Computer Network Layers

CIS748 Class Notes

Alex S.∗

1 The Layers
Normally, the task of communication from computer to computer is broken up into layers.
Each device attached to the network has a corresponding ‘stack’ of these layers, where each
layer conceptually talks to the corresponding layer on the other computer. Each layer only
needs to know how to interface to the layers above and below it.
The ‘standard’ layered model used for illustrative purposes only is the OSI Model1 . It
consists of seven layers:

7 Application Layer (Provide end-user services, like e-mail)

6 Presentation Layer (Data compression, and other data conversion)

5 Session Layer (Authentication/Authorization)

4 Transport Layer (Guarantee end-to-end data transfer—from machine to machine)

3 Network Layer (Routing, accounting).

2 Data Link Layer (Transmit/recieve packets, resolve hardware addresses)

1 Physical Layer (Physical cable, medium, air)

There is an interesting quote from UNIX System Administration Handbook: “Some think
a financial layer and a political layer should be added to these. There is no magic about
the number seven; seven committees were involved in the specification, and one layer was
created for each.”
In any case, there have been some attempts to implement the OSI model, and some sorta
actually worked, etc., but it never managed to catch on.
∗
[email protected]
1
http://en.wikipedia.org/wiki/OSI model

1
1.1 PDU, Protocol Data Unit 2 OSI MODEL

1.1 PDU, Protocol Data Unit

Layers communicate using a thing called a PDU, or Protocol Data Unit. Each layer may
have a different PDU format, etc. It is only required that corresponding layers be able to
understand their PDU (the interface between layers). For this purpose, we often attach
a number to the PDU, to know what layer it belongs to. For example, a 3-PDU would
corresponds to the 3rd layer PDU, and so on for any N -PDU.
Layer N needs to be able to talk to layer N + 1 and layer N − 1, but not to any other
layer.

1.2 Encapsulation
Layers are arranged in a way that each layer only has to know about layers that are directly
below or directly above. All data from the higher layers is wrapped (and subsequently
unwrapped at the destination) so that, for example, the data-link layer doesn’t have to
know whether you’re browsing the web or transferring files—all it sees are just a bunch of
bits that need to be transferred.

1.3 Services
We can also view the layered model as higher layers getting services from lower layers and
lower layers providing a service to higher layers.

1.3.1 Service Access Points

An entity at layer N may service more than one entity at layer N + 1. Layer N may use a
Service Access Point address to determine which entity at a higher layer will get serviced.
An example of that are Port addresses in TCP/IP.

2 OSI Model
Even though the OSI Model isn’t the actual ‘real’ model, it is still useful for illustration
purposes as an idealized network model. It consists of seven layers (descriptions taken from
Wikipedia):

2.1 Application Layer

Layer 7, the highest layer: This layer interfaces directly to and performs common application
services for the application processes. The common application services provide semantic
conversion between associated application processes. Examples of common application ser-
vices include the virtual file, virtual terminal (for example, Telnet), and ”Job transfer and
Manipulation protocol” (JTM, standard ISO/IEC 8832).

2
2.2 Presentation Layer 2 OSI MODEL

2.2 Presentation Layer

Layer 6: The Presentation layer relieves the Application layer of concern regarding syn-
tactical differences in data representation within the end-user systems. MIME encoding,
encryption and similar manipulation of the presentation of data is done at this layer. An
example of a presentation service would be the conversion of an EBCDIC-coded text file to
an ASCII-coded file.

2.3 Session Layer

Layer 5: The Session layer provides the mechanism for managing the dialogue between
end-user application processes. It provides for either duplex or half-duplex operation and
establishes checkpointing, adjournment, termination, and restart procedures. This layer is
responsible for setting up and tearing down TCP/IP sessions.

2.4 Transport Layer

Layer 4: The purpose of the Transport layer is to provide transparent transfer of data
between end users, thus relieving the upper layers from any concern with providing reliable
and cost-effective data transfer. The transport layer controls the reliability of a given link.
Some protocols are stateful and connection oriented. This means that the transport layer
can keep track of the packets and retransmit those that fail. The best known example of a
layer 4 protocol is TCP.

2.5 Network Layer

Layer 3: The Network layer provides the functional and procedural means of transferring
variable length data sequences from a source to a destination via one or more networks while
maintaining the quality of service requested by the Transport layer. The Network layer
performs network routing, flow control, segmentation/desegmentation, and error control
functions. The router operates at this layer – sending data throughout the extended network
and making the Internet possible, although there are layer 3 (or IP) switches. This is a logical
addressing scheme - values are chosen by the network engineer. The addressing scheme is
hierarchical.

2.6 Data Link Layer

Layer 2: The Data link layer provides the functional and procedural means to transfer data
between network entities and to detect and possibly correct errors that may occur in the
Physical layer. The addressing scheme is physical which means that the addresses (MAC)
are hard-coded into the network cards at the time of manufacture. The addressing scheme
is flat. Note: The best known example of this is Ethernet. Other examples of data link
protocols are HDLC and ADCCP for point-to-point or packet-switched networks and LLC

3
2.7 Physical Layer 4 IEEE STANDARDS

and Aloha for local area networks. This is the layer at which bridges and switches operate.
Connectivity is provided only among locally attached network nodes.

