Internet Protocol Architectures and Network Programming

310/1779-6
Fourth Workshop on Distributed Laboratory Instrumentation

Systems
(30 October - 24 November 2006)
_____________________________________________________________________
Internet Protocol Architectures

and
Network Programming
Abhaya S. INDURUWA
Department of Computing
Canterbury Christ Church University
North Holmes Road
Canterbury CT1 QU
U.K.
________________________________________________________________
These lecture notes are intended only for distribution to participants
Internet Protocol Architectures
and
Network Programming
Abhaya Induruwa∗
Department of Computing
Canterbury Christ Church University
Canterbury
United Kingdom
Supporting material for the

lectures given at:
The Fourth Trieste Workshop on
Distributed Laboratory Instrumentation Systems
Trieste, Italy 30 October – 24 November 2006
LNS
∗
[email protected]
Abstract
This Chapter is intended to give a broad overview of the Internet and its
architecture comprising network protocols, standards, hardware and sup-
porting technologies. Various protocol architectures are presented in order
to be able to understand the most desirable features that are required in a
protocol to support the characteristics of modern day communication sys-
tems. The Internet Protocol (IP) architecture and its components used in time
critical data communication are discussed.
Also discussed are High Speed LANs based on ATM technology and the
new variants of Ethernets running at Gigabit speeds. More recent attempts
based on WAP to deliver Internet content and services to hand-held mobile
devices and the deployment of ADSL to deliver Internet to the home user are
introduced.
Bluetooth, a wireless communication technology capable of location/ de-
vice aware connection and switching, and is aiming to take the place of wires
for connecting peripherals, telephones, computers, is also briefly discussed.
The lectures also briefly cover the features in JavaTM that enable todays
network programmers to implement sophisticated networking applications
using Java APIs.
Contents
1 Introduction 1
2 The Internet 2
3 Network Classification 2
3.1 Geographical Coverage . . . . . . . . . . . . . . . . . . . . . . 2
3.2 Network Topology . . . . . . . . . . . . . . . . . . . . . . . . . 3
4 Network Architecture 5
4.1 What is a Network Protocol? . . . . . . . . . . . . . . . . . . . 5
4.2 Transmission Mechanism . . . . . . . . . . . . . . . . . . . . 7
4.2.1 Packet Switching . . . . . . . . . . . . . . . . . . . . 8
4.2.2 Frame Relay . . . . . . . . . . . . . . . . . . . . . . 8
4.2.3 Cell Switching . . . . . . . . . . . . . . . . . . . . . 8
4.3 Physical Media . . . . . . . . . . . . . . . . . . . . . . . . . . 9
5 Internetworking 9
5.1 Building Internets and Intranets . . . . . . . . . . . . . . . . 10
5.2 Repeaters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
5.3 Bridges and Switches . . . . . . . . . . . . . . . . . . . . . . . 11
5.4 Routers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
5.5 Gateways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
5.6 Multiport-Multiprotocol Devices . . . . . . . . . . . . . . . . 12
6 A word about the Internet 12
7 Internet Protocol Architecture 14

7.1 IP Addressing . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
7.2 Resolving IP Addresses . . . . . . . . . . . . . . . . . . . . . . 17
7.3 The Internet Protocol . . . . . . . . . . . . . . . . . . . . . . . 17
7.4 Internetworking with IP . . . . . . . . . . . . . . . . . . . . . 17
7.5 IP Datagram Format . . . . . . . . . . . . . . . . . . . . . . . 17
7.6 Brief Description of TCP and UDP . . . . . . . . . . . . . . . 19
7.7 Client/Server Computing . . . . . . . . . . . . . . . . . . . . 21
7.7.1 Port Numbers . . . . . . . . . . . . . . . . . . . . . . 21
7.7.2 Sockets . . . . . . . . . . . . . . . . . . . . . . . . . 24
7.8 IP Multicasting . . . . . . . . . . . . . . . . . . . . . . . . . . 25
7.9 Resource Reservation Protocol – RSVP . . . . . . . . . . . . . 26
7.10 TCP/IP Over Serial Lines – SLIP and PPP . . . . . . . . . . . 27
7.10.1 SLIP – Serial Line Internet Protocol . . . . . . . . . 27
7.10.2 PPP – The Point to Point Protocol . . . . . . . . . . 28
4 Internet Protocols and Network Programming
8 IPv6 – The New Generation Internet Protocol 29

8.1 The Design of IPv6 . . . . . . . . . . . . . . . . . . . . . . . . 29
8.2 IPv6 Header Format . . . . . . . . . . . . . . . . . . . . . . . 30
8.3 Simplifications . . . . . . . . . . . . . . . . . . . . . . . . . . 30
8.4 New Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
8.5 Special Services . . . . . . . . . . . . . . . . . . . . . . . . . . 32
8.6 IPv6 Address Space . . . . . . . . . . . . . . . . . . . . . . . . 32
8.7 Making IPv6 Compliant . . . . . . . . . . . . . . . . . . . . . 33
8.7.1 IPv6 Tunnelling . . . . . . . . . . . . . . . . . . . . . 33
9 Data Communication in Real Time 34

9.1 RTP Data Transfer Protocol . . . . . . . . . . . . . . . . . . . 34
9.1.1 Characteristics of RTP . . . . . . . . . . . . . . . . . 35
9.1.2 Definitions in RTP . . . . . . . . . . . . . . . . . . . 36
9.1.3 RTP Fixed Header Fields . . . . . . . . . . . . . . . 37
9.1.4 Multiplexing RTP Sessions . . . . . . . . . . . . . . 38
9.1.5 Real time Transport Control Protocol (RTCP) . . . 38
9.2 Real Time Data Transfer using ATM . . . . . . . . . . . . . . 39
9.2.1 ATM Protocol Structure . . . . . . . . . . . . . . . . 40
9.2.2 ATM Services . . . . . . . . . . . . . . . . . . . . . . 43
9.2.3 IP over ATM . . . . . . . . . . . . . . . . . . . . . . . 43
9.3 IP/TV – A Real life Example . . . . . . . . . . . . . . . . . . . 44
9.4 Delivering Real Time Data to the Desk Top . . . . . . . . . . 45
9.4.1 Ethernet Based Solutions . . . . . . . . . . . . . . . 45
9.4.2 Signal Topology and Timing . . . . . . . . . . . . . 45
9.4.3 High Performance Ethernets . . . . . . . . . . . . . 46
9.4.4 Switched Ethernets . . . . . . . . . . . . . . . . . . 46
9.4.5 Fast Ethernet . . . . . . . . . . . . . . . . . . . . . . 48
9.4.6 Gigabit Ethernet . . . . . . . . . . . . . . . . . . . . 49
9.4.7 Wireless LANs . . . . . . . . . . . . . . . . . . . . . 50
9.4.8 Bluetooth Technology . . . . . . . . . . . . . . . . . 52
9.4.9 Future of Wireless Technology . . . . . . . . . . . . 53
9.5 ADSL – delivering RT multimedia to the home and small business 53
9.5.1 ADSL Standardisation . . . . . . . . . . . . . . . . . 55
10 WAP – Wireless Application Protocol 55

10.1 WAP Specification . . . . . . . . . . . . . . . . . . . . . . . . . 56
10.2 Architecture of the WAP Gateway . . . . . . . . . . . . . . . . 57
10.3 Mobile Web Services . . . . . . . . . . . . . . . . . . . . . . . 58
11 Brief Introduction to Programming the Internet with JavaTM 60

11.1 InetAddress Class . . . . . . . . . . . . . . . . . . . . . . . . . 60
11.2 URL Class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
11.3 Socket and ServerSocket Classes . . . . . . . . . . . . . . . . 64
11.4 DatagramSocket and DatagramPacket Classes . . . . . . . . 66
5
12 Programming Multitasking Applications 72

