32 bit
integer, which is sometimes represented as an IP address. Among the OSPF areas, area 0, also called the backbone
area has a special role. The backbone area groups all the area border routers (routers RA, RB and RC in the figure
below) and the routers that are directly connected to the backbone routers but do not belong to another area (router
RD in the figure below). An important restriction imposed by OSPF is that the path between two routers that
belong to two different areas (e.g. R1 and R8 in the figure below) must pass through the backbone area.
Inside each non-backbone area, routers distribute the topology of the area by exchanging link state packets with
the other routers in the area. The internal routers do not know the topology of other areas, but each router knows
how to reach the backbone area. Inside an area, the routers only exchange link-state packets for all destinations
that are reachable inside the area. In OSPF, the inter-area routing is done by exchanging distance vectors. This is
illustrated by the network topology shown below.
Let us first consider OSPF routing inside area 2. All routers in the area learn a route towards 2001:db8:1234::/48
and 2001:db8:5678::/48. The two area border routers, RB and RC, create network summary advertisements.
Assuming that all links have a unit link metric, these would be:
44 OSPF can support virtual links to connect routers together that belong to the same area but are not directly connected. However, this goes
beyond this introduction to OSPF.
3.14. Intradomain routing 191

Figure 3.62: OSPF areas
Figure 3.63: Hierarchical routing with OSPF

RB advertises 2001:db8:1234::/48 at a distance of 2 and 2001:db8:5678::/48 at a distance of 3

RC advertises 2001:db8:5678::/48 at a distance of 2 and 2001:db8:1234::/48 at a distance of 3
These summary advertisements are flooded through the backbone area attached to routers RB and RC. In its
routing table, router RA selects the summary advertised by RB to reach 2001:db8:1234::/48 and the summary
advertised by RC to reach 2001:db8:5678::/48. Inside area 1, router RA advertises a summary indicating that
2001:db8:1234::/48 and 2001:db8:5678::/48 are both at a distance of 3 from itself.
On the other hand, consider the prefixes 2001:db8:aaaa:0000::/64 and 2001:db8:aaaa:0001::/64 that are inside
area 1. Router RA is the only area border router that is attached to this area. This router can create two different
network summary advertisements :
2001:db8:aaaa:0000::/64 at a distance of 1 and 2001:db8:aaaa:0001::/64 at a distance of 2 from RA
2001:db8:aaaa:0000::/63 at a distance of 2 from RA
The first summary advertisement provides precise information about the distance used to reach each prefix. How-
ever, all routers in the network have to maintain a route towards 2001:db8:aaaa:0000::/64 and a route towards
2001:db8:aaaa:0001::/64 that are both via router RA. The second advertisement would improve the scalability
of OSPF by reducing the number of routes that are advertised across area boundaries. However, in practice this
requires manual configuration on the border routers.
The second OSPF particularity that is worth discussing is the support of Local Area Networks (LAN). As shown
in the example below, several routers may be attached to the same LAN.
R1 R2 R3 R4
2001:db8:1234::11/48 2001:db8:1234::22/48 2001:db8:1234::33/48 2001:db8:1234::44/48
lan
A first solution to support such a LAN with a link-state routing protocol would be to consider that a LAN is
equivalent to a full-mesh of point-to-point links as if each router can directly reach any other router on the LAN.
However, this approach has two important drawbacks :
1. Each router must exchange HELLOs and link state packets with all the other routers on the LAN. This
increases the number of OSPF packets that are sent and processed by each router.
2. Remote routers, when looking at the topology distributed by OSPF, consider that there is a full-mesh of
links between all the LAN routers. Such a full-mesh implies a lot of redundancy in case of failure, while in
practice the entire LAN may completely fail. In case of a failure of the entire LAN, all routers need to detect
the failures and flood link state packets before the LAN is completely removed from the OSPF topology by
remote routers.
To better represent LANs and reduce the number of OSPF packets that are exchanged, OSPF handles LAN differ-
ently. When OSPF routers boot on a LAN, they elect 45 one of them as the Designated Router (DR) RFC 2328.
The DR router represents the local area network, and advertises the LANs subnet. Furthermore, LAN routers
only exchange HELLO packets with the DR. Thanks to the utilisation of a DR, the topology of the LAN appears
as a set of point-to-point links connected to the DR router.
Note: How to quickly detect a link failure ?

Network operators expect an OSPF network to be able to quickly recover from link or router failures [VPD2004].
In an OSPF network, the recovery after a failure is performed in three steps [FFEB2005] :
the routers that are adjacent to the failure detect it quickly. The default solution is to rely on the regular
exchange of HELLO packets. However, the interval between successive HELLOs is often set to 10 sec-
onds... Setting the HELLO timer down to a few milliseconds is difficult as HELLO packets are created and
45 The OSPF Designated Router election procedure is defined in RFC 2328. Each router can be configured with a router priority that
influences the election process since the router with the highest priority is preferred when an election is run.
3.14. Intradomain routing 193

processed by the main CPU of the routers and these routers cannot easily generate and process a HELLO
packet every millisecond on each of their interfaces. A better solution is to use a dedicated failure detection
protocol such as the Bidirectional Forwarding Detection (BFD) protocol defined in [KW2009] that can be
implemented directly on the router interfaces. Another solution to be able to detect the failure is to instru-
ment the physical and the datalink layer so that they can interrupt the router when a link fails. Unfortunately,
such a solution cannot be used on all types of physical and datalink layers.
the routers that have detected the failure flood their updated link state packets in the network
all routers update their routing table
A last, but operationally important, point needs to be discussed about intradomain routing protocols such as OSPF
and IS-IS. Intradomain routing protocols always select the shortest path for each destination. In practice, there are
often several equal paths towards the same destination. When a router computes several equal cost paths towards
one destination, it can use these paths in different ways.
A first approach is to select one of the equal cost paths (e.g. the first or the last path found by the SPF computation)
and install it in the forwarding table. In this case, only one path is used to reach each destination.
A second approach is to install all equal cost paths 46 in the forwarding table and load-balance the packets on the
different paths. Consider the case where a router has N different outgoing interfaces to reach destination d. A first
possibility to load-balance the traffic among these interfaces is to use round-robin. Round-robin allows to equally
balance the packets among the N outgoing interfaces. This equal load-balancing is important in practice because
it allows to better spread the load throughout the network. However, few networks use this round-robin strategy
to load-balance traffic on routers. The main drawback of round-robin is that packets that belong to the same flow
(e.g. TCP connection) may be forwarded over different paths. If packets belonging to the same TCP connection
are sent over different paths, they will probably experience different delays and arrive out-of-sequence at their
destination. When a TCP receiver detects out-of-order segments, it sends duplicate acknowledgements that may
cause the sender to initiate a fast retransmission and enter congestion avoidance. Thus, out-of-order segments may
lead to lower TCP performance. This is annoying for a load-balancing technique whose objective is to improve
the network performance by spreading the load.
To efficiently spread the load over different paths, routers need to implement per-flow load-balancing. This implies
that they must forward all the packets that belong to the same flow on the same path. Since a TCP connection is
always identified by the four-tuple (source and destination addresses, source and destination ports), one possibility
would be to select an outgoing interface upon arrival of the first packet of the flow and store this decision in the
routers memory. Unfortunately, such a solution does not scale since the required memory grows with the number
of TCP connections that pass through the router.
Fortunately, it is possible to perform per-flow load balancing without maintaining any state on the router. Most
routers today use hash functions for this purpose RFC 2991. When a packet arrives, the router extracts the Next
Header information and the four-tuple from the packet and computes :
hash(N extHeader, IPsrc , IPdst , P ortsrc , P ortdst ) (mod N )
In this formula, N is the number of outgoing interfaces on the equal cost paths towards the packets destination.
Various hash functions are possible, including CRC, checksum or MD5 RFC 2991. Since the hash function is
computed over the four-tuple, the same hash value will be computed for all packets belonging to the same flow.
This prevents reordering due to load balancing inside the network. Most routers support this kind of load-balancing
today [ACO+2006].
46 In some networks, there are several dozens of paths towards a given destination. Some routers, due to hardware limitations, cannot install
more than 8 or 16 paths in their forwarding table. In this case, a subset of the computed paths is installed in the forwarding table.

3.15 Interdomain routing
As explained earlier, the Internet is composed of more than 45,000 different networks 47 called domains. Each
domain is composed of a group of routers and hosts that are managed by the same organisation. Example domains
include belnet, sprint, level3, geant, abilene, cisco or google ...
Each domain contains a set of routers. From a routing point of view, these domains can be divided into two
classes : the transit and the stub domains. A stub domain sends and receives packets whose source or destination
are one of its own hosts. A transit domain is a domain that provides a transit service for other domains, i.e. the
routers in this domain forward packets whose source and destination do not belong to the transit domain. As of
this writing, about 85% of the domains in the Internet are stub domains 48 . A stub domain that is connected to a
single transit domain is called a single-homed stub. A multihomed stub is a stub domain connected to two or more
transit providers.
Figure 3.64: Transit and stub domains
The stub domains can be further classified by considering whether they mainly send or receive packets. An access-
rich stub domain is a domain that contains hosts that mainly receive packets. Typical examples include small
ADSL- or cable modem-based Internet Service Providers or enterprise networks. On the other hand, a content-
rich stub domain is a domain that mainly produces packets. Examples of content-rich stub domains include google,
yahoo, microsoft, facebook or content distribution networks such as akamai or limelight For the last few years, we
have seen a rapid growth of these content-rich stub domains. Recent measurements [ATLAS2009] indicate that a
growing fraction of all the packets exchanged on the Internet are produced in the data centers managed by these
content providers.
Domains need to be interconnected to allow a host inside a domain to exchange IP packets with hosts located
in other domains. From a physical perspective, domains can be interconnected in two different ways. The first
solution is to directly connect a router belonging to the first domain with a router inside the second domain. Such
links between domains are called private interdomain links or private peering links. In practice, for redundancy or
performance reasons, distinct physical links are usually established between different routers in the two domains
that are interconnected.
Figure 3.65: Interconnection of two domains via a private peering link
Such private peering links are useful when, for example, an enterprise or university network needs to be connected
to its Internet Service Provider. However, some domains are connected to hundreds of other domains 49 . For some
47 An analysis of the evolution of the number of domains on the global Internet during the last ten years may be found in
http://www.potaroo.net/tools/asn32/
48 Several web sites collect and analyse data about the evolution of BGP in the global Internet. http://bgp.potaroo.net provides lots of
statistics and analyses that are updated daily.

49 See http://as-rank.caida.org/ for an analysis of the interconnections between domains based on measurements collected in the global
Internet
3.15. Interdomain routing 195

of these domains, using only private peering links would be too costly. A better solution to allow many domains
to interconnect cheaply are the Internet eXchange Points (IXP). An IXP is usually some space in a data center that
hosts routers belonging to different domains. A domain willing to exchange packets with other domains present
at the IXP installs one of its routers on the IXP and connects it to other routers inside its own network. The IXP
contains a Local Area Network to which all the participating routers are connected. When two domains that are
present at the IXP wish 50 to exchange packets, they simply use the Local Area Network. IXPs are very popular
in Europe and many Internet Service Providers and Content providers are present in these IXPs.
Figure 3.66: Interconnection of two domains at an Internet eXchange Point
In the early days of the Internet, domains would simply exchange all the routes they know to allow a host inside
one domain to reach any host in the global Internet. However, in todays highly commercial Internet, this is no
longer true as interdomain routing mainly needs to take into account the economical relationships between the
domains. Furthermore, while intradomain routing usually prefers some routes over others based on their technical
merits (e.g. prefer route with the minimum number of hops, prefer route with the minimum delay, prefer high
bandwidth routes over low bandwidth ones, etc) interdomain routing mainly deals with economical issues. For
interdomain routing, the cost of using a route is often more important than the quality of the route measured by its
delay or bandwidth.
There are different types of economical relationships that can exist between domains. Interdomain routing converts
these relationships into peering relationships between domains that are connected via peering links.
The first category of peering relationship is the customer->provider relationship. Such a relationship is used when
a customer domain pays an Internet Service Provider to be able to exchange packets with the global Internet over
an interdomain link. A similar relationship is used when a small Internet Service Provider pays a larger Internet
Service Provider to exchange packets with the global Internet.
Figure 3.67: A simple Internet with peering relationships
To understand the customer->provider relationship, let us consider the simple internetwork shown in the figure
above. In this internetwork, AS7 is a stub domain that is connected to one provider : AS4. The contract between
AS4 and AS7 allows a host inside AS7 to exchange packets with any host in the internetwork. To enable this
exchange of packets, AS7 must know a route towards any domain and all the domains of the internetwork must
50 Two routers that are attached to the same IXP only exchange packets when the owners of their domains have an economical incentive to
exchange packets on this IXP. Usually, a router on an IXP is only able to exchange packets with a small fraction of the routers that are present
on the same IXP.

know a route via AS4 that allows them to reach hosts inside AS7. From a routing perspective, the commercial
contract between AS7 and AS4 leads to the following routes being exchanged :
over a customer->provider relationship, the customer domain advertises to its provider all its routes and all
the routes that it has learned from its own customers.
over a provider->customer relationship, the provider advertises all the routes that it knows to its customer.
The second rule ensures that the customer domain receives a route towards all destinations that are reachable via
its provider. The first rule allows the routes of the customer domain to be distributed throughout the Internet.
Coming back to the figure above, AS4 advertises to its two providers AS1 and AS2 its own routes and the routes
learned from its customer, AS7. On the other hand, AS4 advertises to AS7 all the routes that it knows.
The second type of peering relationship is the shared-cost peering relationship. Such a relationship usually does
not involve a payment from one domain to the other in contrast with the customer->provider relationship. A
shared-cost peering relationship is usually established between domains having a similar size and geographic
coverage. For example, consider the figure above. If AS3 and AS4 exchange many packets via AS1, they both need
to pay AS1. A cheaper alternative for AS3 and AS4 would be to establish a shared-cost peering. Such a peering
can be established at IXPs where both AS3 and AS4 are present or by using private peering links. This shared-cost
peering should be used to exchange packets between hosts inside AS3 and hosts inside AS4. However, AS3 does
not want to receive on the AS3-AS4 shared-cost peering links packets whose destination belongs to AS1 as AS3
would have to pay to send these packets to AS1.
From a routing perspective, over a shared-cost peering relationship a domain only advertises its internal routes
and the routes that it has learned from its customers. This restriction ensures that only packets destined to the
local domain or one of its customers is received over the shared-cost peering relationship. This implies that the
routes that have been learned from a provider or from another shared-cost peer is not advertised over a shared-cost
peering relationship. This is motivated by economical reasons. If a domain were to advertise the routes that it
learned from a provider over a shared-cost peering relationship that does not bring revenue, it would have allowed
its shared-cost peer to use the link with its provider without any payment. If a domain were to advertise the routes
it learned over a shared cost peering over another shared-cost peering relationship, it would have allowed these
shared-cost peers to use its own network (which may span one or more continents) freely to exchange packets.
Finally, the last type of peering relationship is the sibling. Such a relationship is used when two domains exchange
all their routes in both directions. In practice, such a relationship is only used between domains that belong to the
same company.
These different types of relationships are implemented in the interdomain routing policies defined by each domain.
The interdomain routing policy of a domain is composed of three main parts :
the import filter that specifies, for each peering relationship, the routes that can be accepted from the neigh-
bouring domain (the non-acceptable routes are ignored and the domain never uses them to forward packets)
the export filter that specifies, for each peering relationship, the routes that can be advertised to the neigh-
bouring domain
the ranking algorithm that is used to select the best route among all the routes that the domain has received
towards the same destination prefix
A domains import and export filters can be defined by using the Route Policy Specification Language (RPSL)
specified in RFC 2622 [GAVE1999] . Some Internet Service Providers, notably in Europe, use RPSL to document
51
their import and export policies. Several tools help to easily convert a RPSL policy into router commands.
The figure below provides a simple example of import and export filters for two domains in a simple internetwork.
In RPSL, the keyword ANY is used to replace any route from any domain. It is typically used by a provider to
indicate that it announces all its routes to a customer over a provider->customer relationship. This is the case
for AS4s export policy. The example below clearly shows the difference between a provider->customer and a
shared-cost peering relationship. AS4s export filter indicates that it announces only its internal routes (AS4) and
the routes learned from its clients (AS7) over its shared-cost peering with AS3, while it advertises all the routes
that it uses (including the routes learned from AS3) to AS7.
51 See ftp://ftp.ripe.net/ripe/dbase for the RIPE database that contains the import and export policies of many European ISPs

Figure 3.68: Import and export policies
3.15.1 The Border Gateway Protocol
The Internet uses a single interdomain routing protocol : the Border Gateway Protocol (BGP). The current ver-
sion of BGP is defined in RFC 4271. BGP differs from the intradomain routing protocols that we have already
discussed in several ways. First, BGP is a path-vector protocol. When a BGP router advertises a route towards a
prefix, it announces the IP prefix and the interdomain path used to reach this prefix. From BGPs point of view,
each domain is identified by a unique Autonomous System (AS) number 52 and the interdomain path contains the
AS numbers of the transit domains that are used to reach the associated prefix. This interdomain path is called the
AS Path. Thanks to these AS-Paths, BGP does not suffer from the count-to-infinity problems that affect distance
vector routing protocols. Furthermore, the AS-Path can be used to implement some routing policies. Another dif-
ference between BGP and the intradomain routing protocols is that a BGP router does not send the entire contents
of its routing table to its neighbours regularly. Given the size of the global Internet, routers would be overloaded
by the number of BGP messages that they would need to process. BGP uses incremental updates, i.e. it only
announces the routes that have changed to its neighbors.
The figure below shows a simple example of the BGP routes that are exchanged between domains. In this example,
prefix 2001:db8:1234/48 is announced by AS1. AS1 advertises a BGP route towards this prefix to AS2. The AS-
Path of this route indicates that AS1 is the originator of the prefix. When AS4 receives the BGP route from AS1,
it re-announces it to AS2 and adds its AS number to the AS-Path. AS2 has learned two routes towards prefix
2001:db8:1234/48. It compares the two routes and prefers the route learned from AS4 based on its own ranking
algorithm. AS2 advertises to AS5 a route towards 2001:db8:1234/48 with its AS-Path set to AS2:AS4:AS1. Thanks
to the AS-Path, AS5 knows that if it sends a packet towards 2001:db8:1234/48 the packet first passes through AS2,
then through AS4 before reaching its destination inside AS1.
BGP routers exchange routes over BGP sessions. A BGP session is established between two routers belonging to
two different domains that are directly connected. As explained earlier, the physical connection between the two
routers can be implemented as a private peering link or over an Internet eXchange Point. A BGP session between
two adjacent routers runs above a TCP connection (the default BGP port is 179). In contrast with intradomain
routing protocols that exchange IP packets or UDP segments, BGP runs above TCP because TCP ensures a reliable
delivery of the BGP messages sent by each router without forcing the routers to implement acknowledgements,
checksums, etc. Furthermore, the two routers consider the peering link to be up as long as the BGP session and
the underlying TCP connection remain up 53 . The two endpoints of a BGP session are called BGP peers.
In practice, to establish a BGP session between routers R1 and R2 in the figure above, the network administrator
of AS3 must first configure on R1 the IP address of R2 on the R1-R2 link and the AS number of R2. Router R1 then
regularly tries to establish the BGP session with R2. R2 only agrees to establish the BGP session with R1 once it
52 In this text, we consider Autonomous System and domain as synonyms. In practice, a domain may be divided into several Autonomous
Systems, but we ignore this detail.

53 The BGP sessions and the underlying TCP connection are typically established by the routers when they boot based on information found
in their configuration. The BGP sessions are rarely released, except if the corresponding peering link fails or one of the endpoints crashes or
needs to be rebooted.

2001:db8:cafe::/48
AS Path : AS2:AS4:AS1
AS2 AS5
2001:db8:cafe::/48
AS Path : AS1
2001:db8:cafe::/48
AS Path : AS4:AS1
AS1
2001:db8:cafe::/48
2001:db8:cafe::/48 AS4
AS Path : AS1
Figure 3.69: Simple exchange of BGP routes
Figure 3.70: A BGP peering session between two directly connected routers
has been configured with the IP address of R1 and its AS number. For security reasons, a router never establishes
a BGP session that has not been manually configured on the router.
The BGP protocol RFC 4271 defines several types of messages that can be exchanged over a BGP session :
OPEN : this message is sent as soon as the TCP connection between the two routers has been established.
It initialises the BGP session and allows the negotiation of some options. Details about this message may
be found in RFC 4271
NOTIFICATION : this message is used to terminate a BGP session, usually because an error has been de-
tected by the BGP peer. A router that sends or receives a NOTIFICATION message immediately shutdowns
the corresponding BGP session.
UPDATE: this message is used to advertise new or modified routes or to withdraw previously advertised
routes.
KEEPALIVE : this message is used to ensure a regular exchange of messages on the BGP session, even
when no route changes. When a BGP router has not sent an UPDATE message during the last 30 seconds,
it shall send a KEEPALIVE message to confirm to the other peer that it is still up. If a peer does not receive
any BGP message during a period of 90 seconds 54 , the BGP session is considered to be down and all the
routes learned over this session are withdrawn.
As explained earlier, BGP relies on incremental updates. This implies that when a BGP session starts, each router
first sends BGP UPDATE messages to advertise to the other peer all the exportable routes that it knows. Once
all these routes have been advertised, the BGP router only sends BGP UPDATE messages about a prefix if the
route is new, one of its attributes has changed or the route became unreachable and must be withdrawn. The BGP
UPDATE message allows BGP routers to efficiently exchange such information while minimising the number of
bytes exchanged. Each UPDATE message contains :
54 90 seconds is the default delay recommended by RFC 4271. However, two BGP peers can negotiate a different timer during the
establishment of their BGP session. Using a too small interval to detect BGP session failures is not recommended. BFD [KW2009] can be
used to replace BGPs KEEPALIVE mechanism if fast detection of interdomain link failures is required.

a list of IP prefixes that are withdrawn

a list of IP prefixes that are (re-)advertised
the set of attributes (e.g. AS-Path) associated to the advertised prefixes
In the remainder of this chapter, and although all routing information is exchanged using BGP UPDATE messages,
we assume for simplicity that a BGP message contains only information about one prefix and we use the words :
Withdraw message to indicate a BGP UPDATE message containing one route that is withdrawn
Update message to indicate a BGP UPDATE containing a new or updated route towards one destination
prefix with its attributes
From a conceptual point of view, a BGP router connected to N BGP peers, can be described as being composed
of four parts as shown in the figure below.
Figure 3.71: Organisation of a BGP router
In this figure, the router receives BGP messages on the left part of the figure, processes these messages and possibly
sends BGP messages on the right part of the figure. A BGP router contains three important data structures :
the Adj-RIB-In contains the BGP routes that have been received from each BGP peer. The routes in the
Adj-RIB-In are filtered by the import filter before being placed in the BGP-Loc-RIB. There is one import
filter per BGP peer.
the Local Routing Information Base (Loc-RIB) contains all the routes that are considered as acceptable by
the router. The Loc-RIB may contain several routes, learned from different BGP peers, towards the same
destination prefix.
the Forwarding Information Base (FIB) is used by the dataplane to forward packets towards their destination.
The FIB contains, for each destination, the best route that has been selected by the BGP decision process.
This decision process is an algorithm that selects, for each destination prefix, the best route according to the
routers ranking algorithm that is part of its policy.
the Adj-RIB-Out contains the BGP routes that have been advertised to each BGP peer. The Adj-RIB-Out for
a given peer is built by applying the peers export filter on the routes that have been installed in the FIB.
There is one export filter per BGP peer. For this reason, the Adj-RIB-Out of a peer may contain different
routes than the Adj-RIB-Out of another peer.
When a BGP session starts, the routers first exchange OPEN messages to negotiate the options that apply through-
out the entire session. Then, each router extracts from its FIB the routes to be advertised to the peer. It is important
to note that, for each known destination prefix, a BGP router can only advertise to a peer the route that it has itself
installed inside its FIB. The routes that are advertised to a peer must pass the peers export filter. The export filter
is a set of rules that define which routes can be advertised over the corresponding session, possibly after having
modified some of its attributes. One export filter is associated to each BGP session. For example, on a shared-cost
peering, the export filter only selects the internal routes and the routes that have been learned from a customer.
The pseudo-code below shows the initialisation of a BGP session.
def initialize_BGP_session( RemoteAS, RemoteIP):
# Initialize and start BGP session
# Send BGP OPEN Message to RemoteIP on port 179
# Follow BGP state machine

# advertise local routes and routes learned from peers*/

for d in BGPLocRIB :
B=build_BGP_Update(d)
S=Apply_Export_Filter(RemoteAS,B)
if (S != None) :
send_Update(S,RemoteAS,RemoteIP)
# entire RIB has been sent
# new Updates will be sent to reflect local or distant
# changes in routers
In the above pseudo-code, the build_BGP_UPDATE(d) procedure extracts from the BGP Loc-RIB the best path
towards destination d (i.e. the route installed in the FIB) and prepares the corresponding BGP UPDATE message.
This message is then passed to the export filter that returns NULL if the route cannot be advertised to the peer or
the (possibly modified) BGP UPDATE message to be advertised. BGP routers allow network administrators to
specify very complex export filters, see e.g. [WMS2004]. A simple export filter that implements the equivalent of
split horizon is shown below.
def apply_export_filter(RemoteAS, BGPMsg) :
# check if RemoteAS already received route
if RemoteAS is BGPMsg.ASPath :
BGPMsg=None
# Many additional export policies can be configured :
# Accept or refuse the BGPMsg
# Modify selected attributes inside BGPMsg
return BGPMsg
At this point, the remote router has received all the exportable BGP routes. After this initial exchange, the router
only sends BGP UPDATE messages when there is a change (addition of a route, removal of a route or change in
the attributes of a route) in one of these exportable routes. Such a change can happen when the router receives a
BGP message. The pseudo-code below summarizes the processing of these BGP messages.
def Recvd_BGPMsg(Msg, RemoteAS) :
B=apply_import_filter(Msg,RemoteAS)
if (B== None): # Msg not acceptable
return
if IsUPDATE(Msg):
Old_Route=BestRoute(Msg.prefix)
Insert_in_RIB(Msg)
Run_Decision_Process(RIB)
if (BestRoute(Msg.prefix) != Old_Route) :
# best route changed
B=build_BGP_Message(Msg.prefix);
S=apply_export_filter(RemoteAS,B);
if (S!=None) : # announce best route
send_UPDATE(S,RemoteAS,RemoteIP);
else if (Old_Route != None) :
send_WITHDRAW(Msg.prefix,RemoteAS, RemoteIP)
else : # Msg is WITHDRAW
Old_Route=BestRoute(Msg.prefix)
Remove_from_RIB(Msg)
Run_Decision_Process(RIB)
if (Best_Route(Msg.prefix) !=Old_Route):
# best route changed
B=build_BGP_Message(Msg.prefix)
S=apply_export_filter(RemoteAS,B)
if (S != None) : # still one best route towards Msg.prefix
send_UPDATE(S,RemoteAS, RemoteIP);
else if(Old_Route != None) : # No best route anymore
send_WITHDRAW(Msg.prefix,RemoteAS,RemoteIP);
When a BGP message is received, the router first applies the peers import filter to verify whether the message is
acceptable or not. If the message is not acceptable, the processing stops. The pseudo-code below shows a simple
import filter. This import filter accepts all routes, except those that already contain the local AS in their AS-Path.

If such a route was used, it would cause a routing loop. Another example of an import filter would be a filter used
by an Internet Service Provider on a session with a customer to only accept routes towards the IP prefixes assigned
to the customer by the provider. On real routers, import filters can be much more complex and some import filters
modify the attributes of the received BGP UPDATE [WMS2004] .
def apply_import_filter(RemoteAS, BGPMsg):
if MysAS in BGPMsg.ASPath :
BGPMsg=None
# Many additional import policies can be configured :
# Accept or refuse the BGPMsg
# Modify selected attributes inside BGPMsg
return BGPMsg
Note: The bogon filters

Another example of frequently used import filters are the filters that Internet Service Providers use to ignore
bogon routes. In the ISP community, a bogon route is a route that should not be advertised on the global Internet.
Typical examples include the documentation IPv6 prefix (2001:db8::/32 used for most examples in this book),
the loopback address (::1/128) or the IPv6 prefixes that have not yet been allocated by IANA. A well managed
BGP router should ensure that it never advertises bogons on the global Internet. Detailed information about these
bogons may be found in [IMHM2013].
If the import filter accepts the BGP message, the pseudo-code distinguishes two cases. If this is an Update
message for prefix p, this can be a new route for this prefix or a modification of the routes attributes. The router
first retrieves from its RIB the best route towards prefix p. Then, the new route is inserted in the RIB and the BGP
decision process is run to find whether the best route towards destination p changes. A BGP message only needs to
be sent to the routers peers if the best route has changed. For each peer, the router applies the export filter to verify
whether the route can be advertised. If yes, the filtered BGP message is sent. Otherwise, a Withdraw message
is sent. When the router receives a Withdraw message, it also verifies whether the removal of the route from its
RIB caused its best route towards this prefix to change. It should be noted that, depending on the content of the
RIB and the export filters, a BGP router may need to send a Withdraw message to a peer after having received an
Update message from another peer and conversely.
Let us now discuss in more detail the operation of BGP in an IPv6 network. For this, let us consider the simple
network composed of three routers located in three different ASes and shown in the figure below.
Figure 3.72: Utilisation of the BGP nexthop attribute
This network contains three routers : R1, R2 and R3. Each router is attached to a local IPv6 subnet that it
advertises using BGP. There are two BGP sessions, one between R1 and R2 and the second between R2 and
R3. A /127 subnet is used on each interdomain link (2001:db8::4/127 on R1-R2 and 2001:db8::0/127 on R2-
R3) in conformance with the latest recommendation RFC 6164. The BGP sessions run above TCP connections
established between the neighboring routers (e.g. 2001:db8::5 - 2001:db8::6 for the R1-R2 session).
Let us assume that the R1-R2 BGP session is the first to be established. A BGP Update message sent on such a
session contains three fields :
the advertised prefix
the BGP nexthop
the attributes including the AS-Path
We use the notation U(prefix, nexthop, attributes) to represent such a BGP Update message in this
section. Similarly, W(prefix) represents a BGP withdraw for the specified prefix. Once the R1-R2
session has been established, R1 sends U(2001:db8:1234::/48,2001:db8::5,AS10) to R2 and R2 sends

U(2001:db8:5678:/48,2001:db8::6,AS20). At this point, R1 can reach 2001:db8:5678::/48 via 2001:db8::6 and

R2 can reach 2001:db8:1234::/48 via 2001:db8::5.
Once the R2-R3 has been established, R3 sends U(2001:db8:acbd::/48,2001:db8::2,AS30). R2 announces on the
R2-R3 session all the routes inside its RIB. It thus sends to R3 : U(2001:db8:1234::/48,2001:db8::1,AS20:AS10)
and U(2001:db8:5678::/48,2001:db8::1,AS20). Note that when R2 advertises the route that it learned
from R1, it updates the BGP nexthop and adds its AS number to the AS-Path. R2 also sends
U(2001:db8:abcd::48,2001:db8::6,AS20:AS30) to R1 on the R1-R3 session. At this point, all BGP routes have
been exchanged and all routers can reach 2001:db8::1234/48, 2001:db8:5678::/48 and 2001:db8:abcd::/48.
If the link between R2 and R3 fails, R3 detects the failure as it did not receive KEEPALIVE messages recently from
R2. At this time, R3 removes from its RIB all the routes learned over the R2-R3 BGP session. R2 also removes
from its RIB the routes learned from R3. R2 also sends W(2001:db8:acbd::/48) to R1 over the R1-R3 BGP session
since it does not have a route anymore towards this prefix.
Note: Origin of the routes advertised by a BGP router

A frequent practical question about the operation of BGP is how a BGP router decides to originate or advertise a
route for the first time. In practice, this occurs in two situations :
the router has been manually configured by the network operator to always advertise one or several routes
on a BGP session. For example, on the BGP session between UCLouvain and its provider, belnet , UCLou-
vains router always advertises the 2001:6a8:3080/48 IPv6 prefix assigned to the campus network
the router has been configured by the network operator to advertise over its BGP session some of the routes
that it learns with its intradomain routing protocol. For example, an enterprise router may advertise over a
BGP session with its provider the routes to remote sites when these routes are reachable and advertised by
the intradomain routing protocol
The first solution is the most frequent. Advertising routes learned from an intradomain routing protocol is not
recommended, this is because if the route flaps 55 , this would cause a large number of BGP messages being
exchanged in the global Internet.
The BGP decision process
Besides the import and export filters, a key difference between BGP and the intradomain routing protocols is that
each domain can define is own ranking algorithm to determine which route is chosen to forward packets when
several routes have been learned towards the same prefix. This ranking depends on several BGP attributes that can
be attached to a BGP route.
The first BGP attribute that is used to rank BGP routes is the local-preference (local-pref) attribute. This attribute
is an unsigned integer that is attached to each BGP route received over an eBGP session by the associated import
filter.
When comparing routes towards the same destination prefix, a BGP router always prefers the routes with the
highest local-pref. If the BGP router knows several routes with the same local-pref, it prefers among the routes
having this local-pref the ones with the shortest AS-Path.
The local-pref attribute is often used to prefer some routes over others.
A common utilisation of local-pref is to support backup links. Consider the situation depicted in the figure below.
AS1 would always like to use the high bandwidth link to send and receive packets via AS2 and only use the backup
link upon failure of the primary one.
As BGP routers always prefer the routes with the highest local-pref attribute, this policy can be implemented
using the following import filter on R1
import: from AS2 RA at R1 set localpref=100;
from AS2 RB at R1 set localpref=200;
accept ANY
55 A link is said to be flapping if it switches several between an operational state and a disabled state within a short period of time. A router
attached to such a link would need to frequently send routing messages.