2.7 Physical Layer

Layer 1: The physical layer defines all electrical and physical specifications for devices.
This includes the layout of pins, voltages, and cable specifications. Hubs and repeaters are
physical-layer devices. The major functions and services performed by the physical layer are:

• establishment and termination of a connection to a communications medium.

• participation in the process whereby the communication resources are effectively shared
among multiple users. For example, contention resolution and flow control.

• modulation, or conversion between the representation of digital data in user equipment

and the corresponding signals transmitted over a communications channel. These are
signals operating over the physical cabling – copper and fibre optic, for example. SCSI
operates at this level.

3 TCP/IP
The TCP/IP can be thought of consisting of four layers:

4 Application Layer (End-user application programs)

3 Transport Layer (Communication among programs on a network)

2 Network Layer (Communication, addressing, and routing)

1 Link Layer (Network hardware and device level)

The TCP/IP does most of the things the OSI model does, except it does it in four layers,
and is actually a lot more robust and real than the OSI idea.
TCP/IP is setup in such a way that it doesn’t really matter which network is used
underneath. The Link Layer can use physical wire, wireless, serial lines, etc.

4 IEEE Standards
In 1985 the Computer Society of the IEEE started a project called Project 802, to set
standards to enable intercommunications between equipment from variety of manufacturers.
It is a specification for the data-link and physical layer of major LAN protocols.
The 802 standard split the data-link layer into 2 layers:

• Logical Link Control (LLC): IEEE 802.2

4
 5 COMPUTER NETWORKS INTRO (REVIEW; OTHER NOTES)

• Media Access Control (MAC): IEEE 802.3 (CSMA/CD), IEEE 802.4 (Token Bus),
IEEE 802.5 (Token Ring), IEEE 802.5 (DQDB—distributed-queue dual-bus network),
IEEE 802.11 (Wireless).
The physical layer is also split into a few layers, depending on the type of network we’re
dealing with. Generally, we’ll have two primary things: PMI (Physical Medium Independent
and PMD (Physical Medium Dependent). Those are generally broken up into PLS (Physical
Layer Signaling), AUI (Attachment Unit Interface), MAU (Medium Attachment Unit), MDI
(Medium Dependent Interface).

5 Computer Networks Intro (Review; Other Notes)

5.1 Serial Lines
Serial lines are generally the most basic and most common communication medium you can
have between computers and/or equipment. They generally consist of relatively cheap wire
sending bits across at relatively slow speeds. That’s all there is to them.
Because of their simplicity, serial ports are available on pretty much all sorts of hardware
(except many new laptops). Many port and jack configurations are supported (there are
round jacks (DIN-8), 9 pin jacks (DB-9), 25 pin jacks (DB-25), etc.) Most times, only a few
of the wires are actually used. In fact, the most basic serial port communication between
two computers uses only 3 wires: (transmit, receive, and ground). Sometimes an RJ-45 is
used for serial communications—and there are various standards on how pins map to which
wires.
The transmission can use hardware flow control or software flow control. The hardware
flow control generally requires more wires to be connected. Software flow control is harder
on the CPU.

5.2 Null Modem

There is a concept of a ‘null’ modem. This is basically when the serial cable isn’t connected to
a device, but to another computer. It is the same wire as a regular serial cable, except a few
wires have been crossed over... namely the send/receive wires—along with a few hardware
flow control wires if those are used).

5.3 Terminal Devices

In the UNIX world, there is a thing called a ‘terminal’. It is basically a screen/keyboard
connected to the ‘computer’ via a serial cable. You’ve seen a similar idea when you use
Telnet (or SSH) to log into Solaris/Linux computers (except connection doesn’t travel over
serial cables—but it still uses terminal devices).
In fact, there is a program called getty which presents a login prompt on a particular
serial port. Something you could try: connect two computers via a null-modem cable, do a

5
5.4 IP over Serial Lines
5 COMPUTER NETWORKS INTRO (REVIEW; OTHER NOTES)

getty on the serial port on one computer, and on another computer, you can cat the serial
port device to read/write it and actually be logged into the other computer through a serial
port (and obviously use all the command line programs).
UNIX has a bunch of virtual terminals, and when the system starts up, getty is executed
on a bunch of those terminals. That’s why you see a login prompt on the screen (your
screen/keyboard are just treated as command line terminals as far as the computer are
concerned).
There is a program called stty which you can use to adjust the properties of your
terminals. For example, to change the speed of a serial port, or to enable/disable control
flow options, specify special characters, number of rows/columns, etc.

5.4 IP over Serial Lines

It turns out, you can use TCP/IP over serial lines. This shouldn’t come as a surprise, as
many of you might still be using dial-up Internet. Basically, there are two protocols that do
the job: SLIP and PPP.
Both have the basic idea of encapsulating IP packets, and transmitting them (with some
flow control) over the serial line. The process would work something like this: the dial-up
server is setup to pickup a phone line when it rings and start a getty on it. You dial up
(ATDT), see a terminal (or your dial-up networking software sees the login), it proceeds to
login (via the simple UNIX like login prompt), it then runs PPP or SLIP to start a remote
protocol engine—which then configures the already established serial link for transfer of IP
packets.