12.1 Thread per Client . . . . . . . . . . . . . . . . . . . . . . . . . 72
12.2 Thread Pool . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
13 Summary 74
References 76
1
1 Introduction
Computer networks have become an integral and indispensable part of scientific
as well as public life today. Over the last couple of decades data networks have
changed their character from a few slow speed point to point links to a high speed
data communication backbone supporting full multimedia information transfer.
The last few years have seen a phenomenal rise in the interest in the collection
of computers and networks known as the Internet. The Web (World Wide Web)
has truly transformed the access to information content, and their delivery over
the Internet.
More recently the emergence of hand-held PDAs (Personal Data Assistants) in-
cluding wireless phones based on WAP (Wireless Application Protocol), and ADSL
(Asynchronous Digital Subscriber Line) technology that supports high speed data
transfer rates over twisted pair subscriber lines, have revolutionised the provi-
sion of public access to the Internet and its information services.
In parallel with these advances in technologies associated with the content
delivery over the Internet, Java and other object oriented programming language
variants have given the programmer the opportunity to bring to life the hitherto
static multimedia Web presentations by adding ’dynamic content’ to the Web
pages. As a results the Web pages of today have become live; they have become
interactive and animated. The static 2D images on Web pages have now become
3D animated images.
The underlying network architecture supporting the above capability is based
on the TCP/IP protocol suite which lets hundreds of millions of computers and
hundreds of thousands of computer networks spread across hundreds of coun-
tries to be internetworked to form the global Wide Area Network (WAN), the Inter-
net, which is now used by hundreds of millions of users. The familiar application
protocols like FTP, HTTP, SMTP, etc., make use of TCP/IP to achieve a reliable
stream transport across inherently unreliable network links. TCP/IP is also the
basis for client/server computing. Other more specific protocols such as the
RTP (Real-time Transport Protocol), together with RTCP (Real-time Transport
Control Protocol), have been devised to facilitate the communication of Real Time
data over computer networks which have been designed and built to guarantee
the delivery of time insensitive bursty data. IPv6 is emerging as the next gener-
ation Internet Protocol with all the features necessary to support the delivery of
multimedia and other real time data services at high speed.
At the Local Area Network (LAN) level, the popular Ethernet running at 10
Mbps has already been enhanced to operate at 100 Mbps (Fast Ethernet) and
1000 Mbps (Gigabit Ethernet) thus making it suitable as a delivery mechanism
of Real Time traffic to the desktop. Work has already begun to formulate 10
Gbps Ethernet specifications which will initially be deployed to enhance the high
speed local backbones.
Wireless bearer technologies have matured in the last couple of years to an
extent that now they take the place of wired connections. Technologies such as
Bluetooth and WiFi, which are capable of location/devices aware switching have
started to revolutionise the way public access to the Internet is provided.
2 The Internet
The one and only global network of interest to the whole scientific community of
the world today is the Internet, which is based on the Internet Protocol (IP) at
its network layer and Transmission Control Protocol (TCP) at its transport
layer.
The Internet is a network of computer networks and computers. It intercon-
nects computers of all types and specifications without regard to their operating
system or hardware specifics. It interconnects all types of computer networks
irrespective of their topology and other physical characteristics. The Internet
spans across almost all countries in the world disregarding their socio-economic
status and political boundaries. The one thing that binds them onto one single
network is the Internet Protocol suite, TCP/IP.
TCP/IP is designed to provide a reliable stream transport service across unre-
liable network links to the application level protocols that use it. TCP/IP is also
the basis for client server networking and distributed computing.
3 Network Classification
A computer network is a collection of co-operating computers interconnected
by one or more transmission paths for the purpose of transfer and exchange of
data between them. Today these networks span the entire globe and belong to
many different nations and network operators. Such networks can be classified
in many ways depending on the switching mechanism, transmission speed, etc,
[3, 40, 42]. They can also be classified on the basis of their geographical coverage
or network topology.
3.1 Geographical Coverage

Networks can be classified into many categories depending on their geographical
coverage. The following lists them in the order from the smallest to the largest.
1. Desktop Area Networks (DANs)1
2. Controller Area Networks (CANs)2
3. Home Area Networks (HANs)3
4. Storage Area Networks (SANs)4
5. Local Area Networks (LANs)
6. Metropolitan Area Networks (MANs)
7. Wide Area Networks (WANs)
8. Global Area Networks (GANs)5 .
1
a classification arising out of delivering ATM or IP to the desktop at 25 Mbps or above. DANs
are used to interconnect devices such as a camera, a telephone and workstations.
2
ISO 11898 specifies CAN bus up to 1 Mbps and ISO 11519 specifies CAN bus up to 125 Kbps.
3
mainly due to the proliferation of Bluetooth technology.
4
networked ultra high capacity disk storage devices.
5
the result of worldwide GRID initiatives.
3
There is a significant level of deployment of all or some of the above even in

developing countries (most notably LANs and WANs with DANs just appearing to
support desk top video conferencing) and hence should be of interest to research
scientists.
3.2 Network Topology

Networks can be classified according to their topology in the following manner
[39, 40]:
• Bus Topology (eg. CSMA/CD; Ethernet)

• Ring Topology (eg. Token Ring, FDDI)
• Star Topology (eg. ArcNet, Switched networks such as Fast Ethernet)
• Mesh Topology (eg. Telephone network).
The bus topology has been used initially in the Ethernet (broadcast) and later
in Token Bus and the DQDB (Distributed Queue Dual Bus).
The token passing mechanism shown in Figure 1 has originally been used in
the Token Ring and later in FDDI.
Data
Token
D Header
Figure 1: Token Passing Ring
If however networks are categorised according to their transfer mechanism,

then the following classification results.
1. Broadcast Networks
Figure 2 shows a broadcast network which is used to broadcast from 1 to
many. CSMA/CD (Carrier Sense Multiple Access/Collision Detect), com-
monly known as Ethernet, is a Medium Access Control (MAC) protocol
which is used to broadcast on a bus topology.
Figure 2: Broadcast Network
2. Switched Networks
In Figure 3 is shown a switched network which is used to switch data from
1 to 1, 1 to many, or many to many. The telephone network is an example
of a circuit switched network. It can be used to support applications such
as tele conferencing and video conferencing (using ISDN) which involves the
switching of many to many.
Figure 3: Switched Network
Although initially switched networks were mainly used in telecommunica-

tion networks, today because of its superior performance, switched tech-
nologies are increasingly used in LANs (for example switched Ethernet) and
ATM networks.
3. Hybrid Networks
Figure 4 shows a hybrid network which consists of a switched part and a
broadcast part.
Figure 4: Hybrid Network

5
4 Network Architecture
The topology, transmission mechanism, and a protocol which manages the trans-
mission mechanism together define a network architecture.
4.1 What is a Network Protocol?

A network protocol is used to facilitate the transmission of data between a sender
(transmitter) and a recipient (receiver) across a data communication network in
an agreed manner. Over the years several different network protocol architec-
tures have evolved, the most notable being:
• ISO - OSI (Open Systems Interconnect) Reference Model (OSI-RM)

• IP (Internet Protocol suite comprising TCP/IP)
• ISO 8802.X (for LANs and MANs - same as IEEE 802.X)
In addition to the above there are hundreds of vendor specific protocols such as:
• IBM’s SNA (Systems Network Architecture)

• DEC’s DNA (DEC Network Architecture),
which are proprietary and hence are not truly interoperable.