Figure 3.73: How to create a backup link with BGP ?
With this import filter, all the BGP routes learned from RB over the high bandwidth links are preferred over the
routes learned over the backup link. If the primary link fails, the corresponding routes are removed from R1s RIB
and R1 uses the route learned from RA. R1 reuses the routes via RB as soon as they are advertised by RB once the
R1-RB link comes back.
The import filter above modifies the selection of the BGP routes inside AS1. Thus, it influences the route followed
by the packets forwarded by AS1. In addition to using the primary link to send packets, AS1 would like to receive
its packets via the high bandwidth link. For this, AS2 also needs to set the local-pref attribute in its import filter.
import: from AS1 R1 at RA set localpref=100;
from AS1 R1 at RB set localpref=200;
accept AS1
Sometimes, the local-pref attribute is used to prefer a cheap link compared to a more expensive one. For example,
in the network below, AS1 could wish to send and receive packets mainly via its interdomain link with AS4.
Figure 3.74: How to prefer a cheap link over an more expensive one ?
AS1 can install the following import filter on R1 to ensure that it always sends packets via R2 when it has learned
a route via AS2 and another via AS4.
import: from AS2 RA at R1 set localpref=100;
from AS4 R2 at R1 set localpref=200;
accept ANY
However, this import filter does not influence how AS3 , for example, prefers some routes over others. If the link
between AS3 and AS2 is less expensive than the link between AS3 and AS4, AS3 could send all its packets via AS2
and AS1 would receive packets over its expensive link. An important point to remember about local-pref is that
it can be used to prefer some routes over others to send packets, but it has no influence on the routes followed by
received packets.
Another important utilisation of the local-pref attribute is to support the customer->provider and shared-cost peer-
ing relationships. From an economic point of view, there is an important difference between these three types of
peering relationships. A domain usually earns money when it sends packets over a provider->customer relation-
ship. On the other hand, it must pay its provider when it sends packets over a customer->provider relationship.
Using a shared-cost peering to send packets is usually neutral from an economic perspective. To take into account
these economic issues, domains usually configure the import filters on their routers as follows :
insert a high local-pref attribute in the routes learned from a customer
insert a medium local-pref attribute in the routes learned over a shared-cost peering
insert a low local-pref attribute in the routes learned from a provider

With such an import filter, the routers of a domain always prefer to reach destinations via their customers whenever
such a route exists. Otherwise, they prefer to use shared-cost peering relationships and they only send packets
via their providers when they do not know any alternate route. A consequence of setting the local-pref attribute
like this is that Internet paths are often asymmetrical. Consider for example the internetwork shown in the figure
below.
Figure 3.75: Asymmetry of Internet paths
Consider in this internetwork the routes available inside AS1 to reach AS5. AS1 learns the AS4:AS6:AS7:AS5 path
from AS4, the AS3:AS8:AS5 path from AS3 and the AS2:AS5 path from AS2. The first path is chosen since it was
learned from a customer. AS5 on the other hand receives three paths towards AS1 via its providers. It may select
any of these paths to reach AS1 , depending on how it prefers one provider over the others.
BGP convergence
In the previous sections, we have explained the operation of BGP routers. Compared to intradomain routing
protocols, a key feature of BGP is its ability to support interdomain routing policies that are defined by each
domain as its import and export filters and ranking process. A domain can define its own routing policies and
router vendors have implemented many configuration tweaks to support complex routing policies. However, the
routing policy chosen by a domain may interfere with the routing policy chosen by another domain. To understand
this issue, let us first consider the simple internetwork shown below.
Figure 3.76: The disagree internetwork
In this internetwork, we focus on the route towards 2001:db8::1234/48 which is advertised by AS1. Let us also
assume that AS3 (resp. AS4) prefers, e.g. for economic reasons, a route learned from AS4 (AS3) over a route
learned from AS1. When AS1 sends U(2001:db8::1234/48,AS1) to AS3 and AS4, three sequences of exchanges of
BGP messages are possible :
1. AS3 sends first U(2001:db8:1234/48,AS3:AS1) to AS4. AS4 has learned two routes towards
2001:db8:1234/48. It runs its BGP decision process and selects the route via AS3 and does not advertise a
route to AS3
2. AS4 first sends U(2001:db8:1234/48,AS3:AS1) to AS3. AS3 has learned two routes towards
2001:db8:1234/48. It runs its BGP decision process and selects the route via AS4 and does not advertise a
route to AS4
3. AS3 sends U(2001:db8:1234/48,AS3:AS1) to AS4 and, at the same time, AS4 sends
U(2001:db8:1234/48,AS4:AS1). AS3 prefers the route via AS4 and thus sends W(2001:db8:1234/48)
to AS4. In the mean time, AS4 prefers the route via AS3 and thus sends W(2001:db8:1234/48) to AS3. Upon

reception of the BGP Withdraws, AS3 and AS4 only know the direct route towards 2001:db8:1234/48. AS3
(resp. AS4) sends U(2001:db8:1234/48,AS3:AS1) (resp. U(2001:db8:1234/48,AS4:AS1)) to AS4 (resp.
AS3). AS3 and AS4 could in theory continue to exchange BGP messages for ever. In practice, one of them
sends one message faster than the other and BGP converges.
The example above has shown that the routes selected by BGP routers may sometimes depend on the ordering of
the BGP messages that are exchanged. Other similar scenarios may be found in RFC 4264.
From an operational perspective, the above configuration is annoying since the network operators cannot easily
predict which paths are chosen. Unfortunately, there are even more annoying BGP configurations. For example,
let us consider the configuration below which is often named Bad Gadget [GW1999]
Figure 3.77: The bad gadget internetwork
In this internetwork, there are four ASes. AS0 advertises one route towards one prefix and we only analyse the
routes towards this prefix. The routing preferences of AS1, AS3 and AS4 are the following :
AS1 prefers the path AS3:AS0 over all other paths
AS0 sends U(p,AS0) to AS1, AS3 and AS4. As this is the only route known by AS1, AS3 and AS4 towards p, they
all select the direct path. Let us now consider one possible exchange of BGP messages :
1. AS1 sends U(p, AS1:AS0) to AS3 and AS4. AS4 selects the path via AS1 since this is its preferred path. AS3
still uses the direct path.
2. AS4 advertises U(p,AS4:AS1:AS0) to AS3.
3. AS3 sends U(p, AS3:AS0) to AS1 and AS4. AS1 selects the path via AS3 since this is its preferred path. AS4
still uses the path via AS1.
4. As AS1 has changed its path, it sends U(p,AS1:AS3:AS0) to AS4 and W(p) to AS3 since its new path is via
AS3. AS4 switches back to the direct path.
5. AS4 sends U(p,AS4:AS0) to AS1 and AS3. AS3 prefers the path via AS4.
6. AS3 sends U(p,AS3:AS4:AS0) to AS1 and W(p) to AS4. AS1 switches back to the direct path and we are
back at the first step.
This example shows that the convergence of BGP is unfortunately not always guaranteed as some interdomain
routing policies may interfere with each other in complex ways. [GW1999] have shown that checking for global
convergence is either NP-complete or NP-hard. See [GSW2002] for a more detailed discussion.
Fortunately, there are some operational guidelines [GR2001] [GGR2001] that can guarantee BGP convergence in
the global Internet. To ensure that BGP will converge, these guidelines consider that there are two types of peering
relationships : customer->provider and shared-cost. In this case, BGP convergence is guaranteed provided that
the following conditions are fulfilled :
1. The topology composed of all the directed customer->provider peering links is an acyclic graph
2. An AS always prefers a route received from a customer over a route received from a shared-cost peer or a
provider.

The first guideline implies that the provider of the provider of ASx cannot be a customer of ASx. Such a relationship
would not make sense from an economic perspective as it would imply circular payments. Furthermore, providers
are usually larger than customers.
The second guideline also corresponds to economic preferences. Since a provider earns money when sending
packets to one of its customers, it makes sense to prefer such customer learned routes over routes learned from
providers. [GR2001] also shows that BGP convergence is guaranteed even if an AS associates the same preference
to routes learned from a shared-cost peer and routes learned from a customer.
From a theoretical perspective, these guidelines should be verified automatically to ensure that BGP will always
converge in the global Internet. However, such a verification cannot be performed in practice because this would
force all domains to disclose their routing policies (and few are willing to do so) and furthermore the problem is
known to be NP-hard [GW1999].
In practice, researchers and operators expect that these guidelines are verified 56 in most domains. Thanks to the
large amount of BGP data that has been collected by operators and researchers 57 , several studies have analysed
the AS-level topology of the Internet. [SARK2002] is one of the first analysis. More recent studies include
[COZ2008] and [DKF+2007]
Based on these studies and [ATLAS2009], the AS-level Internet topology can be summarised as shown in the
figure below.
Figure 3.78: The layered structure of the global Internet
The domains on the Internet can be divided in about four categories according to their role and their position in
the AS-level topology.
the core of the Internet is composed of a dozen-twenty Tier-1 ISPs. A Tier-1 is a domain that has no provider.
Such an ISP has shared-cost peering relationships with all other Tier-1 ISPs and provider->customer rela-
tionships with smaller ISPs. Examples of Tier-1 ISPs include sprint, level3 or opentransit
the Tier-2 ISPs are national or continental ISPs that are customers of Tier-1 ISPs. These Tier-2 ISPs have
smaller customers and shared-cost peering relationships with other Tier-2 ISPs. Example of Tier-2 ISPs
include France Telecom, Belgacom, British Telecom, ...
the Tier-3 networks are either stub domains such as enterprise or campus networks networks and smaller
ISPs. They are customers of Tier-1 and Tier-2 ISPs and have sometimes shared-cost peering relationships
the large content providers that are managing large datacenters. These content providers are producing a
growing fraction of the packets exchanged on the global Internet [ATLAS2009]. Some of these content
providers are customers of Tier-1 or Tier-2 ISPs, but they often try to establish shared-cost peering relation-
ships, e.g. at IXPs, with many Tier-1 and Tier-2 ISPs.
56 Researchers such as [MUF+2007] have shown that modelling the Internet topology at the AS-level requires more than the shared-cost and
customer->provider peering relationships. However, there is no publicly available model that goes beyond these classical peering relationships.
57 BGP data is often collected by establishing BGP sessions between Unix hosts running a BGP daemon and BGP routers in differ-
ent ASes. The Unix hosts stores all BGP messages received and regular dumps of its BGP routing table. See http://www.routeviews.org,
http://www.ripe.net/ris, http://bgp.potaroo.net or http://irl.cs.ucla.edu/topology/

Due to this organisation of the Internet and due to the BGP decision process, most AS-level paths on the Internet
have a length of 3-5 AS hops.
3.16 Datalink layer technologies
In this section, we review the key characteristics of several datalink layer technologies. We discuss in more detail
the technologies that are widely used today. A detailed survey of all datalink layer technologies would be outside
the scope of this book.
3.16.1 The Point-to-Point Protocol
Many point-to-point datalink layers 58 have been developed, starting in the 1960s [McFadyen1976]. In this section,
we focus on the protocols that are often used to transport IP packets between hosts or routers that are directly
connected by a point-to-point link. This link can be a dedicated physical cable, a leased line through the telephone
network or a dial-up connection with modems on the two communicating hosts.
The first solution to transport IP packets over a serial line was proposed in RFC 1055 and is known as Serial
Line IP (SLIP). SLIP is a simple character stuffing technique applied to IP packets. SLIP defines two special
characters : END (decimal 192) and ESC (decimal 219). END appears at the beginning and at the end of each
transmitted IP packet and the sender adds ESC before each END character inside each transmitted IP packet.
SLIP only supports the transmission of IP packets and it assumes that the two communicating hosts/routers have
been manually configured with each others IP address. SLIP was mainly used over links offering bandwidth of
often less than 20 Kbps. On such a low bandwidth link, sending 20 bytes of IP header followed by 20 bytes of
TCP header for each TCP segment takes a lot of time. This initiated the development of a family of compression
techniques to efficiently compress the TCP/IP headers. The first header compression technique proposed in RFC
1144 was designed to exploit the redundancy between several consecutive segments that belong to the same TCP
connection. In all these segments, the IP addresses and port numbers are always the same. Furthermore, fields
such as the sequence and acknowledgement numbers do not change in a random way. RFC 1144 defined simple
techniques to reduce the redundancy found in successive segments. The development of header compression
techniques continued and there are still improvements being developed now RFC 5795.
While SLIP was implemented and used in some environments, it had several limitations discussed in RFC 1055.
The Point-to-Point Protocol (PPP) was designed shortly after and is specified in RFC 1548. PPP aims to support
IP and other network layer protocols over various types of serial lines. PPP is in fact a family of three protocols
that are used together :
1. The Point-to-Point Protocol defines the framing technique to transport network layer packets.
2. The Link Control Protocol that is used to negotiate options and authenticate the session by using username
and password or other types of credentials
3. The Network Control Protocol that is specific for each network layer protocol. It is used to negotiate options
that are specific for each protocol. For example, IPv4s NCP RFC 1548 can negotiate the IPv4 address to
be used, the IPv4 address of the DNS resolver. IPv6s NCP is defined in RFC 5072.
The PPP framing RFC 1662 was inspired by the datalink layer protocols standardised by ITU-T and ISO. A typical
PPP frame is composed of the fields shown in the figure below. A PPP frame starts with a one byte flag containing
01111110. PPP can use bit stuffing or character stuffing depending on the environment where the protocol is used.
The address and control fields are present for backward compatibility reasons. The 16 bit Protocol field contains
the identifier 59 of the network layer protocol that is carried in the PPP frame. 0x002d is used for an IPv4 packet
compressed with RFC 1144 while 0x002f is used for an uncompressed IPv4 packet. 0xc021 is used by the Link
Control Protocol, 0xc023 is used by the Password Authentication Protocol (PAP). 0x0057 is used for IPv6 packets.
58 LAPB and HDLC were widely used datalink layer protocols.
59 The IANA maintains the registry of all assigned PPP protocol fields at : http://www.iana.org/assignments/ppp-numbers

PPP supports variable length packets, but LCP can negotiate a maximum packet length. The PPP frame ends with
a Frame Check Sequence. The default is a 16 bits CRC, but some implementations can negotiate a 32 bits CRC.
The frame ends with the 01111110 flag.
Figure 3.79: PPP frame format
PPP played a key role in allowing Internet Service Providers to provide dial-up access over modems in the late
1990s and early 2000s. ISPs operated modem banks connected to the telephone network. For these ISPs, a key
issue was to authenticate each user connected through the telephone network. This authentication was performed
by using the Extensible Authentication Protocol (EAP) defined in RFC 3748. EAP is a simple, but extensible
protocol that was initially used by access routers to authenticate the users connected through dialup lines. Several
authentication methods, starting from the simple username/password pairs to more complex schemes have been
defined and implemented. When ISPs started to upgrade their physical infrastructure to provide Internet access
over Asymmetric Digital Subscriber Lines (ADSL), they tried to reuse their existing authentication (and billing)
systems. To meet these requirements, the IETF developed specifications to allow PPP frames to be transported over
other networks than the point-to-point links for which PPP was designed. Nowadays, most ADSL deployments
use PPP over either ATM RFC 2364 or Ethernet RFC 2516.
3.16.2 Ethernet
Ethernet was designed in the 1970s at the Palo Alto Research Center [Metcalfe1976]. The first prototype 60 used
a coaxial cable as the shared medium and 3 Mbps of bandwidth. Ethernet was improved during the late 1970s
and in the 1980s, Digital Equipment, Intel and Xerox published the first official Ethernet specification [DIX]. This
specification defines several important parameters for Ethernet networks. The first decision was to standardise
the commercial Ethernet at 10 Mbps. The second decision was the duration of the slot time. In Ethernet, a long
slot time enables networks to span a long distance but forces the host to use a larger minimum frame size. The
compromise was a slot time of 51.2 microseconds, which corresponds to a minimum frame size of 64 bytes.
The third decision was the frame format. The experimental 3 Mbps Ethernet network built at Xerox used short
frames containing 8 bit source and destination addresses fields, a 16 bit type indication, up to 554 bytes of payload
and a 16 bit CRC. Using 8 bit addresses was suitable for an experimental network, but it was clearly too small
for commercial deployments. Although the initial Ethernet specification [DIX] only allowed up to 1024 hosts on
an Ethernet network, it also recommended three important changes compared to the networking technologies that
were available at that time. The first change was to require each host attached to an Ethernet network to have a
globally unique datalink layer address. Until then, datalink layer addresses were manually configured on each host.
[DP1981] went against that state of the art and noted Suitable installation-specific administrative procedures are
also needed for assigning numbers to hosts on a network. If a host is moved from one network to another it may
be necessary to change its host number if its former number is in use on the new network. This is easier said
than done, as each network must have an administrator who must record the continuously changing state of the
system (often on a piece of paper tacked to the wall !). It is anticipated that in future office environments, hosts
locations will change as often as telephones are changed in present-day offices. The second change introduced
by Ethernet was to encode each address as a 48 bits field [DP1981]. 48 bit addresses were huge compared to
the networking technologies available in the 1980s, but the huge address space had several advantages [DP1981]
including the ability to allocate large blocks of addresses to manufacturers. Eventually, other LAN technologies
opted for 48 bits addresses as well [IEEE802] . The third change introduced by Ethernet was the definition of
60 Additional information about the history of the Ethernet technology may be found at http://ethernethistory.typepad.com/
3.16. Datalink layer technologies 209

broadcast and multicast addresses. The need for multicast Ethernet was foreseen in [DP1981] and thanks to the
size of the addressing space it was possible to reserve a large block of multicast addresses for each manufacturer.
The datalink layer addresses used in Ethernet networks are often called MAC addresses. They are structured as
shown in the figure below. The first bit of the address indicates whether the address identifies a network adapter
or a multicast group. The upper 24 bits are used to encode an Organisation Unique Identifier (OUI). This OUI
identifies a block of addresses that has been allocated by the secretariat 61 that is responsible for the uniqueness
of Ethernet addresses to a manufacturer. Once a manufacturer has received an OUI, it can build and sell products
with one of the 16 million addresses in this block.
Figure 3.80: 48 bits Ethernet address format
The original 10 Mbps Ethernet specification [DIX] defined a simple frame format where each frame is composed
of five fields. The Ethernet frame starts with a preamble (not shown in the figure below) that is used by the physical
layer of the receiver to synchronise its clock with the senders clock. The first field of the frame is the destination
address. As this address is placed at the beginning of the frame, an Ethernet interface can quickly verify whether
it is the frame recipient and if not, cancel the processing of the arriving frame. The second field is the source
address. While the destination address can be either a unicast or a multicast/broadcast address, the source address
must always be a unicast address. The third field is a 16 bits integer that indicates which type of network layer
packet is carried inside the frame. This field is often called the EtherType. Frequently used EtherType values 62
include 0x0800 for IPv4, 0x86DD for IPv6 63 and 0x806 for the Address Resolution Protocol (ARP).
The fourth part of the Ethernet frame is the payload. The minimum length of the payload is 46 bytes to ensure a
minimum frame size, including the header of 512 bits. The Ethernet payload cannot be longer than 1500 bytes.
This size was found reasonable when the first Ethernet specification was written. At that time, Xerox had been
using its experimental 3 Mbps Ethernet that offered 554 bytes of payload and RFC 1122 required a minimum
MTU of 572 bytes for IPv4. 1500 bytes was large enough to support these needs without forcing the network
adapters to contain overly large memories. Furthermore, simulations and measurement studies performed in
Ethernet networks revealed that CSMA/CD was able to achieve a very high utilization. This is illustrated in the
figure below based on [SH1980], which shows the channel utilization achieved in Ethernet networks containing
different numbers of hosts that are sending frames of different sizes.
The last field of the Ethernet frame is a 32 bit Cyclical Redundancy Check (CRC). This CRC is able to catch a
much larger number of transmission errors than the Internet checksum used by IP, UDP and TCP [SGP98]. The
format of the Ethernet frame is shown below.
Note: Where should the CRC be located in a frame ?

The transport and datalink layers usually chose different strategies to place their CRCs or checksums. Transport
layer protocols usually place their CRCs or checksums in the segment header. Datalink layer protocols sometimes
place their CRC in the frame header, but often in a trailer at the end of the frame. This choice reflects implementa-
tion assumptions, but also influences performance RFC 893. When the CRC is placed in the trailer, as in Ethernet,
the datalink layer can compute it while transmitting the frame and insert it at the end of the transmission. All Eth-
ernet interfaces use this optimisation today. When the checksum is placed in the header, as in a TCP segment,
61 Initially, the OUIs were allocated by Xerox [DP1981]. However, once Ethernet became an IEEE and later an ISO standard, the allocation
of the OUIs moved to IEEE. The list of all OUI allocations may be found at http://standards.ieee.org/regauth/oui/index.shtml
62 The official list of all assigned Ethernet type values is available from http://standards.ieee.org/regauth/ethertype/eth.txt
63 The attentive reader may question the need for different EtherTypes for IPv4 and IPv6 while the IP header already contains a version
field that can be used to distinguish between IPv4 and IPv6 packets. Theoretically, IPv4 and IPv6 could have used the same EtherType.
Unfortunately, developers of the early IPv6 implementations found that some devices did not check the version field of the IPv4 packets that
they received and parsed frames whose EtherType was set to 0x0800 as IPv4 packets. Sending IPv6 packets to such devices would have caused
disruptions. To avoid this problem, the IETF decided to apply for a distinct EtherType value for IPv6. Such a choice is now mandated by RFC
6274 (section 3.1), although we can find a funny counter-example in RFC 6214.

Figure 3.81: Impact of the frame length on the maximum channel utilisation [SH1980]
Figure 3.82: Ethernet DIX frame format

it is impossible for the network interface to compute it while transmitting the segment. Some network interfaces
provide hardware assistance to compute the TCP checksum, but this is more complex than if the TCP checksum
were placed in the trailer 64 .
The Ethernet frame format shown above is specified in [DIX]. This is the format used to send both IPv4 RFC 894
and IPv6 packets RFC 2464. After the publication of [DIX], the Institute of Electrical and Electronic Engineers
(IEEE) began to standardise several Local Area Network technologies. IEEE worked on several LAN technolo-
gies, starting with Ethernet, Token Ring and Token Bus. These three technologies were completely different, but
they all agreed to use the 48 bits MAC addresses specified initially for Ethernet [IEEE802] . While develop-
ing its Ethernet standard [IEEE802.3], the IEEE 802.3 working group was confronted with a problem. Ethernet
mandated a minimum payload size of 46 bytes, while some companies were looking for a LAN technology that
could transparently transport short frames containing only a few bytes of payload. Such a frame can be sent by an
Ethernet host by padding it to ensure that the payload is at least 46 bytes long. However since the Ethernet header
[DIX] does not contain a length field, it is impossible for the receiver to determine how many useful bytes were
placed inside the payload field. To solve this problem, the IEEE decided to replace the Type field of the Ethernet
[DIX] header with a length field 65 . This Length field contains the number of useful bytes in the frame payload.
The payload must still contain at least 46 bytes, but padding bytes are added by the sender and removed by the
receiver. In order to add the Length field without significantly changing the frame format, IEEE had to remove
the Type field. Without this field, it is impossible for a receiving host to identify the type of network layer packet
inside a received frame. To solve this new problem, IEEE developed a completely new sublayer called the Logi-
cal Link Control [IEEE802.2]. Several protocols were defined in this sublayer. One of them provided a slightly
different version of the Type field of the original Ethernet frame format. Another contained acknowledgements
and retransmissions to provide a reliable service... In practice, [IEEE802.2] is never used to support IP in Ethernet
networks. The figure below shows the official [IEEE802.3] frame format.
Figure 3.83: Ethernet 802.3 frame format
Note: What is the Ethernet service ?

An Ethernet network provides an unreliable connectionless service. It supports three different transmission modes
[unicast, multicast and broadcast. While the Ethernet service is unreliable in theory, a good Ethernet
network should, in practice, provide a service that :]
delivers frames to their destination with a very high probability of successful delivery
does not reorder the transmitted frames
The first property is a consequence of the utilisation of CSMA/CD. The second property is a consequence of
the physical organisation of the Ethernet network as a shared bus. These two properties are important and all
evolutions of the Ethernet technology have preserved them.
64 These network interfaces compute the TCP checksum while a segment is transferred from the host memory to the network interface
[SH2004].
65 Fortunately, IEEE was able to define the [IEEE802.3] frame format while maintaining backward compatibility with the Ethernet [DIX]
frame format. The trick was to only assign values above 1500 as EtherType values. When a host receives a frame, it can determine whether the
frames format by checking its EtherType/Length field. A value lower smaller than 1501 is clearly a length indicator and thus an [IEEE802.3]
frame. A value larger than 1501 can only be type and thus a [DIX] frame.

Several physical layers have been defined for Ethernet networks. The first physical layer, usually called 10Base5,
provided 10 Mbps over a thick coaxial cable. The characteristics of the cable and the transceivers that were used
then enabled the utilisation of 500 meter long segments. A 10Base5 network can also include repeaters between
segments.
The second physical layer was 10Base2. This physical layer used a thin coaxial cable that was easier to install
than the 10Base5 cable, but could not be longer than 185 meters. A 10BaseF physical layer was also defined
to transport Ethernet over point-to-point optical links. The major change to the physical layer was the support
of twisted pairs in the 10BaseT specification. Twisted pair cables are traditionally used to support the telephone
service in office buildings. Most office buildings today are equipped with structured cabling. Several twisted pair
cables are installed between any room and a central telecom closet per building or per floor in large buildings.
These telecom closets act as concentration points for the telephone service but also for LANs.
The introduction of the twisted pairs led to two major changes to Ethernet. The first change concerns the physical
topology of the network. 10Base2 and 10Base5 networks are shared buses, the coaxial cable typically passes
through each room that contains a connected computer. A 10BaseT network is a star-shaped network. All the
devices connected to the network are attached to a twisted pair cable that ends in the telecom closet. From
a maintenance perspective, this is a major improvement. The cable is a weak point in 10Base2 and 10Base5
networks. Any physical damage on the cable broke the entire network and when such a failure occurred, the
network administrator had to manually check the entire cable to detect where it was damaged. With 10BaseT,
when one twisted pair is damaged, only the device connected to this twisted pair is affected and this does not
affect the other devices. The second major change introduced by 10BaseT was that is was impossible to build a
10BaseT network by simply connecting all the twisted pairs together. All the twisted pairs must be connected to
a relay that operates in the physical layer. This relay is called an Ethernet hub. A hub is thus a physical layer
relay that receives an electrical signal on one of its interfaces, regenerates the signal and transmits it over all its
other interfaces. Some hubs are also able to convert the electrical signal from one physical layer to another (e.g.
10BaseT to 10Base2 conversion).
Figure 3.84: Ethernet hubs in the reference model
Computers can directly be attached to Ethernet hubs. Ethernet hubs themselves can be attached to other Ethernet
hubs to build a larger network. However, some important guidelines must be followed when building a complex
network with hubs. First, the network topology must be a tree. As hubs are relays in the physical layer, adding
a link between Hub2 and Hub3 in the network below would create an electrical shortcut that would completely
disrupt the network. This implies that there cannot be any redundancy in a hub-based network. A failure of a
hub or of a link between two hubs would partition the network into two isolated networks. Second, as hubs are
relays in the physical layer, collisions can happen and must be handled by CSMA/CD as in a 10Base5 network.
This implies that the maximum delay between any pair of devices in the network cannot be longer than the 51.2
microseconds slot time. If the delay is longer, collisions between short frames may not be correctly detected. This
constraint limits the geographical spread of 10BaseT networks containing hubs.
In the late 1980s, 10 Mbps became too slow for some applications and network manufacturers developed several
LAN technologies that offered higher bandwidth, such as the 100 Mbps FDDI LAN that used optical fibers. As the
development of 10Base5, 10Base2 and 10BaseT had shown that Ethernet could be adapted to different physical
layers, several manufacturers started to work on 100 Mbps Ethernet and convinced IEEE to standardise this new
technology that was initially called Fast Ethernet. Fast Ethernet was designed under two constraints. First,
Fast Ethernet had to support twisted pairs. Although it was easier from a physical layer perspective to support
higher bandwidth on coaxial cables than on twisted pairs, coaxial cables were a nightmare from deployment
and maintenance perspectives. Second, Fast Ethernet had to be perfectly compatible with the existing 10 Mbps

Figure 3.85: A hierarchical Ethernet network composed of hubs
Ethernets to allow Fast Ethernet technology to be used initially as a backbone technology to interconnect 10
Mbps Ethernet networks. This forced Fast Ethernet to use exactly the same frame format as 10 Mbps Ethernet.
This implied that the minimum Fast Ethernet frame size remained at 512 bits. To preserve CSMA/CD with this
minimum frame size and 100 Mbps instead of 10 Mbps, the duration of the slot time was decreased to 5.12
microseconds.
The evolution of Ethernet did not stop. In 1998, the IEEE published a first standard to provide Gigabit Ethernet
over optical fibers. Several other types of physical layers were added afterwards. The 10 Gigabit Ethernet standard
appeared in 2002. Work is ongoing to develop standards for 40 Gigabit and 100 Gigabit Ethernet and some are
thinking about Terabit Ethernet. The table below lists the main Ethernet standards. A more detailed list may be
found at http://en.wikipedia.org/wiki/Ethernet_physical_layer
Standard Comments
10Base5 Thick coaxial cable, 500m
10Base2 Thin coaxial cable, 185m
10BaseT Two pairs of category 3+ UTP
10Base-F 10 Mb/s over optical fiber
100Base-Tx Category 5 UTP or STP, 100 m maximum
100Base-FX Two multimode optical fiber, 2 km maximum
1000Base-CX Two pairs shielded twisted pair, 25m maximum
1000Base-SX Two multimode or single mode optical fibers with lasers
10 Gbps Optical fiber but also Category 6 UTP
40-100 Gbps Optical fiber (experiences are performed with copper)
Ethernet Switches
Increasing the physical layer bandwidth as in Fast Ethernet was only one of the solutions to improve the perfor-
mance of Ethernet LANs. A second solution was to replace the hubs with more intelligent devices. As Ethernet
hubs operate in the physical layer, they can only regenerate the electrical signal to extend the geographical reach
of the network. From a performance perspective, it would be more interesting to have devices that operate in the
datalink layer and can analyse the destination address of each frame and forward the frames selectively on the link
that leads to the destination. Such devices are usually called Ethernet switches 66 . An Ethernet switch is a relay
that operates in the datalink layer as is illustrated in the figure below.
An Ethernet switch understands the format of the Ethernet frames and can selectively forward frames over each
interface. For this, each Ethernet switch maintains a MAC address table. This table contains, for each MAC
address known by the switch, the identifier of the switchs port over which a frame sent towards this address must
66 The first Ethernet relays that operated in the datalink layers were called bridges. In practice, the main difference between switches and
bridges is that bridges were usually implemented in software while switches are hardware-based devices. Throughout this text, we always use
switch when referring to a relay in the datalink layer, but you might still see the word bridge.

Figure 3.86: Ethernet switches and the reference model
be forwarded to reach its destination. This is illustrated below with the MAC address table of the bottom switch.
When the switch receives a frame destined to address B, it forwards the frame on its South port. If it receives a
frame destined to address D, it forwards it only on its North port.
Figure 3.87: Operation of Ethernet switches
One of the selling points of Ethernet networks is that, thanks to the utilisation of 48 bits MAC addresses, an
Ethernet LAN is plug and play at the datalink layer. When two hosts are attached to the same Ethernet segment or
hub, they can immediately exchange Ethernet frames without requiring any configuration. It is important to retain
this plug and play capability for Ethernet switches as well. This implies that Ethernet switches must be able to build
their MAC address table automatically without requiring any manual configuration. This automatic configuration
is performed by the MAC address learning algorithm that runs on each Ethernet switch. This algorithm extracts
the source address of the received frames and remembers the port over which a frame from each source Ethernet
address has been received. This information is inserted into the MAC address table that the switch uses to forward
frames. This allows the switch to automatically learn the ports that it can use to reach each destination address,
provided that this host has previously sent at least one frame. This is not a problem since most upper layer
protocols use acknowledgements at some layer and thus even an Ethernet printer sends Ethernet frames as well.
The pseudo-code below details how an Ethernet switch forwards Ethernet frames. It first updates its MAC address
table with the source address of the frame. The MAC address table used by some switches also contains a
timestamp that is updated each time a frame is received from each known source address. This timestamp is
used to remove from the MAC address table entries that have not been active during the last n minutes. This limits

the growth of the MAC address table, but also allows hosts to move from one port to another. The switch uses its
MAC address table to forward the received unicast frame. If there is an entry for the frames destination address
in the MAC address table, the frame is forwarded selectively on the port listed in this entry. Otherwise, the switch
does not know how to reach the destination address and it must forward the frame on all its ports except the port
from which the frame has been received. This ensures that the frame will reach its destination, at the expense of
some unnecessary transmissions. These unnecessary transmissions will only last until the destination has sent its
first frame. Multicast and Broadcast frames are also forwarded in a similar way.
# Arrival of frame F on port P
# Table : MAC address table dictionary : addr->port
# Ports : list of all ports on the switch
src=F.SourceAddress
dst=F.DestinationAddress
Table[src]=P #src heard on port P
if isUnicast(dst) :
if dst in Table:
ForwardFrame(F,Table[dst])
else:
for o in Ports :
if o!= P : ForwardFrame(F,o)
else:
# multicast or broadcast destination
for o in Ports :
if o!= P : ForwardFrame(F,o)
Note: Security issues with Ethernet hubs and switches

From a security perspective, Ethernet hubs have the same drawbacks as the older coaxial cable. A host attached to
a hub will be able to capture all the frames exchanged between any pair of hosts attached to the same hub. Ethernet
switches are much better from this perspective thanks to the selective forwarding, a host will usually only receive
the frames destined to itself as well as the multicast, broadcast and unknown frames. However, this does not imply
that switches are completely secure. There are, unfortunately, attacks against Ethernet switches. From a security
perspective, the MAC address table is one of the fragile elements of an Ethernet switch. This table has a fixed
size. Some low-end switches can store a few tens or a few hundreds of addresses while higher-end switches can
store tens of thousands of addresses or more. From a security point of view, a limited resource can be the target
of Denial of Service attacks. Unfortunately, such attacks are also possible on Ethernet switches. A malicious
host could overflow the MAC address table of the switch by generating thousands of frames with random source
addresses. Once the MAC address table is full, the switch needs to broadcast all the frames that it receives. At
this point, an attacker will receive unicast frames that are not destined to its address. The ARP attack discussed in
the previous chapter could also occur with Ethernet switches [Vyncke2007]. Recent switches implement several
types of defences against these attacks, but they need to be carefully configured by the network administrator. See
[Vyncke2007] for a detailed discussion on security issues with Ethernet switches.
The MAC address learning algorithm combined with the forwarding algorithm work well in a tree-shaped network
such as the one shown above. However, to deal with link and switch failures, network administrators often add
redundant links to ensure that their network remains connected even after a failure. Let us consider what happens
in the Ethernet network shown in the figure below.
When all switches boot, their MAC address table is empty. Assume that host A sends a frame towards host C.
Upon reception of this frame, switch1 updates its MAC address table to remember that address A is reachable
via its West port. As there is no entry for address C in switch1s MAC address table, the frame is forwarded to
both switch2 and switch3. When switch2 receives the frame, its updates its MAC address table for address A
and forwards the frame to host C as well as to switch3. switch3 has thus received two copies of the same frame.
As switch3 does not know how to reach the destination address, it forwards the frame received from switch1 to
switch2 and the frame received from switch2 to switch1... The single frame sent by host A will be continuously
duplicated by the switches until their MAC address table contains an entry for address C. Quickly, all the available
link bandwidth will be used to forward all the copies of this frame. As Ethernet does not contain any TTL or
HopLimit, this loop will never stop.
The MAC address learning algorithm allows switches to be plug-and-play. Unfortunately, the loops that arise

Figure 3.88: Ethernet switches in a loop
when the network topology is not a tree are a severe problem. Forcing the switches to only be used in tree-shaped
networks as hubs would be a severe limitation. To solve this problem, the inventors of Ethernet switches have
developed the Spanning Tree Protocol. This protocol allows switches to automatically disable ports on Ethernet
switches to ensure that the network does not contain any cycle that could cause frames to loop forever.
The Spanning Tree Protocol (802.1d)
The Spanning Tree Protocol (STP), proposed in [Perlman1985], is a distributed protocol that is used by switches
to reduce the network topology to a spanning tree, so that there are no cycles in the topology. For example,
consider the network shown in the figure below. In this figure, each bold line corresponds to an Ethernet to which
two Ethernet switches are attached. This network contains several cycles that must be broken to allow Ethernet
switches that are using the MAC address learning algorithm to exchange frames.
Figure 3.89: Spanning tree computed in a switched Ethernet network