5.4.1 SLIP
SLIP (Serial Line Internet Protocol) is probably the simplest one, and at one point, you’d
use it if you wanted to setup a dial-up connection to the Internet from your Linux box.

5.4.2 PPP
PPP (Point-to-Point Protocol) is a huge bloated standard that is capable of quite a bit more
than just sending packets. When you dial into the your ISP, this is what you’re using.

5.4.3 Others
There may be other such encapsulation protocols used—so you can’t just dial up and expect
things to work (via standard networking configuration). Some ISPs require you to have
special “dialer” software that enables your Windows box to use other encapsulation protocol.
AOL was notorious for that.

6
5.5 TCP/IP 5 COMPUTER NETWORKS INTRO (REVIEW; OTHER NOTES)

5.5 TCP/IP
TCP/IP is a whole bunch of various protocols stuck into one:

5.5.1 IP
Internet Protocol, or IP, is used to transport RAW data from one machine to the next. This
involves hopping from computer to computer until the data gets there.

5.5.2 ICMP
Internet Control Message Protocol, or ICMP, is used as a helper protocol for IP. It provides
assistance with error messages, routing, and echo requests.

5.5.3 ARP
Address Resolution Protocol, or ARP, provides services to resolve logical network addresses
to hardware network addresses. Generally, figuring out which IP addresses correspond to
which MAC addresses.
It basically screams out to the network “does anyone know the hardware address for this
IP”. Then some machine (if it is alive) will respond, “Yes, here I am.”

5.5.4 UDP
User Datagram Protocol, or UDP, sends a packet from one program to another program. It’s
unreliable, and connectionless.

5.5.5 TCP
Transmission Control Protocol, or TCP, sends a stream of data from one program to another
program. It is reliable, and connection oriented (establishes a ‘session’).

5.6 Segmentation and Packets

TCP/IP breaks up all traffic into packets. A packet is just a structure with a header, and
some data. When referring to a packet on a LAN, we often use the term “frame” (even
though when a packet travels in a frame it has a bit more info attached to it).
Every network has MTU, or Maximum Transfer Unit. This is the maximum size of a
packet that can traverse the network. The IP layer, breaks up your data into packets of
MTU size, transfers those to the destination machine, where they are reassembled.
Every packet, obviously has to have source IP, destination IP, size of data, etc. When it
is encapsulated in some other protocol, those are also included. For example, a UDP packet
carrying 100 bytes of data on an Ethernet would have:
• 14 bytes Ethernet Header (includes things like MAC addresses)

7
5.7 Internet Addresses
5 COMPUTER NETWORKS INTRO (REVIEW; OTHER NOTES)

• 20 bytes IP Header (includes things like source/destination IP addresses)

• 8 bytes UDP Header (includes things like source/destination port addresses)

• 100 bytes of user data

• 4 bytes Ethernet Trailer (terminates the frame)

So while we were just sending 100 bytes via UDP, over the Ethernet, that sent 146 bytes.

5.7 Internet Addresses

Since we are primarily concerned with TCP/IP, let’s discuss network addresses. An IP
addresses is just a 32 bit number, written out with each byte (0-255) separated by dots. For
example, 192.168.0.1 is an IP address.
There are several classes of networks (and thus addresses). Each address belongs to a
class. Part of the IP address indicates the network address. For example, N.N.N.H has 3
bytes dedicated to the network address, and 1 byte to the host address. Various classes of
networks:

• Class A: Starts with 1-126. Major networks. N.H.H.H

• Class B: Starts with 128-191. Large networks. N.N.H.H

• Class C: Starts with 192-223. Small networks (easy to get). N.N.N.H

• Class D: Starts with 224-239. Multi-cast addresses.

• Class E: Starts with 240-254. Experimental addresses.

In any case, most networks that most people are concerned with are B and C.

5.8 Routing
Routing is the process of directing a packet through the network to the destination computer.
You can view it as asking for directions. You are sent to deliver something to an unknown
city. You ask for directions, and someone may point you towards a general direction (or
towards someone who may known the general direction). You soon get to the city, and
you start asking for more specific information concerning the street address, etc. Same idea
works for routing.
For routing to work, the operating system maintains a routing table. It is just a table
of network masks and destinations. Whenever the kernel sees a packet, it figures out which
network it corresponds to, and sends the packet to that network—that’s all there is to it.
There are also ‘default’ routes; when everything else fails (you don’t know what to do with
a packet) you send it off to the default route.

8
5.8 Routing 5 COMPUTER NETWORKS INTRO (REVIEW; OTHER NOTES)

There is static routing and dynamic routing. Static means that you setup the routing
table manually, at boot-time, etc., and it stays around. For small networks that is quite
efficient.
Dynamic routing involves the use of routing daemons like routed or gated to discover
the network topology and how to reach distant networks. There are many approaches to
this, which we won’t get to at this time.