Because todays networks span the globe, it is important to utilise standard
protocols to ensure interoperability and to facilitate seamless data communi-
cation over networks belonging to many different network operators. This has
been the major objective of the many national, international and professional
bodies such as the American National Standards Institute (ANSI), International
Standards Organisation (ISO), International Telecommunication Union (ITU), In-
stitute of Electrical and Electronic Engineers (IEEE)1, Internet Engineering Task
Force (IETF) and the ATM Forum which are involved in preparing standards.
Figure 5 shows the layered architectures of the protocols developed by some
of the standards bodies above.
ISO−OSI TCP/IP ITU−T IEEE802.X

Application
Application
Presentation
Session
Transport TCP
Network IP X.25−3
Data Link Physical X.25−2 LLC/MAC
Physical Link X.25−1 Physical
Figure 5: Layered Architectures of different Protocols

1
http://grouper.ieee.org/groups/802/index.html
In Figure 6 is shown the architecture of the IEEE 802 family of standards

for LANs and MANs. These standards cover only LLC (Logical Link Control) and
MAC (Medium Access Control) protocols. It is seen that over the years the copper
COAX (coaxial cable) has been replaced by UTP (Unshielded Twisted Pair), STP
(Shielded Twisted Pair) and Optical Fibre (OF) for wired connections. The recent
standards are mainly wireless in order to support mobile connectivity.
802.1 Network Control and Management
802.2 Logical Link Control (LLC) LLC
802.3 802.3a 802.3z 802.4 802.5 802.6 802.7 802.8 802.9 802.10 802.11 802.12 802.15 802.16 802.20
VG−Any Wireless
SMDS/ Broadband Mobile
CSMA/CD Fast Gigabit Token Token Broadband Optical Integrated Secure Wireless
LAN Personal Wireless Broadband MAC
Ethernet Ethernet Bus Ring DQDB Fibre Voice/ Data Tech
Tech (Demand Area Access Wireless
(FDDI) MAN Tech Data Exchange CSMA/CA Access
Priority) Networks
COAX UTP UTP COAX STP COAX 2.4 − 5 UTP 2.4 GHz 2−11 GHz 3.5 GHz
1−10Mbps 4−16Mbps OF OF Shortrange
UTP STP GHz WMANs 1+ Mbps
OF (Bluetooth @ speeds Physical
OF OF OF OF Tech) 10−66 GHz upto
45−155
10Mbps 100Mbps 1Gbps 5−20Mbps 100Mbps Mbps 1−54 Mbps 100Mbps 22Mbps 100+ Mbps 250 kmph
Figure 6: The Architecture of IEEE 802 LAN/MAN Standards
FDDI is a standard developed by ANSI-ASC X3T9 (Accredited Standards Com-

mittee) and provides services specified by the ISO Data link and Physical layers
(ISO 9314). The key features of FDDI are 100 Mbps data rate, use of optical
fibre (multi mode fibre, single mode fibre, low cost fibre), the token ring style MAC
protocol and the reconfiguration concept (automatic healing property in case of
faults). The FDDI standard however allows the FDDI traffic to be carried on other
physical media such as twisted pair copper (CDDI). FDDI-II supports protocols
for transmission of real time data [22].
IEEE 802.6 MAN standard specifies a DQDB (Distributed Queue Dual Bus)
protocol which can support data, voice and video traffic. It can also serve as a
LAN. DQDB MAN operates on a shared medium with two unidirectional buses
that supports data flow in opposite directions. Two methods of gaining ac-
cess to the medium depending on the type of traffic have been specified. In
the first method a node on the DQDB subnetwork can queue to gain access to
the medium by using a distributed queue or by requesting a fixed bandwidth
through a pre-arbitrated access method. Data is transmitted in fixed size units
called slots of length 53 bytes (52 bytes data + 1 byte access control field).
DQDB private networks are connected to the public network by point to point
links. A DQDB MAN can typically range up to more than 50 km in diameter and
can operate at a variety of speeds ranging from 34 Mbps to 150 Mbps.
DQDB has been conceived to integrate data and voice over a common set
of equipment, thereby reducing maintenance and administrative costs. B-ISDN
(Broadband Integrated Services Digital Network) is an attempt to provide uni-
versal and seamless connectivity for multimedia services. IEEE 802.6 MAN has
been designed to provide an interim solution and to act as a migration path to
B-ISDN.
7
Nodes in a DQDB subnetwork are connected to a pair of buses flowing in

opposite directions and can operate in one of two topologies, namely; open bus
(Figure 7) or looped bus (Figure 8) (open bus topology is similar to Ethernet and
looped bus topology is similar to token ring).
Head
Head
Figure 7: Open Bus DQDB Network
Although the DQDB access layer is independent of the physical medium, the
speeds at which DQDB MANs operate demand the use of fibre or coaxial cables.
For example, ANSI-DS3 operates at 44.736 Mbps over 75 Ω coax or fibre and
ANSI SONET STS3 (Synchronous Optical NETwork) operates at 155.52 Mbps
over single mode fibre. The ITU-T G.703 operates at 34.368 Mbps and 139.264
Mbps over a metallic medium.
Node 1
Head
Node
Node 2
Figure 8: Looped Bus DQDB Network– DQDB Ring
SMDS (Switched Multi-megabit Data Service) which provides a connectionless

service has been designed specifically to carry data instead of voice traffic on high
speed public wide area networks.
A complete treatment of FDDI, DQDB and SMDS along with LANs and MANs
is found in references [39, 40].
4.2 Transmission Mechanism

From the beginning, the voice data in a telephone network had been transmitted
in real time using circuit switching techniques. Since circuit switching is not
an efficient transmission mechanism for the communication of non-voice data,
packet switching techniques have been devised.
4.2.1 Packet Switching
Transmission mechanisms based on packet switching allow the multiplexing of

data packets from different sources on the same transmission path thereby mak-
ing use of the channel bandwidth more efficiently. It also allows the transmission
of packets from one source along different paths thus taking care of line conges-
tion and availability problems. When the packets reach the intended destination
in whatever path they may have taken, they are reassembled and presented to
the user application.
The most widely used protocols which manage the packet transfer across a
data network belong to the family of standards conceived by the ITU-T1 (X.25)
and Internet Architecture Board (IP). However the variable packet lengths and
the multiplexing technique introduce jitter making their Quality of Service
(QOS) unacceptable for real time applications. The inherent limitation of X.25
in high speed data transfer has been removed to some extent in the Frame Relay
[37] transfer mechanism.
4.2.2 Frame Relay
Frame Relay offers a high speed version of packet switching and has the potential
of operating effectively at much higher speeds compared to X.25, reaching speeds
of 45 Mbps. Frame relay is well suited to high speed data applications, but not
suited for delay sensitive applications such as voice and video because of the
variable length of frames [37, 38].
4.2.3 Cell Switching
Cell Relay is a transmission mechanism that combines the benefits of time di-
vision multiplexing with packet switching. It operates on the packet switching
principle of statistically interleaving cells on a link, on an ‘as required basis’,
rather than on the basis of a permanently allocated time slot. The fixed cell size
used enables a reasonably deterministic delay to be achieved across a network.
This deterministic nature of cell relay guarantees Quality Of Service (QOS) and
hence makes it suitable for all traffic types including real time traffic, within a
single network.
ATM (Asynchronous Transfer Mode) has become the dominant form of cell re-
lay. It uses a cell of 53 bytes long (5 bytes header and 48 bytes data) and typically
operates at speeds of 155 and 622 Mbps. ATM is delivered to the desktop at 25
Mbps and Gbps platforms are being tested [25]. ITU-T has selected ATM as the
transport technique for B-ISDN (Broadband Integrated Services Digital Network)
[10, 17, 38].
1
formerly CCITT – Consultative Committee for International Telephony and Telegraphy
9
4.3 Physical Media

The most commonly used physical media for wired connections are:
• Optical fibre cables

• Coaxial cables
• Unshielded Twisted Pair (UTP) cables
• Shielded Twisted Pair (STP) cables
Today all of the above media are used to carry data in excess of 100 Mbps
speeds. Only the distances they cover are different (for example, optical fibre can
operate at gigabit speeds for a few km whereas UTP can operate at 100 Mbps
over a distance of 100 m without repeaters).
With the spread of handheld mobile communication devices such as PDAs
and 3G mobile phones there is a rapidly growing demand for the delivery of
high bandwidth services to these devices. This is reflected in the recent work on
wireless standards carried out by various standard bodies and industry consortia
[13].
5 Internetworking
By internetworking is meant the process of interconnection of computers and
their networks to form a single internet. In other words to make two or more
physical networks to logically look and work like one. The Internet (note the
uppercase ‘I’) is such an internet and uses TCP/IP (Transmission Control Proto-
col /Internet Protocol) protocol suite [45].
TCP is connection oriented and provides a reliable stream transport service.
Although TCP is commonly associated with IP (as its underlying network proto-
col), it is an independent, general purpose protocol that can be adapted to use
with other delivery systems. The popularity of TCP has resulted in the develop-
ment of ISO – TP4, which is a TCP derivative. TCP, together with IP, provides a
reliable stream delivery for data traffic.
TCP/IP protocol suite has become the de facto standard for open system in-
terconnection in the computer industry. It is used world wide in academic,
government, private and public institutions. Some of the reasons for its wide
acceptance can be attributed to the following:
• It provides the highest degree of interoperability.