In this network, the STP will compute the following spanning tree. Switch1 will be the root of the tree. All the
interfaces of Switch1, Switch2 and Switch7 are part of the spanning tree. Only the interface connected to LANB
will be active on Switch9. LANH will only be served by Switch7 and the port of Switch44 on LANG will be
disabled. A frame originating on LANB and destined for LANA will be forwarded by Switch7 on LANC, then by
Switch1 on LANE, then by Switch44 on LANF and eventually by Switch2 on LANA.
Switches running the Spanning Tree Protocol exchange BPDUs. These BPDUs are always sent as frames with
destination MAC address as the ALL_BRIDGES reserved multicast MAC address. Each switch has a unique 64
bit identifier. To ensure uniqueness, the lower 48 bits of the identifier are set to the unique MAC address allocated
to the switch by its manufacturer. The high order 16 bits of the switch identifier can be configured by the network
administrator to influence the topology of the spanning tree. The default value for these high order bits is 32768.
The switches exchange BPDUs to build the spanning tree. Intuitively, the spanning tree is built by first selecting
the switch with the smallest identifier as the root of the tree. The branches of the spanning tree are then composed
of the shortest paths that allow all of the switches that compose the network to be reached. The BPDUs exchanged
by the switches contain the following information :
the identifier of the root switch (R)
the cost of the shortest path between the switch that sent the BPDU and the root switch (c)
the identifier of the switch that sent the BPDU (T)
the number of the switch port over which the BPDU was sent (p)
We will use the notation <R,c,T,p> to represent a BPDU whose root identifier is R, cost is c and that was sent
on the port p of switch T. The construction of the spanning tree depends on an ordering relationship among the
BPDUs. This ordering relationship could be implemented by the python function below.
# returns True if bpdu b1 is better than bpdu b2
def better( b1, b2) :
return ( (b1.R < b2.R) or
( (b1.R==b2.R) and (b1.c<b2.c) ) or
( (b1.R==b2.R) and (b1.c==b2.c) and (b1.T<b2.T) ) or
( (b1.R==b2.R) and (b1.c==b2.c) and (b1.T==b2.T) and (b1.p<b2.p) ) )
In addition to the identifier discussed above, the network administrator can also configure a cost to be associated to
each switch port. Usually, the cost of a port depends on its bandwidth and the [IEEE802.1d] standard recommends
the values below. Of course, the network administrator may choose other values. We will use the notation cost[p]
to indicate the cost associated to port p in this section.
Bandwidth Cost
10 Mbps 2000000
100 Mbps 200000
1 Gbps 20000
10 Gbps 2000
100 Gbps 200
The Spanning Tree Protocol uses its own terminology that we illustrate in the figure above. A switch port can be
in three different states : Root, Designated and Blocked. All the ports of the root switch are in the Designated
state. The state of the ports on the other switches is determined based on the BPDU received on each port.
The Spanning Tree Protocol uses the ordering relationship to build the spanning tree. Each switch listens to
BPDUs on its ports. When BPDU=<R,c,T,p> is received on port q, the switch computes the ports root priority
vector: V[q]=<R,c+cost[q],T,p,q> , where cost[q] is the cost associated to the port over which the BPDU was
received. The switch stores in a table the last root priority vector received on each port. The switch then compares
its own identifier with the smallest root identifier stored in this table. If its own identifier is smaller, then the switch
is the root of the spanning tree and is, by definition, at a distance 0 of the root. The BPDU of the switch is then
<R,0,R,p>, where R is the switch identifier and p will be set to the port number over which the BPDU is sent.
Otherwise, the switch chooses the best priority vector from its table, bv=<R,c,T,p>. The port over which this
best root priority vector was learned is the switch port that is closest to the root switch. This port becomes the
Root port of the switch. There is only one Root port per switch (except for the Root switches whose ports are all
Designated). The switch can then compute its own BPDU as BPDU=<R,c,S,p> , where R is the root identifier, c

the cost of the best root priority vector, S the identifier of the switch and p will be replaced by the number of the
port over which the BPDU will be sent.
To determine the state of its other ports, the switch compares its own BPDU with the last BPDU received on each
port. Note that the comparison is done by using the BPDUs and not the root priority vectors. If the switchs
BPDU is better than the last BPDU of this port, the port becomes a Designated port. Otherwise, the port becomes
a Blocked port.
The state of each port is important when considering the transmission of BPDUs. The root switch regularly sends
its own BPDU over all of its (Designated) ports. This BPDU is received on the Root port of all the switches that
are directly connected to the root switch. Each of these switches computes its own BPDU and sends this BPDU
over all its Designated ports. These BPDUs are then received on the Root port of downstream switches, which
then compute their own BPDU, etc. When the network topology is stable, switches send their own BPDU on
all their Designated ports, once they receive a BPDU on their Root port. No BPDU is sent on a Blocked port.
Switches listen for BPDUs on their Blocked and Designated ports, but no BPDU should be received over these
ports when the topology is stable. The utilisation of the ports for both BPDUs and data frames is summarised in
the table below.
Port state Receives BPDUs Sends BPDU Handles data frames
Blocked yes no no
Root yes no yes
Designated yes yes yes
To illustrate the operation of the Spanning Tree Protocol, let us consider the simple network topology in the figure
below.
Figure 3.90: A simple Spanning tree computed in a switched Ethernet network
Assume that Switch4 is the first to boot. It sends its own BPDU=<4,0,4,?> on its two ports. When Switch1
boots, it sends BPDU=<1,0,1,1>. This BPDU is received by Switch4, which updates its table and computes a
new BPDU=<1,3,4,?>. Port 1 of Switch4 becomes the Root port while its second port is still in the Designated
state.
Assume now that Switch9 boots and immediately receives Switch1 s BPDU on port 1. Switch9 computes its own
BPDU=<1,1,9,?> and port 1 becomes the Root port of this switch. This BPDU is sent on port 2 of Switch9 and
reaches Switch4. Switch4 compares the priority vector built from this BPDU (i.e. <1,2,9,2>) and notices that it is
better than Switch4 s BPDU=<1,3,4,2>. Thus, port 2 becomes a Blocked port on Switch4.
During the computation of the spanning tree, switches discard all received data frames, as at that time the network
topology is not guaranteed to be loop-free. Once that topology has been stable for some time, the switches again
start to use the MAC learning algorithm to forward data frames. Only the Root and Designated ports are used

to forward data frames. Switches discard all the data frames received on their Blocked ports and never forward
frames on these ports.
Switches, ports and links can fail in a switched Ethernet network. When a failure occurs, the switches must be
able to recompute the spanning tree to recover from the failure. The Spanning Tree Protocol relies on regular
transmissions of the BPDUs to detect these failures. A BPDU contains two additional fields : the Age of the
BPDU and the Maximum Age. The Age contains the amount of time that has passed since the root switch initially
originated the BPDU. The root switch sends its BPDU with an Age of zero and each switch that computes its own
BPDU increments its Age by one. The Age of the BPDUs stored on a switchs table is also incremented every
second. A BPDU expires when its Age reaches the Maximum Age. When the network is stable, this does not
happen as BPDU s are regularly sent by the root switch and downstream switches. However, if the root fails or
the network becomes partitioned, BPDU will expire and switches will recompute their own BPDU and restart the
Spanning Tree Protocol. Once a topology change has been detected, the forwarding of the data frames stops as
the topology is not guaranteed to be loop-free. Additional details about the reaction to failures may be found in
[IEEE802.1d]
Virtual LANs
Another important advantage of Ethernet switches is the ability to create Virtual Local Area Networks (VLANs).
A virtual LAN can be defined as a set of ports attached to one or more Ethernet switches. A switch can support
several VLANs and it runs one MAC learning algorithm for each Virtual LAN. When a switch receives a frame
with an unknown or a multicast destination, it forwards it over all the ports that belong to the same Virtual LAN
but not over the ports that belong to other Virtual LANs. Similarly, when a switch learns a source address on a
port, it associates it to the Virtual LAN of this port and uses this information only when forwarding frames on this
Virtual LAN.
The figure below illustrates a switched Ethernet network with three Virtual LANs. VLAN2 and VLAN3 only require
a local configuration of switch S1. Host C can exchange frames with host D, but not with hosts that are outside of
its VLAN. VLAN1 is more complex as there are ports of this VLAN on several switches. To support such VLANs,
local configuration is not sufficient anymore. When a switch receives a frame from another switch, it must be able
to determine the VLAN in which the frame originated to use the correct MAC table to forward the frame. This
is done by assigning an identifier to each Virtual LAN and placing this identifier inside the headers of the frames
that are exchanged between switches.
Figure 3.91: Virtual Local Area Networks in a switched Ethernet network
IEEE defined in the [IEEE802.1q] standard a special header to encode the VLAN identifiers. This 32 bit header
includes a 20 bit VLAN field that contains the VLAN identifier of each frame. The format of the [IEEE802.1q]
header is described below.
Figure 3.92: Format of the 802.1q header

The [IEEE802.1q] header is inserted immediately after the source MAC address in the Ethernet frame (i.e. before
the EtherType field). The maximum frame size is increased by 4 bytes. It is encoded in 32 bits and contains four
fields. The Tag Protocol Identifier is set to 0x8100 to allow the receiver to detect the presence of this additional
header. The Priority Code Point (PCP) is a three bit field that is used to support different transmission priorities
for the frame. Value 0 is the lowest priority and value 7 the highest. Frames with a higher priority can expect to
be forwarded earlier than frames having a lower priority. The C bit is used for compatibility between Ethernet and
Token Ring networks. The last 12 bits of the 802.1q header contain the VLAN identifier. Value 0 indicates that
the frame does not belong to any VLAN while value 0xFFF is reserved. This implies that 4094 different VLAN
identifiers can be used in an Ethernet network.
3.16.3 802.11 wireless networks
The radio spectrum is a limited resource that must be shared by everyone. During most of the twentieth century,
governments and international organisations have regulated most of the radio spectrum. This regulation controls
the utilisation of the radio spectrum, in order to ensure that there are no interferences between different users.
A company that wants to use a frequency range in a given region must apply for a license from the regulator.
Most regulators charge a fee for the utilisation of the radio spectrum and some governments have encouraged
competition among companies bidding for the same frequency to increase the license fees.
In the 1970s, after the first experiments with ALOHANet, interest in wireless networks grew. Many experiments
were done on and outside the ARPANet. One of these experiments was the first mobile phone , which was de-
veloped and tested in 1973. This experimental mobile phone was the starting point for the first generation analog
mobile phones. Given the growing demand for mobile phones, it was clear that the analog mobile phone technol-
ogy was not sufficient to support a large number of users. To support more users and new services, researchers in
several countries worked on the development of digital mobile telephones. In 1987, several European countries
decided to develop the standards for a common cellular telephone system across Europe : the Global System for
Mobile Communications (GSM). Since then, the standards have evolved and more than three billion users are
connected to GSM networks today.
While most of the frequency ranges of the radio spectrum are reserved for specific applications and require a spe-
cial licence, there are a few exceptions. These exceptions are known as the Industrial, Scientific and Medical (ISM)
radio bands. These bands can be used for industrial, scientific and medical applications without requiring a licence
from the regulator. For example, some radio-controlled models use the 27 MHz ISM band and some cordless tele-
phones operate in the 915 MHz ISM. In 1985, the 2.400-2.500 GHz band was added to the list of ISM bands.
This frequency range corresponds to the frequencies that are emitted by microwave ovens. Sharing this band with
licensed applications would have likely caused interferences, given the large number of microwave ovens that are
used. Despite the risk of interferences with microwave ovens, the opening of the 2.400-2.500 GHz allowed the
networking industry to develop several wireless network techniques to allow computers to exchange data without
using cables. In this section, we discuss in more detail the most popular one, i.e. the WiFi [IEEE802.11] family of
wireless networks. Other wireless networking techniques such as BlueTooth or HiperLAN use the same frequency
range.
Today, WiFi is a very popular wireless networking technology. There are more than several hundreds of millions
of WiFi devices. The development of this technology started in the late 1980s with the WaveLAN proprietary
wireless network. WaveLAN operated at 2 Mbps and used different frequency bands in different regions of the
world. In the early 1990s, the IEEE created the 802.11 working group to standardise a family of wireless network
technologies. This working group was very prolific and produced several wireless networking standards that use
different frequency ranges and different physical layers. The table below provides a summary of the main 802.11
standards.
Standard Frequency Typical throughput Max bandwidth Range (m) indoor/outdoor
802.11 2.4 GHz 0.9 Mbps 2 Mbps 20/100
802.11a 5 GHz 23 Mbps 54 Mbps 35/120
802.11b 2.4 GHz 4.3 Mbps 11 Mbps 38/140
802.11g 2.4 GHz 19 Mbps 54 Mbps 38/140
802.11n 2.4/5 GHz 74 Mbps 150 Mbps 70/250
When developing its family of standards, the IEEE 802.11 working group took a similar approach as the IEEE
802.3 working group that developed various types of physical layers for Ethernet networks. 802.11 networks use

the CSMA/CA Medium Access Control technique described earlier and they all assume the same architecture and
use the same frame format.
The architecture of WiFi networks is slightly different from the Local Area Networks that we have discussed until
now. There are, in practice, two main types of WiFi networks : independent or adhoc networks and infrastructure
networks 67 . An independent or adhoc network is composed of a set of devices that communicate with each other.
These devices play the same role and the adhoc network is usually not connected to the global Internet. Adhoc
networks are used when for example a few laptops need to exchange information or to connect a computer with a
WiFi printer.
Figure 3.93: An 802.11 independent or adhoc network
Most WiFi networks are infrastructure networks. An infrastructure network contains one or more access points
that are attached to a fixed Local Area Network (usually an Ethernet network) that is connected to other networks
such as the Internet. The figure below shows such a network with two access points and four WiFi devices. Each
WiFi device is associated to one access point and uses this access point as a relay to exchange frames with the
devices that are associated to another access point or reachable through the LAN.
An 802.11 access point is a relay that operates in the datalink layer like switches. The figure below represents the
layers of the reference model that are involved when a WiFi host communicates with a host attached to an Ethernet
network through an access point.
802.11 devices exchange variable length frames, which have a slightly different structure than the simple frame
format used in Ethernet LANs. We review the key parts of the 802.11 frames. Additional details may be found
in [IEEE802.11] and [Gast2002] . An 802.11 frame contains a fixed length header, a variable length payload that
may contain up 2324 bytes of user data and a 32 bits CRC. Although the payload can contain up to 2324 bytes,
most 802.11 deployments use a maximum payload size of 1500 bytes as they are used in infrastructure networks
attached to Ethernet LANs. An 802.11 data frame is shown below.
The first part of the 802.11 header is the 16 bit Frame Control field. This field contains flags that indicate the type
of frame (data frame, RTS/CTS, acknowledgement, management frames, etc), whether the frame is sent to or from
a fixed LAN, etc [IEEE802.11]. The Duration is a 16 bit field that is used to reserve the transmission channel. In
data frames, the Duration field is usually set to the time required to transmit one acknowledgement frame after a
SIFS delay. Note that the Duration field must be set to zero in multicast and broadcast frames. As these frames
67 The 802.11 working group defined the basic service set (BSS) as a group of devices that communicate with each other. We continue to
use network when referring to a set of devices that communicate.

Figure 3.94: An 802.11 infrastructure network
Figure 3.95: An 802.11 access point
Figure 3.96: 802.11 data frame format

are not acknowledged, there is no need to reserve the transmission channel after their transmission. The Sequence
control field contains a 12 bits sequence number that is incremented for each data frame.
The astute reader may have noticed that the 802.11 data frames contain three 48-bits address fields 68 . This is
surprising compared to other protocols in the network and datalink layers whose headers only contain a source and
a destination address. The need for a third address in the 802.11 header comes from the infrastructure networks. In
such a network, frames are usually exchanged between routers and servers attached to the LAN and WiFi devices
attached to one of the access points. The role of the three address fields is specified by bit flags in the Frame
Control field.
When a frame is sent from a WiFi device to a server attached to the same LAN as the access point, the first address
of the frame is set to the MAC address of the access point, the second address is set to the MAC address of the
source WiFi device and the third address is the address of the final destination on the LAN. When the server
replies, it sends an Ethernet frame whose source address is its MAC address and the destination address is the
MAC address of the WiFi device. This frame is captured by the access point that converts the Ethernet header into
an 802.11 frame header. The 802.11 frame sent by the access point contains three addresses : the first address is
the MAC address of the destination WiFi device, the second address is the MAC address of the access point and
the third address the MAC address of the server that sent the frame.
802.11 control frames are simpler than data frames. They contain a Frame Control, a Duration field and one or
two addresses. The acknowledgement frames are very small. They only contain the address of the destination of
the acknowledgement. There is no source address and no Sequence Control field in the acknowledgement frames.
This is because the acknowledgement frame can easily be associated to the previous frame that it acknowledges.
Indeed, each unicast data frame contains a Duration field that is used to reserve the transmission channel to ensure
that no collision will affect the acknowledgement frame. The Sequence Control field is mainly used by the receiver
to remove duplicate frames. Duplicate frames are detected as follows. Each data frame contains a 12 bits Sequence
Control field and the Frame Control field contains the Retry bit flag that is set when a frame is transmitted. Each
802.11 receiver stores the most recent sequence number received from each source address in frames whose Retry
bit is reset. Upon reception of a frame with the Retry bit set, the receiver verifies its sequence number to determine
whether it is a duplicated frame or not.
Figure 3.97: IEEE 802.11 ACK and CTS frames
802.11 RTS/CTS frames are used to reserve the transmission channel, in order to transmit one data frame and its
acknowledgement. The RTS frames contain a Duration and the transmitter and receiver addresses. The Duration
field of the RTS frame indicates the duration of the entire reservation (i.e. the time required to transmit the CTS,
the data frame, the acknowledgements and the required SIFS delays). The CTS frame has the same format as the
acknowledgement frame.
Note: The 802.11 service

Despite the utilization of acknowledgements, the 802.11 layer only provides an unreliable connectionless service
like Ethernet networks that do not use acknowledgements. The 802.11 acknowledgements are used to minimize
the probability of frame duplication. They do not guarantee that all frames will be correctly received by their
recipients. Like Ethernet, 802.11 networks provide a high probability of successful delivery of the frames, not a
guarantee. Furthermore, it should be noted that 802.11 networks do not use acknowledgements for multicast and
broadcast frames. This implies that in practice such frames are more likely to suffer from transmission errors than
unicast frames.
68 In fact, the [IEEE802.11] frame format contains a fourth optional address field. This fourth address is only used when an 802.11 wireless
network is used to interconnect bridges attached to two classical LAN networks.

Figure 3.98: IEEE 802.11 RTS frame format
In addition to the data and control frames that we have briefly described above, 802.11 networks use several types
of management frames. These management frames are used for various purposes. We briefly describe some of
these frames below. A detailed discussion may be found in [IEEE802.11] and [Gast2002].
A first type of management frames are the beacon frames. These frames are broadcasted regularly by access
points. Each beacon frame contains information about the capabilities of the access point (e.g. the supported
802.11 transmission rates) and a Service Set Identity (SSID). The SSID is a null-terminated ASCII string that can
contain up to 32 characters. An access point may support several SSIDs and announce them in beacon frames. An
access point may also choose to remain silent and not advertise beacon frames. In this case, WiFi stations may
send Probe request frames to force the available access points to return a Probe response frame.
Note: IP over 802.11

Two types of encapsulation schemes were defined to support IP in Ethernet networks : the original encapsulation
scheme, built above the Ethernet DIX format is defined in RFC 894 and a second encapsulation RFC 1042
scheme, built above the LLC/SNAP protocol [IEEE802.2]. In 802.11 networks, the situation is simpler and only
the RFC 1042 encapsulation is used. In practice, this encapsulation adds 6 bytes to the 802.11 header. The first
four bytes correspond to the LLC/SNAP header. They are followed by the two bytes Ethernet Type field (0x800
for IP and 0x806 for ARP). The figure below shows an IP packet encapsulated in an 802.11 frame.
Figure 3.99: IP over IEEE 802.11
The second important utilisation of the management frames is to allow a WiFi station to be associated with an
access point. When a WiFi station starts, it listens to beacon frames to find the available SSIDs. To be allowed to
send and receive frames via an access point, a WiFi station must be associated to this access point. If the access
point does not use any security mechanism to secure the wireless transmission, the WiFi station simply sends an
Association request frame to its preferred access point (usually the access point that it receives with the strongest
radio signal). This frame contains some parameters chosen by the WiFi station and the SSID that it requests to
join. The access point replies with an Association response frame if it accepts the WiFI station.


CHAPTER 4
Part 3: Practice
4.1 Exercises
This section gathers several types of exercises that have been designed to enable the students to improve their
understanding of the topics presented in the ebook. There are three types of exercises :
Multiple Choice Questions. These are simple exercises that allow the students to verify that they have
understood the basics. The students should usually be able to answer these questions after a first detailed
reading of the associated chapter. The HTML version of the ebook includes javascript code that allows the
student to verify his/her answers immediately.
Code/Practice exercises. To better understand a networking concept, it is sometimes important to imple-
ment it or experiment with it. Some of these practice questions will be automatically graded by using test
cases running on http://inginious.info.ucl.ac.be while others will be open questions that could lead to class
projects.
Open questions and Discussion questions. These questions are more challenging. They typically contain a
small problem that needs to be solved by the student. These questions are designed so that they can lead to
discussions with teaching assistants.
A good understanding of the topics covered by this ebook can only be obtained by solving the proposed exer-
cises. Reading the ebook from the first to the last page is not sufficient to get a detailed knowledge of computer
networking.
4.2 Reliable transfer
4.2.1 Open questions
1. Consider a b bits per second link between two hosts that has a propagation delay of t seconds. Derive a
formula that computes the time elapsed between the transmission of the first bit of a d bytes segment from
a sending host and the reception of the last bit of this segment on the receiving host.
2. Consider two hosts connected by a physical cable. The two hosts are separated by a distance of 1000
kilometers and the propagation delay is 5 microseconds per kilometer. The bandwidth of the physical cable
is 1 Mbps. If the data frames have a length of 1000 bits and the acknowledgements a length of 40 bits, what
is the exact value of the round-trip-time, i.e. the delay in seconds between the transmission of the first bit
of a data frame and the reception of the last bit of the acknowledgement ? To compute this delay, you can
assume that the receiver needs one millisecond to verify a received frame and prepare the corresponding
acknowledgement.
3. Transmission links have sometimes different upstream and downstream bandwidths. A typical example are
access networks that use ADSL (Assymetric Digital Subscriber Lines). Consider two hosts connected via
227
an ADSL link having an upstream bandwidth of 1 Mbps and a downstream bandwidth of 50 Mbps. The
propagation delay between the two hosts is 10 milliseconds. What is the maximum throughput, expressed
in frames/second, that the alternating bit protocol can obtain on this link if each data frame has a length of
125 bytes and acknowledgements are 25 bytes long. Same question if the protocol is modified to support
1500 bytes long data frames.
4. How would you set the duration of the retransmission timer in the alternating bit protocol ?
5. A version of the Alternating Bit Protocol supporting variable length frames uses a header that contains the
following fields :
a number (0 or 1)
a length field that indicates the length of the data
a Cyclic Redundancy Check (CRC)
To speedup the transmission of the frames, a student proposes to compute the CRC over the data part of the
segment but not over the header. What do you think of this optimisation ?
6. Derive a mathematical expression that provides the goodput, i.e. the amount of payload bytes that have been
transmitted during a period of time, achieved by the Alternating Bit Protocol assuming that :
Each frame contains D bytes of data and c bytes of control information
Each acknowledgement contains c bytes of control information
The bandwidth of the two directions of the link is set to B bits per second
The delay between the two hosts is s seconds in both directions
there are no transmission errors
7. Consider a go-back-n sender and a go-back receiver that are directly connected with a 10 Mbps link that
has a propagation delay of 100 milliseconds. Assume that the retransmission timer is set to three seconds.
If the window has a length of 4 segments, draw a time-sequence diagram showing the transmission of 10
segments (each segment contains 10000 bits):
when there are no losses
when the third and seventh segments are lost
when every second acknowledgement is discarded due to transmission errors
8. Same question when using selective repeat instead of go-back-n. Note that the answer is not necessarily the
same.
4.2.2 Practice
Reliable protocols depend on error detection algorithms to detect transmission errors. The following questions
will reinforce your understanding of these algorithms.
1. Reliable protocols rely on different types of checksums to verify whether frames have been affected by
transmission errors. The most frequently used checksums are :
the Internet checksum used by UDP, TCP and other Internet protocols which
is defined in RFC 1071 and implemented in various libraries. See e.g.
http://ilab.cs.byu.edu/cs460/code/ftp/ichecksum.py for a python implementation
the 16 bits or the 32 bits Cyclical Redundancy Checks (CRC) that are often used on disks, in zip
archives and in datalink layer protocols. See http://rosettacode.org/wiki/CRC-32 for CRC-32
implementations in various languages.
the Fletcher checksum [Fletcher1982], see http://drdobbs.com/database/184408761 for imple-
mentation details
By using your knowledge of the Internet checksum, can you find a transmission error that will not be
detected by this checksum ?
228 Chapter 4. Part 3: Practice

2. The Cyclic Redundancy Checks (CRCs) are efficient error detection codes that are able to detect :
all errors that affect an odd number of bits
all errors that affect a sequence of bits which is shorter than the length of the CRC
Implement a small software that computes the CRC-32 for a text file. Then, modify the contents of
the file to change an even number of bits or an odd number of bits inside the file. When modifying
the file, remember that an ASCII file is composed of 8 bits characters that are encoded by using the
ASCII table that you can find at : http://en.wikipedia.org/wiki/ASCII . You can also write a small
program that produces binary files that are a small variation of each other.
3. Checksums and CRCs should not be confused with secure hash functions such as MD5 defined in RFC
1321 or SHA-1 described in RFC 4634. Secure hash functions are used to ensure that files or sometimes
packets/segments have not been modified. Secure hash functions aim at detecting malicious changes while
checksums and CRCs only detect random transmission errors. Use the shasum or md5sum programs on
Linux to perform the same tests as above.
4.2.3 Discussion questions
1. Consider two high-end servers connected back-to-back by using a 10 Gbps interface. If the delay between
the two servers is one millisecond, what is the throughput that can be achieved by a reliable protocol that is
using 10,000 bits frames and a window of
one frame
ten frames
hundred frames
2. Is it possible for a go-back-n receiver to inter-operate with a selective-repeat sender ? Justify your answer.
3. Is it possible for a selective-repeat receiver to inter-operate with a go-back-n sender ? Justify your answer.
4. A go-back-n receiver has sent 2n data segments. All the segments have been received correctly and in-order
by the receiver, but all the returned acknowledgements have been lost. Show by using a time sequence
diagram (e.g. by considering a window of four segments) what happens in this case. Can you fix the
problem on the go-back-n sender ?
5. Same question as above, but assume now that both the sender and the receiver implement selective repeat.
Note that the answer can be different from the above question.
4.3 Open questions
1. In your daily life, you also use hierarchical and flat address spaces. Can you provide examples of these two
types of addresses and discuss the benefits of using a hierarchical or flat addressing space in this particular
context ?
2. The network below uses port forwarding with flat addresses. The network boots and all hosts start one after
the other. Explain at each step how the packets are forwarded and how the port forwarding tables of the
network nodes are modified. Host C sends a packet to host B. Some time later, host A sends a packet to host
C. Finally, host B sends a packet to host A.
4.3. Open questions 229

A R1 R2
R3 B
3. Same question as above, but the network is modified as shown in the figure below.
A R1 R2
R3 B
4. Routing protocols used in data networks only use positive link weights. What would happen with a distance
vector routing protocol in the network below that contains a negative link weight ?
Figure 4.1: A simple network
5. When a network specialist designs a network, one of the problems that he needs to solve is to set the metrics
the links in his network. In the USA, the Abilene network interconnects most of the research labs and
universities. The figure below shows the topology of this network in 2009.
Figure 4.2: The Abilene network
In this network, assume that all the link weights are set to 1. What is the paths followed by a packet

sent by the router located in Los Angeles to reach :

the router located in New York
the router located in Washington ?
Is it possible to configure the link metrics so that the packets sent by the router located in Los Angeles
to the routers located in respectively New York and Washington do not follow the same path ?
Is it possible to configure the link weights so that the packets sent by the router located in Los Angeles
to router located in New York follow one path while the packets sent by the router located in New York
to the router located in Los Angeles follow a completely different path ?
Assume that the routers located in Denver and Kansas City need to exchange lots of packets. Can you
configure the link metrics such that the link between these two routers does not carry any packet sent
by another router in the network ?
6. In the five nodes network shown below, can you configure the link metrics so that the packets sent by router
E to router A use link B->A while the packets sent by router B use links B->D and D->A?
Figure 4.3: Simple five nodes network
7. In the five nodes network shown above, can you configure the link weights so that the packets sent by router
E (resp. F) follow the E->B->A path (resp. F->D->B->A) ?
8. Consider the network shown in the figure below.
R2
R1 R3
R4
Assuming that the network uses source routing, what are the possible paths from R1 to R4 ?
4.3. Open questions 231

R3 R4
R1 R2 R5
The network operator uses would like to have the following paths in this network :
R3->R2->R4->R5 and R1->R2->R5
Is it possible to achieve these paths and if so what are the required forwarding tables ?
R3 R4
R1 R2 R5
The network operator would like to use the following paths :

R1->R2->R4 and R3->R2->R5->R4
Are these paths possible with link-state or distance vector routing ? If yes, how do configure the link
weights. If no, explain your answer.
Same question with virtual circuits.
R3 R4
R1 R2 R5
The network operator would like to use the following paths :

R1->R5->R4 and R3->R2->R4
Are these paths possible with link-state or distance vector routing ? If yes, how do configure the link
weights. If no, explain your answer.
4.4 Discussion questions
1. The network below uses port forwarding tables. It has been running for several hours and all hosts have
exchanged packets. What is the content of the port forwarding tables ?

A R1 R2
R3 B
At this point, a new link is added between R1 and R3. What happens for the forwarding of packets ?
2. The network below uses port forwarding tables. What happens if host A moves by removing its link with
R1 and replacing it with a link with R3? How should networks using port forwarding deal with such mobile
hosts ?
A R1 R2
R3 B
3. Some hosts need to be multihomed, i.e. attached to two different network nodes as shown in the figure
below.
A R1 R2
R3 B
Would this network work correctly with port-forwarding tables if :

1. Host A uses the same flat address for both links.
2. Host A uses a different flat address on each of its links
4. What are the advantages and drawbacks of flat addresses versus hierarchical addresses ?
5. Let us now consider the transient problems that mainly happen when the network topology changes. For
this, consider the network topology shown in the figure below and assume that all routers use a distance
vector protocol that uses split horizon.
If you compute the routing tables of all routers in this network, you would obtain a table such as the
table below :
Destination Routes on A Routes on B Routes on C Routes on D Routes on E
A 0 1 via A 2 via B 3 via C 4 via D
B 1 via B 0 1 via B 2 via C 3 via D
C 2 via B 1 via C 0 1 via C 2 via D
D 3 via B 2 via C 1 via D 0 1 via D
E 4 via B 3 via C 2 via D 1 via E 0
4.4. Discussion questions 233

Figure 4.4: Simple network with redundant links
Distance vector protocols can operate in two different modes : periodic updates and triggered updates.
Periodic updates is the default mode for a distance vector protocol. For example, each router could
advertise its distance vector every thirty seconds. With the triggered updates a router sends its distance
vector when its routing table changes (and periodically when there are no changes).
Consider a distance vector protocol using split horizon and periodic updates. Assume that the
link B-C fails. B and C update their local routing table but they will only advertise it at the end
of their period. Select one ordering for the periodic updates and every time a router sends its
distance vector, indicate the vector sent to each neighbor and update the table above. How many
periods are required to allow the network to converge to a stable state ?
Consider the same distance vector protocol, but now with triggered updates. When link B-C
fails, assume that B updates its routing table immediately and sends its distance vector to A and
D. Assume that both A and D process the received distance vector and that A sends its own
distance vector, ... Indicate all the distance vectors that are exchanged and update the table
above each time a distance vector is sent by a router (and received by other routers) until all
routers have learned a new route to each destination. How many distance vector messages must
be exchanged until the network converges to a stable state ?
6. Consider again the network shown above. After some time, link state routing converges and all routers
compute the following routing tables :
Destination Routes on A Routes on B Routes on C Routes on D Routes on E
A 0 2 via C 1 via A 3 via B,E 2 via C
B 2 via C 0 1 via B 1 via B 2 via D,C
C 1 via C 1 via C 0 2 via B,E 1 via C
D 3 via C 1 via D 2 via B,E 0 1 via D
E 2 via C 2 via C,D 1 via E 1 via E 0
An important difference between OSPF and RIP is that OSPF routers flood link state packets that
allow the other routers to recompute their own routing tables while RIP routers exchange distance
vectors. Consider that link B-C fails and that router B is the first to detect the failure. At this point,
B cannot reach anymore A, C and 50% of its paths towards E have failed. C cannot reach B anymore
and half of its paths towards D have failed.
Router B will flood its updated link state packet through the entire network and all routers will re-
compute their forwarding table. Upon reception of a link state packet, routers usually first flood the
received link-state packet and then recompute their forwarding table. Assume that B is the first to
recompute its forwarding table, followed by D, A, C and finally E
7. After each update of a forwarding table, verify which pairs of routers are able to exchange packets. Provide
your answer using a table similar to the one shown above.

4.5 Serving applications
1. What are the mechanisms that should be included in a transport protocol that provides an unreliable con-
nectionless transport service that can detect transmission errors but not correct them ?
2. A reliable connection oriented transport places a 32 bits sequence number inside the segment header to
number the segments. This sequence number is incremented for each data segment. The connection starts
as shown in the figure below :
Host A Host B
CONNECT.req()
CR(seq=1341)
CONNECT.ind()
CONNECT.resp()
CR(ack=1341,seq=2141)
CONNECT.conf()
CA(seq=1341,ack=2141)
DATA.req(a)
Continue the connection so that Host B sends Hello as data and Host A replies by sending Pleased
to meet you. After havng received the response, Host B closes the connection gracefully and Host A
does the same. Discuss on the state that needs to be maintained inside each host.
3. A transport connection that provides a message-mode service has been active for some time and all data
has been exchanged and acknowledged in both directions. As in the exercise above, the sequence number
is incremented after the transmission of each segment. At this time, Host A sends two DATA segments as
Host A Host B
DATA.req(abc)
DATA(seq=1123,abc)
DATA.req(def)
DATA(seq=1124,def)
DISCONNECT.req(graceful,A->B)
What are the acknowledgements sent by Host B, how does Host A react and how does it terminate
the connection ?
4. Consider a reliable connection-oriented transport protocol that provides the bytestream service. In this
transport protocol, the sequence number that is placed inside each DATA segment reflects the position of
the bytes in the bytestream. Considering the connection shown below, provide the DATA segments that are
sent by Host A in response to the DATA.request, assuming that one segment is sent for each DATA.request.
4.5. Serving applications 235

Host A Host B
CONNECT.req()
CR(seq=8765)
CONNECT.ind()
CONNECT.resp()
CR(ack=8765,seq=4321)
CONNECT.conf()
CA(seq=8765,ack=4321)
DATA.req(a)
DATA.req(bcdefg)
DATA.req(ab)
DATA.req(bbbbbbbbbbbb)
5. Same question as above, but consider now that the transport protocol tries to send large DATA segments
whenever possible. For this exercise, we consider that a DATA segment can contain up to 8 bytes of data in
the payload. Do not forget to show the acknowledgements in your answer.
6. Consider a transport protocol that provides a reliable connection-oriented bystream service. You observe
the segments sent by a host that uses this protocol. Does the time-sequence diagram below reflects a valid
implementation of this protocol ? Justify your answer.
Host A Host B
DATA.req(abc)
DATA(seq=1123,abc)
DATA.req(def)
DATA(seq=1126,def)
DATA(seq=1123,abcdef)
7. In the above example, the two DATA segments were lost before arriving at the destination. Discuss the
following scenario and explain how the receiver should react to the reception of the last DATA segment.
Host A Host B
DATA.req(abc)
DATA(seq=1123,abc)
DATA.req(def)
DATA(seq=1126,def)
DATA(seq=1123,abcdef)
8. A network layer service guarantees that network will never live during more than 100 seconds inside the
network. A reliable connection-oriented transport protocol places a 32 bits sequence number inside each
segment. What is the maximum rate (in segments per second) at which is should sent data segments to
prevent having two segments with the same sequence number inside the network ?
4.5.2 Using the socket API
Networked applications were usually implemented by using the socket API. This API was designed when TCP/IP
was first implemented in the Unix BSD operating system [Sechrest] [LFJLMT], and has served as the model for

many APIs between applications and the networking stack in an operating system. Although the socket API is
very popular, other APIs have also been developed. For example, the STREAMS API has been added to several
Unix System V variants [Rago1993]. The socket API is supported by most programming languages and several
textbooks have been devoted to it. Users of the C language can consult [DC2009], [Stevens1998], [SFR2004] or
[Kerrisk2010]. The Java implementation of the socket API is described in [CD2008] and in the Java tutorial. In
this section, we will use the python implementation of the socket API to illustrate the key concepts. Additional
information about this API may be found in the socket section of the python documentation .
The socket API is quite low-level and should be used only when you need a complete control of the network
access. If your application simply needs, for instance, to retrieve data from a web server, there are much simpler
and higher-level APIs.
A detailed discussion of the socket API is outside the scope of this section and the references cited above provide
a detailed discussion of all the details of the socket API. As a starting point, it is interesting to compare the
socket API with the service primitives that we have discussed in the previous chapter. Let us first consider the
connectionless service that consists of the following two primitives :
DATA.request(destination,message) is used to send a message to a specified destination. In this socket API,
this corresponds to the send method.
DATA.indication(message) is issued by the transport service to deliver a message to the application. In the
socket API, this corresponds to the return of the recv method that is called by the application.
The DATA primitives are exchanged through a service access point. In the socket API, the equivalent to the service
access point is the socket. A socket is a data structure which is maintained by the networking stack and is used by
the application every time it needs to send or receive data through the networking stack. The socket method in the
python API takes two main arguments :
an address family that specifies the type of address family and thus the underlying networking
stack that will be used with the socket. This parameter can be either socket.AF_INET or
socket.AF_INET6. socket.AF_INET, which corresponds to the TCP/IPv4 protocol stack is the
default. socket.AF_INET6 corresponds to the TCP/IPv6 protocol stack.
a type indicates the type of service which is expected from the networking stack. socket.STREAM (the
default) corresponds to the reliable bytestream connection-oriented service. socket.DGRAM corresponds
to the connectionless service.
A simple client that sends a request to a server is often written as follows in descriptions of the socket API.
import socket
import sys
HOSTIP=sys.argv[1]
PORT=int(sys.argv[2])
MSG="Hello, World!"
s = socket.socket( socket.AF_INET6, socket.SOCK_DGRAM )
s.sendto( MSG, (HOSTIP, PORT,0,0) )
A typical usage of this application would be

python client.py ::1 12345
where ::1 is the IPv6 address of the host (in this case the localhost) where the server is running and 12345 the
port of the server.
The first operation is the creation of the socket. Two parameters must be specified while creating a socket.
The first parameter indicates the address family and the second the socket type. The second operation is the
transmission of the message by using sendto to the server. It should be noted that sendto takes as arguments
the message to be transmitted and a tuple that contains the IPv6 address of the server and its port number.
The code shown above supports only the TCP/IPv6 protocol stack. To use the TCP/IPv4 protocol stack the
socket must be created by using the socket.AF_INET address family. Forcing the application developer
to select TCP/IPv4 or TCP/IPv6 when creating a socket is a major hurdle for the deployment and usage of
TCP/IPv6 in the global Internet [Cheshire2010]. While most operating systems support both TCP/IPv4 and
TCP/IPv6, many applications still only use TCP/IPv4 by default. In the long term, the socket API should

be able to handle TCP/IPv4 and TCP/IPv6 transparently and should not force the application developer to always
specify whether it uses TCP/IPv4 or TCP/IPv6.
Another important issue with the socket API as supported by python is that it forces the application to deal with IP
addresses instead of dealing directly with domain names. This limitation dates from the early days of the socket
API in Unix 4.2BSD. At that time, name resolution was not widely available and only IP addresses could be
used. Most applications rely on DNS names to interact with servers and this utilisation of the DNS plays a very
important role to scale web servers and content distribution networks. To use domain names, the application needs
to perform the DNS resolution by using the getaddrinfo method. This method queries the DNS and builds
the sockaddr data structure which is used by other methods of the socket API. In python, getaddrinfo takes
several arguments :
a name that is the domain name for which the DNS will be queried
an optional port number which is the port number of the remote server
an optional address family which indicates the address family used for the DNS request.
socket.AF_INET (resp. socket.AF_INET6) indicates that an IPv4 (IPv6) address is expected. Fur-
thermore, the python socket API allows an application to use socket.AF_UNSPEC to indicate that it is
able to use either IPv4 or IPv6 addresses.
an optional socket type which can be either socket.SOCK_DGRAM or socket.SOCK_STREAM
In todays Internet hosts that are capable of supporting both IPv4 and IPv6, all applications should be
able to handle both IPv4 and IPv6 addresses. When used with the socket.AF_UNSPEC parameter, the
socket.getaddrinfo method returns a list of tuples containing all the information to create a socket.
import socket
socket.getaddrinfo(www.example.net,80,socket.AF_UNSPEC,socket.SOCK_STREAM)
[ (30, 1, 6, , (2001:db8:3080:3::2, 80, 0, 0)),
(2, 1, 6, , (203.0.113.225, 80))]
In the example above, socket.getaddrinfo returns two tuples. The first one corresponds to the sockaddr
containing the IPv6 address of the remote server and the second corresponds to the IPv4 information. Due to some
peculiarities of IPv6 and IPv4, the format of the two tuples is not exactly the same, but the key information in
both cases are the network layer address (2001:db8:3080:3::2 and 203.0.113.225) and the port number
(80). The other parameters are seldom used.
socket.getaddrinfo can be used to build a simple client that queries the DNS and contact the server by
using either IPv4 or IPv6 depending on the addresses returned by the socket.getaddrinfo method. The
client below iterates over the list of addresses returned by the DNS and sends its request to the first destination
address for which it can create a socket. Other strategies are of course possible. For example, a host running in
an IPv6 network might prefer to always use IPv6 when IPv6 is available 1 .
import socket
import sys
HOSTNAME=sys.argv[1]
MSG="Hello, World!"
for a in socket.getaddrinfo(HOSTNAME, PORT, socket.AF_UNSPEC,socket.SOCK_DGRAM,0, socket.AI_PASSIV
address_family,sock_type,protocol,canonicalname, sockaddr=a
try:
s = socket.socket(address_family, sock_type)
except socket.error:
s = None
print "Could not create socket"
continue
if s is not None:
s.sendto(MSG, sockaddr)
break
1 Most operating systems today by default prefer to use IPv6 when the DNS returns both an IPv4 and an IPv6 address for a name. See
http://ipv6int.net/systems/ for more detailed information.