• It encompasses the widest set of vendors’ systems.
• It runs over more network technologies than any other protocol suite.
Over the years Unix (and hence Linux) and TCP/IP have become almost syn-
onymous. It has now become part of the operating system kernel. The OS runs
a separate process for IP, TCP input/output and UDP input/output.
The proliferation of TCP/IP is such that today there is hardly a single hard-
ware/software platform that does not support TCP/IP.
5.1 Building Internets and Intranets

In building internets and intranets, following hardware devices used to intercon-
nect networks are of major interest [3, 6].
1. Repeaters
2. Bridges and Switches
3. Routers and Brouters
4. Gateways.
Figure 9 shows the use of these components in relation to the ISO-OSI Refer-
ence Model.
Application Gateway Application
Presentation Presentation
Session Session
Transport Transport
Network Router Network
Data link Data link

Bridge
Physical Repeater Physical
Figure 9: Interconnection Components
5.2 Repeaters
When two networks are to be connected at the lowest level ie. the physical
level, an interconnecting device known as a repeater is used. A repeater simply
takes bits arriving from one network and just forwards (repeats) them on to the
other. In some cases a repeater might have to translate between two different
physical layer formats, for example from optical fibre to UTP (Unshielded Twisted
Pair) cable. This may involve some processing such as signal regeneration of
the received signal for noise elimination. However, repeaters pass on the data
received without paying attention to the address information.
11
5.3 Bridges and Switches

Bridges and switches are used to interconnect networks at the medium access
control (MAC) layer. Typically this requires the interconnected networks to have
identical MAC layers although networks with different but related MAC layers
can be interconnected. Since Bridges operate at the MAC layer, they can be used
to effectively segment the traffic by filtering the traffic entering one segment from
another thereby reducing the unwanted traffic flow on network segments.
Unlike a repeater which replicates electrical signals, bridges replicate packets.
They are superior in their function, because they do not replicate noise and
errors or malformed frames. Moreover, bridges implement CSMA/CD rules and
hence collisions and propagation delays remain isolated without affecting the
other segments. In other words, every port on an Ethernet bridge or a switch
is on a separate Collision Domain or physical network. As a result an almost
arbitrary number of Ethernet segments can be connected together with bridges
whereas with repeaters the maximum number of 10base5 segments is five giving
a total length of 2.5 km. Since bridges hide details of interconnection, a set of
bridged segments acts like a single Ethernet.
A bridge can be used to make a decision on which frames are to be for-
warded from one segment to another. Such bridges are called adaptive or learn-
ing bridges. They learn over time, about which hosts are connected to which
segment. Thus an adaptive bridge builds up the address table automatically
without human intervention.
Bridges are often used to improve the performance of an overloaded network
by effectively partitioning the network into segments.
The type of bridges used in CSMA/CD LANs are known as transparent bridges
(IEEE 802.1D) since their presence is not visible to the stations. The type of
bridges used to interconnect token rings are called source routing bridges (IEEE
802.5) and the routing information is provided by the source station.
5.4 Routers
Routers interconnect networks at the network layer (level 3 in the OSI-RM and
the IP layer of the TCP/IP suite) and perform routing functions. This is the main
building block in internetworking, that is building internets or intranets, using
IP. Most routers now support at least one of the multicast routing protocols,
which is an essential functionality to support the delivery of Real Time data over
IP.
For performance reasons the functionality of a bridge and a router is com-
bined to form a brouter. Whenever possible a brouter bridges data for better
efficiency rather than routing.
5.5 Gateways
Gateways are used when networks based on completely different network archi-
tectures have to be interconnected at layers higher than the network layer and
when protocol translation is necessary. A typical example is interconnecting two

networks based on TCP/IP suite and OSI suite of protocols. An application level
gateway is required to support FTP on TCP/IP and FTAM on OSI. From this it
should be clear that a different application level gateway is required for every
application supported across the interconnected networks.
5.6 Multiport-Multiprotocol Devices

The multiplicity of transmission media and protocols used in today’s networks
require the use of multiport repeaters and bridges, as well as multiprotocol
routers which support more than one protocol stack.
6 A word about the Internet

Internet Protocol is a truly scalable protocol that is used to connect computers
to small LANs and to WANs forming a global internetwork. IP is available for
almost all computing platforms ranging from the smallest palmtop to the largest
ultrasuper computer. Nowadays IP is increasingly becoming available in hand-
held mobile phones and personal data assistants (PDAs) and also on stand alone
micro–controller boards and embedded systems.
The Internet is an ever expanding network of networks. As of this writing
(November 2004), it interconnects an estimated 200 million host computers and
2 million computer networks in more than 200 countries. It experiences a stag-
gering growth rate of 100% per year. Figure 10 shows the Internet connectivity
in the world in 1997. Today, as a result of the emergence of various transmission
technologies including optical fibre, satellite and enhanced communication tech-
nologies based on robust high speed modems and DSL (Digital Subscriber Line),
almost all the countries in the world have some Internet connectivity. Although
it is difficult to keep track of the growth of the Internet in real time any more,
Figures 11 and 12 serve to show the scale of world wide Internet deployment.
The DSL, specially ADSL (Asynchronous Digital Subscriber Line) is an innova-
tive modem based communication technology that allows the use of the ordinary
telephone line to enjoy the full multimedia capability of the Internet at home de-
livered at a reasonably high speed (a download speed of ∼1 Mbps compared to
56 Kbps using an ordinary modem).
The world wide popularity of the Internet is largely due to the World Wide Web.
This is the most pervasive application on the Internet today. Figure 11 shows
the staggering growth in the number of web hosts in the world. Over a period
of five years this number has risen from near zero to over 35 million (in March
2002). Also this number has more than doubled in a year in the initial period.
Figure 12 shows the explosive growth of the Internet, measured in terms of
the number of Internet hosts connected to the Internet.
Both these show evidence of exponential growth in the usage of the Internet.
This has required some rethinking of the Internet strategy, specially in relation
13
INTERNATIONAL CONNECTIVITY
Version 16 - 6/15/97
Internet
Bitnet but not Internet Copyright © 1997
Larry Landweber
EMail Only (UUCP, FidoNet) and the Internet Society.
Unlimited permission to
No Connectivity copy or use is hereby granted
subject to inclusion of
This map may be obtained via anonymous ftp this copyright notice.
from ftp.cs.wisc.edu, connectivity_table directory
Figure 10: Worldwide Internet Connectivity
Figure 11: Growth of the number of Web hosts on the Internet
to security, address space, quality of service, etc, in order to make it sustainable

in the face of this rapid growth.
Due to this unprecedented popularity and the growth of usage of IP, today IP
suite is included as a standard component of all Unix and Unix like (and hence
Linux) distributions, making the networking of any computer running Linux much
easier. IP is also incorporated into the OSI-RM thus ensuring compliance with
ISO standards.
Figure 12: Growth of the number of Internet hosts
7 Internet Protocol Architecture