Now that we have described the utilisation of the socket API to write a simple client using the connectionless
transport service, let us have a closer look at the reliable byte stream transport service. As explained above, this
service is invoked by creating a socket of type socket.SOCK_STREAM. Once a socket has been created, a
client will typically connect to the remote server, send some data, wait for an answer and eventually close the
connection. These operations are performed by calling the following methods :
socket.connect : this method takes a sockaddr data structure, typically returned by
socket.getaddrinfo, as argument. It may fail and raise an exception if the remote server cannot
be reached.
socket.send : this method takes a string as argument and returns the number of bytes that were actually
sent. The string will be transmitted as a sequence of consecutive bytes to the remote server. Applications
are expected to check the value returned by this method and should resend the bytes that were not send.
socket.recv : this method takes an integer as argument that indicates the size of the buffer that has been
allocated to receive the data. An important point to note about the utilisation of the socket.recv method
is that as it runs above a bytestream service, it may return any amount of bytes (up to the size of the buffer
provided by the application). The application needs to collect all the received data and there is no guarantee
that some data sent by the remote host by using a single call to the socket.send method will be received
by the destination with a single call to the socket.recv method.
socket.shutdown : this method is used to release the underlying connection. On some platforms, it is
possible to specify the direction of transfer to be released (e.g. socket.SHUT_WR to release the outgoing
direction or socket.SHUT_RDWR to release both directions).
socket.close: this method is used to close the socket. It calls socket.shutdown if the underlying
connection is still open.
With these methods, it is now possible to write a simple HTTP client. This client operates over both IPv6 and IPv4
and writes the main page of the remote server on the standard output. It also reports the number of socket.recv
calls that were used to retrieve the homepage 2 . We will provide more details on the HTTP protocol that is used
in this example later.
#!/usr/bin/python
# A simple http client that retrieves the first page of a web site
import socket, sys
if len(sys.argv)!=3 and len(sys.argv)!=2:

print "Usage : ",sys.argv[0]," hostname [port]"
hostname = sys.argv[1]
if len(sys.argv)==3 :
port=int(sys.argv[2])
else:
port = 80
READBUF=16384 # size of data read from web server

s=None
for res in socket.getaddrinfo(hostname, port, socket.AF_UNSPEC, socket.SOCK_STREAM):

af, socktype, proto, canonname, sa = res
# create socket
try:
s = socket.socket(af, socktype, proto)
except socket.error:
s = None
continue
# connect to remote host
try:
2 Experiments with the client indicate that the number of socket.recv calls can vary at each run. There are various factors that influence
the number of such calls that are required to retrieve some information from a server. Well discuss some of them after having explained the
operation of the underlying transport protocol.

print "Trying "+sa[0]

s.connect(sa)
except socket.error, msg:
# socket failed
s.close()
s = None
continue
if s :
print "Connected to "+sa[0]
s.send(GET / HTTP/1.1\r\nHost:+hostname+\r\n\r\n)
finished=False
count=0
while not finished:
data=s.recv(READBUF)
count=count+1
if len(data)!=0:
print repr(data)
else:
finished=True
s.shutdown(socket.SHUT_WR)
s.close()
print "Data was received in ",count," recv calls"
break
The second type of applications that can be written by using the socket API are the servers. A server is typically
runs forever waiting to process requests coming from remote clients. A server using the connectionless will
typically start with the creation of a socket with the socket.socket. This socket can be created above the
TCP/IPv4 networking stack (socket.AF_INET) or the TCP/IPv6 networking stack (socket.AF_INET6),
but not both by default. If a server is willing to use the two networking stacks, it must create two threads, one to
handle the TCP/IPv4 socket and the other to handle the TCP/IPv6 socket. It is unfortunately impossible to define
a socket that can receive data from both networking stacks at the same time with the python socket API.
A server using the connectionless service will typically use two methods from the socket API in addition to those
that we have already discussed.
socket.bind is used to bind a socket to a port number and optionally an IP address. Most servers will
bind their socket to all available interfaces on the servers, but there are some situations where the server
may prefer to be bound only to specific IP addresses. For example, a server running on a smartphone might
want to be bound to the IP address of the WiFi interface but not on the 3G interface that is more expensive.
socket.recvfrom is used to receive data from the underlying networking stack. This method returns
both the senders address and the received data.
The code below illustrates a very simple server running above the connectionless transport service that simply
prints on the standard output all the received messages. This server uses the TCP/IPv6 networking stack.
import socket, sys
BUFF_LEN=8192
s=socket.socket(socket.AF_INET6, socket.SOCK_DGRAM)
s.bind((,PORT,0,0))
while True:
data, addr = s.recvfrom( BUFF_LEN )
if data=="STOP" :
print "Stopping server"
sys.exit(0)
print "received from ", addr, " message:", data
A server that uses the reliable byte stream service can also be built above the socket API. Such a server starts
by creating a socket that is bound to the port that has been chosen for the server. Then the server calls the
socket.listen method. This informs the underlying networking stack of the number of transport connection

attempts that can be queued in the underlying networking stack waiting to be accepted and processed by the
server. The server typically has a thread waiting on the socket.accept method. This method returns as soon
as a connection attempt is received by the underlying stack. It returns a socket that is bound to the established
connection and the address of the remote host. With these methods, it is possible to write a very simple web server
that always returns a 404 error to all GET requests and a 501 errors to all other requests.
# An extremely simple HTTP server
import socket, sys, time
# Server runs on all IP addresses by default

HOST=
# 8080 can be used without root priviledges
PORT=8080
BUFLEN=8192 # buffer size
s = socket.socket(socket.AF_INET6, socket.SOCK_STREAM)
try:
print "Starting HTTP server on port ", PORT
s.bind((HOST,PORT,0,0))
except socket.error :
print "Cannot bind to port :",PORT
sys.exit(-1)
s.listen(10) # maximum 10 queued connections
while True:
# a real server would be multithreaded and would catch exceptions
conn, addr = s.accept()
print "Connection from ", addr
data=
while not \n in data : # wait until first line has been received
data = data+conn.recv(BUFLEN)
if data.startswith(GET):
# GET request
conn.send(HTTP/1.0 404 Not Found\r\n)
# a real server should serve files
else:
# other type of HTTP request
conn.send(HTTP/1.0 501 Not implemented\r\n)
now = time.strftime("%a, %d %b %Y %H:%M:%S", time.localtime())

conn.send(Date: + now +\r\n)
conn.send(Server: Dummy-HTTP-Server\r\n)
conn.send(\r\n)
conn.shutdown(socket.SHUT_RDWR)
conn.close()
This server is far from a production-quality web server. A real web server would use multiple threads and/or
non-blocking IO to process a large number of concurrent requests 3 . Furthermore, it would also need to handle
all the errors that could happen while receiving data over a transport connection. These are outside the scope
of this section and additional information on more complex networked applications may be found elsewhere.
For example, [RG2010] provides an in-depth discussion of the utilisation of the socket API with python while
[SFR2004] remains an excellent source of information on the socket API in C.
4.5.3 Practice
1. The socket interface allows you to use the UDP protocol that provides the connectionless service on a Unix
host. UDP, in theory, allows you to send SDUs of up to 64 KBytes.
3 There are many production quality web servers software available. apache is a very complex but widely used one. thttpd and lighttpd are
less complex and their source code is probably easier to understand.

Implement a small UDP client and a small UDP server in C

run the client and the servers on different workstations to determine experimentally the largest SDU that is
supported by your language and OS.
2. The time protocol, defined in RFC 868 allows to read the current time on a remote host. Implement this
very simple protocol on top of the UDP and TCP sockets. Compare the time required to retrieve this time
information over UDP and TCP.
3. By using the socket interface, implement on top of the connectionless unreliable service provided by UDP
a simple client that sends the message shown in the figure below.
In this message, the bit flags should be set to 01010011b, the value of the 16 bits field must be
the square root of the value contained in the 32 bits field, the character string must be an ASCII
representation (without any trailing 0) of the number contained in the 32 bits character field. The last
16 bits of the message contain an Internet checksum that has been computed over the entire message.
Upon reception of a message, the server verifies that :
the flag has the correct value
the 32 bits integer is the square of the 16 bits integer
the character string is an ASCII representation of the 32 bits integer
the Internet checksum is correct
If the verification succeeds, the server returns a SDU containing 11111111b. Otherwise it returns
01010101b
Your implementation must be able to run on both low endian and big endian machines. If you have
access to different types of machines (e.g. x86 laptops and SPARC servers), try to run your imple-
mentation on both types of machines.
4. The socket library is also used to develop applications above the reliable bytestream service provided by
TCP. We have implemented in C language a small application that sends information to a server on port
62141. This server should operate as follows :
the server listens on port 62141 for a TCP connection
upon the establishment of a TCP connection, the server sends an integer by using the following
TLV format :
the first two bits indicate the type of information (01 for ASCII, 10 for boolean)
the next six bits indicate the length of the information (in bytes)
An ASCII TLV has a variable length and the next bytes contain one ASCII character per
byte. A boolean TLV has a length of one byte. The byte is set to 00000000b for true and
00000001b for false.
the client replies by sending the received integer encoded as a 32 bits integer in network byte
order
the server returns a TLV containing true if the integer was correct and a TLV containing false
otherwise and closes the TCP connection
Implement a server in C that interacts with our client available at /exercises/c/tlv_ex.c
5. The Trivial File Transfer Protocol (TFTP), defined in RFC 1350 is a simple file transfer protocol that runs
on top of UDP. Read the specification to answer the following questions :
What is the maximum length of the DATA segments exchanged by this protocol ?
What is the legnth of the header in DATA segments ?
What is the first sequence number for DATA segments ?

Do sequence number count the segments or the bytes that are transmitted ?
Does this protocol uses a sliding window ?
How does the data transfer ends ? Consider two different files. The first one has a length of exactly
1024 bytes, the second 513 bytes. Explain what is the last segment sent in each direction in each case.
1. In the transport layer, the receive window advertised by a receiver can vary during the lifetime of the con-
nection. What are the causes for these variations ?
2. A reliable connection-oriented protocol can provide a message-mode service or a byte stream service.
Which of the following usages of the sequence numbers is the best suited for each of these services ?
1. DATA segments contain a sequence number that is incremented for each byte transmitted
2. DATA segments contain a sequence number that is incremented for each DATA segment transmitted
3. Some transport protocols use 32 bits sequence numbers while others use 64 bits sequence number. What
are the advantages and drawbacks of each approach ?
4. Consider a transport protocol that provides the bytestream service and uses 32 bits sequence number to
represent the position of the first byte of the payload of DATA segments in the bytestream. How would you
modify this protocol so that it can provide a message-mode service ? Consider first short messages that
always fit inside a single segment. In a second step, discuss how you could support messages of unlimited
size.
5. What is piggybacking and what are the benefits of this technique ?
4.6 Sharing resources
To understand the operation of Medium Access Control algorithms, it is often interesting to use a geometric repre-
sentation of the transmission of frames on a shared medium. This representation is suitable if the communicating
devices are attached to a single cable. Consider a simple scenario with a host connected at one end of a cable.
For simplicity, let us consider a cable that has a length of one kilometer. Let us also consider that the propagation
delay of the electrical signal is five microseconds per kilometer. The figure below shows the transmission of a
2000 bits frame at 100 Mbps by host A on the cable.

A C B
time
If the transmitting host is located at another position on the shared medium than one of the edges, then the
geometrical pattern that represents the transmission of a frame is slightly different. If the transmitting host is
placed in the middle of the cable, then the signal is transmitted in both directions on the cable. The figure below
shows the transmission of one 100 bits frame at 100 Mbps by host C on the same cable.
A C B
time

In a shared medium, a collision may happen if two hosts transmit at almost the same time as shown in the example
below.
A C B
time
1. Consider the following scenario for the ALOHA medium access control algorithm. Three hosts are attached
to a one-kilometer long cable and transmit 1000 bits frames at 1 Mbps. Each arrow represents a request
to transmit a frame on the corresponding host. Each square represents 250 microseconds in the figure.
Represent all the transmitted frames and list the frames that collide.

A C B
A
250sec
time
2. Same question as above, but now consider that the hosts transmit 1000 bits frames at 100 Mbps. The cable
has a length of 2 kilometers. C is in the middle of the cable. Each square in the figure below corresponds to
10 microseconds.
3. In ALOHA, the hosts rely on acknowledgements to detect whether their frame has been received correctly
by the destination. Consider a network running at 100 Mbps where the host exchange 1000 bits frames and
acknowledgements of 100 bits. Draw the frames sent by hosts A and B in the figure below. Assume that a
square corresponds to 10 microseconds and that the cable has a length of 2 kilometers.

A C B
A
10sec
C B
time
Figure 4.5: :libs: positioning, matrix, arrows
A C B
AB [1000 bits]
10sec
BA [1000 bits]
BA [1000 bits]
time
4. Same question as above, but now assume that the retransmission timer of each host is set to 50 microseconds.

A C B
AB [1000 bits]
10sec
CA [1000 bits]
BA [1000 bits]
time
Figure 4.6: :libs: positioning, matrix, arrows
5. In practice, hosts transmit variable length frames. Consider a cable having a bandwidth of 100 Mbps and a
length of 2 kilometers.

A C B
AB [2000 bits]
10sec
CA [1000 bits]
BA [2000 bits]
time
6. With CSMA, hosts need to listen to the communication channel before starting their transmission. Consider
again a 2 kilometers long cable where hosts send frames at 100 Mbps. Show in the figure below the correct
transmission of frames with CSMA.
A C B
AB [2000 bits]
10sec
CA [1000 bits]
BA [1000 bits]
time
7. CSMA/CD does not use acknowledgements but instead assumes that each host can detect collisions by

listening while transmitting. Consider a 2 kilometers long cable running at 10 Mbps. Show in the figure
below the utilisation of the communication channel and the collisions that would occur. For this exercise,
do not attempt to retransmit the frames that have collided.
A C B
AB [200 bits] CA [100 bits]

10sec
BA [100 bits]
time
8. Consider again a network that uses CSMA/CD. This time, the bandwidth is set to 1 Gbps and the cable has
a length of two kilometers. When a collision occurs, consider that the hosts B and C retransmit immediately
while host A waits for the next slot.

A C B
AB [10000 bits]
5sec
CA [10000 bits]
BA [10000 bits]
AB [10000 bits]
time
9. An important part of the CSMA/CD algorithm is the exponential backoff. To illustrate the operation of
this algorithm, let us consider a cable that has a length of one kilometer. The bandwidth of the network is
set to 10 Mbps. Assume that when a collision occurs, host A always selects the highest possible random
delay according to the exponential backoff algorithm while host B always selects the shortest one. In this
network, the slot time is equal to the time required to transmit 100 bits. We further assume that a host can
detect collision immediately (i.e. as soon as the other frame arrives).

A C B
5sec
BC [100 bits]
AC [100 bits]
BC [100 bits]
time
10. Consider the network below. Compute the max-min fair allocation for the hosts in this network assuming
that nodes Sx always send traffic towards node Dx. Furthermore, link R1-R2 has a bandwidth of 10 Mpbs
while link R2-R3 has a bandwidth of 20 Mbps.
D4 D5
S1 R1 R2 D2
S2 R3 D1
S3 D3
S4
S5
To understand congestion control algorithms, it can also be useful to represent the exchange of packets by using
a graphical representation. As a first example, let us consider a very simple network composed of two hosts
interconnected through a switch.

3 Mbps 1 Mbps
A R B
Suppose now that host A uses a window of three segments and sends these three segments immediately. The
segments will be queued in the router before being transmitted on the output link and delivered to their destination.
The destination will reply with a short acknowledgement segment. A possible visualisation of this exchange of
packets is represented in the figure below. We assume for this figure that the router marks the packets to indicate
congestion as soon as its buffer is non-empty when its receives a packet on its input link. In the figure, a (c) sign
is added to each packet to indicate that it has been explicitly marked.
Input Output
P1
P2
P3 P1
P2(c)
P3(c)
time
In practice, a router is connected to multiple input links. The figure below shows an example with two hosts.
2 Mbps 1 Mbps
A R C
2 Mbps
B

InputA InputB Output
P1 P1
P2 P2
P1
P1(c)
P2(c)
P2(c)
time
In general, the links have a non-zero delay. This is illustrated in the figure below where a delay has been added on
the link between R and C.
InputA InputB Output
P1 P1
P2 P2
P1
P1(c)
P2(c)
P2(c)
time
11. Consider the network depicted in the figure below.

1 Mbps
0 msec
A R1 R2 D
1 Mbps 100 Mbps
10 msec 0 msec
10 Mbps
10 msec
B
1. In this network, compute the minimum round-trip-time between A (resp. B) and D. Perform the
computation if the hosts send segments containing 1000 bits.
2. How is the maximum round-trip-time influenced if the buffers of router R1 store 10 packets ?
3. If hosts A and B send to D 1000 bits segments and use a sending window of four segments, what
is the maximum throughput that they can achieve ?
4. Assume now that R1 is using round-robin scheduling instead of a FIFO buffer. One queue is
used to store the packets sent by A and another for the packets sent by B. A sends one 1000 bits
packet every second while B sends packets at 10 Mbps. What is the round-trip-time measured
by each of these two hosts if each of the two queues of R1 can store 5 packets ?
12. When analyzing the reaction of a network using round-robin schedulers, it is sometimes useful to consider
that the packets sent by each source are equivalent to a fluid and that each scheduler acts as a tap. Using
this analogy, consider the network below. In this network, all the links are 100 Mbps and host B is sending
packets at 100 Mbps. If A sends at 1, 5, 10, 20, 30, 40, 50, 60, 80 and 100 Mbps, what is the throughput
that destination D will receive from A. Use this data to plot a graph that shows the portion of the traffic sent
by host A which is received by host D.
A R1 D
13. Compute the max-min fair bandwidth allocation in the network below.
Figure 4.7: Simple network topology
14. Consider the simple network depicted in the figure below.

1 Mbps
A R1 R2 D
250 kbps 100 Mbps
1. In this network, a 250 Kbps link is used between the routers. The propagation delays in the net-
work are negligible. Host A sends 1000 bits long segments so that it takes one msec to transmit
one segment on the A-R1 link. Neglecting the transmission delays for the acknowledgements,
what is the minimum round-trip time measured on host A with such segments ?
2. If host A uses a window of two segments and needs to transmit five segments of data. How long
does the entire transfer lasts ?
3. Same question as above, but now host A uses the simple DECBIT congestion control mechanism
and a maximum window size of four segments.
15. Consider the network depicted in the figure below.
1 Mbps
10 msec
A R1 D
1 Mbps 500 kbps
0 msec 0 msec
Hosts A and B use the simple congestion control scheme described in the book and router R1 uses the
DECBIT mechanism to mark packets as soon as its buffers contain one packet. Hosts A and B need
to send five segments and start exactly at the same time. How long does each hosts needs to wait to
receive the acknowledgement for its fifth segment ?
1. In a deployed CSMA/CD network, would it be possible to increase or decrease the duration of the slottime
? Justify your answer
2. Consider a CSMA/CD network that contains hosts that generate frames at a regular rate. When the trans-
mission rate increases, the amount of collisions increases. For a given network load, measured in bits/sec,
would the number of collisions be smaller, equal or larger with short frames than with long frames ?
3. Slotted ALOHA improves the performance of ALOHA by dividing the time in slots. However, this basic
idea raises two interested questions. First how would you enforce the duration of these slots ? Second,
should a slot include the time to transmit a data frame or the time to transmit a data frame and the corre-
sponding acknowledgement ?
4. Like ALOHA, CSMA relies on acknowledgements to detect where a frame has been correctly received.
When a host senses an idle channel, if should transmit its frame immediately. How should it react if it
detects that another host is already transmitting ? Consider two options :
the host continues to listen until the communication channel becomes free. It transmits as soon as the
communication channel becomes free.
the host stops to listen and waits for a random time before sensing again the communication channel to
check whether it is free.
4.7 Application layer

4.7.1 The DNS
The Domain Name System (DNS) plays a key role in the Internet today as it allows applications to use fully
qualified domain names (FQDN) instead of IPv4 or IPv6 addresses. Many tools enable queries through DNS
servers. For this exercise, we will use dig which is installed on most Unix systems.
A typical usage of dig is as follows
dig @server -t type fqdn
where
server is the IP address or the name of a DNS server or resolver
type is the type of DNS record that is requested by the query such as NS for a nameserver, A for an IPv4
address, AAAA for an IPv6 address, MX for a mail relay, ...
fqdn is the fully qualified domain name being queried
dig also contains some additional parameters and flags that are described in the manpage. Among these, the +trace
flag allows to trace all requests that are sent when recursing through DNS servers.
1. What are the IP addresses of the resolvers that the dig implementation you are using relies on 4 ?
2. What is the IPv6 address that corresponds to inl.info.ucl.ac.be ? Which type of DNS query does dig send to
obtain this information ?
3. Which type of DNS request do you need to send to obtain the nameservers that are responsible for a given
domain ?
4. What are the nameservers that are responsible for the be top-level domain ? Is it possible to use IPv6 to
query them ?
5. When run without any parameter, dig queries one of the root DNS servers and retrieves the list of the
names of all root DNS servers. For technical reasons, there are only 13 different root DNS servers. This
information is also available as a text file from http://www.internic.net/zones/named.root . What are the
IPv6 addresses of all these servers.
6. Assume now that you are residing in a network where there is no DNS resolver and that you need to perform
your query manualla starting from the DNS root.
Use dig to send a query to one of these root servers to find the IPv6 address of the DNS server(s) (NS
record) responsible for the org top-level domain
Use dig to send a query to one of these DNS servers to find the IP address of the DNS server(s) (NS
record) responsible for root-servers.org
Continue until you find the server responsible for www.root-servers.org
What is the lifetime associated to this IPv6 address ?
7. Perform the same analysis for a popular website such as www.google.com. What is the lifetime associated
to the corresponding IPv6 address ? If you perform the same request several times, do you always receive
the same answer ? Can you explain why a lifetime is associated to the DNS replies ?
8. Use dig to find the mail relays used by the uclouvain.be and student.uclouvain.be domains. What is the TTL
of these records ? Can you explain the preferences used by the MX records. You can find more information
about the MX records in RFC 5321
9. When dig is run, the header section in its output indicates the id the DNS identifier used to send the query.
Does your implementation of dig generates random identifiers ?
dig -t MX gmail.com
; <<>> DiG 9.4.3-P3 <<>> -t MX gmail.com

;; global options: printcmd
4 On a Linux machine, the Description section of the dig manpage tells you where dig finds the list of nameservers to query.
4.7. Application layer 257

;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 25718
10. A DNS implementation such as dig and more importantly a name resolver such as bind or unbound, always
checks that the received DNS reply contains the same identifier as the DNS request that it sent. Why is this
so important ?
Imagine an attacker who is able to send forged DNS replies to, for example, associate www.bigbank.com to
his own IP address. How could he attack a DNS implementation that
sends DNS requests containing always the same identifier
sends DNS requests containing identifiers that are incremented by one after each request
sends DNS requests containing random identifiers
11. The DNS protocol can run over UDP and over TCP. Most DNS servers prefer to use UDP because it
consumes fewer resources on the server. However, TCP is useful when a large answer is expected or when
a large answer is expected. Use time dig +tcp to query a root DNS server. Is it faster to receive an answer
via TCP or via UDP ?
4.7.2 Internet email protocols
Many Internet protocols are ASCII-based protocols where the client sends requests as one line of ASCII text
terminated by CRLF and the server replies with one of more lines of ASCII text. Using such ASCII messages has
several advantages compared to protocols that rely on binary encoded messages
the messages exchanged by the client and the server can be easily understood by a developer or network
engineer by simply reading the messages
it is often easy to write a small prototype that implements a part of the protocol
it is possible to test a server manually by using telnet Telnet is a protocol that allows to obtain a terminal
on a remote server. For this, telnet opens a TCP connection with the remote server on port 23. However,
most telnet implementations allow the user to specify an alternate port as telnet hosts port When used with
a port number as parameter, telnet opens a TCP connection to the remote host on the specified port. telnet
can thus be used to test any server using an ASCII-based protocol on top of TCP. Note that if you need to
stop a running telnet session, Ctrl-C will not work as it will be sent by telnet to the remote host over the
TCP connection. On many telnet implementations you can type Ctrl-] to freeze the TCP connection and
return to the telnet interface.
1. Use your preferred email tool to send an email message to yourself containing a single line of text. Most
email tools have the ability to show the source of the message, use this function to look at the message that
you sent and the message that you received. Can you find an explanation for all the lines that have been
added to your single line email ?
2. The TCP protocol supports 65536 different ports numbers. Many of these port numbers have been reserved
for some applications. The official repository of the reserved port numbers is maintained by the Internet
Assigned Numbers Authority (IANA) on http://www.iana.org/assignments/port-numbers 5 Using this in-
formation, what is the default port number for the POP3 protocol ? Does it run on top of UDP or TCP
?
3. The Post Office Protocol (POP) is a rather simple protocol described in RFC 1939. POP operates in three
phases. The first phase is the authorization phase where the client provides a username and a password. The
second phase is the transaction phase where the client can retrieve emails. The last phase is the update phase
where the client finalises the transaction. What are the main POP commands and their parameters ? When
a POP server returns an answer, how can you easily determine whether the answer is positive or negative ?
4. On smartphones, users often want to avoid downloading large emails over a slow wireless connection. How
could a POP client only download emails that are smaller than 5 KBytes ?
5 On Unix hosts, a subset of the port assignments is often placed in /etc/services

4.7.3 The HyperText Transfer Protocol
1. System administrators who are responsible for web servers often want to monitor these servers and check
that they are running correctly. As a HTTP server uses TCP on port 80, the simplest solution is to open a
TCP connection on port 80 and check that the TCP connection is accepted by the remote host. However, as
HTTP is an ASCII-based protocol, it is also very easy to write a small script that downloads a web page on
the server and compares its content with the expected one. Use telnet to verify that a web server is running
on host cnp3book.info.ucl.ac.be 6
2. Instead of using telnet on port 80, it is also possible to use a command-line tool such as curl Use curl with
the trace-ascii tracefile option to store in tracefile all the information exchanged by curl when accessing
the server.
what is the version of HTTP used by curl ?
can you explain the different headers placed by curl in the request ?
can you explain the different headers found in the response ?
3. HTTP 1.1, specified in RFC 2616 forces the client to use the Host: in all its requests. HTTP 1.0 does not
define the Host: header, by most implementations support it. By using telnet and curl retrieve the first page
of the http://cnp3book.info.ucl.ac.be webserver by sending http requests with and without the Host: header.
Explain the difference between the two.
4. The headers sent in a HTTP request allow the client to provide additional information to the server. One of
these headers is the Accept-Language header that allows to indicate the preferred language of the client 7 .
For example, curl -HAccept-Language:en http://www.google.be will send to http://www.google.be a HTTP
request indicating English (en) as the preferred language. Does google provides a different page in French
(fr) and Walloon (wa) ? Same question for http://www.uclouvain.be (given the size of the homepage, use
diff to compare the different pages retrieved from www.uclouvain.be)
5. Compare the size of the http://www.yahoo.com and http://www.google.com web pages by downloading
them with curl
6. The ipvfoo extension on google chrome allows to visually detect whether a website is using IPv6 and IPv4,
but also to see which web sites have been contacted when rendering a given webpage. Some websites are
distributed over several dozens of different servers. Can you find one ?
4.8 Configuring DNS and HTTP servers
Configuring DNS and HTTP servers can be complex on real hosts. To allow you to learn network configurations
without risking breaking anything, we will use Netkit. Netkit is network emulator based on User Mode Linux.
It allows to easily set up an emulated network composed of virtual Linux machines, that can act as end-host or
routers.
Note: Where can I find Netkit?

Netkit is available at http://www.netkit.org. For the labs, we have built a custom netkit image which is installed
on the INGI servers 8 .
There are two ways to use Netkit : The manual way, and by using pre-configured labs. In the first case, you
boot and control each machine individually, using the commands starting with a v (for virtual machine). In the
second case, you can start a whole network in a single operation. The commands for controlling the lab start with
a l. The man pages of those commands is available from http://wiki.netkit.org/man/man7/netkit.7.html
6 The minimum command sent to a HTTP server is GET / HTTP/1.0 followed by CRLF and a blank line
7 The list of available language tags can be found at http://www.iana.org/assignments/language-subtag-registry Versions in other formats
are available at http://www.langtag.net/registries.html Additional information about the support of multiple languages in Internet protocols
may be found in rfc5646
8 The image that we use is a custom Netkit filesystem with a recent 64 bits Linux kernel. If you wish to test it on your
own Linux machine, you can install Netkit as explained on http://wiki.netkit.org and download the filesystem and kernel images from
http://cnp3book.info.ucl.ac.be/netkit/netkit-images.zip Note that this archive has a compressed size of more than 1 GByte !
4.8. Configuring DNS and HTTP servers 259

A netkit lab is simply a directory containing at least a configuration file called lab.conf, and one directory for each
virtual machine.
The directory of each device is initially empty, but will be used by Netkit to store their filesystem.
The lab directory can contain optional files. Some labs will include configuration files while others will use scripts
that are executed when the virtual machines boot.
Starting a lab consists thus simply in unpacking the provided archive, going into the lab directory and typing lstart
to start the network.
Note: netkit tools

As the virtual machines run Linux, standard networking tools such as ping(8), tcpdump, netstat etc. are avail-
able.
Another useful hint is that is it possible to share files between the Netkit virtual machines and the local host. Virtual
machines can access to the directory of the lab they belong to. This directory is mounted in their filesystem at the
path /hostlab.
4.8.1 Starting Netkit in the lab
Netkit has been installed in the INGI labs. In order to run the Netkit network emulator, launch the following
commands:
ssh -Y <ingilogin>@permeke.info.ucl.ac.be
export PATH=$PATH:/etinfo/applications/netkit/bin
To launch a single host instance, use the command vstart:

vstart hostname
To launch the DNS lab[#fdnslab]_, use the following commands:

cp -r /etinfo/applications/netkit/dnslab/ $HOME/ # do not forget the trailing /s
lstart -f -d $HOME/dnslab
An interesting option for lstart is this one in order to use Gnome-Terminal instead of Xterm windows:
lstart -f -d $HOME/dnslab -o --xterm=gnome
To stop the lab, please stop all the involved instances by using the command halt inside each virtual machine.
Do not forget to cleanup the virtual disks when you are finished:
rm -f $HOME/dnslab/\*.disk
Note that you can also use lcrash command in order to immediately stop all your instances and remove all
virtual disks.
lcrash -d $HOME/dnslab
4.8.2 Exploring DNS
In this lab, you will experiment with the Domain Name system. Several DNS servers and resolvers are preconfig-
ured in the Netkit which is provided.
Below, you can find a graph where the DNS topology we will use is depicted.

To begin experimentation, start the lab by using the commands explained above. In this lab, the DNS servers are
correctly configured. We ask you to find the IP address of the following fully qualified domain names (FQDN):
pc2.nanoinside.net
dnsorg.org
dnsroot
For this, you should use the dig command whose syntax is :
dig @server -t type FQDN
If no server is specified, dig uses the default resolver that you can find in the configuration file
/etc/resolv.conf.
While doing these requests, observe the packets that are exchanged between the differents DNS server with the
+trace option. Is this what you expected? Sketch the Questions/responses on the figure below.
You have learned that DNS can work in two ways: Forward and Reverse. We will now resolve IPv6 addresses into
their corresponding DNS names. Find the FQDN domain name of the following IPv6 addresses :
2001:db8:ba1:b0a::22
2001:db8:ba1:b0a::2
Again, you should use the dig command but with the -x option.
dig @server -x ipv6
with as parameter the IPv6 address you want to resolve.
4.8.3 Using DNS to access a website
Now that you have played a bit with deployed DNS servers and resolvers, we will now try to add a DNS entry that
will point to some IP address and setup a website that can be reached through the added DNS entry.
We will create the website on pc2 and we will call it helloworld.nanoinside.net. You thus have
to add a DNS entry so that helloworld.nanoinside.net points to the IP address of pc2. See
https://help.ubuntu.com/community/BIND9ServerHowto for a tutorial on how to configure bind9.
Once the DNS entry is set up, it is time to configure the web server. Apache2 is installed. See
http://tuxtweaks.com/2009/07/how-to-configure-apache-linux/ for a tutorial. The final goal is to see Hello world
! when accessing the website:
$ curl -s helloworld.nanoinside.net
Hello world !
The configuration files of apache are located in /etc/apache2/

Enjoy !
4.8. Configuring DNS and HTTP servers 261

4.9 Experimenting with Internet transport protocols
Transport protocols play a very important role on todays Internet. Most applications interact directly with them
to exchange data. In the previous lab, you have performed DNS queries using dig. Each DNS request that you
sent or reply that you received were placed inside UDP segments.
4.9.1 Packet trace analysis
When debugging networking problems or to analyse performance problems, it is sometimes useful to capture the
segments that are exchanged between two hosts and to analyse them.
Several packet trace analysis tools are available, either as commercial or open-source tools. These tools are able
to capture all the packets exchanged on a link. Of course, capturing packets require administrator privileges. They
can also analyse the content of the captured packets and display information about them. The captured packets
can be stored in a file for offline analysis.
tcpdump is probably one of the most well known packet capture software. It is able to both capture packets and
display their content. tcpdump is a text-based tool that can display the value of the most important fields of the
captured packets. Additional information about tcpdump may be found in tcpdump(1).
The text below is an example of the output of tcpdump for all the TCP segments exchanged during the download
of a short webpage over HTTP.
11:58:54.207193 IP6 2001:db8:b0:15:da:b055:0:2.52305 > 2001:db8:be:600d::2.80: Flags [S], seq 5445
11:58:54.207628 IP6 2001:db8:be:600d::2.80 > 2001:db8:b0:15:da:b055:0:2.52305: Flags [S.], seq 190
11:58:54.208344 IP6 2001:db8:b0:15:da:b055:0:2.52305 > 2001:db8:be:600d::2.80: Flags [.], ack 1, w
You can easily recognize in the output above the SYN segment containing the MSS option, the SYN+ACK seg-
ment returned by the server. Note that tcpdump shows relative sequence numbers by default, except in the SYN
segments. Thus, the acknowledgment that you observe in the third segment is an acknowledgement for the SYN
returned by the server.
11:58:54.208360 IP6 2001:db8:b0:15:da:b055:0:2.52305 > 2001:db8:be:600d::2.80: Flags [P.], seq 1:1
11:58:54.208750 IP6 2001:db8:be:600d::2.80 > 2001:db8:b0:15:da:b055:0:2.52305: Flags [.], ack 110,
The two lines above correspond to the request sent by the client and the acknowledgement returned by the server.
Note that the first byte sent by the client has 1 as relative sequence number. In this example, the HTTP request has
a total length of 109 bytes and a second segment is sent by the server.
11:58:54.227126 IP6 2001:db8:be:600d::2.80 > 2001:db8:b0:15:da:b055:0:2.52305: Flags [P.], seq 1:4
11:58:54.227526 IP6 2001:db8:b0:15:da:b055:0:2.52305 > 2001:db8:be:600d::2.80: Flags [.], ack 491,
The TCP connection is then closed by exchanging three segments, the first two having the FIN flag set.
11:58:54.234242 IP6 2001:db8:b0:15:da:b055:0:2.52305 > 2001:db8:be:600d::2.80: Flags [F.], seq 110
11:58:54.234921 IP6 2001:db8:be:600d::2.80 > 2001:db8:b0:15:da:b055:0:2.52305: Flags [F.], seq 491
11:58:54.235245 IP6 2001:db8:b0:15:da:b055:0:2.52305 > 2001:db8:be:600d::2.80: Flags [.], ack 492,
wireshark is more recent than tcpdump. It evolved from the ethereal packet trace analysis software. It can be used
as a text tool like tcpdump. For a TCP connection, wireshark would provide almost the same output as tcpdump.
The main advantage of wireshark is that it also includes a graphical user interface that allows to perform various
types of analysis on a packet trace.
The wireshark window is divided in three parts. The top part of the window is a summary of the first packets from
the trace. By clicking on one of the lines, you can show the detailed content of this packet in the middle part of
the window. The middle of the window allows you to inspect all the fields of the captured packet. The bottom part
of the window is the hexadecimal representation of the packet, with the field selected in the middle window being
highlighted.

Figure 4.8: Wireshark : default window
wireshark is very good at displaying packets, but it also contains several analysis tools that can be very useful.
The first tool is Follow TCP stream. It is part of the Analyze menu and allows you to reassemble and display all
the payload exchanged during a TCP connection. This tool can be useful if you need to analyse for example the
commands exchanged during a SMTP session.
The second tool is the flow graph that is part of the Statistics menu. It provides a time sequence diagram of the
packets exchanged with some comments about the packet contents. See blow for an example.
Figure 4.9: Wireshark : flow graph
The third set of tools are the TCP stream graph tools that are part of the Statistics menu. These tools allow you
to plot various types of information extracted from the segments exchanged during a TCP connection. A first
interesting graph is the sequence number graph that shows the evolution of the sequence number field of the
captured segments with time. This graph can be used to detect graphically retransmissions.
Figure 4.10: Wireshark : sequence number graph
4.9. Experimenting with Internet transport protocols 263

A second interesting graph is the round-trip-time graph that shows the evolution of the round-trip-time in function
of time. This graph can be used to check whether the round-trip-time remains stable or not. Note that from a
packet trace, wireshark can plot two round-trip-time graphs, One for the flow from the client to the server and
the other one. wireshark will plot the round-trip-time graph that corresponds to the selected packet in the top
wireshark window.
Figure 4.11: Wireshark : round-trip-time graph
4.9.2 Experimenting with UDP
For these UDP experiments, you will use a simple lab that contains four hosts.
Client1
router
webserver
Client2
This lab can be downloaded from /netkit/netkit-lab_tcpudp.zip

Two hosts, Client1 and Client2 are clients that you will use to send information to the webserver host through an
intermediate router. Thanks to the router, you will be able to easily observe the packets that are exchanged and
delay or discard some of them to see how the protocol reacts to these events.

Note: Discovering IP addresses

On a netkit lab such as this one, it is sometimes necessary to discover the IP addresses of the different hosts. On
Linux, the IP addresses associated to an interface are configured by using ifconfig(8). The output below
shows the configuration of the interfaces of the router
#ifconfig
eth0 Link encap:Ethernet HWaddr 02:bb:88:79:c7:fb
inet6 addr: fe80::bb:88ff:fe79:c7fb/64 Scope:Link
inet6 addr: 2001:db8:b0:15:da:b055:0:1/96 Scope:Global
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:15 errors:0 dropped:0 overruns:0 frame:0
TX packets:6 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:1052 (1.0 KiB) TX bytes:636 (636.0 B)
Interrupt:5
eth1 Link encap:Ethernet HWaddr 5a:f7:20:7e:4e:9d

inet6 addr: 2001:db8:be:600d::1/64 Scope:Global
inet6 addr: fe80::58f7:20ff:fe7e:4e9d/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX bytes:488 (488.0 B) TX bytes:636 (636.0 B)
Interrupt:5
lo Link encap:Local Loopback

inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:65536 Metric:1
RX bytes:200 (200.0 B) TX bytes:200 (200.0 B)
The output of ifconfig(8) shows that this router has three interfaces. The loopback interface (lo on Linux)
is the default software-based interface of all hosts. The eth0 interface is connected to the two clients while
the eth1 interface is connected to the webserver. The IP address of the router on eth0 (resp. eth1) is
2001:db8:b0:15:da:b055:0::1 (resp. 2001:db8:be:600d::1).
The webserver has been configured as a server that supports the following services :
http over TCP on port 80
echo over both UDP and TCP on port 7
discard over both UDP and TCP on port 9
daytime over both UDP and TCP on port 13
telnet over TCP on port 23
The last three services were popular services installed on all TCP/IP hosts. However, some of them caused security
problems and nowadays they are rarely enabled on real servers. Despite of these security concerns, they are very
useful to perform simple tests with TCP implementations.
echo is a very simple service. When a server receives some information, over UDP or TCP, it simply returns it
to the client.
discard is a kind of blackhole. All the information, sent over UDP or TCP, to a discard server is simply
discarded upon reception.
4.9. Experimenting with Internet transport protocols 265

daytime is a very simple protocol that allows to query the current time on the server. The format of the response
is described in RFC 867.
Several tools allow to send information over UDP and TCP. telnet is very useful to interact with TCP servers.
nc (or netcat) is another tool which can be very useful when debugging network problems. It allows to easily
contact servers over UDP or TCP, but can also be used to create simple but powerful servers from the command
line. Several versions of nc or netcat have been written. See http://en.wikipedia.org/wiki/Netcat for additional
details.
Start tcpdump on router to capture all UDP segments. The tcpdump manpage will show you how to only
capture UDP segments.
1. Using nc on Client1, send data to the discard server running on webserver. Observe the segments that
are exchanged. How does the client select its source port number ? Can you explain why this method of
allocating source ports is used.
2. Using nc on Client2, send data to the echo server running on webserver. Use tcpdump to verify whether
the data returned by the server is the same as the one sent by the client.
Note: Some useful tcpdump options

tcpdump contains many options as described in the tcpdump. Among these, the following ones could prove
useful :
-n : instructs tcpdump to print the addresses of the captured packets and do not try to resolve their names.
To resolve names, tcpdump needs to query the reverse DNS servers and this may interfere with the packet
capture or introduce delays.
-w filename : instructs tcpdump to save the captured packets into a file for further postprocessing.
The packet trace can then be read by using tcpdump -r filename or with Wireshark.
-v, or -vv or even -vvv : use different levels of verbosity when printing information extracted from the
packet
-S forces TCP print the exact sequence/acknowledgements numbers found in the captured segments. By
default, tcpdump prints sequence numbers that are relative to the beginning of the connection
-s snaplen indicates the default size for the captured packets. Some versions of tcpdump use a default
snaplength of 64 or 96 bytes, i.e. they only capture the beginning of the packets. This usually includes all
useful headers. You might want to increase this value to capture long data segments.
4.9.3 Experiments with TCP
nc can also be used to interact with TCP servers. TCP is a complex protocol and a TCP implementation such
the Linux kernel contains a large number of configuration parameters. To ease your understanding of the basic
mechanisms of TCP, we have disabled most of the TCP extensions that are used by a modern TCP implementation
on Client1 and Client2.
Start by using tcpdump on router to capture all the packets sent on the interface attached to webserver
1. Using nc, try to open a TCP connection to a port on webserver where there is no listening server, e.g. port
5. How does tcpdump shows the first segment of the three way handshake. How does the TCP stack on
webserver answers to this segment ? What are the TCP options used ?
2. Using nc, open a TCP connection to the echo port on webserver and send some information. By looking
at the tcpdump output, explain the method used by nc to close the TCP connection.
3. Perform the same experiment as above, but now by using the daytime server.
4. Perform the same experiment as above, but now by using wget to retrieve the homepage on webserver.
Which version of HTTP is used ? How is the TCP connection closed ?
5. The MSS option is the first option that was specified in TCP. It is used during the three-way-handshake to
announce the Maximum Segment Size supported by a host. On Linux, the MSS value is computed from

the maximum packet size of the underlying network. You can change the maximum packet size of the
underlying network (or Maximum Transmission Unit - MTU) by using the command ifconfig(8) :
ifconfig eth0 mtu 1300
This command reduces the MTU of interface eth0 to 1300 bytes. Use tcpdump(8) to observe
whether this change affects the segments sent by the client or by the server when nc(1) is used with
the echo service. To see an impact, you need to send more than several KBytes of data with nc(1).
6. The TCP stack on Client1 was configured to disable all recent TCP extensions, including Window Scale
defined in RFC 1323, Timestamps defined in RFC 1323 and Selective acknowledgements defined in RFC
2018. Enable the Timestamp option by setting the configuration variable net.ipv4.tcp_timestamps
to 1 using sysctl -w and verify with tcpdump(8) that this extension is actually used.
7. The main benefit of TCP is that it can react to delays, losses and packet duplications. In a netkit lab,
there are usually no delay and no losses or duplications. Fortunately, various tools can be used on the
Linux kernel to emulate various network properties. Netem is one of these tools. It can be used on a
router to add delay, losses or duplications when a router sends packets. Using the commands described
in http://www.linuxfoundation.org/collaborate/workgroups/networking/netem, configure the interface be-
tween router and websever with :
1. A fixed delay of 100 milliseconds
2. Packet losses of 10%, 50% and 95%
3. Packet corruption
4. Packet reordering
Using nc(1) with the discard or echo service, observe by using tcpdump(8) how TCP reacts
to these events during :
the three-way handshake
the data transmission phase
the connection release phase
8. Perform the same experiment with the discard service, but this time introduce errors on the link between
router and Client1. Is TCP more affected from errors on the data segments or on the acknowledgements ?
9. Using a configuration with netem that includes a non-zero delay, packet losses and reordering, observe
the benefits of using Selective Acknowledgements. For this, configure netem on the link between router
and webserver and enable the selective acknowledgements with the tcp_sack configuration variable with
sysctl on Client2. Observe the difference between Client1, which does not use the selective acknowl-
edgements and Client2.
4.10 Injecting TCP segments
task_id packetdrill
Packet capture tools like tcpdump and Wireshark are very useful to observe the segments that transport protocols
exchange. They are also very useful to understand and debug network problems as well discuss in subsequent
labs. TCP is a complex protocol that has evolved a lot since its first specification RFC 793. TCP includes a
large number of heuristics that influence the reaction of a TCP implementation to various types of events. A
TCP implementation interacts with the application through the socket API. Recently, several researchers from
Google proposed packetdrill [CCB+2013]. packetdrill is a TCP test suite that was designed to develop unit tests
to verify the correct operation of a TCP implementation. A detailed description of packetdrill in [CCB+2013].
packetdrill uses a syntax which is a mix between the C language and the tcpdump syntax. To understand the
operation of packetdrill, it is useful to study several examples in details. The TCP implementation in the Linux
4.10. Injecting TCP segments 267

kernel supports all the recent TCP extensions to improve its performance. For pedagogical reasons, we disable 9
most of these extensions to use a simple TCP stack.
Let us start with a very simple example that uses packetdrill to open a TCP connection on a server running on the
Linux kernel. A packetdrill script is a sequence of lines that are executed one after the other. Each of these lines
can specify :
packetdrill executes a system call and verifies its return value
packetdrill injects 10 a packet in the instrumented Linux kernel as if it were received from the network
packetdrill compares a packet transmitted by the instrumented Linux kernel with the packet that the script
expects
Each line starts with a timing parameter that indicates at what time the event specified on this line should happen.
packetdrill supports absolute and relative timings. An absolute timing is simply a number that indicates the delay
in seconds between the start of the script and the event. A relative timing is indicated by using + followed by a
number. This number is then the delay in seconds between the previous event and the current line. Additional
informations may be found in [CCB+2013].
For this first example, we will program packetdrill to behave as a client that attempts to create a connection. The
first step is thus to prepare a socket that can be used to accept this connection. This socket can be created by
using the four system calls below.
// create a TCP socket. Since stdin, stdout and stderr are already defined,
// the kernel will assign file descriptor 3 to this socket
// 0 is the absolute time at which the socket is created
0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
// Enable reuse of addresses
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
// binds the created socket to the available addresses
+0 bind(3, ..., ...) = 0
// configure the socket to accept incoming connections
+0 listen(3, 1) = 0
At this point, the socket is ready to accept incoming TCP connections. packetdrill needs to inject a TCP segment
in the instrumented Linux stack. This can be done with the line below.
+0 < S 0:0(0) win 1000 <mss 1000>
packetdrill uses a syntax that is very close to the tcpdump syntax. The +0 timing indicates that the line is executed
immediately after the previous event. The < sign indicates that packetdrill injects a TCP segment and the S
character indicates that the SYN flag must be set. Like tcpdump, packetdrill uses sequence numbers that are
relative to initial sequence number. The three numbers that follow are the sequence number of the first byte of
the payload of the segment (0), the sequence number of the last byte of the payload of the segment (0 after the
semi-column) and the length of the payload (0 between brackets) of the SYN segment. This segment does not
contain a valid acknowledgement but advertises a window of 1000 bytes. All SYN segments must also include the
MSS option. In this case, we set the MSS to 1000 bytes. The next line of the packetdrill script is to verify the reply
sent by the instrumented Linux kernel.
+0 > S. 0:0(0) ack 1 <...>
This TCP segment is sent immediately by the stack. The SYN flag is set and the dot next to the S character
indicates that the ACK flag is also set. The SYN+ACK segment does not contain any data but its acknowledgement
number is set to 1 (relative to the initial sequence number). The packetdrill script does not match the window size
advertised in the TCP segment nor the TCP options (<...>).
The third segment of the three-way handshake is sent by packetdrill after a delay of 0.1 seconds. The connection
is now established and the accept system call will succeed.
9 On Linux, most of the parameters to tune the TCP stack are accessible via sysctl. The
/exercises/packetdrill_scripts/sysctl-cnp3.conf file contains all the sysctl variables that we change to disable
these various TCP extensions.
10 By default, packetdrill uses port 8080 when creating TCP segments. You can thus capture the packets injected by packetdrill and the
responses from the stack by using tcpdump -i any -n port 8080

+.1 < . 1:1(0) ack 1 win 1000

+0 accept(3, ..., ...) = 4
The accept system call returns a new file descriptor, in this case value 4. At this point, packetdrill can write
data on the socket or inject packets.
+0 write(4, ..., 10)=10
+0 > P. 1:11(10) ack 1
+.1 < . 1:1(0) ack 11 win 1000
packetdrill writes 10 bytes of data through the write system call. The stack immediately sends these 10 bytes
inside a segment whose Push flag is set 11 . The payload starts at sequence number 1 and ends at sequence number
10. packetdrill replies by injecting an acknowledgement for the entire data after 100 milliseconds.
packetdrill can also inject data that will be read by the stack as shown by the lines below.
+.1 < P. 1:3(2) ack 11 win 4000
+0 > . 11:11(0) ack 3
+.2 read(4,...,1000)=2
In the example above, packetdrill injects a segment containing two bytes. This segment is acknowledged and after
that the read system call succeeds and reads the available data with a buffer of 1000 bytes. It returns the amount
of read bytes, i.e. 2.
We can now close the connection gracefully. Let us first issue inject a segment with the FIN flag set.
//Packetdrill closes connection gracefully
+0 < F. 3:3(0) ack 11 win 4000
+0 > . 11:11(0) ack 4
packetdrill injects the FIN segment and the instrumented kernel returns an acknowledgement. If packetdrill issues
the close system call, the kernel will send a FIN segment to terminate the connection. packetdrill injects an
acknowledgement to confirm the end of the connection.
+0 close(4) = 0
+0 > F. 11:11(0) ack 4
+0 < . 4:4(0) ack 12 win 4000
The complete packetdrill script is available from /exercises/packetdrill_scripts/connect.pkt

packetdrill can be used to explore in details the operation of the Linux TCP implementation to understand how it
reacts to system calls and the reception of packets.
1. A first interesting point to explore is how TCP reacts with out-of-order segments. Consider the packetdrill
script shown below :
+0 bind(3, ..., ...) = 0
+0 listen(3, 1) = 0
//TCP three-way handshake

+0 < S 0:0(0) win 4000 <mss 1000>
+0 > S. 0:0(0) ack 1 <...>
+.1 < . 1:1(0) ack 1 win 1000
+0 accept(3, ..., ...) = 4
+0 < P. 1:201(200) win 4000

+0 > . 1:1(0) ack 201
+0 < P. 231:251(20) win 4000

+0 > . 1:1(0) ack 201
11 The Push flag is one of the TCP flags defined in RFC 793. TCP stacks usually set this flag when transmitting a segment that empties the
send buffer. This is the reason why we observe this push flag in our example.
4.10. Injecting TCP segments 269

a. At this step of the script, which would be the result of the read system call ?
b. packetdrill_ now issues a FIN segment to indicate that all data has been transmitted.
c. What is the acknowledgement that will be returned by the TCP stack ?
d. At this stage, what are the packets that need to be exchanged to allow the read system call on
2. A second topic that we can explore with packetdrill are the retransmissions when there are packet losses.
TCP uses a mix of go-back-n and selective repeat to retransmit the missing segments. When the retransmis-
sion timer expires, it retransmits one segment due to the congestion control scheme, see below :
+0 bind(3, ..., ...) = 0
+0 listen(3, 1) = 0
+0 < S 0:0(0) win 4000 <mss 1000>

+0 > S. 0:0(0) ack 1 <...>
+.1 < . 1:1(0) ack 1 win 4000
+0 accept(3, ..., ...) = 4
+0 write(4, ..., 1000) = 1000

+0 > P. 1:1001(1000) ack 1
+.1 < . 1:1(0) ack 1001 win 4000
+0 write(4, ..., 2000) = 2000

+0 > . 1001:2001(1000) ack 1
+0 > P. 2001:3001(1000) ack 1
// timeout
+0.3 > . 1001:2001(1000) ack 1

+0.6 > . 1001:2001(1000) ack 1
+1.2 > . 1001:2001(1000) ack 1
Note that TCP applies an exponential backoff to the retransmission timer that doubles after each expiration.
3. The TCP state machine allows two hosts to simultaneously open a TCP connection. In this case, both the
clients and the server start the connection by sending a SYN segment. The following packetdrill script
demonstrates this simultaneous establishment of a connection.
+0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+0 fcntl(3, F_GETFL) = 0x2 (flags O_RDWR)
+0 fcntl(3, F_SETFL, O_RDWR|O_NONBLOCK) = 0
// Establish connection
+0 connect(3, ..., ...) = -1 EINPROGRESS (Operation now in progress)
+0 > S 0:0(0) <...>
+0 < S 0:0(0) win 5792 <mss 1000>
+0 > S. 0:0(0) ack 1 <...>
+0 < . 1:1(0) ack 1 win 5792
+0 < F. 1:1(0) ack 1 win 5792

+0 > . 1:1(0) ack 2
//Kernel closes connection gracefully

+0 close(3) = 0
+0 > F. 1:1(0) ack 2
+0 < . 2:2(0) ack 2 win 5792

4. A TCP connection can be terminated gracefully by exchaning FIN segments. In practice, since these seg-
ments can be exchanged at any time, there are multiple ways to express a graceful connection release in
packetdrill
Consider a TCP connection where no data has been exchanged that needs to be gracefully closed. The connection
starts as follows :
+0 bind(3, ..., ...) = 0
+0 listen(3, 1) = 0
+0 < S 0:0(0) win 1000
+0 > S. 0:0(0) ack 1 <...>
+.1 < . 1:1(0) ack 1 win 1000
+0 accept(3, ..., ...) = 4
Select all the packetdrill_ scripts below that correspond to a correct graceful release of this co
4.11 Experimenting with Internet congestion control
For this lab, you will reuse a slightly modified version of the TCP/UDP lab. The filesystem used for this lab
contains several new software packages. Make sure to start the lab by using lstart from the labs directory. The
archive containing the lab directory can be downloaded from /netkit/netkit-lab_congestion.zip
Your objective with this lab is to better understand the TCP congestion control scheme and how it influences the
performance of TCP.
Client1 [2001:db8:be:feed::11]
router
server [2001:db8:be:b00b::11]
Client2 [2001:db8:be:beef::11]
During this lab, you will use three software packages that are very useful to understand TCP performance in real
IP networks. To simulate network congestion, the bandwidth of the link between the router and the server has
been shaped at 1 Mbps.
iperf (version 3) is a frequently used network performance testing tool. It is often used by network administrators
who need to test the performance of a network between two hosts. To use iperf, you first need to start the server
process by using iperf -s on host server on the lab. The server listens on 5201 for measurements requests
coming from the server. When a measurement starts, the client sends data to the server.
The iperf manpage lists all the options of the server and the client. The most interesting ones are :
-6 forces the utilisation of IPv6
--udp uses UDP for the measurements instead of TCP which is the default
4.11. Experimenting with Internet congestion control 271

--set-mss x sets the TCP MSS at x bytes

--bandwidth b iperf tries to send data at b bits/sec (default for UDP is 1 Mbps, unlimited for TCP)
--verbose provides much more detailed output, including TCP_INFO statistics with TCP on Linux
--window x specifies the TCP window size in bytes on both the client and the server
--reverse forces the measurement to be done from the server to the client
--parallel n iperf uses n parallel flows to perform the measurement
4.11.1 Experiments
The experiments below will allow you to verify experimentally the key factors that influence the performance of
the TCP congestion control scheme. We use a TCP implementation that supports the TCP congestion control
scheme, but not the window scale, timestamp and selective acknowledgement options. You can, of course, enable
these options if you want to experiment with them.
1. The round-trip-time is a key factor that influences the performance of TCP in a network. TCP maintains its
own estimate of the round-trip-time over a connection. Other tools like ping6(8) can be used to measure
the round-trip-time. Start a lab and use ping6(8) to measure the round-trip-time. Then, start iperf sessions
from client1 and client2 and capture the TCP segments on router with tcpdump(8). Analyze
the collected trace with tcptrace or wireshark and observe the evolution of the measured round-trip-time.
2. An important factor to achieve a high goodput with TCP is the window size. Using the --window param-
eter of iperf, compare the performance achieved by a client with a window of 4 KBytes, 16 KBytes and 32
KBytes. Compare this with the bandwidth delay product in the emulated network.
3. The first factor that influences the performance of the TCP congestion control scheme is the round-trip-
time. A TCP connection with a longer round-trip-time will react slower than a connection with a shorter
round-trip time. Start an iperf server and use netem to add a delay of
10 msec on the link between the router and client1
200 msec on the link between the router and client2
Using iperf, verify experimentally that the client1 obtains a higher goodput 12 than client2
1. The TCP congestion control scheme operates on a per TCP connection basis. This implies that a client that
uses several parallel TCP connections should be favored compared to a client that uses a single TCP connec-
tion. Using the --parallel parameter of iperf_, verify that this is indeed
the case in a lab where client1 and client2 have the same round-trip-time.
2. Another factor that influences the performance of TCP is the size of the transmitted segments. Using the
-mss parameter of iperf, change the MSS on client1 and verify whether this reduces its performance
compared to client2 (assuming, of course, that the round-trip-times are the same for the two clients)
3. In the book, we have explained that the TCP goodput was inversely proportional to the square root of the
packet loss ratio. Using netem, simulate different packet loss ratios and verify this formula.
4. In some cases, the goodput obtained by an application also depends on the performance of the application
itself. For example, iperf provides the --file option that reads the data to be sent from a file instead of
from memory to verify whether the disk is the bottleneck 13 . In our emulated lab, this option cannot be
used. However, you can emulate a bottleneck on the client/server by using the --bandwidth parameter
of iperf. Use this option and analyze the captured packet trace to see how you can identify this behavior
from the trace.
12 The goodput is defined as the total number of bytes transmitted by an application divided by the duration of the transfer. This is the
measurement reported by iperf. It should not be confused with the throughput which is measured on the network interfaces and usually
includes all the overheads of the different layers.
13 See http://fasterdata.es.net/performance-testing/network-troubleshooting-tools/iperf-and-iperf3/disk-testing-using-iperf/ for additional
information and an example.

4.11.2 Packet traces
Network administrators often need to debug performance problems in a network by looking only at packet traces.
The three packet traces below were collected in an emulated network that is similar to the one we used in the lab.
Can you analyze the packet traces and :
order them by increasing TCP goodput
identify the performance bottleneck inside each trace
The three traces can be downloaded from :
/exercises/traces/congestion1.pcap
wireshark and tcptrace should help you to analyze these three traces.
4.12 Configuring IPv6
During the previous labs, you have mainly used existing netkit labs to understand transport layer protocols such
as TCP and UDP. This is one usage of netkit, but not the most interesting one. An advantage of netkit compared
to the utilisation of real networking devices is that it allows you to build quickly a network, configure it and test
its operation. The netkit manpages, available from http://wiki.netkit.org/man/man7/netkit.7.html provide many
details on the operation of netkit
netkit is composed of a set of scripts that read configuration files and allow you to run scripts and copy files on
the virtual machines that represent the routers and the switches. All the information about a netkit lab resides in
a directory. The first file in such a directory is the lab.conf file. This file is a text file that describes the network
topology that you want to emulate.
A r13
r1 r12
hostA [2001:db8:be:feed::AA]
r23
B
r2 C r3
hostB [2001:db8:be:beef::BB]
hostC [2001:db8:be:b00b::CC]
The network above, with three hosts and three routers can be represented by the following lab.conf file.
# Header
LAB_DESCRIPTION="Lab to understand the basics of IPv6"
LAB_VERSION=1
LAB_AUTHOR="you"
# List of virtual machines in the lab
machines="hostA hostB hostC r1 r2 r3"
# First host, 64 MBytes of RAM
hostA[M]=64
# The link between hostA and r1 ends on interface 0 on hostA and r1
hostA[0]=A
r1[0]=A
4.12. Configuring IPv6 273

# The link between hostB and r2

hostB[0]=B
r2[0]=B
# The link between hostC and r3
hostC[0]=C
r3[0]=C
# The link between r1 and r2
r1[1]=r12
r2[1]=r12
r1[2]=r13
r3[2]=r13
r2[2]=r23
r3[1]=r23
To build your own lab, you need to first define the subnetworks that compose the lab. Each subnetwork is identified
by a label (A for the subnetwork between hostA and r1). Then, you connect each subnetwork to an interface
on a host or router. Each interface is identified by a unique integer. The first interface has number 0, the second
number 1, ... In the virtual machine, interface 0 corresponds to interface eth0. By default, a virtual machine has
32 Mbytes of virtual memory. If you need more memory, for example to run tcpdump, you can extend it with the
[M] parameter. Other parameters are described in the netkit man pages.
Note: Big brother in your emulated network

In some scenarios, it might be interesting to have a virtual machine that is present on all links inside the emulated
network. This virtual machine could allow you to easily collect packets on any link in the network. For example, if
you would like to add such a monitoring machine in the network above, you could add the following configuration.
# Add nsa to list of machines
machines="hostA hostB hostC r1 r2 r3 nsa"
# NSA can capture all packets
nsa[0]=A
nsa[1]=B
nsa[2]=C
nsa[3]=r12
nsa[4]=r23
nsa[5]=r13
# To analyze all these packets, NSA needs memory
nsa[M]=128
The above lab.conf file defines the network topology and the configuration of the virtual machines. To allow
lstart to start the lab, you need to create a directory for each virtual machine and a startup script. The directory is
named machine where machine is the name chosen for the virtual machine. The startup script is specific for
each virtual machine and is called machine.startup. For example, to create these directories and files, you
could run the following bash script in the directory where the lab.conf file is stored.
for vm in hostA hostB hostC r1 r2 r3
do
mkdir $vm
touch $vm.startup
done
With these files, you can start the lab, but the virtual machines need to be configured before you can exchange
packets. For this, you need first to manually assign IPv6 addresses. On Linux, IP addresses are configured by
using the ifconfig(8) command 14 . This command takes a lot of parameters. A typical usage is the following
:
# enable interface eth0
ifconfig eth0 up
14 You can use the ip command instead of ifconfig(8) or route(8). See the Linux IPv6 Howto for additional information.

# add one IPv6 address on this interface

ifconfig eth0 add 2001:db8:be:feed::AA/64
The first command above activates interface eth0. This command is mandatory before the interface can
send/receive packets. The second command configures the IPv6 address associated to this interface. An IPv6
address is always composed of two parts :
a subnet identifier
a host identifier
The subnet identifier are the high order bits of the IPv6 address. They identify the subnet to which the interface
is connected. In the example above, the subnet is 64 bits long and is 2001:db8:be:feed. All IPv6 addresses
that belong to this subnet can be reached by using the attached subnetwork (i.e. through the datalink layer) without
passing through an intermediate router. The low order bits of the address represent the host identifier (::AA in
the above configuration) inside the subnet.
Your first objective is to configure the IPv6 addresses of the three endhosts based on the information in the figure
above. For this lab, we use only 64 bits subnets. Once an IPv6 address has been configured, you can verify that it
is reachable from the host where it has been configured by issuing a ping6(8) command towards this address
on this host.
Once IPv6 has been configured on the endhosts, you need to configure the IPv6 addresses on the three routers.
Start by configuring the IPv6 addresses on the interfaces eth0 of these routers. For this, select one identifier in
the subnetwork attached to each host and assign one host identifier to the attached router.
The next step is to configure the IPv6 addresses on the links between the routers. Select one subnetwork identifier
starting from 2001:db8: for each inter-router subnetwork and configure the IPv6 addresses on the two attached
routers.
At this point, all IPv6 addresses should have been configured. Make sure that you recorded in a text file all the
commands that you typed. They will be necessary later on to automate the creation of the lab. A good idea would
be to create an /etc/hosts file that contains the mapping between names and all assigned IPv6 addresses. This
file is a text file such as the one below.
##
# Host Database
#
# localhost is used to configure the loopback interface
# when the system is booting. Do not change this entry.
##
127.0.0.1 localhost
255.255.255.255 broadcasthost
::1 localhost
2001:db8:be:beef::AA hostA
2001:db8:be:feed::BB hostB
2001:db8:be:b00b::CC hostC
Add to the file above the IPv6 addresses that you have configured for the routers. Make sure that a different name
is used for each address chosen for a router. This file has the format of the /etc/hosts that provides name to
address mappings when DNS is not in operation (as in this lab since we did not configure any DNS server). You
can copy this file on all virtual hosts to have the list of all adresses on each host.
At this point, you have configured all IPv6 addresses, but there are still no routes. Without routes, packets will not
be forwarded in the network. You need to manually configure the forwarding tables on the hosts and the routers.
On the hosts, configuring the routing table is simple. You simply need to add a default route towards the router
that is directly connected to the host. This can be done by using the route(8) command 1 .
#route -A inet6 add default gw 2001:db8:be:feed::11
The commande above adds a default route, i.e. a route towards ::/0 in the IPv6 routing table (parameter
-A inet6) that points to router (gw or gateway is an old synonym for router) 2001:db8:be:feed::11.
This command can be issued on hostA and router r1 must have been configured with address
2001:db8:be:feed::11. You can now add a default route on all virtual hosts.
4.12. Configuring IPv6 275

You can verify that these routes have been inserted in the routing tables by inspecting them with the ip(8) or
route(8) command.
The next step is to configure the routes on the three routers. For this, you need to first decide the paths that you
want to use and make sure that all routers have a route towards at least the subnets attached to the endhosts.
A first approach is to use shortest path routing. In this case, you need to make sure that :
r1 has a route to 2001:db8:be:beef/64 via r2
r2 has a route to 2001:db8:be:feed/64 via r1
r1 has a route to 2001:db8:be:b00b/64 via r3
When you configure such a route, make sure that you use the correct IPv6 address of the gateway. For example,
to configure the first route above on r1, you might issue a command such as :
/sbin/route -A inet6 add 2001:db8:be:feed/64 gw 2001:db8:bad:1212::22
Assuming that subnet 2001:db8:bad:1212/64 has been used on the link between r1 and r2 and that the
address of r2 on this subnet is 2001:db8:bad:1212::22.
Note: Asymmetric paths

Note that when manually configuring the routes as above, nothing forces you to use symmetric routes. For exam-
ple, the following paths could be configured in the network above.
Feel free to configure such paths in a different lab.
Before testing the lab, make sure that the three routers are configured to forward IPv6 packets, i.e. act as routers.
By default, Linux virtual machines are configured as endhosts and do not forward IPv6 packets. You can change
this configuration by issuing the following sysctl :
sysctl -w net.ipv6.conf.all.forwarding=1
Now, you are ready to test the correct operation of your emulated network. For this, you can use the two most
common network debugging tools :
ping6(8)
traceroute6(8)
ping6(8) sends an ICMP Echo request to a given destination address. If ping6(8) succeeds, this indicates
that both the forward and backward paths operate correctly. traceroute6(8) sends UDP segments with
different HopLimit values to discover the routers on a path towards a given destination address. If ping6(8)
and traceroute6(8) succeed between all pairs of hosts, your network is correctly configured. Otherwise, try
to find the missing or buggy configuration and correct it.
Configuring a network manually takes some time and requires many commands. You probably do not want to
issue all these commands each time you start a netkit lab. netkit helps you to automate the configuration of the
virtual machines with two simple tools.

The first way to automate a lab is the startup script. A startup script is a simple shell script that is launched
automatically by netkit once a virtual machine has booted. This script is named machine.startup where
machine is the name of the virtual machine. It is placed in the directory that contains the lab.conf file.
The second way to automate a lab is by automatically copying files on the virtual machines. For each virtual
machine, you can provide a hierarchy of directories and files that will be copied by netkit when the virtual machine
starts. For example, if you want to place the hosts file automatically as /etc/hosts on virtual machine
hostA, issue the following commands from the directory that contains lab.conf.
mkdir hostA
cd hostA
mkdir etc
cd etc
cp ../../hosts .
These commands create the hostA/etc directory and copy the hosts file that we created earlier at its final location.
This file will be copied as /etc/hosts in the filesystem used by virtual machine hostA.
4.12.1 Exercise
ICMPv6 RFC 4443, the Internet Control Message Protocol, is a key companion to IPv6. ICMPv6 can report to
the sender various types of errors that can occur during the transmission of a packet. traceroute6(8) exploits
one of these messages to determine the path followed by packets towards a given destination.
To demonstrate your understanding of ICMPv6, prepare a lab with a few hosts and routers, prepare and test a
scenario that uses a few commands that would cause a host or router to generate one of the following ICMPv6
error messages :
Destination Unreachable (but not from a router directly connected to the source of the packet)
Code 0 : No route to destination
Code 3 : Address unreachable
Code 4 : Port unreachable
Packet Too Big (with UDP segments and TCP segments)
Time Exceeded message (but only Code 1 - Fragment reassembly time exceeded)
4.13 Inter-domain routing
4.13.1 Exercises
1. Consider the interdomain topology shown in the figure below.
4.13. Inter-domain routing 277

AS1 AS4
$ $
$ AS2 =
AS3
In this network, what are the paths :

from AS1 to AS4
from AS4 to AS2
from AS4 to AS1
2. Consider the interdomain topology shown in the figure below. Assuming, that AS1 advertises prefix
2001:db8:1::/48, AS2 prefix 2001:db8:2::/48, ... compute the routing tables of the different
ASes.

AS2
= AS3
AS1 $
AS4
Are all ASes capable of reaching all the other ASes in this simple Internet ?
3. Consider the interdomain topology shown in the figure below. Assuming, that AS1 advertises prefix
2001:db8:1::/48, AS2 prefix 2001:db8:2::/48, ... compute the routing tables of the different
ASes.