By far the most successful and the most widely used protocol architecture for
LANs and WANs alike is the Internet Protocol (IP) [3, 6, 43].
IP is a layered architecture (see Figure 5 on page 5) consisting of only 4 layers
(ISO-OSI RM has 7!). TCP (Transmission Control Protocol) and UDP (User Data-
gram Protocol) are the two transport protocols supported over IP. Both TCP and
UDP make the assumption that the link is of acceptable reliability (measured in
terms of its BER - Bit Error Rate) and is capable of delivering a packet to the
intended destination without the intervention of the upper layers. Some of the
services supported in TCP/IP, including RTP are shown in Figure 13.
1111
0000
0000
1111
0000
1111
HTTP FTP SMTP DNS SNMP RTP
TCP UDP
IP
Physical Link
Figure 13: Some Services Supported over IP
In IP there are two transport layer protocols available to programmers (see

section 11 on Networking Programming – page 60):
TCP - connection oriented, stream oriented protocol providing a reliable data
flow between computers
UDP - connectionless protocol which sends independent packages (called
Datagrams) with no guarantees of arrival.
The current version of IP is version 4 (IPv4) and work is underway to define
the next generation IP (IPng). Already IPv6 testbeds are operational in many
countries providing useful experience with this new protocol which is expected
to replace IPv4.
15
7.1 IP Addressing
IPv4 uses 5 classes of addresses as shown in Figure 14. These map on to IP
address ranges shown in Figure 15.
0 12 3 4 7 15 23 31
Class A 0 netid hostid
Class B 10 netid hostid
Class C 1 10 netid hostid
Class D 1 11 0 multicast addresses
Class E 1 11 1 0 reserved for future use
Figure 14: IP Addresses
Address Network No of No of
Address Range
Class ID Networks Hosts
Class A 1.0.0.0 − 126.255.255.255 1 − 126 126 16,777,214
Class B 128.0.0.0 − 191.255.255.255 128 − 191 16,384 65,534
Class C 192.0.0.0 − 223.255.255.255 192 − 223 2,097,151 254
Figure 15: IP Address Ranges
An example of IP address allocation in a typical network is shown in Figure

16.
Many hosts have a host name associated with the IP address which can be
used to address them. However it must be understood that an IP address does
not identify a host. It identifies a network connection to a host because an IP
address encodes both a network and a host on that network. Hence if the host
is moved to another network, its IP address has to be changed.
No two devices on the global Internet should be allocated the same IP address,
although on a private network with no connection to the outside world, arbitrary
addresses can be allocated. The address allocation has been initially handled by
the Network Information Centre (NIC). Now it is handled by the NIC as well as
RARE in Europe and APNIC in the Asia- Pacific region. The addresses must be
officially obtained from one of the above before using them on your network.
Nowadays there are Internet Service Providers (ISPs) in many countries (most
of the time more than one!) who are allocated blocks of IP addresses by the
relevant NIC. These ISPs in turn allocate the IP addresses to their customers.
The explosive growth of the Internet has resulted in the near exhaustion of
the address space (a total of 4 giga addresses) provided in IPv4 which uses a 32
bit (4 byte) addressing scheme. IPng (IP new generation), also known as IPv6, is
designed to provide among other things an enhanced address space (256 x 1036
addresses) using 16 byte (128 bit) long addresses [30]1 , [19, 14, 24].
1
RFCs are available from ftp://rs.internic.net/rfc
Ethernet 128.10.0.0
128.10.2.3 128.10.2.8 128.10.2.70 128.10.2.26
Ethernet Ethernet Ethernet

T/R G
Host Host
Host
192.5.48.3 192.5.48.7
T/R Token Internet

G
Host Ring 192.5.48.6 10.2.0.37
10.0.0.0
192.5.48.1
192.5.48.0
One address for each network / segment

Internet 10.0.0.0 ( Class A )
Ethernet 128.10.0.0 ( Class B )
Token Ring 192.5.48.0 ( Class C )
Network Hosts
128.10.2.8
192.5.48.1
128.10.2.3
192.5.48.3
Two addresses for each gateway.
Figure 16: Allocation of IP Addresses on an internet: an Example

17
7.2 Resolving IP Addresses

An IP address uniquely identifies a host (machine) on a network anywhere on the
Internet. Most of the time these IP addresses (in dotted decimal quad format) are
associated with a host name (or IP name) which is more convenient to human
beings but not directly useful to the machines themselves.
The Internet provides DNS (Domain Name System – a hierarchical data base
of host names and IP addresses) service which maps a given host name to its IP
address (address resolution) and an IP address to its host name (reverse lookup).
The Unix utility nslookup converts host names to IP addresses and vice versa.
7.3 The Internet Protocol

The Internet Protocol is a connectionless network layer protocol. TCP is a higher
layer protocol which sits on top of IP layer. IP provides a connectionless (or
datagram type) service with Best Effort Delivery to its user. In other words data
given to IP (by the layer above) is not guaranteed to be delivered.
The IP datagrams are the encapsulation of data packets passed from the
higher layer with an IP header as shown in Figure 17.
TCP Header Data
IP Header Data
Figure 17: IP Datagram
7.4 Internetworking with IP

Figure 18 shows how IP is used to internetwork two networks running different
medium access control mechanisms namely, CSMA/CD and Token Ring.
7.5 IP Datagram Format

The IP datagram format which has a preamble of ten 16 bit words and an option
field of variable length is shown in Figure 19.
• Version: 4 bits; current version is IPv4

• IHL: 4 bits; Internet Header Length in 32 bit words. The minimum IHL for
a valid header is 5.
• Type of Service: 8 bits; indicates the type and quality of service desired. The
first 3 bits are used to define 8 different types including Network Control,
Internetwork Control, etc, and the next 3 bits are used to set the Delay
(Normal/Low), Throughput (Normal/High) and Reliability (Normal/High).
The last 2 bits are reserved for future use.
USER A USER B
TCP TCP
IP IP IP
LLC LLC LLC LLC
CSMA/CD CSMA/CD Token Ring Token Ring

MAC MAC MAC MAC
Ethernet Ethernet Token Ring Token Ring

NET 1 NET 2
Figure 18: Internetworking Using IP
• Total Length: 16 bits; is the length of the datagram including header and
data.
• Identification: 16 bits; a value assigned by the sender to aid in assembling
the fragments of a datagram.
• Flags: 3 bits; Bit 0 is reserved and must be 0. Bit 1 is used to set fragment-
ing policy (May/Don’t) and Bit 2 to indicate either last fragment or more
fragments to follow.
• Fragment Offset: 13 bits; indicates where in the datagram this fragment
belongs. The first fragment has zero offset.
• Time To Live: 8 bits; sets the maximum time the datagram is allowed to
remain in the internet. Prevents continuous looping of datagrams.
• Protocol: 8 bits; indicates the Transport Layer Protocol that the data portion
of this datagram is passed to. The values for various protocols were origi-
nally specified in the Assigned Numbers RFCs until October 1994 [29], but
now they are available in an on-line database1 (RFC 3232 [36]).
• Header Checksum: 16 bits; for header only. Since some header fields
change (eg. TTL), this is recomputed and verified at each point that the
header is processed.
• Source Address: 32 bits
• Destination Address: 32 bits
• Options: variable; may or may not appear in a datagram.
1
http://www.iana.org
19
# of 16 bit
16 bits words
Internet
Version (4b) Header Length Type of Service (8b) 1
IHL (4b)
Total Length (16b) 1
Identification (16b) 1
Flags (3b) Fragment Offset (13b) 1
Time to Live (8b) Protocol (8b) 1
Header Checksum (16b) 1
Source Address (32b) 2
Destination Address (32b) 2
Options (varibale) variable
Padding(variable)
Figure 19: IP Datagram Format
7.6 Brief Description of TCP and UDP