AS4 AS3
$ = = =
$ AS2 AS5 =
AS1
In this internet, some ASes cannot reach all other ASes. Can you fix the problem by adding one
shared-cost peering link or one customer-provider peering link ?
4. Consider the network below in which a stub domain, AS456, is connected to two providers
AS123 and AS789. AS456 advertises its prefix to both its providers. On the other hand,
AS123 advertises 2001:db8:dead::/48 while AS789 advertises 2001:db8:beef::/48
and 2001:db8:dead:cafe::/63. Via which provider will the packets destined to
2001:db8:dead:cafe::1 will be received by AS456 ?

Should AS123 change its configuration ?

5. Consider that the AS stub (AS456) shown in the figure below decides to advertise two /48 prefixes instead
of its allocated /47 prefix.
Via which provider does AS456 receive the packets destined to 2001:db8:caff::bb and
2001:db8:cafe::aa ?
How is the reachability of these addresses affected when link R1-R3 fails ?
Propose a configuration on R1 that achieves the same objective as the one shown in the figure but
also preserves the reachability of all IP addresses inside AS456 if one of AS456s interdomain
links fails.
6. Consider the network shown below. In this network, the metric of each link is set to 1 except link A-B whose
metric is set to 4 in both directions. In this network, there are two paths with the same cost between D and
C. Old routers would randomly select one of these equal cost paths and install it in their forwarding table.
Recent routers are able to use up to N equal cost paths towards the same destination.
On recent routers, a lookup in the forwarding table for a destination address returns a set of outgoing
interfaces. How would you design an algorithm that selects the outgoing interface used for each
packet, knowing that to avoid reordering, all segments of a given TCP connection should follow the
same path ?
7. A traceroute6 towards ipv6.google.com provides the following output :
traceroute6 to ipv6.l.google.com (2a00:1450:4009:800::1001) from 2a02:2788:2c4:16f:5099:ccba:
1 2a02:2788:2c4:16f:a221:b7ff:fed8:aa90 4.777 ms 1.189 ms 1.023 ms
2 2a02:2788:2c0::1 8.746 ms 7.934 ms 10.024 ms
3 2a02:2788:2c0:3::1 10.039 ms 14.967 ms 8.943 ms
4 2a02:2788:ffff:12::1 9.808 ms 11.076 ms 13.658 ms
5 xe-4-2.r00.brslbe01.be.bb.gin.ntt.net 10.043 ms 10.408 ms 9.551 ms
6 ae-10.r02.amstnl02.nl.bb.gin.ntt.net 15.591 ms 18.416 ms 15.665 ms
7 core1.ams.net.google.com 21.259 ms 24.261 ms 20.826 ms
8 2001:4860::1:0:4b3 19.134 ms

2001:4860::1:0:8 22.208 ms
2001:4860::1:0:4b3 19.713 ms
9 2001:4860::8:0:519f 26.712 ms
2001:4860::8:0:51a0 25.313 ms 19.392 ms
10 2001:4860::8:0:5bb8 24.197 ms
2001:4860::8:0:5bb9 25.337 ms 26.264 ms
11 2001:4860::1:0:3067 29.431 ms 31.585 ms 29.260 ms
12 2001:4860:0:1::9 24.806 ms 24.297 ms 23.819 ms
13 lhr14s23-in-x01.1e100.net 29.406 ms 25.729 ms 29.160 ms
Can you explain why at the eighth, ninth and tenth hopes several IPv6 addresses are reported in the
traceroute6 output ?
8. Section 3.3 of RFC 4443 explains two different reasons why an IPv6 enabled device could generate an
ICMPv6 Time Exceeded message. Explain when a router could generate such a message with Code==0
and when a host could generate such a message with Code==1.
9. Section 3.1 of RFC 4443 seven different Codes for the ICMPv6 Destination Unreachable Message. Under
which circumstances would a router generate such an ICMPv6 message with :
Code==0
10. An ICMPv6 error message includes in its message body the beginning of the IPv6 packet that triggered
this error. How many bytes of the original packet must be returned to allow the host to recover the original
source and destination addresses and source and destination ports of the packet that caused the error ?
4.14 IP Address Assignment Methods and Intradomain Routing
9. Consider the network shown below. In this network, the metric of each link is set to 1 except link A-B whose
metric is set to 4 in both directions. In this network, there are two paths with the same cost between D and
C. Old routers would randomly select one of these equal cost paths and install it in their forwarding table.
Recent routers are able to use up to N equal cost paths towards the same destination.
On recent routers, a lookup in the forwarding table for a destination address returns a set of outgoing
interfaces. How would you design an algorithm that selects the outgoing interface used for each
packet, knowing that to avoid reordering, all segments of a given TCP connection should follow the
same path ?

This lab allows you to experiment with the main IPv6 address techniques and illustrates the
operation of two intradomain routing protocols (OSPFv3 and RIP). For this, you will use 4
netkit labs: /netkit/netkit-lab_slaac.zip, /netkit/netkit-lab_dhcpv6.zip,
/netkit/netkit-lab_ospfv3.zip and /netkit/netkit-lab_rip.zip.
You will need to observe into the packets exchanged in the emulated network to understand the operation of the in-
volved protocols and mechanisms. For this, you can use packet capture tools (such as tcpdump or wireshark).
Some tcpdump options (for more details and options, check man tcpdump):
-i allows to choose a specific interface (any for all interfaces).
-v (or -vv) provides a verbose more with more details on the contents of each packet
-s allows to capture the entire packets (not only the first 68 bytes).
Wireshark also allows to capture packets but provides a graphical user interface that is useful to analyse
long packet traces. Wireshark is installed on the Linux machines in the lab and can be downloaded from
http://www.wireshark.org
Note: If you have root access to the machine, you can also use wireshark to visualize traffic captured within
a netkit lab :
When you have launched the lab, you can access to your $HOME directory or the lab directory (in the host machine)
from a netkit virtual machine. These directories are located in /hosthome and /hostlab in netkit . Go into
/hostlab :
cd /hostlab
Now you can launch a tcpdump capture and save the captured traffic on a file, in the hostlab (or
hosthome) directory (option -w). This allows you to start a capture from this file with wireshark.
tcpdump -n -i IF -w aaa.pcap &
where aaa.out is the output file, IF the interface we want to listen on (use any for all interfaces)
and we add the & symbol to run the sniffer in the background, so we can continue to work in the netkit
shell.
You can then launch wireshark on your computer with the input file aaa.out :
wireshark -k -i<(tail -f aaa.pcap)&
Note that this is not possible if you are a non-privileged user in the machine. Therefore, if you are
running netkit in the INGI lab you should use tcpdump to look at on-the-fly traffic.
4.14. IP Address Assignment Methods and Intradomain Routing 283

4.14.1 IPv6 Address Assignment Methods
The SLAAC lab
The first lab (/netkit/netkit-lab_slaac.zip) illustrates the operation of the Stateless Address Auto-
configuration (SLAAC) in IPv6. This mechanism allows hosts connecting to an IPv6 link to dynamically acquire
a link-local address and, in case a router is present, to be assigned a global-scope IPv6. The link-local address en-
ables a host to exchange packets with any other host in the same IPv6 subnet. Global-scope IPv6 address enables
to communicate outside the local subnet.
Start the /netkit/netkit-lab_slaac.zip lab in netkit. This lab consists of 4 virtual machines: 3 hosts
(hostA, hostB and hostC) and router r, connected as shown in the figure below
Router r uses IPv6 address 2001:db8:dead:beef::11/64 and is configured to send periodically Router
Advertisements and reply to Router Sollicitations, via the radvd daemon. You can stop/restart this dae-
mon by using the /etc/init.d/radvd script. The configuration of this daemon is detailed in file
/etc/radvdv.conf. The configuration used for the lab is provided below.
interface eth0 {
AdvSendAdvert on;
MinRtrAdvInterval 3;
MaxRtrAdvInterval 10;
prefix 2001:db8:dead:beef::/64
{
AdvOnLink on;
AdvAutonomous on;
AdvRouterAddr on;
};
};
Unlike previous labs, IPv6 addresses are not configured with ifconfig(8) on the virtual hosts. These addresses
will be obtained dynamically. All host interfaces are down at the startup, so you need to activate them with
ifconfig eth0 up when you are ready to monitor the address assignment process. Use tcpdump on a host
(for instance, hostA) to capture the packets in the link (or wireshark if you have root access to your machine).
Look at the exchanged packets when the interfaces are activated.
1. How are global-scope addresses assigned ? Describe the observed router discovery process.
2. How do hosts acquire their link-local addresses ? How do these link-local addresses look like ? For this
question, you can set an interface down and up to observe the acquisition of a new link-local address without
restarting the lab.

The DHCPv6 lab
The goal of the /netkit/netkit-lab_dhcpv6.zip lab is to observe the operation of DHCPv6. In this
lab, you will work on an emulated network with one router (r1) running a DHCPv6 server and 2 emulated hosts
(pc1 and pc2) running DHCPv6 clients.
Here is the topology of the network:
To use DHCPv6, these virtual machines uses the dibbler daemon.

When the lab is launched, run the dibbler server daemon on the router, then run the client daemon on the hosts
(pc1 and pc2). Startup scripts are provided for each of these daemons.
/etc/init.d/dibbler-server start
/etc/init.d/dibbler-client start
Based on the observed packets, try to figure out:

1. Why the server sends periodically multicast packets ?
2. What kind of packets are exchanged between the server and the clients? (You should identify 4 different
types)
4.14.2 Intradomain routing
We focus on two intradomain routing protocols : RIP and OSPFv3.
The OSPFv3 lab
The Open Shortest Path First (OSPF) protocol is a link-state routing protocol, widely used in todays Internet for
intra-domain routing. OSPFv3 (OSPF for IPv6) is specified in RFC 5340. OSPF control packets are sent directly
over the IP layer, with protocol number 89, so they can be filtered with tcpdump with the command tcpdump
ip[9]==89.
In the /netkit/netkit-lab_ospfv3.zip lab, you will work on a network with routers that use OSPFv3
to compute their routing tables. Here is the topology of the network:
To use OSPF, these routers uses daemons called zebra and ospf6d . Start the lab. Note that, if you try to
ping6 from a router to a non-adjacent one, you will see a destination unreachable. This is because the OSPF
deamon is not launched yet. It is interesting to run tcpdump (in the background) on at least one machine, to
capture the exchanged packets.

You should launch the ospf6d daemon on every router, looking at how every new OSPF-enabled router impacts
in the monitored traffic. Launch first the daemon on bb1. To do that enter the following command line in the bb1
terminal :
/etc/init.d/zebra start
Then launch the daemon on bb2, ...

1. Which packets have you observed in the packet traces ?
2. Were the routing tables modified ? If so, how ?
Launch now the deamon on the others routers while looking at the packets captured.
3. Perform traceroutes from/to different interfaces. Think about the path the traceroute is expected to take, and
the path ICMP replies are expected to take. Does the traceroute6(8) confirm your expectations?
Now we will access the ospf6d daemon. This will help us to see the OSPF link-state database (LSDB), neighbors
and routes.
In netkit, type :
telnet ::1 ospf6d
Note: Reminder: ::1 is the IPv6 address for localhost. ospf6d is the name of the dae-
mon[#fservices]_ that runs OSPF in our router.
The daemon asks for a password. Use the default one, zebra.
Now you can interact with the OSPF daemon and observe its current state and the datastructures that it maintains.
Some useful commands are :
show ipv6 ospf6 database
show ipv6 ospf6 neighbor
show ipv6 ospf6 route
show ipv6 ospf6 interface
exit
4. What is the information returned by each of the above commands ?

5. Is the LSDB the same for all routers ? Should it be ?

Now it is time to play with the topology. You can request the shortest path tree computed by a router (and
monitor how it changes) with the command show ipv6 ospf6 spf tree when connected via telnet to
the ospf6d daemon.
5. Try to disable some link sand observe what is happening. You can disable a link with the ifconfig(8)
command :
ifconfig IF down
where IF is the name of your interface. (If you want to set up an interface down, remember that
manually assigned static IP addresses vanish when the interface is down, so you need to assign
them manually when you set it up again)
6. When you are in the daemon (telnet ::1 ospf6d) , change the link cost and try some traceroute.
Below, the line you should enter in your console:
telnet ::1 ospf6d

zebra
enable
configure terminal
router ospf6
interface IF
ipv6 ospf6 cost X
where IF is the interface and X the new cost.
The RIP lab
The Routing Information Protocol (RIP) is a popular distance-vector protocol. It was widely used for intra-domain
routing in the Internet, before being replaced by link-state protocols such as OSPF or IS-IS. RIPng (for IPv6) is
specified in RFC 2080. RIPng packets are UDP packets and are sent to port 521, so you can filter RIPng packets
with this command: tcpdump udp and port 521.
In the /netkit/netkit-lab_rip.zip lab, you will work on a network composed of routers that use RIPng
to build their routing tables. Here is the topology of the network:
To use RIP, these routers use daemons called zebra and ripngd .
After launching the lab, use tcpdump at the machine sniffer. This machine has 5 interfaces, each of them
connected to a different network link (see the topology description in the file lab.conf and the interfaces
configured in the sniffer.startup file).
First of all, launch the ripngd and zebra daemons. To do that, type on each router the command :
/etc/init.d/zebra start
Observe the evolution of the routing table of one router. After a while, all destinations are available. Why it is not
instantaneous?
1. Check routing tables. Are they updated ?
2. Sniff the RIP packets using tcpdump and observe them. Is this consistent with what you expected ?
Now it is time to modify the topology.
3. Try to make some links fail and observe what is happening. You can do that by stoping one interface on a
router :
ifconfig IF down
where IF is the name of your interface.

4. Observe what is happening. Is the network recovering fast ? Why ?

4.14.3 Assignment
Some networks still rely on manually configured static routes. Static routing provides some flexibility compared
to distance vector routing, but suffers from two important problems. First it does not react to failure. Second,
configuration errors can cause forwarding loops and blackholes.
/netkit/netkit-lab_err1.zip, /netkit/netkit-lab_err2.zip and
/netkit/netkit-lab_err3.zip are three netkit labs describing networks with four routers (r1 to
r4). The routes in each network are configured with static routes, but routing is not correct. You can extract the
netkit labs from the corresponding ZIP files (laberr1.zip for laberr1, and so on).
For each lab, find the errors by using traceroute6(8) or observing the routing tables of the routers. Then,
propose a fix for each problem so that every router can reach successfully (e.g. with ping6) every other router in
the network.
4.15 Inter-domain routing and BGP
4.15.1 Exercises
1. The interdomain topology below is composed of four domains. For this exercise, we assume that there
are no routing policies, i.e. each domain advertise all its best paths to its peers. We focus on the prefix p
advertised by AS1.

AS1
AS3 AS2
AS4
Assume that the BGP sessions are activated in the following order :
AS1-AS2
AS2-AS1
AS3-AS4
AS1-AS3
AS1-AS4
At each step of this activation sequence, show the BGP messages for prefix p that are exchanged and provide the
BGP routing table of the different ASes. Assume that BGP always prefers the short AS-Path.
Once the interdomain network has fully converged, analyze the consequence of a failure of the following BGP
sessions on the routes towards prefix p :
AS1-AS2
AS1-AS4
2. Consider the network below in which a stub domain, AS456, is connected to two providers AS123
and AS789. AS456 advertises its prefix to both its providers. On the other hand, AS123 advertises
2001:db8:dead://48 while AS789 advertises 2001:db8:beef::/48 and 2001:db8:dead:cafe::/63. Via which
provider will the packets destined to 2001:db8:dead:cafe::1 will be received by AS456 ?
Should AS123 change its configuration ?

3. Consider that the AS stub (AS456) shown in the figure below decides to advertise two /48 prefixes instead
of its allocated /47 prefix.
Via which provider does AS456 receive the packets destined to 2001:db8:caff::bb and
2001:db8:cafe::aa ?
How is the reachability of these addresses affected when link R1-R3 fails ?
Propose a configuration on R1 that achieves the same objective as the one shown in the figure but
also preserves the reachability of all IP addresses inside AS456 if one of AS456s interdomain
links fails.
4.15. Inter-domain routing and BGP 289


4. Consider the network shown in the figure below. In this network, each AS contains a single BGP router.
Assume that R1 advertises a single prefix. R1 receives a lot of packets from R9. Without any help from R2,
R9 or R4, how could R1 configure its BGP advertisement such that it receives the packets from R9 via R3 ?
What happens when a link fails ?
5. Consider the network shown in the figure below where R1 advertises a single prefix. In this network, the
link between R1 and R2 is considered as a backup link. It should only be used only when the primary link
(R1-R4) fails.
Can you implement this in R2 ? How ?
Assuming that R1-R2 is a backup link, what are the paths used by all routers to reach R1 ?
Assume now that the link R1-R4 fails. Which BGP messages are exchanged and what are now
the paths used to reach R1 ?
Link R1 - R4 comes back. Which BGP messages are exchanged and what do the paths used to
reach R1 become ?
6. Consider the small Internet show in the figure below.



AS2
AS1 $
= =
AS3
$ =
$ AS5
AS4
In this small Internet, AS5 advertises prefix 2001:db8:5/48. Which BGP messages are exchanged
concerning this prefix when the following events occur (one after the other) :
the link between AS3 and AS5 fails
7. Consider the small Internet show in the figure below.

AS5
AS1 =
AS3 $ AS4
$ =
AS2
1. AS3 advertises prefix 2001:db8:3/48. Compute the BGP routing tables of all ASes towards
this prefix.
2. After a few months, AS3 decides to become a customer of AS4. Explain which BGP messages
are exchanged when the link between AS3 and AS4 becomes active.
3. After a few months, AS3 decides to become a customer of AS5. Explain which BGP messages
are exchanged when the link between AS3 and AS5 becomes active.
4. AS3 is satisfied with its connectivity via AS4 and AS5 and decides to stop the BGP session with
AS2. Explain the BGP messages that are exchanged when this link is disabled.
4.15.2 Netkit BGP lab
This lab, /netkit/netkit-lab_bgp.zip, allows you to experiment with the BGP operation. The simulated
internetwork consists of 8 BGP routers, each one corresponding to a different Autonomous System (AS). They
have the peering relations shown in the figure:
To run BGP, these routers uses daemons called zebra and bgpd .
You can launch the lab using lstart in the lab folder. For monitoring traffic, you can use tcpdump and
wireshark.
In this lab, router r9 announces via BGP an IPv6 prefix throughout the network. You can observe in the routing
tables of the other routers that there are entries for the local prefixes and for the prefix announced by r9, in which
the nexthop is indicated. But routers have no addional information about prefixes from other routers.


You can have more advanced informations about how BGP runs on the routers by accessing to the bgpd daemon
via telnet:
telnet localhost bgpd
The password is zebra, as usual.

In the telnet terminal you can use:
show ipv6 bgp summary
show ip bgp neighbors
show ip community-list
to get more infos. (Note that neighbors and community list queries use the ip command instead of the ipv6
command). See the quagga manual for bgpd for a more complete description of available commands.
You can find the configuration files of the running daemons in the routers folders. For instance, consider router
r1. You can find 3 configuration files in lab/r1/etc/quagga:
The first one is daemons. This file contains informations about which daemon should be started on our
router.
The second one is zebra.conf. This file contains the password that we use to connect to the zebra
daemon when we are on the router. (The password asked when accessing telnet localhost zebra)
The third one is bgpd.conf. This is the configuration file of our bgpd daemon. The following picture
details the meaning of Lets see what all these lines means.
With this in mind, you are able to play with the topology and even create new routers that use BGP. Try some
different configurations, try to change how the filters work and observe what happens. On the original lab, for
instance, you can cause a failure on the AS9-AS1 link (with the command ifconfig ... down). Observe
which BGP messages are exchanged and how the state of router r7 changes. What are your expectations? Are
your observations consistent with what you expected ?
4.16 Local Area Networks: The Spanning Tree Protocol and Virtual
LANs
4.16.1 Exercises
1. Consider the switched network shown in Fig. 1. What is the spanning tree that will be computed by 802.1d
in this network assuming that all links have a unit cost ? Indicate the state of each port.
2. Consider the switched network shown in Fig. 1. In this network, assume that the LAN between switches S3
and S12 fails. How should the switches update their port/address tables after the link failure ?
3. Consider the switched network shown in the figure below. Compute the Spanning Tree of this network.
4.16. Local Area Networks: The Spanning Tree Protocol and Virtual LANs 297

Figure 4.14: Fig. 1. A small network composed of Ethernet switches
S2
S9 S5 S10
S4
4. Many enterprise networks are organized with a set of backbone devices interconnected by using a full mesh
of links as shown in Fig.2. In this network, what are the benefits and drawbacks of using Ethernet switches
and IP routers running OSPF ?
Figure 4.15: Fig. 2. A typical enterprise backbone network
5. In the network depicted in Fig. 3, the host H0 performs a traceroute toward its peer H1 (designated by its
name) through a network composed of switches and routers. Explain precisely the frames, packets, and
segments exchanged since the network was turned on. You may assign addresses if you need to.
Figure 4.16: Fig. 3. Host H0 performs a traceroute towards its peer H1 through a network composed of switches
and routers
6. In the network represented in Fig. 4, can the host H0 communicate with H1 and vice-versa? Explain. Add
whatever you need in the network to allow them to communicate.
Figure 4.17: Fig. 4. Can H0 and H1 communicate ?
7. Consider the network depicted in Fig. 5. Both of the hosts H0 and H1 have two interfaces: one connected
to the switch S0 and the other one to the switch S1. Will the link between S0 and S1 ever be used? If so,
under which assumptions? Provide a comprehensive answer.
8. Most commercial Ethernet switches are able to run the Spanning tree protocol independently on each
VLAN. What are the benefits of using per-VLAN spanning trees ?
4.16.2 Netkit STP lab
In the lab lab_stp (/netkit/netkit-lab_stp.zip), you can explore the behavior of a network with
switches that use the Spanning Tree Protocol (STP). This protocol allows switches to automatically disable ports
on Ethernet switches to ensure that the network does not contain any cycle that could cause frames to loop forever.
Here is the topology of the network:

Figure 4.18: Fig. 5. Will the link between S0 and S1 ever be used?
To use STP, these switches uses brctl, a tool that allows to configure devices as Ethernet bridges and build the
spanning tree.
For this lab, you can use the wireshark or tcpdump as packet sniffers.
To launch the lab you have to go in the directory of the lab and launch it with netkit using command lstart
(you can add the option -f for a quick launch).
You can see that the 6 machines are launched. For the moment, no one of them runs STP. You will run it (via the
brctl command) on two routers and activate wireshark on one of them as explained above. To activate the STP
on one switch type in his terminal:
brctl stp br0 on
ifconfig br0 up
With these two routers you can see what messages are exchanged for the root bridge election. You can see the
state of a bridge by typing :
brctl showstp br0
This command brings information about the designated root of the tree, the root port of the switch and the cost to
the root switch.
Now, you can launch some other switches. By doing that you change the topology. With wireshark you can
observe the packets of the spanning tree protocol that are exchanged. The switches already launched will generate
a topology change notification, then others switches will acknowlegde these changes.
When all switches are launched, you can look at the bridge state of each switches:
brctl showstp br0
You can see wich ports are in blocking state, wich are in forwarding state.
You can also look at the port-station table by entering :
brctl showmacs br0
You can make some links fail and observe what is happening. You can do that by stoping one interface on a switch
or the entire bridge (if=br0) :
ifconfig IF down
where IF is the name of your interface.

CHAPTER 5
Appendices
5.1 Glossary
AIMD Additive Increase, Multiplicative Decrease. A rate adaption algorithm used notably by TCP where a host
additively increases its transmission rate when the network is not congested and multiplicatively decreases
when congested is detected.
anycast a transmission mode where an information is sent from one source to one receiver that belongs to a
specified group
API Application Programming Interface
ARP The Address Resolution Protocol is a protocol used by IPv4 devices to obtain the datalink layer address
that corresponds to an IPv4 address on the local area network. ARP is defined in RFC 826
ARPANET The Advanced Research Project Agency (ARPA) Network is a network that was built by network
scientists in USA with funding from the ARPA of the US Ministry of Defense. ARPANET is considered as
the grandfather of todays Internet.
ascii The American Standard Code for Information Interchange (ASCII) is a character-encoding scheme that
defines a binary representation for characters. The ASCII table contains both printable characters and
control characters. ASCII characters were encoded in 7 bits and only contained the characters required to
write text in English. Other character sets such as Unicode have been developed later to support all written
languages.
ASN.1 The Abstract Syntax Notation One (ASN.1) was designed by ISO and ITU-T. It is a standard and flexible
notation that can be used to describe data structures for representing, encoding, transmitting, and decoding
data between applications. It was designed to be used in the Presentation layer of the OSI reference model
but is now used in other protocols such as SNMP.
ATM Asynchronous Transfer Mode
BGP The Border Gateway Protocol is the interdomain routing protocol used in the global Internet.
BNF A Backus-Naur Form (BNF) is a formal way to describe a language by using syntactic and lexical rules.
BNFs are frequently used to define programming languages, but also to define the messages exchanged
between networked applications. RFC 5234 explains how a BNF must be written to specify an Internet
protocol.
broadcast a transmission mode where is same information is sent to all nodes in the network
CIDR Classless Inter Domain Routing is the current address allocation architecture for IPv4. It was defined in
RFC 1518 and RFC 4632.
dial-up line A synonym for a regular telephone line, i.e. a line that can be used to dial any telephone number.
DNS The Domain Name System is a distributed database that allows to map names on IP addresses.
303
DNS The Domain Name System is defined in RFC 1035

DNS The Domain Name System is a distributed database that can be queried by hosts to map names onto IP
addresses
eBGP An eBGP session is a BGP session between two directly connected routers that belong to two different
Autonomous Systems. Also called an external BGP session.
EGP Exterior Gateway Protocol. Synonym of interdomain routing protocol
EIGRP The Enhanced Interior Gateway Routing Protocol (EIGRP) is a proprietary intradomain routing protocol
that is often used in enterprise networks. EIGRP uses the DUAL algorithm described in [Garcia1993].
frame a frame is the unit of information transfer in the datalink layer
Frame-Relay A wide area networking technology using virtual circuits that is deployed by telecom operators.
ftp The File Transfer Protocol defined in RFC 959 has been the de facto protocol to exchange files over the
Internet before the widespread adoption of HTTP RFC 2616
FTP The File Transfer Protocol is defined in RFC 959
hosts.txt A file that initially contained the list of all Internet hosts with their IPv4 address. As the network grew,
this file was replaced by the DNS, but each host still maintains a small hosts.txt file that can be used when
DNS is not available.
HTML The HyperText Markup Language specifies the structure and the syntax of the documents that are ex-
changed on the world wide web. HTML is maintained by the HTML working group of the W3C
HTTP The HyperText Transport Protocol is defined in RFC 2616
hub A relay operating in the physical layer.
IANA The Internet Assigned Numbers Authority (IANA) is responsible for the coordination of the DNS Root,
IP addressing, and other Internet protocol resources
iBGP An iBGP session is a BGP between two routers belonging to the same Autonomous System. Also called
an internal BGP session.
ICANN The Internet Corporation for Assigned Names and Numbers (ICANN) coordinates the allocation of
domain names, IP addresses and AS numbers as well protocol parameters. It also coordinates the operation
and the evolution of the DNS root name servers.
IETF The Internet Engineering Task Force is a non-profit organisation that develops the standards for the proto-
cols used in the Internet. The IETF mainly covers the transport and network layers. Several application layer
protocols are also standardised within the IETF. The work in the IETF is organised in working groups. Most
of the work is performed by exchanging emails and there are three IETF meetings every year. Participation
is open to anyone. See http://www.ietf.org
IGP Interior Gateway Protocol. Synonym of intradomain routing protocol
IGRP The Interior Gateway Routing Protocol (IGRP) is a proprietary intradomain routing protocol that uses
distance vector. IGRP supports multiple metrics for each route but has been replaced by EIGRP
IMAP The Internet Message Access Protocol is defined in RFC 3501
IMAP The Internet Message Access Protocol (IMAP), defined in RFC 3501, is an application-level protocol
that allows a client to access and manipulate the emails stored on a server. With IMAP, the email messages
remain on the server and are not downloaded on the client.
Internet a public internet, i.e. a network composed of different networks that are running IPv4 or IPv6
internet an internet is an internetwork, i.e. a network composed of different networks. The Internet is a very
popular internetwork, but other internets have been used in the path.
inverse query For DNS servers and resolvers, an inverse query is a query for the domain name that corresponds
to a given IP address.
IP Internet Protocol is the generic term for the network layer protocol in the TCP/IP protocol suite. IPv4 is
widely used today and IPv6 is expected to replace IPv4
304 Chapter 5. Appendices

IPv4 is the version 4 of the Internet Protocol, the connectionless network layer protocol used in most of the
Internet today. IPv4 addresses are encoded as a 32 bits field.
IPv6 is the version 6 of the Internet Protocol, the connectionless network layer protocol which is intended to
replace IPv4 . IPv6 addresses are encoded as a 128 bits field.
IS-IS Intermediate System- Intermediate System. A link-state intradomain routing that was initially defined for
the ISO CLNP protocol but was extended to support IPv4 and IPv6. IS-IS is often used in ISP networks. It
is defined in [ISO10589]
ISN The Initial Sequence Number of a TCP connection is the sequence number chosen by the client ( resp. server)
that is placed in the SYN (resp. SYN+ACK) segment during the establishment of the TCP connection.
ISO The International Standardization Organisation is an agency of the United Nations that is based in Geneva
and develop standards on various topics. Within ISO, country representatives vote to approve or reject stan-
dards. Most of the work on the development of ISO standards is done in expert working groups. Additional
information about ISO may be obtained from http://www.iso.int
ISO The International Standardization Organisation
ISO-3166 An ISO standard that defines codes to represent countries and their subdivisions. See
http://www.iso.org/iso/country_codes.htm
ISP An Internet Service Provider, i.e. a network that provides Internet access to its clients.
ITU The International Telecommunication Union is a United Nations agency whose purpose is to develop stan-
dards for the telecommunication industry. It was initially created to standardise the basic telephone system
but expanded later towards data networks. The work within ITU is mainly done by network specialists from
the telecommunication industry (operators and vendors). See http://www.itu.int for more information
IXP Internet eXchange Point. A location where routers belonging to different domains are attached to the same
Local Area Network to establish peering sessions and exchange packets. See http://www.euro-ix.net/ or
http://en.wikipedia.org/wiki/List_of_Internet_exchange_points_by_size for a partial list of IXPs.
LAN Local Area Network
leased line A telephone line that is permanently available between two endpoints.
MAN Metropolitan Area Network
MIME The Multipurpose Internet Mail Extensions (MIME) defined in RFC 2045 are a set of extensions to the
format of email messages that allow to use non-ASCII characters inside mail messages. A MIME message
can be composed of several different parts each having a different format.
MIME document A MIME document is a document, encoded by using the MIME format.
minicomputer A minicomputer is a multi-user system that was typically used in the 1960s/1970s
to serve departments. See the corresponding wikipedia article for additional information :
http://en.wikipedia.org/wiki/Minicomputer
modem A modem (modulator-demodulator) is a device that encodes (resp. decodes) digital information by mod-
ulating (resp. demodulating) an analog signal. Modems are frequently used to transmit digital information
over telephone lines and radio links. See http://en.wikipedia.org/wiki/Modem for a survey of various types
of modems
MSS A TCP option used by a TCP entity in SYN segments to indicate the Maximum Segment Size that it is able
to receive.
multicast a transmission mode where an information is sent efficiently to all the receivers that belong to a given
group
nameserver A server that implements the DNS protocol and can answer queries for names inside its own domain.
NAT A Network Address Translator is a middlebox that translates IP packets.
NBMA A Non Broadcast Mode Multiple Access Network is a subnetwork that supports multiple hosts/routers
but does not provide an efficient way of sending broadcast frames to all devices attached to the subnetwork.
ATM subnetworks are an example of NBMA networks.
5.1. Glossary 305

network-byte order Internet protocol allow to transport sequences of bytes. These sequences of bytes are suf-
ficient to carry ASCII characters. The network-byte order refers to the Big-Endian encoding for 16 and 32
bits integer. See http://en.wikipedia.org/wiki/Endianness
NFS The Network File System is defined in RFC 1094
NTP The Network Time Protocol is defined in RFC 1305
OSI Open Systems Interconnection. A set of networking standards developed by ISO including the 7 layers OSI
reference model.
OSPF Open Shortest Path First. A link-state intradomain routing protocol that is often used in enterprise and
ISP networks. OSPF is defined in and RFC 2328 and RFC 5340
packet a packet is the unit of information transfer in the network layer
PBL Problem-based learning is a teaching approach that relies on problems.
POP The Post Office Protocol is defined in RFC 1939
POP The Post Office Protocol (POP), defined RFC 1939, is an application-level protocol that allows a client to
download email messages stored on a server.
resolver A server that implements the DNS protocol and can resolve queries. A resolver usually serves a set
of clients (e.g. all hosts in campus or all clients of a given ISP). It sends DNS queries to nameservers
everywhere on behalf of its clients and stores the received answers in its cache. A resolver must know the
IP addresses of the root nameservers.
RIP Routing Information Protocol. An intradomain routing protocol based on distance vectors that is sometimes
used in enterprise networks. RIP is defined in RFC 2453.
RIR Regional Internet Registry. An organisation that manages IP addresses and AS numbers on behalf of IANA.
root nameserver A name server that is responsible for the root of the domain names hierarchy. There are
currently a dozen root nameservers and each DNS resolver See http://www.root-servers.org/ for more infor-
mation about the operation of these root servers.
round-trip-time The round-trip-time (RTT) is the delay between the transmission of a segment and the reception
of the corresponding acknowledgement in a transport protocol.
router A relay operating in the network layer.
RPC Several types of remote procedure calls have been defined. The RPC mechanism defined in RFC 5531 is
used by applications such as NFS
SDU (Service Data Unit) a Service Data Unit is the unit information transferred between applications
segment a segment is the unit of information transfer in the transport layer
SMTP The Simple Mail Transfer Protocol is defined in RFC 821
SNMP The Simple Network Management Protocol is a management protocol defined for TCP/IP networks.
socket A low-level API originally defined on Berkeley Unix to allow programmers to develop clients and servers.
spoofed packet A packet is said to be spoofed when the sender of the packet has used as source address a
different address than its own.
SSH The Secure Shell (SSH) Transport Layer Protocol is defined in RFC 4253
standard query For DNS servers and resolvers, a standard query is a query for a A or a AAAA record. Such a
query typically returns an IP address.
switch A relay operating in the datalink layer.
SYN cookie The SYN cookies is a technique used to compute the initial sequence number (ISN)
TCB The Transmission Control Block is the set of variables that are maintained for each established TCP con-
nection by a TCP implementation.
TCP The Transmission Control Protocol is a protocol of the transport layer in the TCP/IP protocol suite that
provides a reliable bytestream connection-oriented service on top of IP

TCP/IP refers to the TCP and IP protocols

telnet The telnet protocol is defined in RFC 854
TLD A Top-level domain name. There are two types of TLDs. The ccTLD are the TLD that correspond to a two
letters ISO-3166 country code. The gTLD are the generic TLDs that are not assigned to a country.
TLS Transport Layer Security, defined in RFC 5246 is a cryptographic protocol that is used to provide commu-
nication security for Internet applications. This protocol is used on top of the transport service but a detailed
description is outside the scope of this book.
UDP User Datagram Protocol is a protocol of the transport layer in the TCP/IP protocol suite that provides an
unreliable connectionless service that includes a mechanism to detect corruption
unicast a transmission mode where an information is sent from one source to one recipient
vnc A networked application that allows to remotely access a computers Graphical User Interface. See
http://en.wikipedia.org/wiki/Virtual_Network_Computing
W3C The world wide web consortium was created to standardise the protocols and mechanisms used in the
global www. It is thus focused on a subset of the application layer. See http://www.w3c.org
WAN Wide Area Network
X.25 A wide area networking technology using virtual circuits that was deployed by telecom operators.
X11 The XWindow system and the associated protocols are defined in [SG1990]
XML The eXtensible Markup Language (XML) is a flexible text format derived from SGML. It was originally
designed for the electronic publishing industry but is now used by a wide variety of applications that need
to exchange structured data. The XML specifications are maintained by several working groups of the W3C
5.2 Bibliography
Whenever possible, the bibliography includes stable hypertext links to the references cited.
5.3 Indices and tables
genindex
search
5.2. Bibliography 307


Bibliography
[IEEE802.11] LAN/MAN Standards Committee of the IEEE Computer Society. IEEE Standard for Information
Technology - Telecommunications and information exchange between systems - local and metropolitan area
networks - specific requirements - Part 11 : Wireless LAN Medium Access Control (MAC) and Physical
Layer (PHY) Specifications. IEEE, 1999.
[IEEE802.1d] LAN/MAN Standards Committee of the IEEE Computer Society, IEEE Standard for Local and
metropolitan area networks Media Access Control (MAC) Bridges , IEEE Std 802.1DTM-2004, 2004,
[IEEE802.1q] LAN/MAN Standards Committee of the IEEE Computer Society, IEEE Standard for Local and
metropolitan area networks Virtual Bridged Local Area Networks, 2005,
[IEEE802.2] IEEE 802.2-1998 (ISO/IEC 8802-2:1998), IEEE Standard for Information technology
Telecommunications and information exchange between systemsLocal and metropoli-
tan area networksSpecific requirementsPart 2: Logical Link Control. Available from
http://standards.ieee.org/getieee802/802.2.html
[IEEE802.3] LAN/MAN Standards Committee of the IEEE Computer Society. IEEE Standard for Informa-
tion Technology - Telecommunications and information exchange between systems - local and metropoli-
tan area networks - specific requirements - Part 3 : Carrier Sense multiple access with collision
detection (CSMA/CD) access method and physical layer specification. IEEE, 2000. Available from
http://standards.ieee.org/getieee802/802.3.html
[IEEE802.5] LAN/MAN Standards Committee of the IEEE Computer Society. IEEE Standard for Information
technologyTelecommunications and information exchange between systemsLocal and metropolitan area
networksSpecific requirementsPart 5: Token Ring Access Method and Physical Layer Specification. IEEE,
1998. available from http://standards.ieee.org/getieee802
[IEEE802] IEEE, Std 802-2001 : IEEE Standard for Local and Metropolitan Area Networks: Overview and
Architecture, Available from http://standards.ieee.org/getieee802/download/802-2001.pdf
[ACO+2006] Augustin, B., Cuvellier, X., Orgogozo, B., Viger, F., Friedman, T., Latapy, M., Magnien, C., Teix-
eira, R., Avoiding traceroute anomalies with Paris traceroute, Internet Measurement Conference, October
2006, See also http://www.paris-traceroute.net/
[AS2004] Androutsellis-Theotokis, S. and Spinellis, D.. 2004. A survey of peer-to-peer content distribution tech-
nologies. ACM Comput. Surv. 36, 4 (December 2004), 335-371.
[ATLAS2009] Labovitz, C., Iekel-Johnson, S., McPherson, D., Oberheide, J. and Jahanian, F., Internet inter-
domain traffic. In Proceedings of the ACM SIGCOMM 2010 conference on SIGCOMM (SIGCOMM 10).
ACM, New York, NY, USA, 75-86.
309
[AW05] Arlitt, M. and Williamson, C. 2005. An analysis of TCP reset behaviour on the internet. SIGCOMM
Comput. Commun. Rev. 35, 1 (Jan. 2005), 37-44.
[Abramson1970] Abramson, N., THE ALOHA SYSTEM: another alternative for computer communications. In
Proceedings of the November 17-19, 1970, Fall Joint Computer Conference (Houston, Texas, November 17 -
19, 1970). AFIPS 70 (Fall). ACM, New York, NY, 281-285.
[B1989] Berners-Lee, T., Information Management: A Proposal, March 1989
[Baran] Baran, P., On distributed communications series, http://www.rand.org/about/history/baran.list.html,
[BE2007] Biondi, P. and A. Ebalard, IPv6 Routing Header Security, CanSecWest Security Conference 2007,
April 2007.
[BF1995] Bonomi, F. and Fendick, K.W., The rate-based flow control framework for the available bit rate ATM
service, IEEE Network, Mar/Apr 1995, Volume: 9, Issue: 2, pages : 25-39
[BG1992] Bertsekas, D., Gallager, G., Data networks, second edition, Prentice Hall, 1992
[BMO2006] Bhatia, M., Manral, V., Ohara, Y., IS-IS and OSPF Difference Discussions, work in progress, Internet
draft, Jan. 2006
[BMvB2009] Bagnulo, M., Matthews, P., van Beijnum, I., NAT64: Network Address and Protocol Translation
from IPv6 Clients to IPv4 Servers, Internet draft, work in progress, October 2009,
[BNT1997] Beech, W., Nielsen, D., Taylor, J., AX.25 Link Access Protocol for Amateur Packet Radio, version
2.2, Revision: July 1998
[BOP1994] Brakmo, L. S., OMalley, S. W., and Peterson, L. L., TCP Vegas: new techniques for congestion
detection and avoidance. In Proceedings of the Conference on Communications Architectures, Protocols and
Applications (London, United Kingdom, August 31 - September 02, 1994). SIGCOMM 94. ACM, New York,
NY, 24-35.
[Benvenuti2005] Benvenuti, C., Understanding Linux Network Internals, OReilly Media, 2005
[BH2013] Bormann, C., Hoffman, P., Concise Binary Object Representation (CBOR), Internet draft, draft-
bormann-cbor-09, work in progress, 2013
[Bush1945] Bush, V. As we may think The Atlantic Monthly 176 (July 1945), pp. 101108
[Bush1993] Bush, R., FidoNet: technology, tools, and history. Commun. ACM 36, 8 (Aug. 1993), 31-35.
[Bux1989] Bux, W., Token-ring local-area networks and their performance, Proceedings of the IEEE, Vol 77, No
2, p. 238-259, Feb. 1989
[BYL2008] Buford, J., Yu, H., Lua, E.K., P2P Networking and Applications, Morgan Kaufmann, 2008
[CB2003] Cheswick, William R., Bellovin, Steven M., Rubin, Aviel D., Firewalls and internet security - Second
edition - Repelling the Wily Hacker, Addison-Wesley 2003
[CCB+2013] Cardwell, N., Cheng, Y., Brakmo, L., Mathis, M., Raghavan, B., Dukkipati, N., Chu, H., Terzis,
A., and Herbert, T., packetdrill: scriptable network stack testing, from sockets to packets. In Proceedings of
the 2013 USENIX conference on Annual Technical Conference (USENIX ATC13). USENIX Association,
Berkeley, CA, USA, 213-218.
[CD2008] Calvert, K., Donahoo, M., TCP/IP Sockets in Java : Practical Guide for Programmers, Morgan Kauf-
man, 2008
[CJ1989] Chiu, D., Jain, R., Analysis of the Increase and Decrease Algorithms for Congestion Avoidance in
Computer Networks, Computer Networks and ISDN Systems Vol 17, pp 1-14, 1989
[CK74] Cerf, V., Kahn, R., A Protocol for Packet Network Intercommunication, IEEE Transactions on Commu-
nications, May 1974
[CNPI09] Gont, F., Security Assessment of the Transmission Control Protocol (TCP),Security Assessment of the
Transmission Control Protocol (TCP), Internet draft, work in progress, Jan. 2011
[COZ2008] Chi, Y., Oliveira, R., Zhang, L., Cyclops: The Internet AS-level Observatory, ACM SIGCOMM
Computer Communication Review (CCR), October 2008
310 Bibliography
[CSP2009] Carr, B., Sury, O., Palet Martinez, J., Davidson, A., Evans, R., Yilmaz, F., Wijte, Y., IPv6 Address
Allocation and Assignment Policy, RIPE document ripe-481, September 2009
[CT1980] Crane, R., Taft, E., Practical considerations in Ethernet local network design, Proc. of the 13th Hawaii
International Conference on Systems Sciences, Honolulu, January, 1980, pp. 166174
[Cheshire2010] Cheshire, S., Connect-By-Name for IPv6, presentation at IETF 79th, November 2010
[Cheswick1990] Cheswick, B., An Evening with Berferd In Which a Cracker is Lured, Endured, and Studied,
Proc. Winter USENIX Conference, 1990, pp. 163-174
[Clark88] Clark D., The Design Philosophy of the DARPA Internet Protocols, Computer Communications Re-
view 18:4, August 1988, pp. 106-114
[Comer1988] Comer, D., Internetworking with TCP/IP : principles, protocols & architecture, Prentice Hall, 1988
[Comer1991] Comer D., Internetworking With TCP/IP : Design Implementation and Internals, Prentice Hall,
1991
[Cohen1980] Cohen, D., On Holy Wars and a Plea for Peace, IEN 137, April 1980,
http://www.ietf.org/rfc/ien/ien137.txt
[DC2009] Donahoo, M., Calvert, K., TCP/IP Sockets in C: Practical Guide for Programmers , Morgan Kaufman,
2009,
[DIX] Digital, Intel, Xerox, The Ethernet: a local area network: data link layer and physical layer specifications.
SIGCOMM Comput. Commun. Rev. 11, 3 (Jul. 1981), 20-66.
[DKF+2007] Dimitropoulos, X., Krioukov, D., Fomenkov, M., Huffaker, B., Hyun, Y., Claffy, K., Riley, G., AS
Relationships: Inference and Validation, ACM SIGCOMM Computer Communication Review (CCR), Jan.
2007
[DP1981] Dalal, Y. K. and Printis, R. S., 48-bit absolute internet and Ethernet host numbers. In Proceedings of
the Seventh Symposium on Data Communications (Mexico City, Mexico, October 27 - 29, 1981). SIGCOMM
81. ACM, New York, NY, 240-245.
[DRC+2010] Dukkipati, N., Refice, T., Cheng, Y., Chu, J., Herbert, T., Agarwal, A., Jain, A., Sutin, N., An
Argument for Increasing TCPs Initial Congestion Window, ACM SIGCOMM Computer Communications
Review, vol. 40 (2010), pp. 27-33
[Dubuisson2000] 15. Dubuisson, ASN.1 : Communication between Heterogeneous Systems
<http://www.oss.com/asn1/resources/books-whitepapers-pubs/asn1-books.html#dubuisson>, Mor-
gan Kauffman, 2000
[Dunkels2003] Dunkels, A., Full TCP/IP for 8-Bit Architectures. In Proceedings of the first international confer-
ence on mobile applications, systems and services (MOBISYS 2003), San Francisco, May 2003.
[DT2007] Donnet, B. and Friedman, T., Internet Topology Discovery: a Survey. IEEE Communications Surveys
and Tutorials, 9(4):2-15, December 2007
[DYGU2004] Davik, F. Yilmaz, M. Gjessing, S. Uzun, N., IEEE 802.17 resilient packet ring tutorial, IEEE
Communications Magazine, Mar 2004, Vol 42, N 3, p. 112-118
[Dijkstra1959] Dijkstra, E., A Note on Two Problems in Connection with Graphs. Numerische Mathematik,
1:269- 271, 1959
[FDDI] ANSI. Information systems - Fiber Distributed Data Interface (FDDI) - token ring media access control
(MAC). ANSI X3.139-1987 (R1997), 1997
[Fletcher1982] Fletcher, J., An Arithmetic Checksum for Serial Transmissions, Communications, IEEE Transac-
tions on, Jan. 1982, Vol. 30, N. 1, pp. 247-252
[FFEB2005] Francois, P., Filsfils, C., Evans, J., and Bonaventure, O., Achieving sub-second IGP convergence in
large IP networks. SIGCOMM Comput. Commun. Rev. 35, 3 (Jul. 2005), 35-44.
[NGB+1997] Nielsen, H., Gettys, J., Baird-Smith, A., Prudhommeaux, E., Wium Lie, H., and Lilley, C. Network
performance effects of HTTP/1.1, CSS1, and PNG. SIGCOMM Comput. Commun. Rev. 27, 4 (October 1997),
155-166.
Bibliography 311
[FJ1993] Sally Floyd and Van Jacobson. 1993. Random early detection gateways for congestion avoidance.
IEEE/ACM Trans. Netw. 1, 4 (August 1993), 397-413.
[FJ1994] Floyd, S., and Jacobson, V., The Synchronization of Periodic Routing Messages, IEEE/ACM Transac-
tions on Networking, V.2 N.2, p. 122-136, April 1994
[FLM2008] Fuller, V., Lear, E., Meyer, D., Reclassifying 240/4 as usable unicast address space, Internet draft,
March 2008, workin progress
[FRT2002] Fortz, B. Rexford, J. ,Thorup, M., Traffic engineering with traditional IP routing protocols, IEEE
Communication Magazine, October 2002
[FTY99] Theodore Faber, Joe Touch, and Wei Yue, The TIME-WAIT state in TCP and Its Effect on Busy Servers,
Proc. Infocom 99, pp. 1573
[Feldmeier95] Feldmeier, D. C., Fast software implementation of error detection codes. IEEE/ACM Trans. Netw.
3, 6 (Dec. 1995), 640-651.
[GAVE1999] Govindan, R., Alaettinoglu, C., Varadhan, K., Estrin, D., An Architecture for Stable, Analyzable
Internet Routing, IEEE Network Magazine, Vol. 13, No. 1, pp. 2935, January 1999
[GC2000] Grier, D., Campbell, M., A social history of Bitnet and Listserv, 1985-1991, Annals of the History of
Computing, IEEE, Volume 22, Issue 2, Apr-Jun 2000, pp. 32 - 41
[Genilloud1990] Genilloud, G., X.400 MHS: first steps towards an EDI communication standard. SIGCOMM
Comput. Commun. Rev. 20, 2 (Apr. 1990), 72-86.
[GGR2001] Gao, L., Griffin, T., Rexford, J., Inherently safe backup routing with BGP, Proc. IEEE INFOCOM,
April 2001
[GN2011] Gettys, J., Nichols, K., Bufferbloat: dark buffers in the internet. Communications of the ACM 55, no.
1 (2012): 57-65.
[GR2001] Gao, L., Rexford, J., Stable Internet routing without global coordination, IEEE/ACM Transactions on
Networking, December 2001, pp. 681-692
[GSW2002] Griffin, T. G., Shepherd, F. B., and Wilfong, G., The stable paths problem and interdomain routing.
IEEE/ACM Trans. Netw. 10, 2 (Apr. 2002), 232-243
[GW1999] Griffin, T. G. and Wilfong, G., An analysis of BGP convergence properties. SIGCOMM Comput.
Commun. Rev. 29, 4 (Oct. 1999), 277-288.
[GW2002] Griffin, T. and Wilfong, G. T., Analysis of the MED Oscillation Problem in BGP. In Proceedings of the
10th IEEE international Conference on Network Protocols (November 12 - 15, 2002). ICNP. IEEE Computer
Society, Washington, DC, 90-99
[Garcia1993] Garcia-Lunes-Aceves, J., Loop-Free Routing Using Diffusing Computations, IEEE/ACM Transac-
tions on Networking, Vol. 1, No, 1, Feb. 1993
[Gast2002] Gast, M., 802.11 Wireless Networks : The Definitive Guide, OReilly, 2002
[Gill2004] Gill, V. , Lack of Priority Queuing Considered Harmful, ACM Queue, December 2004
[Goralski2009] Goralski, W., The Illustrated network : How TCP/IP works in a modern network, Morgan Kauf-
mann, 2009
[HFPMC2002] Huffaker, B., Fomenkov, M., Plummer, D., Moore, D., Claffy, K., Distance Metrics in the Internet,
Presented at the IEEE International Telecommunications Symposium (ITS) in 2002.
[HRX2008] Ha, S., Rhee, I., and Xu, L., CUBIC: a new TCP-friendly high-speed TCP variant. SIGOPS Oper.
Syst. Rev. 42, 5 (Jul. 2008), 64-74.
[HV2008] Hogg, S. Vyncke, E., IPv6 Security, Cisco Press, 2008
[IMHM2013] Ishihara, K., Mukai, M., Hiromi, R., Mawatari, M., Packet Filter and Route Filter Recommendation
for IPv6 at xSP routers, 2013
[ISO10589] ISO, Intermediate System to Intermediate System intra-domain routeing information exchange pro-
tocol for use in conjunction with the protocol for providing the connectionless-mode network service (ISO
8473) , 2002
312 Bibliography
[Jacobson1988] Jacobson, V., Congestion avoidance and control. In Symposium Proceedings on Communica-
tions Architectures and Protocols (Stanford, California, United States, August 16 - 18, 1988). V. Cerf, Ed.
SIGCOMM 88. ACM, New York, NY, 314-329.
[Jain1990] Jain, R., Congestion control in computer networks : Issues and trends, IEEE Network Magazine, May
1990, pp. 24-30
[JLT2013] Jesup, R., Loreto, S., Tuexen, M., RTCWeb Data Channels, Internet draft, draft-ietf-rtcweb-data-
channel, work in progress, 2013
[JSBM2002] Jung, J., Sit, E., Balakrishnan, H., and Morris, R. 2002. DNS performance and the effectiveness of
caching. IEEE/ACM Trans. Netw. 10, 5 (Oct. 2002), 589-603.
[JSON-RPC2] JSON-RPC Working group, JSON-RPC 2.0 Specification, available on http://www.jsonrpc.org,
2010
[Kerrisk2010] Kerrisk, M., The Linux Programming Interface, No Starch Press, 2010
[KM1995] Kent, C. A. and Mogul, J. C., Fragmentation considered harmful. SIGCOMM Comput. Commun. Rev.
25, 1 (Jan. 1995), 75-87.
[KNT2013] Khlewind, M., Neuner, S., Trammell, B., On the state of ECN and TCP Options on the Internet.
Proceedings of the 14th Passive and Active Measurement conference (PAM 2013), Hong Kong, March 2013
[KP91] Karn, P. and Partridge, C., Improving round-trip time estimates in reliable transport protocols. ACM
Trans. Comput. Syst. 9, 4 (Nov. 1991), 364-373.
[KPD1985] Karn, P., Price, H., Diersing, R., Packet radio in amateur service, IEEE Journal on Selected Areas in
Communications, 3, May, 1985
[KPS2003] Kaufman, C., Perlman, R., and Sommerfeld, B. DoS protection for UDP-based protocols. In Pro-
ceedings of the 10th ACM Conference on Computer and Communications Security (Washington D.C., USA,
October 27 - 30, 2003). CCS 03. ACM, New York, NY, 2-7.
[KR1995] Kung, N.T. Morris, R., Credit-based flow control for ATM networks, IEEE Network, Mar/Apr 1995,
Volume: 9, Issue: 2, pages: 40-48
[KT1975] Kleinrock, L., Tobagi, F., Packet Switching in Radio Channels: Part ICarrier Sense Multiple-Access
Modes and their Throughput-Delay Characteristics, IEEE Transactions on Communications, Vol. COM-23,
No. 12, pp. 1400-1416, December 1975.
[KW2009] Katz, D., Ward, D., Bidirectional Forwarding Detection, RFC 5880, June 2010
[KZ1989] Khanna, A. and Zinky, J. 1989. The revised ARPANET routing metric. SIGCOMM Comput. Commun.
Rev. 19, 4 (Aug. 1989), 45-56.
[KuroseRoss09] Kurose J. and Ross K., Computer networking : a top-down approach featuring the Internet,
Addison-Wesley, 2009
[Licklider1963] Licklider, J., Memorandum For Members and Affiliates of the Intergalactic Computer Network,
1963
[LCCD09] Leiner, B. M., Cerf, V. G., Clark, D. D., Kahn, R. E., Kleinrock, L., Lynch, D. C., Postel, J., Roberts,
L. G., and Wolff, S., A brief history of the internet. SIGCOMM Comput. Commun. Rev. 39, 5 (Oct. 2009),
22-31.
[LCP2005] Eng Keong Lua, Crowcroft, J., Pias, M., Sharma, R., Lim, S., A survey and comparison of peer-to-
peer overlay network schemes, Communications Surveys & Tutorials, IEEE, Volume: 7 , Issue: 2, 2005, pp.
72-93
[LeB2009] Leroy, D. and O. Bonaventure, Preparing network configurations for IPv6 renumbering, International
of Network Management, 2009
[LFJLMT] Leffler, S., Fabry, R., Joy, W., Lapsley, P., Miller, S., Torek, C., An Advanced 4.4BSD Interprocess
Communication Tutorial, 4.4 BSD Programmers Supplementary Documentation
[LNO1996] 20. (a) Lakshman, Arnold Neidhardt, and Teunis J. Ott. 1996. The drop from front strategy in
TCP and in TCP over ATM. INFOCOM96, Vol. 3. IEEE Computer Society, Washington, DC,
USA, 1242-1250.
Bibliography 313
[LSP1982] Lamport, L., Shostak, R., and Pease, M., The Byzantine Generals Problem. ACM Trans. Program.
Lang. Syst. 4, 3 (Jul. 1982), 382-401.
[Leboudec2008] Leboudec, J.-Y., Rate Adaptation Congestion Control and Fairness : a tutorial, Dec. 2008
[Malamud1991] Malamud, C., Analyzing DECnet/OSI phase V, Van Nostrand Reinhold, 1991
[McFadyen1976] McFadyen, J., Systems Network Architecture: An overview, IBM Systems Journal, Vol. 15, N.
1, pp. 4-23, 1976
[McKusick1999] McKusick, M., Twenty Years of Berkeley Unix : From AT&T-Owned to Freely
Redistributable, in Open Sources: Voices from the Open Source Revolution, Oreilly, 1999,
http://oreilly.com/catalog/opensources/book/toc.html
[ML2011] Minei I. and Lucek J. ,MPLS-Enabled Applications: Emerging Developments and New
Technologies <http://www.amazon.com/MPLS-Enabled-Applications-Developments-Technologies-
Communications/dp/0470665459>_ (Wiley Series on Communications Networking & Distributed Systems),
Wiley, 2011
[MRR1979] McQuillan, J. M., Richer, I., and Rosen, E. C., An overview of the new routing algorithm for the
ARPANET. In Proceedings of the Sixth Symposium on Data Communications (Pacific Grove, California,
United States, November 27 - 29, 1979). SIGCOMM 79. ACM, New York, NY, 63-68.
[MRR1980] McQuillan, J.M., Richer, I., Rosen, E., The New Routing Algorithm for the ARPANET Communi-
cations, IEEE Transactions on , vol.28, no.5, pp.711,719, May 1980
[MSMO1997] Mathis, M., Semke, J., Mahdavi, J., and Ott, T. 1997. The macroscopic behavior of the TCP
congestion avoidance algorithm. SIGCOMM Comput. Commun. Rev. 27, 3 (Jul. 1997), 67-82.
[MSV1987] Molle, M., Sohraby, K., Venetsanopoulos, A., Space-Time Models of Asynchronous CSMA Proto-
cols for Local Area Networks, IEEE Journal on Selected Areas in Communications, Volume: 5 Issue: 6, Jul
1987 Page(s): 956 -96
[MUF+2007] Mhlbauer, W., Uhlig, S., Fu, B., Meulle, M., and Maennel, O., In search for an appropriate gran-
ularity to model routing policies. In Proceedings of the 2007 Conference on Applications, Technologies, Ar-
chitectures, and Protocols For Computer Communications (Kyoto, Japan, August 27 - 31, 2007). SIGCOMM
07. ACM, New York, NY, 145-156.
[Malkin1999] Malkin, G., RIP: An Intra-Domain Routing Protocol, Addison Wesley, 1999
[Metcalfe1976] Metcalfe R., Boggs, D., Ethernet: Distributed packet-switching for local computer networks.
Communications of the ACM, 19(7):395404, 1976.
[Mills2006] Mills, D.L., Computer Network Time Synchronization: the Network Time Protocol. CRC Press,
March 2006, 304 pp.
[Miyakawa2008] Miyakawa, S., From IPv4 only To v4/v6 Dual Stack, IETF72 IAB Technical Plenary, July 2008
[Mogul1995] Mogul, J. , The case for persistent-connection HTTP. In Proceedings of the Conference on Applica-
tions, Technologies, Architectures, and Protocols For Computer Communication (Cambridge, Massachusetts,
United States, August 28 - September 01, 1995). D. Oran, Ed. SIGCOMM 95. ACM, New York, NY, 299-
313.
[Moore] Moore, R., Packet switching history, http://rogerdmoore.ca/PS/
[Moy1998] Moy, J., OSPF: Anatomy of an Internet Routing Protocol, Addison Wesley, 1998
[Myers1998] Myers, B. A., A brief history of human-computer interaction technology. interactions 5, 2 (Mar.
1998), 44-54.
[Nelson1965] Nelson, T. H., Complex information processing: a file structure for the complex, the changing and
the indeterminate. In Proceedings of the 1965 20th National Conference (Cleveland, Ohio, United States,
August 24 - 26, 1965). L. Winner, Ed. ACM 65. ACM, New York, NY, 84-100.
[Paxson99] Paxson, V. , End-to-end Internet packet dynamics. SIGCOMM Comput. Commun. Rev. 27, 4 (Oct.
1997), 139-152.
[Perlman1985] Perlman, R., An algorithm for distributed computation of a spanning tree in an extended LAN.
SIGCOMM Comput. Commun. Rev. 15, 4 (Sep. 1985), 44-53.
314 Bibliography
[Perlman2000] Perlman, R., Interconnections : Bridges, routers, switches and internetworking protocols, 2nd
edition, Addison Wesley, 2000
[Perlman2004] Perlman, R., RBridges: Transparent Routing, Proc. IEEE Infocom , March 2004.
[Pouzin1975] Pouzin, L., The CYCLADES Network - Present state and development trends, Symposium on
Computer Networks, 1975 pp 8-13.
[Rago1993] Rago, S., UNIX System V network programming, Addison Wesley, 1993
[RE1989] Rochlis, J. A. and Eichin, M. W., With microscope and tweezers: the worm from MITs perspective.
Commun. ACM 32, 6 (Jun. 1989), 689-698.
[RFC20] Cerf, V., ASCII format for network interchange, RFC 20, Oct. 1969
[RFC768] Postel, J., User Datagram Protocol, RFC 768, Aug. 1980
[RFC789] Rosen, E., Vulnerabilities of network control protocols: An example, RFC 789, July 1981
[RFC791] Postel, J., Internet Protocol, RFC 791, Sep. 1981
[RFC792] Postel, J., Internet Control Message Protocol, RFC 792, Sep. 1981
[RFC793] Postel, J., Transmission Control Protocol, RFC 793, Sept. 1981
[RFC813] Clark, D., Window and Acknowledgement Strategy in TCP, RFC 813, July 1982
[RFC819] Su, Z. and Postel, J., Domain naming convention for Internet user applications, RFC 819, Aug. 1982
[RFC821] Postel, J., Simple Mail Transfer Protocol, RFC 821, Aug. 1982
[RFC822] Crocker, D., Standard for the format of ARPA Internet text messages, :rfc:822, Aug. 1982
[RFC826] Plummer, D., Ethernet Address Resolution Protocol: Or Converting Network Protocol Addresses to
48.bit Ethernet Address for Transmission on Ethernet Hardware, RFC 826, Nov. 1982
[RFC879] Postel, J., TCP maximum segment size and related topics, RFC 879, Nov. 1983
[RFC893] Leffler, S. and Karels, M., Trailer encapsulations, RFC 893, April 1984
[RFC894] Hornig, C., A Standard for the Transmission of IP Datagrams over Ethernet Networks, RFC 894, April
1984
[RFC896] Nagle, J., Congestion Control in IP/TCP Internetworks, RFC 896, Jan. 1984
[RFC952] Harrenstien, K. and Stahl, M. and Feinler, E., DoD Internet host table specification, RFC 952, Oct.
1985
[RFC959] Postel, J. and Reynolds, J., File Transfer Protocol, RFC 959, Oct. 1985
[RFC974] Partridge, C., Mail routing and the domain system, RFC 974, Jan. 1986
[RFC1032] Stahl, M., Domain administrators guide, RFC 1032, Nov. 1987
[RFC1035] Mockapteris, P., Domain names - implementation and specification, RFC 1035, Nov. 1987
[RFC1042] Postel, J. and Reynolds, J., Standard for the transmission of IP datagrams over IEEE 802 networks,
RFC 1042, Feb. 1988
[RFC1055] Romkey, J., Nonstandard for transmission of IP datagrams over serial lines: SLIP, RFC 1055, June
1988
[RFC1071] Braden, R., Borman D. and Partridge, C., Computing the Internet checksum, RFC 1071, Sep. 1988
[RFC1122] Braden, R., Requirements for Internet Hosts - Communication Layers, RFC 1122, Oct. 1989
[RFC1144] Jacobson, V., Compressing TCP/IP Headers for Low-Speed Serial Links, RFC 1144, Feb. 1990
[RFC1149] Waitzman, D., Standard for the transmission of IP datagrams on avian carriers, RFC 1149, Apr.
1990
[RFC1169] Cerf, V. and Mills, K., Explaining the role of GOSIP, RFC 1169, Aug. 1990
[RFC1191] Mogul, J. and Deering, S., Path MTU discovery, RFC 1191, Nov. 1990
Bibliography 315
[RFC1195] Callon, R., Use of OSI IS-IS for routing in TCP/IP and dual environments, RFC 1195, Dec. 1990
[RFC1258] Kantor, B., BSD Rlogin, RFC 1258, Sept. 1991
[RFC1321] Rivest, R., The MD5 Message-Digest Algorithm, RFC 1321, April 1992
[RFC1323] Jacobson, V., Braden R. and Borman, D., TCP Extensions for High Performance, RFC 1323, May
1992
[RFC1347] Callon, R., TCP and UDP with Bigger Addresses (TUBA), A Simple Proposal for Internet Addressing
and Routing, RFC 1347, June 1992
[RFC1518] Rekhter, Y. and Li, T., An Architecture for IP Address Allocation with CIDR, RFC 1518, Sept. 1993
[RFC1519] Fuller V., Li T., Yu J. and Varadhan, K., Classless Inter-Domain Routing (CIDR): an Address Assign-
ment and Aggregation Strategy, RFC 1519, Sept. 1993
[RFC1542] Wimer, W., Clarifications and Extensions for the Bootstrap Protocol, RFC 1542, Oct. 1993
[RFC1548] Simpson, W., The Point-to-Point Protocol (PPP), RFC 1548, Dec. 1993
[RFC1550] Bradner, S. and Mankin, A., IP: Next Generation (IPng) White Paper Solicitation, RFC 1550, Dec.
1993
[RFC1561] Piscitello, D., Use of ISO CLNP in TUBA Environments, RFC 1561, Dec. 1993
[RFC1621] Francis, P., PIP Near-term architecture, RFC 1621, May 1994
[RFC1624] Risjsighani, A., Computation of the Internet Checksum via Incremental Update, RFC 1624, May
1994
[RFC1631] Egevang K. and Francis, P., The IP Network Address Translator (NAT), RFC 1631, May 1994
[RFC1661] Simpson, W., The Point-to-Point Protocol (PPP), RFC 1661, Jul. 1994
[RFC1662] Simpson, W., PPP in HDLC-like Framing, RFC 1662, July 1994
[RFC1710] Hinden, R., Simple Internet Protocol Plus White Paper, RFC 1710, Oct. 1994
[RFC1738] Berners-Lee, T., Masinter, L., and McCahill M., Uniform Resource Locators (URL), RFC 1738, Dec.
1994
[RFC1752] Bradner, S. and Mankin, A., The Recommendation for the IP Next Generation Protocol, RFC 1752,
Jan. 1995
[RFC1812] Baker, F., Requirements for IP Version 4 Routers, RFC 1812, June 1995
[RFC1819] Delgrossi, L., Berger, L., Internet Stream Protocol Version 2 (ST2) Protocol Specification - Version
ST2+, RFC 1819, Aug. 1995
[RFC1889] Schulzrinne H., Casner S., Frederick, R. and Jacobson, V., RTP: A Transport Protocol for Real-Time
Applications, RFC 1889, Jan. 1996
[RFC1896] Resnick P., Walker A., The text/enriched MIME Content-type, RFC 1896, Feb. 1996
[RFC1918] Rekhter Y., Moskowitz B., Karrenberg D., de Groot G. and Lear, E., Address Allocation for Private
Internets, RFC 1918, Feb. 1996
[RFC1939] Myers, J. and Rose, M., Post Office Protocol - Version 3, RFC 1939, May 1996
[RFC1945] Berners-Lee, T., Fielding, R. and Frystyk, H., Hypertext Transfer Protocol HTTP/1.0, RFC 1945,
May 1996
[RFC1948] Bellovin, S., Defending Against Sequence Number Attacks, RFC 1948, May 1996
[RFC1951] Deutsch, P., DEFLATE Compressed Data Format Specification version 1.3, RFC 1951, May 1996
[RFC1981] McCann, J., Deering, S. and Mogul, J., Path MTU Discovery for IP version 6, RFC 1981, Aug. 1996
[RFC2003] Perkins, C., IP Encapsulation within IP, RFC 2003, Oct. 1996
[RFC2018] Mathis, M., Mahdavi, J., Floyd, S. and Romanow, A., TCP Selective Acknowledgment Options, RFC
2018, Oct. 1996
316 Bibliography
[RFC2045] Freed, N. and Borenstein, N., Multipurpose Internet Mail Extensions (MIME) Part One: Format of
Internet Message Bodies, RFC 2045, Nov. 1996
[RFC2046] Freed, N. and Borenstein, N., Multipurpose Internet Mail Extensions (MIME) Part Two: Media Types,
RFC 2046, Nov. 1996
[RFC2050] Hubbard, K. and Kosters, M. and Conrad, D. and Karrenberg, D. and Postel, J., Internet Registry IP
Allocation Guidelines, RFC 2050, Nov. 1996
[RFC2080] Malkin, G. and Minnear, R., RIPng for IPv6, RFC 2080, Jan. 1997
[RFC2082] Baker, F. and Atkinson, R., RIP-2 MD5 Authentication, RFC 2082, Jan. 1997
[RFC2131] Droms, R., Dynamic Host Configuration Protocol, RFC 2131, March 1997
[RFC2140] Touch, J., TCP Control Block Interdependence, RFC 2140, April 1997
[RFC2225] Laubach, M., Halpern, J., Classical IP and ARP over ATM, RFC 2225, April 1998
[RFC2328] Moy, J., OSPF Version 2, RFC 2328, April 1998
[RFC2332] Luciani, J. and Katz, D. and Piscitello, D. and Cole, B. and Doraswamy, N., NBMA Next Hop Reso-
lution Protocol (NHRP), RFC 2332, April 1998
[RFC2364] Gross, G. and Kaycee, M. and Li, A. and Malis, A. and Stephens, J., PPP Over AAL5, RFC 2364,
July 1998
[RFC2368] Hoffman, P. and Masinter, L. and Zawinski, J., The mailto URL scheme, RFC 2368, July 1998
[RFC2453] Malkin, G., RIP Version 2, RFC 2453, Nov. 1998
[RFC2460] Deering S., Hinden, R., Internet Protocol, Version 6 (IPv6) Specification, RFC 2460, Dec. 1998
[RFC2464] Crawford, M., Transmission of IPv6 Packets over Ethernet Networks, RFC 2464, Dec. 1998
[RFC2507] Degermark, M. and Nordgren, B. and Pink, S., IP Header Compression, RFC 2507, Feb. 1999
[RFC2516] Mamakos, L. and Lidl, K. and Evarts, J. and Carrel, J. and Simone, D. and Wheeler, R., A Method for
Transmitting PPP Over Ethernet (PPPoE), RFC 2516, Feb. 1999
[RFC2581] Allman, M. and Paxson, V. and Stevens, W., TCP Congestion Control, RFC 2581, April 1999
[RFC2616] Fielding, R. and Gettys, J. and Mogul, J. and Frystyk, H. and Masinter, L. and Leach, P. and Berners-
Lee, T., Hypertext Transfer Protocol HTTP/1.1, RFC 2616, June 1999
[RFC2617] Franks, J. and Hallam-Baker, P. and Hostetler, J. and Lawrence, S. and Leach, P. and Luotonen, A.
and Stewart, L., HTTP Authentication: Basic and Digest Access Authentication, RFC 2617, June 1999
[RFC2622] Alaettinoglu, C. and Villamizar, C. and Gerich, E. and Kessens, D. and Meyer, D. and Bates, T. and
Karrenberg, D. and Terpstra, M., Routing Policy Specification Language (RPSL), RFC 2622, June 1999
[RFC2675] Tsirtsis, G. and Srisuresh, P., Network Address Translation - Protocol Translation (NAT-PT), RFC
2766, Feb. 2000
[RFC2854] Connolly, D. and Masinter, L., The text/html Media Type, RFC 2854, June 2000
[RFC2965] Kristol, D. and Montulli, L., HTTP State Management Mechanism, RFC 2965, Oct. 2000
[RFC2988] Paxson, V. and Allman, M., Computing TCPs Retransmission Timer, RFC 2988, Nov. 2000
[RFC2991] Thaler, D. and Hopps, C., Multipath Issues in Unicast and Multicast Next-Hop Selection, RFC 2991,
Nov. 2000
[RFC3021] Retana, A. and White, R. and Fuller, V. and McPherson, D., Using 31-Bit Prefixes on IPv4 Point-to-
Point Links, RFC 3021, Dec. 2000
[RFC3022] Srisuresh, P., Egevang, K., Traditional IP Network Address Translator (Traditional NAT), RFC 3022,
Jan. 2001
[RFC3031] Rosen, E. and Viswanathan, A. and Callon, R., Multiprotocol Label Switching Architecture, RFC
3031, Jan. 2001
Bibliography 317
[RFC3168] Ramakrishnan, K. and Floyd, S. and Black, D., The Addition of Explicit Congestion Notification
(ECN) to IP, RFC 3168, Sept. 2001
[RFC3243] Carpenter, B. and Brim, S., Middleboxes: Taxonomy and Issues, RFC 3234, Feb. 2002
[RFC3235] Senie, D., Network Address Translator (NAT)-Friendly Application Design Guidelines, RFC 3235,
Jan. 2002
[RFC3309] Stone, J. and Stewart, R. and Otis, D., Stream Control Transmission Protocol (SCTP) Checksum
Change, RFC 3309, Sept. 2002
[RFC3315] Droms, R. and Bound, J. and Volz, B. and Lemon, T. and Perkins, C. and Carney, M., Dynamic Host
Configuration Protocol for IPv6 (DHCPv6), RFC 3315, July 2003
[RFC3330] IANA, Special-Use IPv4 Addresses, RFC 3330, Sept. 2002
[RFC3360] Floyd, S., Inappropriate TCP Resets Considered Harmful, RFC 3360, Aug. 2002
[RFC3390] Allman, M. and Floyd, S. and Partridge, C., Increasing TCPs Initial Window, RFC 3390, Oct. 2002
[RFC3490] Faltstrom, P. and Hoffman, P. and Costello, A., Internationalizing Domain Names in Applications
(IDNA), RFC 3490, March 2003
[RFC3501] Crispin, M., Internet Message Access Protocol - Version 4 rev1, RFC 3501, March 2003
[RFC3513] Hinden, R. and Deering, S., Internet Protocol Version 6 (IPv6) Addressing Architecture, RFC 3513,
April 2003
[RFC3596] Thomson, S. and Huitema, C. and Ksinant, V. and Souissi, M., DNS Extensions to Support IP Version
6, RFC 3596, October 2003
[RFC3748] Aboba, B. and Blunk, L. and Vollbrecht, J. and Carlson, J. and Levkowetz, H., Extensible Authenti-
cation Protocol (EAP), RFC 3748, June 2004
[RFC3819] Karn, P. and Bormann, C. and Fairhurst, G. and Grossman, D. and Ludwig, R. and Mahdavi, J. and
Montenegro, G. and Touch, J. and Wood, L., Advice for Internet Subnetwork Designers, RFC 3819, July 2004
[RFC3828] Larzon, L-A. and Degermark, M. and Pink, S. and Jonsson, L-E. and Fairhurst, G., The Lightweight
User Datagram Protocol (UDP-Lite), RFC 3828, July 2004
[RFC3927] Cheshire, S. and Aboba, B. and Guttman, E., Dynamic Configuration of IPv4 Link-Local Addresses,
RFC 3927, May 2005
[RFC3931] Lau, J. and Townsley, M. and Goyret, I., Layer Two Tunneling Protocol - Version 3 (L2TPv3), RFC
3931, March 2005
[RFC3971] Arkko, J. and Kempf, J. and Zill, B. and Nikander, P., SEcure Neighbor Discovery (SEND), RFC
3971, March 2005
[RFC3972] Aura, T., Cryptographically Generated Addresses (CGA), RFC 3972, March 2005
[RFC3986] Berners-Lee, T. and Fielding, R. and Masinter, L., Uniform Resource Identifier (URI): Generic Syn-
tax, RFC 3986, January 2005
[RFC4033] Arends, R. and Austein, R. and Larson, M. and Massey, D. and Rose, S., DNS Security Introduction
and Requirements, RFC 4033, March 2005
[RFC4193] Hinden, R. and Haberman, B., Unique Local IPv6 Unicast Addresses, RFC 4193, Oct. 2005
[RFC4251] Ylonen, T. and Lonvick, C., The Secure Shell (SSH) Protocol Architecture, RFC 4251, Jan. 2006
[RFC4264] Griffin, T. and Huston, G., BGP Wedgies, RFC 4264, Nov. 2005
[RFC4271] Rekhter, Y. and Li, T. and Hares, S., A Border Gateway Protocol 4 (BGP-4), RFC 4271, Jan. 2006
[RFC4291] Hinden, R. and Deering, S., IP Version 6 Addressing Architecture, RFC 4291, Feb. 2006
[RFC4301] Kent, S. and Seo, K., Security Architecture for the Internet Protocol, RFC 4301, Dec. 2005
[RFC4302] Kent, S., IP Authentication Header, RFC 4302, Dec. 2005
[RFC4303] Kent, S., IP Encapsulating Security Payload (ESP), RFC 4303, Dec. 2005
318 Bibliography
[RFC4340] Kohler, E. and Handley, M. and Floyd, S., Datagram Congestion Control Protocol (DCCP), RFC
4340, March 2006
[RFC4443] Conta, A. and Deering, S. and Gupta, M., Internet Control Message Protocol (ICMPv6) for the Inter-
net Protocol Version 6 (IPv6) Specification, RFC 4443, March 2006
[RFC4451] McPherson, D. and Gill, V., BGP MULTI_EXIT_DISC (MED) Considerations, RFC 4451, March
2006
[RFC4456] Bates, T. and Chen, E. and Chandra, R., BGP Route Reflection: An Alternative to Full Mesh Internal
BGP (IBGP), RFC 4456, April 2006
[RFC4614] Duke, M. and Braden, R. and Eddy, W. and Blanton, E., A Roadmap for Transmission Control Proto-
col (TCP) Specification Documents, RFC 4614, Oct. 2006
[RFC4648] Josefsson, S., The Base16, Base32, and Base64 Data Encodings, RFC 4648, Oct. 2006
[RFC4822] Atkinson, R. and Fanto, M., RIPv2 Cryptographic Authentication, RFC 4822, Feb. 2007
[RFC4838] Cerf, V. and Burleigh, S. and Hooke, A. and Torgerson, L. and Durst, R. and Scott, K. and Fall, K.
and Weiss, H., Delay-Tolerant Networking Architecture, RFC 4838, April 2007
[RFC4861] Narten, T. and Nordmark, E. and Simpson, W. and Soliman, H.,Neighbor Discovery for IP version 6
(IPv6), RFC 4861, Sept. 2007
[RFC4862] Thomson, S. and Narten, T. and Jinmei, T., IPv6 Stateless Address Autoconfiguration, RFC 4862,
Sept. 2007
[RFC4870] Delany, M., Domain-Based Email Authentication Using Public Keys Advertised in the DNS (Do-
mainKeys), RFC 4870, May 2007
[RFC4871] Allman, E. and Callas, J. and Delany, M. and Libbey, M. and Fenton, J. and Thomas, M., DomainKeys
Identified Mail (DKIM) Signatures, RFC 4871, May 2007
[RFC4941] Narten, T. and Draves, R. and Krishnan, S., Privacy Extensions for Stateless Address Autoconfigura-
tion in IPv6, RFC 4941, Sept. 2007
[RFC4944] Montenegro, G. and Kushalnagar, N. and Hui, J. and Culler, D., Transmission of IPv6 Packets over
IEEE 802.15.4 Networks, RFC 4944, Sept. 2007
[RFC4952] Klensin, J. and Ko, Y., Overview and Framework for Internationalized Email, RFC 4952, July 2007
[RFC4953] Touch, J., Defending TCP Against Spoofing Attacks, RFC 4953, July 2007
[RFC4954] Simeborski, R. and Melnikov, A., SMTP Service Extension for Authentication, RFC 4954, July 2007
[RFC4963] Heffner, J. and Mathis, M. and Chandler, B., IPv4 Reassembly Errors at High Data Rates, RFC 4963,
July 2007
[RFC4966] Aoun, C. and Davies, E., Reasons to Move the Network Address Translator - Protocol Translator
(NAT-PT) to Historic Status, RFC 4966, July 2007
[RFC4987] Eddy, W., TCP SYN Flooding Attacks and Common Mitigations, RFC 4987, Aug. 2007
[RFC5004] Chen, E. and Sangli, S., Avoid BGP Best Path Transitions from One External to Another, RFC 5004,
Sept. 2007
[RFC5065] Traina, P. and McPherson, D. and Scudder, J., Autonomous System Confederations for BGP, RFC
5065, Aug. 2007
[RFC5068] Hutzler, C. and Crocker, D. and Resnick, P. and Allman, E. and Finch, T., Email Submission Opera-
tions: Access and Accountability Requirements, RFC 5068, Nov. 2007
[RFC5072] Varada, S. and Haskins, D. and Allen, E., IP Version 6 over PPP, RFC 5072, Sept. 2007
[RFC5095] Abley, J. and Savola, P. and Neville-Neil, G., Deprecation of Type 0 Routing Headers in IPv6, RFC
5095, Dec. 2007
[RFC5227] Cheshire, S., IPv4 Address Conflict Detection, RFC 5227, July 2008
[RFC5234] Crocker, D. and Overell, P., Augmented BNF for Syntax Specifications: ABNF, RFC 5234, Jan. 2008
Bibliography 319
[RFC5321] Klensin, J., Simple Mail Transfer Protocol, RFC 5321, Oct. 2008
[RFC5322] Resnick, P., Internet Message Format, RFC 5322, Oct. 2008
[RFC5340] Coltun, R. and Ferguson, D. and Moy, J. and Lindem, A., OSPF for IPv6, RFC 5340, July 2008
[RFC5598] Crocker, D., Internet Mail Architecture, RFC 5598, July 2009
[RFC5646] Phillips, A. and Davis, M., Tags for Identifying Languages, RFC 5646, Sept. 2009
[RFC5681] Allman, M. and Paxson, V. and Blanton, E., TCP congestion control, RFC 5681, Sept. 2009
[RFC5735] Cotton, M. and Vegoda, L., Special Use IPv4 Addresses, RFC 5735, January 2010
[RFC5795] Sandlund, K. and Pelletier, G. and Jonsson, L-E., The RObust Header Compression (ROHC) Frame-
work, RFC 5795, March 2010
[RFC6077] Papadimitriou, D. and Welzl, M. and Scharf, M. and Briscoe, B., Open Research Issues in Internet
Congestion Control, RFC 6077, February 2011
[RFC6068] Duerst, M., Masinter, L. and Zawinski, J., The mailto URI Scheme , RFC 6068, October 2010
[RFC6144] Baker, F. and Li, X. and Bao, X. and Yin, K., Framework for IPv4/IPv6 Translation, RFC 6144, April
2011
[RFC6265] Barth, A., HTTP State Management Mechanism, RFC 6265, April 2011
[RFC6274] Gont, F., Security Assessment of the Internet Protocol Version 4, RFC 6274, July 2011
[RG2010] Rhodes, B. and Goerzen, J., Foundations of Python Network Programming: The Comprehensive Guide
to Building Network Applications with Python, Second Edition, Academic Press, 2004
[RJ1995] Ramakrishnan, K. K. and Jain, R., A binary feedback scheme for congestion avoidance in computer
networks with a connectionless network layer. SIGCOMM Comput. Commun. Rev. 25, 1 (Jan. 1995), 138-
156.
[RIB2013] Raiciu, C., Iyengar, J., Bonaventure, O., Recent Advances in Reliable Transport Protocols, in H.
Haddadi, O. Bonaventure (Eds.), Recent Advances in Networking, (2013), pp. 59-106.
[RY1994] Ramakrishnan, K.K. and Henry Yang, The Ethernet Capture Effect: Analysis and Solution, Proceed-
ings of IEEE 19th Conference on Local Computer Networks, MN, Oct. 1994.
[Roberts1975] Roberts, L., ALOHA packet system with and without slots and capture. SIGCOMM Comput.
Commun. Rev. 5, 2 (Apr. 1975), 28-42.
[Ross1989] Ross, F., An overview of FDDI: The fiber distributed data interface, IEEE J. Selected Areas in Comm.,
vol. 7, no. 7, pp. 1043-1051, Sept. 1989
[Russel06] Russell A., Rough Consensus and Running Code and the Internet-OSI Standards War, IEEE Annals
of the History of Computing, July-September 2006
[SAO1990] Sidhu, G., Andrews, R., Oppenheimer, A., Inside AppleTalk, Addison-Wesley, 1990
[SARK2002] Subramanian, L., Agarwal, S., Rexford, J., Katz, R.. Characterizing the Internet hierarchy from
multiple vantage points. In IEEE INFOCOM, 2002
[Sechrest] Sechrest, S., An Introductory 4.4BSD Interprocess Communication Tutorial, 4.4BSD Programmers
Supplementary Documentation
[SG1990] Scheifler, R., Gettys, J., X Window System: The Complete Reference to Xlib, X Protocol, ICCCM,
XLFD, X Version 11, Release 4, Digital Press
[SGP98] Stone, J., Greenwald, M., Partridge, C., and Hughes, J., Performance of checksums and CRCs over real
data. IEEE/ACM Trans. Netw. 6, 5 (Oct. 1998), 529-543.
[SH1980] Shoch, J. F. and Hupp, J. A., Measured performance of an Ethernet local network. Commun. ACM 23,
12 (Dec. 1980), 711-721.
[SH2004] Senapathi, S., Hernandez, R., Introduction to TCP Offload Engines, March 2004
320 Bibliography
[SMKKB2001] Stoica, I., Morris, R., Karger, D., Kaashoek, F., and Balakrishnan, H., Chord: A scalable peer-
to-peer lookup service for internet applications. In Proceedings of the 2001 conference on Applications, tech-
nologies, architectures, and protocols for computer communications (SIGCOMM 01). ACM, New York, NY,
USA, 149-160
[SMM1998] Semke, J., Mahdavi, J., and Mathis, M., Automatic TCP buffer tuning. SIGCOMM Comput. Com-
mun. Rev. 28, 4 (Oct. 1998), 315-323.
[SPMR09] Stigge, M., Plotz, H., Muller, W., Redlich, J., Reversing CRC - Theory and Practice. Berlin: Humboldt
University Berlin. pp. 24.
[STBT2009] Sridharan, M., Tan, K., Bansal, D., Thaler, D., Compound TCP: A New TCP Congestion Control
for High-Speed and Long Distance Networks, Internet draft, work in progress, April 2009
[STD2013] Stewart, R., Tuexen, M., Dong, X., ECN for Stream Control Transmission Protocol (SCTP), Internet
draft, draft-stewart-tsvwg-sctpecn-04, April 2013, work in progress
[Seifert2008] Seifert, R., Edwards, J., The All-New Switch Book : The complete guide to LAN switching tech-
nology, Wiley, 2008
[Selinger] Selinger, P., MD5 collision demo, http://www.mscs.dal.ca/~selinger/md5collision/
[SFR2004] Stevens R. and Fenner, and Rudoff, A., UNIX Network Programming: The sockets networking API,
Addison Wesley, 2004
[Sklower89] Sklower, K. 1989. Improving the efficiency of the OSI checksum calculation. SIGCOMM Comput.
Commun. Rev. 19, 5 (Oct. 1989), 32-43.
[SMASU2012] Sarrar, N., Maier, G., Ager, B., Sommer, R. and Uhlig, S., Investigating IPv6 traffic, Passive and
Active Measurements, Lecture Notes in Computer Science vol 7192, 2012, pp.11-20
[SMM98] Semke, J., Mahdavi, J., and Mathis, M., Automatic TCP buffer tuning. SIGCOMM Comput. Commun.
Rev. 28, 4 (Oct. 1998), 315-323.
[Stevens1994] Stevens, R., TCP/IP Illustrated : the Protocols, Addison-Wesley, 1994
[Stevens1998] Stevens, R., UNIX Network Programming, Volume 1, Second Edition: Networking APIs: Sockets
and XTI, Prentice Hall, 1998
[Stewart1998] Stewart, J., BGP4: Inter-Domain Routing In The Internet, Addison-Wesley, 1998
[Stoll1988] Stoll, C., Stalking the wily hacker, Commun. ACM 31, 5 (May. 1988), 484-497.
[SV1995] 13. Shreedhar and G. Varghese. Efficient fair queueing using deficit round robin SIGCOMM Comput.
Commun. Rev. 25, 4 (October 1995), 231-242.
[TE1993] Tsuchiya, P. F. and Eng, T., Extending the IP internet through address reuse. SIGCOMM Comput.
Commun. Rev. 23, 1 (Jan. 1993), 16-33.
[Thomborson1992] Thomborson, C., The V.42bis Standard for Data-Compressing Modems, IEEE Micro,
September/October 1992 (vol. 12 no. 5), pp. 41-53
[Unicode] The Unicode Consortium. The Unicode Standard, Version 5.0.0, defined by: The Unicode Standard,
Version 5.0 (Boston, MA, Addison-Wesley, 2007
[VPD2004] Vasseur, J., Pickavet, M., and Demeester, P., Network Recovery: Protection and Restoration of Opti-
cal, SONET-SDH, IP, and MPLS. Morgan Kaufmann Publishers Inc., 2004
[Varghese2005] Varghese, G., Network Algorithmics: An Interdisciplinary Approach to Designing Fast Net-
worked Devices, Morgan Kaufmann, 2005
[Vyncke2007] Vyncke, E., Paggen, C., LAN Switch Security: What Hackers Know About Your Switches, Cisco
Press, 2007
[WB2008] Waserman, M., Baker, F., IPv6-to-IPv6 Network Address Translation (NAT66), Internet draft, Novem-
ber 2008, http://tools.ietf.org/html/draft-mrw-behave-nat66-02
[WMH2008] Wilson, P., Michaelson, G., Huston, G., Redesignation of 240/4 from Future Use to Private
Use, Internet draft, September 2008, work in progress, http://tools.ietf.org/html/draft-wilson-class-e-02
Bibliography 321
[WMS2004] White, R., Mc Pherson, D., Srihari, S., Practical BGP, Addison-Wesley, 2004
[Watson1981] Watson, R., Timer-Based Mechanisms in Reliable Transport Protocol Connection Management.
Computer Networks 5: 47-56 (1981)
[Williams1993] Williams, R. A painless guide to CRC error detection algorithms, August 1993, unpublished
manuscript, http://www.ross.net/crc/download/crc_v3.txt
[Winston2003] Winston, G., NetBIOS Specification, 2003
[WY2011] Wing, D. and Yourtchenko, A., Happy Eyeballs: Success with Dual-Stack Hosts, Internet draft, work
in progress, July 2011, http://tools.ietf.org/html/draft-ietf-v6ops-happy-eyeballs-03
[X200] ITU-T, recommendation X.200, Open Systems Interconnection - Model and Notation, 1994
[X224] ITU-T, recommendation X.224, Information technology - Open Systems Interconnection - Protocol for
providing the connection-mode transport service, 1995
[XNS] Xerox, Xerox Network Systems Architecture, XNSG058504, 1985
[Zimmermann80] Zimmermann, H., OSI Reference Model - The ISO Model of Architecture for Open Systems
Interconnection, IEEE Transactions on Communications, vol. 28, no. 4, April 1980, pp. 425 - 432.
322 Bibliography
Index
Symbols BGP KEEPALIVE, 199