TCP and UDP are the two transport protocols used in the IP architecture [6, 43].
TCP is a connection oriented transport protocol designed to work in conjunc-
tion with IP. TCP provides its user (application layer) with the ability to transmit
reliably a byte stream to a destination and allows for multiplexing multiple TCP
connections within a transmitting or receiving host computer.
Being connection oriented, TCP requires a connection establishment phase
(like dialling a number to make a phone call) which is followed by the data trans-
mission phase. A connection is terminated when it is no longer in use. TCP/IP
is ideal for the transmission of bursty data. It works on the principle of retrans-
mission of dropped packets which is one of the major contributors to delays in
transmission. However, since voice and video data are time sensitive, packet
technologies such as TCP/IP cannot guarantee the proper delivery of such data.
Figure 20 shows the TCP header format.
• Source Port: 16 bits
• Destination Port: 16 bits
• Sequence Number: 32 bits; is the sequence number of the first data octet
in this segment (except when SYN is present). If SYN is present then the
sequence number is the Initial Sequence Number (ISN) and the first data
octet is ISN+1.
• Acknowledgement Number: 32 bits; If the ACK control bit is set, this field
contains the value of the next sequence number the sender of the segment
is expecting to receive.
• Data Offset: 4 bits; is the number of 32 bit words in the TCP header. This
indicates where the data begins. The TCP header is always an integral
number of 32 bits long.
32 bits
Source Port (16 bits ) Destination Port (16 bits )
Sequence Number (32 bits)
Acknowledgement Number (32 bits)

Data U A P R S F
Reserved
Offset R C S S Y I Window (16 bits)
(6 bits)
(4 bits) G K H T N N
Checksum (16 bits) Urgent Pointer (16 bits)
Options (variable) Padding
Data (variable)
Figure 20: TCP Segment Format
• Reserved: 6 bits; for future use, must be zero.

• Control bits: 6 bits (single value bits).
– URG: Urgent Pointer

– ACK: Acknowledgement
– PSH: Push function
– RST: Reset the connection
– SYN: Synchronise sequence numbers
– FIN: No more data from sender
• Window: 16 bits; the number of data octets the sender of this segment
willing to accept.
• Checksum: 16 bits; is the checksum for header and data.
• Urgent Pointer: 16 bits; contains the current value of the urgent pointer as
a positive offset from the sequence number in this segment. It points to the
sequence number of the octet following the urgent data. This field is only
interpreted in segments with the URG control bit set.
• Options: variable
UDP, on the other hand, is a connectionless transport protocol designed to

operate over IP. Its primary functions are error detection and multiplexing. UDP
does not guarantee the delivery of packets (compare with the ordinary postal
service) but guarantees that if a packet is ever delivered in error, such error will
be detected (use of checksum). It also allows for communicating with multiple
processes residing on the same host computer.
UDP packet format is simple (see Figure 21). It is also fast compared to the
use of TCP, since there is no connection establishment phase. Moreover, UDP is
important since RTP (Real time Transport Protocol) is supported over UDP.
21
32 bits
Source Port Destination Port UDP

Header
Length Checksum
Data
Figure 21: UDP Packet Format
7.7 Client/Server Computing

A computer on a network usually has a single physical network connection but
can have more than one application or network service using the network simul-
taneously to send and receive data from other applications/services resident in
computers connected onto the network. In effect the network protocol architec-
ture provides a form of multiplexing/demultiplexing of data streams related to
different applications or network services.
Network drivers which form part of the network operating system must ensure
that each packet of data or message sent to the destination computer arrives at
the intended application. This is achieved by assigning a Port which is a unique
identifier for an application that uses the network.
App 1 App 2 App 3 App n
Port # 1 Port # 2 Port # 3 Port # n
TCP/UDP
TCP/UDP data packet
Port # Data
Figure 22: Use of Port Numbers
The IP datagram carries the machine identification (its IP address) and the
TCP segment or UDP packet as data. The TCP segment or the UDP packet car-
ries the Port Numbers (16 bits long) and the data for the application (Figure
22). On receipt of the TCP segment, the machine (already identified by its IP
address) directs the data to the correct application/service associated with the
port number attached to that application/service.
7.7.1 Port Numbers

When IP layer passes incoming data to the transport protocol layer (TCP), the TCP
passes the data to the correct application process. The application processes,
also known as network services, are identified by 16 bit long port addresses.
HTTP
port 80
TCP
protocol 6
Internet Protocol (IP)
address 140.125.12.167
IP
Header 6
140.125.12.167
TCP 80
Header
32 bits
Figure 23: Mapping of Protocol and Port Numbers
Once data are routed through the network and delivered to a specific host,
a mechanism must exist to deliver the data to the correct user application or
process. This mechanism needs to be able to deliver data to the correct protocols
in each layer as the data moves up or down the layers of the TCP/IP stack. This
mechanism must also be able to combine data from many applications into a few
transport protocols and from the transport protocol to the Internet Protocol. In
other words the many data streams must be multiplexed into a single data stream
before transporting on the network and the data arriving from the network must
be demultiplexed.
IP uses protocol numbers (field 9 of the IP datagram header – see figure 19
on page 19) to identify transport protocols and the transport protocols use port
numbers (fields 1 and 2 of the TCP segment header – see figure 20 on page 20)
to identify applications.
Figure 23 shows how the IP address, protocol number and the port number
are extracted from the IP datagram header and the TCP segment header.
Some protocol and port numbers are reserved to identify well known services
such as HTTP and FTP. These are documented in the Assigned Numbers RFC.
On a Unix (Linux) system the protocol numbers are defined in /etc/protocols file.
The first few lines of this file from a Linux machine is shown below:
# /etc/protocols:
# $Id: net.tex,v 1.1.1.1 2003/11/14 15:06:22 rtime Exp $
#
# Internet (IP) protocols
#
# from: @(#)protocols 5.1 (Berkeley) 4/17/89
#
23
# Updated for NetBSD

# based on RFC 1340, Assigned Numbers (July 1992).
ip 0 IP # internet protocol, pseudo protocol number

icmp 1 ICMP # internet control message protocol
igmp 2 IGMP # Internet Group Management
ggp 3 GGP # gateway-gateway protocol
ipencap 4 IP-ENCAP # IP encapsulated in IP (officially ‘‘IP’’)
st 5 ST # ST datagram mode
tcp 6 TCP # transmission control protocol
egp 8 EGP # exterior gateway protocol
pup 12 PUP # PARC universal packet protocol
udp 17 UDP # user datagram protocol
hmp 20 HMP # host monitoring protocol
On Unix systems the port numbers are defined in /etc/services file. The 16
bits gives rise to 64 K port addresses. Port numbers below 256 are reserved
for well known services such as FTP and HTTP. Port numbers between 256 and
1024 are used for Unix-specific services (such as rlogin, which is no longer Unix
specific).
The following is a few lines from /etc/services file:
# /etc/services:
# $Id: net.tex,v 1.1.1.1 2003/11/14 15:06:22 rtime Exp $
#
# Network services, Internet style
#
# Note that it is presently the policy of IANA to assign a single
# well-known port number for both TCP and UDP; hence, most entries
# here have two entries even if the protocol doesn’t support UDP
# operations. Updated from RFC 1700, ‘‘Assigned Numbers’’
# (October 1994). Not all ports are included, only the
# more common ones.
tcpmux 1/tcp # TCP port service multiplexer
echo 7/tcp
echo 7/udp
daytime 13/tcp
daytime 13/udp
ftp 21/tcp
fsp 21/udp fspd
ssh 22/tcp # SSH Remote Login Protocol
ssh 22/udp # SSH Remote Login Protocol
telnet 23/tcp
# 24 - private
smtp 25/tcp mail
# 26 - unassigned
time 37/tcp timserver
time 37/udp timserver
whois 43/tcp nicname
mtp 57/tcp # deprecated
finger 79/tcp
www 80/tcp http # WorldWideWeb HTTP
www 80/udp # HyperText Transfer Protocol
link 87/tcp ttylink
kerberos 88/tcp kerberos5 krb5 # Kerberos v5
kerberos 88/udp kerberos5 krb5 # Kerberos v5
Port numbers are unique only within a specific transport protocol. In other
words both TCP and UDP can, and do, assign the same port numbers. It is
combination of protocol and port number that uniquely identifies the specific
application and provides all of the information necessary to deliver the data to
the correct application.
7.7.2 Sockets
In addition to the use of well known port numbers, the operating system can use
dynamically allocated ports. As the name implies, these are not pre-assigned
and are assigned to processes when needed. The operating system ensures that
the same port number is not simultaneously assigned to another application
and that the numbers assigned are above the range of standard port numbers
reserved for well known services.
The dynamically assigned ports provide the flexibility needed to support con-
current multiple users. Once a service on a well known port is assigned to a user
the second user requesting the same service cannot be assigned the same port
number. In order to overcome this problem, the operating system dynamically
allocates a port number for the source port but uses the well known port number
for the destination port.
The dynamically allocated port number in conjunction with the IP address
is used to uniquely identify a single network process within the entire Internet.
This combination of an IP address and a port number is called a socket.
A socket is one endpoint of a two way logical communication link between
two programs associated with a client and a server. A socket is bound to a port
number so that the TCP layer can identify the application that data are destined
to be sent. The client and the server can reside on two different machines or on
the same physical machine.
The server which listens on a socket bound to a well known port (pre assigned)
responds to a client requesting a connection through that port by creating an-
other socket bound to a different port number. The server then informs the client
of this port number and the client communicates with the application running
on the server through this port (Figure 24). The server continues to listen on
25
the first socket/port. At the end of communication session the sockets/ports are
closed down.
P connection
Server o Client
r request
t
P
o
Server r
t
Port
Two way Client−Server P
o Client
Communication Channel r
t
Figure 24: Socket Creation
7.8 IP Multicasting
Multicasting is important in allowing a stream of data to be sent efficiently to