::, 171 BGP local-preference, 203
::1, 171 BGP nexthop, 202
100BaseTX, 213 BGP NOTIFICATION, 199
10Base2, 213 BGP OPEN, 199
10Base5, 213 BGP peer, 198
10BaseT, 213 BGP RIB, 200
802.11 frame format, 222 BGP UPDATE, 199
802.5 data frame, 95 binary exponential back-off (CSMA/CD), 88
802.5 token frame, 94 bit stuffing, 9
black hole, 28
A BNF, 303
abrupt connection release, 58, 68 Border Gateway Protocol, 197
Additive Increase Multiplicative Decrease (AIMD), 99 bridge, 214
address, 27 broadcast, 303
address learning, 215 BSS, 222
Address resolution problem, 185
addressing, 68 C
adhoc network, 222 Carrier Sense Multiple Access, 84
AF_INET, 237 Carrier Sense Multiple Access with Collision Avoid-
AF_INET6, 237 ance, 89
AF_UNSPEC, 238 Carrier Sense Multiple Access with Collision Detec-
AIMD, 303 tion, 85
ALOHA, 83 character stuffing, 10
Alternating Bit Protocol, 16 Checksum computation, 137
anycast, 303 CIDR, 303
API, 303 Clear To Send, 92
Application layer, 107 collision, 82
ARP, 303 collision detection, 86
ARPANET, 303 collision domain, 213
ascii, 303 congestion collapse, 77, 97
ASN.1, 303 congestion control, 73
ATM, 303 congestion window, 103
connection establishment, 56
B connectionless service, 54
Base64 encoding, 118 control plane, 29
Basic Service Set (BSS), 222 count to infinity, 45
beacon frame (802.11), 225 CSMA, 84
BGP, 197, 303 CSMA (non-persistent), 85
BGP Adj-RIB-In, 200 CSMA (persistent), 85
BGP Adj-RIB-Out, 200 CSMA/CA, 89
BGP decision process, 203 CSMA/CD, 85
323
CTS, 92 go-back-n, 19
CTS frame (802.11), 224 graceful connection release, 58, 68
cumulative acknowledgements, 19
customer-provider peering relationship, 196 H
head-of-line blocking, 154
D Hello message, 47
data plane, 28 hidden station problem, 92
Datalink layer, 9, 105 hop-by-hop forwarding, 27
delayed acknowledgements, 149 hosts.txt, 70, 304
Denial of Service, 142 HTML, 304
DHCPv6, 186, 188 HTTP, 304
dial-up line, 303 hub, 304
DIFS, 90
Distance vector, 42 I
Distributed Coordination Function Inter Frame Space, IANA, 304
90 iBGP, 304
DNS, 303, 304 ICANN, 304
DNS message format, 110 IETF, 304
Duplicate Address Detection, 186 IGP, 304
IGRP, 304
E IMAP, 304
EAP, 209 import policy, 197
eBGP, 304 independent network, 222
EGP, 304 infrastructure network, 222
EIFS, 90 interdomain routing policy, 197
EIGRP, 304 Internet, 304
electrical cable, 5 internet, 304
email message format, 114 inverse query, 304
Ending Delimiter (Token Ring), 94 IP, 304
Ethernet bridge, 214 IPv4, 305
Ethernet DIX frame format, 210 IPv4 fragmentation and reassembly, 176
Ethernet hub, 213 IPv6, 305
Ethernet switch, 214 IPv6 fragmentation, 175
Ethernet Type field, 210 IPv6 Renumbering, 187
EtherType, 210 IS-IS, 305
exponential backoff, 149 ISN, 305
export policy, 197 ISO, 305
Extended Inter Frame Space, 90 ISO-3166, 305
Extensible Authentication Protocol, 209 ISP, 305
ITU, 305
F IXP, 305
Fairness, 97
Fast Ethernet, 213 J
FDM, 82 jamming, 86
FECN, 80 jumbogram, 175
Five layers reference model, 105
Forward Explicit Congestion Notification, 80 L
forwarding loop, 28 label switching, 40
forwarding table, 27 LAN, 305
frame, 9, 105, 304 large window, 146
Frame-Relay, 304 leased line, 305
framing, 9 Link Local address, 171
Frequency Division Multiplexing, 82 link-local IPv6 address, 184
FTP, 304 link-state routing, 46
ftp, 304 LLC, 212
Logical Link Control (LLC), 212
G
getaddrinfo, 238
324 Index
M Path MTU discovery, 182

MAC address learning, 215 PBL, 306
MAC address table (Ethernet switch), 214 peer-to-peer, 53
MAN, 305 persistence timer, 25, 66
Manchester encoding, 8 persistent CSMA, 85
max-min fairness, 76, 97 Physical layer, 8, 105
Maximum Segment Lifetime (MSL), 62 physical layer, 105
maximum segment lifetime (MSL), 25, 68 piggybacking, 25
Maximum Segment Size, 143 ping6, 182
Maximum Transmission Unit, 176 Point-to-Point Protocol, 208
medium access control, 73 POP, 306
message-mode data transfer, 57 port-address table, 29
MIME, 305 portmapper, 135
MIME document, 305 Post Office Protocol, 122
minicomputer, 305 PPP, 208
modem, 305 Provider Aggregatable address, 169
Monitor station, 95 Provider Independent address, 169
monomode optical fiber, 5
MSS, 143, 305 R
MTU, 176 record route, 36
multicast, 305 Reference models, 104
multimode optical fiber, 5 reliable connectionless service, 54
Remote Procedure Call, 59
N Request To Send, 92
Nagle algorithm, 145 request-response service, 59
nameserver, 305 resolver, 306
naming, 68 RFC
NAT, 305 RFC 1032, 70, 315
NBMA, 42, 167, 305 RFC 1035, 71, 110, 112, 304, 315
NDP, 185 RFC 1042, 225, 315
Neighbor Discovery Protocol, 185 RFC 1055, 208, 315
Neighbor Solicitation message, 185 RFC 1071, 137, 228, 315
Neighbour Discovery Protocol, 184 RFC 1094, 306
network congestion, 77 RFC 1122, 107, 139, 140, 144, 150, 210, 315
Network Information Center, 70 RFC 1144, 208, 315
Network layer, 106 RFC 1149, 68, 315
network-byte order, 306 RFC 1169, 315
NFS, 306 RFC 1191, 315
Non-Broadcast Multi-Access Networks, 42, 167 RFC 1195, 191, 316
non-persistent CSMA, 85 RFC 1258, 141, 316
NTP, 306 RFC 1305, 306
RFC 1321, 229, 316
O RFC 1323, 144, 146148, 152, 267, 316
Open Shortest Path First, 191 RFC 1347, 167, 168, 316
optical fiber, 5 RFC 1350, 242
Organisation Unique Identifier, 210 RFC 1518, 303, 316
OSI, 306 RFC 1519, 170, 316
OSI reference model, 107 RFC 1542, 316
OSPF, 191, 306 RFC 1548, 208, 316
OSPF area, 191 RFC 1550, 167, 316
OSPF Designated Router, 193 RFC 1561, 167, 316
OUI, 210 RFC 1621, 167, 316
RFC 1624, 316
P RFC 1631, 167, 316
packet, 106, 306 RFC 1661, 26, 166, 316
packet discard mechanism, 79 RFC 1662, 208, 316
packet radio, 83 RFC 1710, 167, 168, 173, 316
packet size distribution, 146 RFC 1738, 125, 131, 316
Index 325
RFC 1752, 167, 316 RFC 3330, 318

RFC 1812, 100, 316 RFC 3360, 153, 318
RFC 1819, 167, 316 RFC 3390, 161, 318
RFC 1831, 133, 135 RFC 3490, 318
RFC 1832, 133, 134 RFC 3501, 122, 304, 318
RFC 1833, 135 RFC 3513, 168, 318
RFC 1889, 316 RFC 3540, 162
RFC 1896, 117, 316 RFC 3550, 136
RFC 1918, 171, 316 RFC 3596, 113, 318
RFC 1939, 122, 123, 258, 306, 316 RFC 3708, 158
RFC 1945, 127, 128, 316 RFC 3736, 189
RFC 1948, 141, 316 RFC 3748, 209, 318
RFC 1951, 178, 316 RFC 3758, 155
RFC 1981, 180, 182, 316 RFC 3782, 165
RFC 20, 10, 52, 118, 315 RFC 3819, 166, 318
RFC 2003, 316 RFC 3828, 318
RFC 2018, 151, 267, 316 RFC 3927, 171, 318
RFC 2045, 116, 118, 305, 317 RFC 3931, 318
RFC 2046, 117, 317 RFC 3971, 188, 318
RFC 2050, 317 RFC 3972, 188, 318
RFC 2080, 190, 191, 287, 317 RFC 3986, 124, 131, 318
RFC 2082, 317 RFC 4033, 112, 318
RFC 2131, 188, 317 RFC 4151, 125
RFC 2140, 145, 148, 317 RFC 4193, 171, 318
RFC 2225, 317 RFC 4251, 123, 318
RFC 2328, 191, 193, 306, 317 RFC 4253, 306
RFC 2332, 317 RFC 4264, 206, 318
RFC 2364, 209, 317 RFC 4271, 198, 199, 318
RFC 2368, 317 RFC 4287, 125
RFC 2402, 175 RFC 4291, 171, 173, 184, 185, 318
RFC 2406, 175 RFC 4301, 318
RFC 2453, 190, 306, 317 RFC 4302, 318
RFC 2460, 137, 173, 175, 178, 179, 317 RFC 4303, 318
RFC 2464, 212, 317 RFC 4340, 319
RFC 2507, 317 RFC 4443, 178, 180, 182, 277, 282, 319
RFC 2516, 209, 317 RFC 4451, 319
RFC 2581, 151, 317 RFC 4456, 319
RFC 2616, 120, 127129, 259, 304, 317 RFC 4614, 138, 319
RFC 2617, 131, 317 RFC 4627, 133, 134
RFC 2622, 197, 317 RFC 4632, 303
RFC 2675, 175 RFC 4634, 229
RFC 2711, 175 RFC 4648, 118, 319
RFC 2766, 317 RFC 4822, 319
RFC 2821, 119, 120 RFC 4838, 319
RFC 2854, 117, 317 RFC 4861, 185, 319
RFC 2965, 317 RFC 4862, 186, 319
RFC 2988, 143, 147149, 317 RFC 4870, 122, 319
RFC 2991, 194, 317 RFC 4871, 122, 319
RFC 3021, 317 RFC 4941, 186, 319
RFC 3022, 317 RFC 4944, 178, 319
RFC 3031, 317 RFC 4952, 118, 319
RFC 3168, 100, 139, 162, 163, 318 RFC 4953, 153, 319
RFC 3187, 125 RFC 4954, 120, 122, 319
RFC 3234, 318 RFC 4960, 136, 154158
RFC 3235, 318 RFC 4963, 182, 319
RFC 3286, 157 RFC 4966, 319
RFC 3309, 318 RFC 4987, 143, 319
RFC 3315, 188, 318 RFC 5004, 319
326 Index
RFC 5065, 319 router, 306

RFC 5068, 122, 319 Routing Information Protocol, 190
RFC 5072, 208, 319 RPC, 59, 306
RFC 5082, 188 RTS, 92
RFC 5095, 178, 319 RTS frame (802.11), 224
RFC 5227, 319
RFC 5234, 52, 303, 319 S
RFC 5246, 307 scheduler, 80
RFC 5321, 119, 120, 257, 320 scheduling algorithm, 80
RFC 5322, 114, 115, 120, 320 SCTP, 154
RFC 5340, 191, 285, 306, 320 SCTP chunk, 155
RFC 5531, 306 SCTP common header, 155
RFC 5598, 114, 320 SCTP CWR chunk, 163
RFC 5646, 320 SCTP data chunk, 157
RFC 5681, 159, 161, 320 SCTP ECN Echo chunk, 163
RFC 5735, 320 SCTP SACK chunk, 157
RFC 5795, 178, 208, 320 SCTP segment, 155
RFC 5880, 313 SCTP Selective Acknowledgement chunk, 157
RFC 5890, 71, 118 SCTP TSN, 157
RFC 6068, 125, 320 SDU (Service Data Unit), 306
RFC 6077, 165, 320 segment, 61, 106, 306
RFC 6106, 188 selective acknowledgements, 24
RFC 6144, 320 selective repeat, 21
RFC 6164, 202 sendto, 237
RFC 6214, 210 sequence number, 16
RFC 6265, 131, 320 Serial Line IP, 208
RFC 6274, 210, 320 Service Set Identity (SSID), 225
RFC 6275, 175 shared-cost peering relationship, 197
RFC 6437, 174 Short Inter Frame Spacing, 89
RFC 6824, 154 sibling peering relationship, 197
RFC 6928, 161 SIFS, 89
RFC 7010, 187 SLAC, 186
RFC 768, 136, 315 slot time (Ethernet), 88
RFC 789, 48, 315 slotted ALOHA, 84
RFC 791, 53, 140, 315 slotTime (CSMA/CA), 90
RFC 792, 315 SMTP, 306
RFC 793, 68, 138141, 144, 145, 148, 149, 153, SNMP, 306
267, 269, 315 SOCK_DGRAM, 237
RFC 813, 150, 315 SOCK_STREAM, 237
RFC 819, 70, 315 socket, 237, 306
RFC 821, 306, 315 socket.bind, 240
RFC 822, 116, 117 socket.close, 239
RFC 826, 303, 315 socket.connect, 239
RFC 854, 307 socket.recv, 239
RFC 867, 266 socket.recvfrom, 240
RFC 868, 242 socket.send, 239
RFC 879, 143, 315 socket.shutdown, 239
RFC 893, 210, 315 source routing, 34
RFC 894, 212, 225, 315 speed of light, 86
RFC 896, 77, 97, 146, 315 split horizon, 45
RFC 952, 70, 315 split horizon with poison reverse, 45
RFC 959, 120, 123, 304, 315 spoofed packet, 306
RFC 974, 315 SSH, 306
RIP, 190, 306 SSID, 225
RIR, 306 standard query, 306
Robustness principle, 153 Starting Delimiter (Token Ring), 94
root nameserver, 306 Stateless Address Autoconfiguration, 186
round-trip-time, 306 stream-mode data transfer, 57
Index 327
stub domain, 195 WAN, 307

stuffing (bit), 9 Wavelength Division Multiplexing, 83
stuffing (character), 10 WDM, 83
switch, 214, 306 WiFi, 221
SYN cookie, 306
SYN cookies, 142 X
X.25, 307
T X11, 307
TCB, 306 XDR, 133
TCP, 138, 306 XML, 307
TCP Connection establishment, 140
TCP connection release, 152
TCP fast retransmit, 150
TCP header, 138
TCP Initial Sequence Number, 141
TCP MSS, 143
TCP Options, 143
TCP RST, 141
TCP SACK, 151
TCP selective acknowledgements, 151
TCP self clocking, 96
TCP SYN, 140
TCP SYN+ACK, 140
TCP/IP, 307
TCP/IP reference model, 107
telnet, 307
Tier-1 ISP, 207
Time Division Multiplexing, 83
time-sequence diagram, 6
TLD, 307
TLS, 307
Token Holding Time, 95
Token Ring data frame, 95
Token Ring Monitor, 94
Token Ring token frame, 94
traceroute6, 182
transit domain, 195
Transmission Control Block, 145
Transmission Sequence Number, 157
transport clock, 62
Transport layer, 106
two-way connectivity, 49
U
UDP, 136, 307
UDP Checksum, 137
UDP segment, 136
unicast, 307
Unique Local Unicast IPv6, 171
unreliable connectionless service, 54
V
Virtual LAN, 220
VLAN, 220
vnc, 307
W
W3C, 307
328 Index

Cnp3bis7dfkcompressed PDF

Uploaded by

Copyright:

Available Formats

Cnp3bis7dfkcompressed PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cnp3bis7dfkcompressed PDF

Uploaded by

Copyright:

Available Formats

Computer Networking : Principles,

Protocols and Practice

September 16, 2015

3 Part 2: Protocols 109

4 Part 3: Practice 227

1.1 About the author

2.1 Connecting two hosts

Figure 2.1: Connecting two hosts together

2.1.1 The physical layer

Note: Bit rate

6 Chapter 2. Part 1: Principles

Host A Physical link Host B

Host A Physical link Host B

Host A Physical link Host B

2.1. Connecting two hosts 7

Host A Physical link Host B

Figure 2.2: Manchester encoding

Figure 2.3: The Physical layer

8 Chapter 2. Part 1: Principles

2.1.2 The datalink layer

Note: Bit rate and bandwidth

2.1. Connecting two hosts 9

10 Chapter 2. Part 1: Principles

Framing-A Phys-A Phys-B Framing-B

Recovering from transmission errors

2.1. Connecting two hosts 11

Note: Services and protocols

12 Chapter 2. Part 1: Principles

Figure 2.4: Finite state machine of the simplest reliable protocol

Reliable data transfer on top of an imperfect link

2.1. Connecting two hosts 13

14 Chapter 2. Part 1: Principles

a payload that contains the user data

Note: Checksums, CRCs, ...

2.1. Connecting two hosts 15

Figure 2.5: Alternating bit protocol : Sender FSM

Note: Dealing with corrupted frames

16 Chapter 2. Part 1: Principles

Figure 2.6: Alternating bit protocol : Receiver FSM

2.1. Connecting two hosts 17

Go-back-n and selective repeat

18 Chapter 2. Part 1: Principles

Figure 2.7: Pipelining improves the performance of reliable protocols

2.1. Connecting two hosts 19

Figure 2.8: The sliding window

Figure 2.9: Sliding window example

20 Chapter 2. Part 1: Principles

Figure 2.10: Utilisation of the sliding window with modulo arithmetic

2.1. Connecting two hosts 21

Figure 2.11: Go-back-n : receiver FSM

Figure 2.12: Go-back-n : sender FSM

22 Chapter 2. Part 1: Principles

Figure 2.13: Go-back-n : example

Figure 2.14: The receiving window with selective repeat

2.1. Connecting two hosts 23

Figure 2.15: Selective repeat : example

24 Chapter 2. Part 1: Principles

Note: Maximum window size with go-back-n and selective repeat

Figure 2.16: Piggybacking example