many receivers. In multicasting, rather than sending a separate stream of data
packets to each intended user (unicasting) or transmitting all packets to everyone
(broadcasting), a stream is transmitted simultaneously to a designated subset of
network users. The concept of multicasting is shown in Figure 25.
IP Multicasting is defined in RFC 1112 [27] of 1989. A key to IP multicasting
is the Internet Group Management Protocol (IGMP) specified in RFC 2236 [34]
which updates RFC 1112. IGMP enables users to sign up for multicast sessions
and allows these multicast groups to be managed dynamically, in a distributed
fashion. Enhancements have been made to existing protocols to direct the traf-
fic to the members of the group. These include Distance Vector Multicasting
Routing Protocol (DVMRP – RFC 1075) and Multicast Open Shortest Path First
(MOSPF – RFC 1585) protocol. An entirely new protocol developed specifically for
multicasting is the Protocol Independent Multicast (PIM – RFC 2362).
At the start of a multicast session group addresses are allocated which are
relinquished at the end of that session and reused later.
Internet MBONE is Internet’s multicast backbone which is a collection of mul-
ticast routers that can distribute multicast traffic. MBONE participants use
class D Internet addresses which identify a group of hosts rather than individual
hosts.
Recipients
Router
Router Router
Router
Recipient
Multicast Server
Figure 25: Multicasting
7.9 Resource Reservation Protocol – RSVP

A key factor in achieving real time quality of service is a reservation set up pro-
tocol, a mechanism for creating and maintaining flow specific state information
in the end point hosts and in routers along the data flow path. The IETF has
developed its Resource Reservation Protocol (RSVP) specifically for the packet
switched multicast environment (RFC 2205 [33]).
RSVP has been designed to meet a number of requirements:
1. support for heterogeneous service needs;

2. flexible control over the way reservations are shared along branches of mul-
ticast delivery trees;
3. scalability to large multicast groups;
4. and the ability to preempt resources to accommodate advance reservations.
The RSVP protocol basically acts according to its name. An RSVP request
specifies the level of resources to be reserved for some or all of the packets in a
particular session. An application requests resources by specifying a flow speci-
fication, which describes the type of traffic anticipated (for example, average and
peak bandwidths and level of burstiness), and a resource class specifying the
type of service required (such as guaranteed delay). A filter specification is also
specified, which determines the sources to which a given reservation applies.
RSVP mandates that a resource reservation be initiated by the receiver rather
than the sender. While the sender knows the properties of the traffic stream it
is transmitting, it has been found that the sender initiated reservation scales
poorly for large, dynamic multicast delivery trees. Receiver initiated reservation
27
deals with this by having each receiver request a reservation appropriate to it-
self; differences among heterogeneous receivers are resolved within the network
by RSVP. After learning sender’s flow specification via a higher level “out of band
mechanism”, the receiver generates its own desired flow specification and prop-
agates it to senders, making reservations in each router along the way.
RSVP itself uses a connectionless approach to multicast distribution. The
reservation state is cached in the router and periodically refreshed by the end
station. If the route changes, these refresh messages automatically install the
necessary state along the new route.
RTP and RTCP information is simply data from the point of view of routers
that move the packets to their destination. RSVP prioritises multimedia traffic
and provides a guaranteed quality of service. Routers that have been upgraded
to support RSVP can reserve carrying capacity for video and audio streams and
prevent unpredictable delays that would interfere with their transmission.
7.10 TCP/IP Over Serial Lines – SLIP and PPP

When data is sent over a serial link using TCP/IP some point to point data link
layer protocol is required to perform framing, error control and other data link
layer functions.
There are two protocols namely, SLIP (Serial Line Internet Protocol) and PPP
(Point to Point Protocol) that are used to communicate over dial up lines using
IP.
7.10.1 SLIP – Serial Line Internet Protocol
Around 1984, Rick Adams implemented SLIP for 4.2 Berkeley Unix and Sun Mi-
crosystems workstations and released it to the world. It quickly caught on as an
easy and reliable way to connect TCP/IP hosts and routers with serial lines.
The Serial Line Internet Protocol (SLIP), documented in RFC 10551 , is a
packet-framing protocol for relaying IP packets over dial-up lines. It defines
an encapsulation mechanism that frames IP packets on a serial line. There is
no support for dynamic address assignment, link testing, multiplexing different
protocols over a single link, packet type identification, error detection/correction
or compression mechanisms.
SLIP is commonly used on dedicated serial links and sometimes for dialup
purposes, and is usually used with line speeds between 1200 bps and 19.2 Kbps.
It is useful for allowing mixes of hosts and routers to communicate with one
another (host-host, host-router and router- router are all common SLIP network
configurations).
The SLIP protocol does not define any link control information that could be
used to dynamically control the characteristics of a connection. Therefore, SLIP
systems must assume certain link characteristics. Because of this limitation,
1
Nonstandard for transmission of IP datagrams over serial lines: SLIP
SLIP can only be used when both hosts know each other’s address, and only
when IP datagrams are being transmitted.
SLIP sends the datagram across the serial line as a series of bytes, and it uses
special characters to mark when a series of bytes should be grouped together as
a datagram. SLIP defines two special characters for this purpose:
1. The SLIP ”END” character, a single byte with the decimal value 192, is
the character that marks the end of a datagram. When the receiving SLIP
encounters the ”END” character, it knows that it has a complete datagram
that can be sent up to IP.
2. The SLIP ”ESC” character, a single byte with the decimal value of 219, is
used to ”escape” the SLIP control characters.
If the sending SLIP encounters a byte value equivalent to either a SLIP ”END”
character or a SLIP ”ESC” character in the datagram it is sending, it converts
that character to a sequence of two characters. The two-character sequences
are ”ESC 220” for the ”END” character, and ”ESC 221” for the ”ESC” character
itself. When the receiving SLIP encounters these two-byte sequences, it converts
them back to single-byte values. This procedure prevents the receiving SLIP from
incorrectly interpreting a data byte as the end of the datagram. Here ”ESC” refers
to the SLIP escape character, and not the ASCII escape character.
7.10.2 PPP – The Point to Point Protocol

The Internet uses PPP (point to point protocol – RFCs 1661, 1662 and 1663)
to carry traffic between routers and that between home users and ISPs (Internet
Service Providers). PPP is designed to handle among other things, error detection,
support for multiple protocols, IP address negotiation at connection time, and
authentication.
In order to achieve these PPP has the following features:
1. A framing method that unambiguously detects the end of one frame and the
start of the next one. The frame format also handles error detection.
2. A Link Control Protocol (LCP) for bringing lines up, testing them, negotiating
options and bringing them down gracefully when they are no longer needed.
LCP supports both synchronous and asynchronous circuit technologies and
byte oriented and bit oriented encodings.
3. A different Network Control Protocol (NCP) for each network layer sup-
ported. This allows to negotiate network layer options in a way that is
independent of the network layer protocol to be used.
The PPP frame format (see figure 26) closely resembles the HDLC (High level
Data Link Control) frame format, the major difference being that PPP is character
oriented whereas HDLC is bit oriented. In particular PPP uses byte stuffing on
dial up lines so all frames are an integral number of bytes. PPP uses the same
29
1 1 1 1−2 variable 2 or 4 1
Flag Address Control Protocol Payload Checksum Flag
01111110 11111111 00000011 01111110
Figure 26: PPP Frame Format
HDLC flags which are byte stuffed if they occur within the payload field. The
address field is always set to 11111111 to indicate that all stations are to accept
the frame thus avoiding the need to assign data link addresses.
PPP does not support reliable transmission using sequence numbers and ac-
knowledgements as the default. Hence the control field is set to 00000011 in-
dicating unnumbered mode. In noisy environments such as wireless networks,
RFC 1663 defines a reliable transmission using numbered mode that is rarely
used.
The default LCP settings allow two parties to negotiate an option to just omit
the transmission of Address and Control fields altogether thus saving 2 bytes per
frame.
PPP defines codes for the protocol field. Protocol codes starting with a 0 bit
are network layer protocols such as IP, IPX and OSI CLNP. Those starting with a
1 bit are used to negotiate other protocols including LCP and a different NCP for
each network layer supported. The default size of the protocol field is 2 bytes.
However, using LCP it can be negotiated down to 1 byte.
Each Network Control Protocol (NCP) on the other hand is specific to some
network layer protocol and allows configuration requests to be made that are
specific to that protocol.
In PPP the default payload length is 1500 bytes but can be negotiated up to a
maximum length using LCP during line set up. The checksum field is normally 2
bytes (16 bits) but a 4 byte (32 bit) checksum can be negotiated.
Most Linux systems include both SLIP and PPP. However, on some Unix sys-
tems such as Solaris, PPP is included and SLIP is not. As a simple rule use PPP
where you can and SLIP where you must.
8 IPv6 – The New Generation Internet Protocol

8.1 The Design of IPv6
IPv4 is a very robust design. If it had a major design flaw the Internet could not
have been so successful. However IPv6 is not a simple derivative of IPv4, but a
definitive improvement [19, 14, 24].
IPv6 is an evolutionary step forward from IPv4. Not only it fixes a number of
shortcomings in IPv4 such as the limited number of available addresses, it also
adds several improvements in such areas as routing and address auto configu-
ration, authentication and privacy capabilities, and IP mobility. IPv6 is expected
to eventually replace IPv4, but until such time both protocols will coexist. It is
therefore important that applications are developed to work transparently with

both IPv4 and IPv6 during this transition period.
8.2 IPv6 Header Format

Compared to IPv4 header format (Figure 19 on page 19) which has 10 fixed
header fields, 2 addresses and some options, the IPv6 header format shown in
Figure 27 with only 6 fields and 2 addresses is in fact much simpler. Conse-
quently the header processing is expected to be much more efficient in IPv6,
thereby cutting down the processing and reducing the transit delays in internet-
working devices.
0 3 7 15 31
Version Priority Flow Label
(4b) (4b) (24bits)
Payload Length (16 bits) Next Header Hop Limit
(8 bits) (8 bits)
Source Address
(128 bits)
Destination Address
(128 bits)
Figure 27: The IPv6 Header Format
8.3 Simplifications
The design of IPv6 has included three major simplifications (based on the expe-
rience gained in operating IPv4 for over 20 years!):
1. assign a fixed format to all headers
2. remove the header checksum
3. remove the hop by hop segmentation procedure.
IPv6 handles options in a different way (see Figure 28) compared to IPv4, ie.
in the form of extension headers and hence there is no need for header length
field in the IPv6 header.
31
IPv6 Header
Next Header = TCP Header + Data
TCP
IPv6 Header Routing Header

Next Header = Next Header = TCP Header + Data
Routing TCP
IPv6 Header Routing Header Fragment Header Fragment of

Next Header = Next Header = Next Header = TCP Header + Data
Routing Fragment TCP
Figure 28: Chaining of IPv6 Headers
IPv6 currently defines six extension headers:

1. Hop by hop options header – special options requiring hop by hop process-
ing
2. Routing header
3. Fragment header – fragmentation and reassembly
4. Authentication header – integrity and authentication
5. Encrypted security payload header – confidentiality
6. Destination options header – optional information to be examined by the
destination node.
In order to improve the performance when handling subsequent option head-
ers and the transport protocol which follows, IPv6 options are always an integer
multiple of 8 octets long. This also helps to retain the alignment for subsequent
headers.
The main advantage of removing header checksum is to diminish the cost of
header processing by removing the need to check and update the checksum at
each intermediate relay. This can however result in misrouted packets. Expe-
rience has shown that the risk is minimal since most encapsulation procedures
include a packet checksum (eg. MAC procedure of IEEE 802.X networks, in the
adaptation layers for ATM and in the framing procedure of the Point to Point
Protocol for serial lines).
The last simplification is the removal of the Type of Service (TOS) field. It
was found that although IPv4 provides TOS, this filed was hardly ever set by
applications.
8.4 New Fields

The Flow Label and Priority have been included mostly to facilitate the handling of
Real Time Traffic so as to ensure the proper treatment of high quality multimedia
communications in the new Internet. Flow labels will allow the stipulation of
severe real time constraints, for example.
8.5 Special Services

In traditional packet switching, a packet is queued at a switch (Figure 29) until
a line is available to carry it thereby giving rise to congestion and delay.
Figure 29: Traditional Packet Switching – One Queue Per Line
Assigning higher priority to the queue of real time packets over that of data
packets is not enough. It is necessary to ensure that the service rate is equal to
or higher than the arrival rate of packets. This can be achieved as follows:
R1
R2
R3 S
R4
data
Figure 30: Real Time Streams and Data in Separate Queues
If each real time queue is serviced at a rate compatible with its requirements
then it will never suffer from unpredictable queueing delays (see Figure 30). The
data queue, however, will only be serviced on a best effort basis.
The proposers of IPv6 suggest that the New Internet is capable of providing
all the services required by its users, including Real Time Audio and Video. It is
their opinion that ATM is ”Another Terrible Mistake”.
8.6 IPv6 Address Space

The address length of 128 bits gives rise to a total of 256 x 1036 different addresses
in its address space. However due to inefficiencies in address allocation and
33
administration (expressed by H Factor1 which has found generally to lie in the

range 0.22 to 0.26), the new Internet is expected to support 1015 (quadrillion)
hosts and 1012 (trillion) networks.
The notation for writing IPv6 addresses is (remember the dot notation in
IPv4!):
FEDC:BA98:7654:3210:FEDC:BA98:7654:3210
The notation allows to skip leading zeros; for instance
0000 can be written as just 0 and 0032 can be written as 32. Further the
notation also allows removing a 0 leaving the colons. Thus an address such
as 1030:0:0:0:80:3210:2c15:417a would become 1030::80:3210:2c15:417a. The
double colon notation can be used at the beginning or at the end of an address
but only once.
In the interim period an IPv4 address will be written as an IPv6 address by
prepending 12 octets of zeros giving 0:0:0:0:0:0:0:0:0:0:0:0:128.145.48.12. It is
also allowed to write this as ::128.145.48.12.
8.7 Making IPv6 Compliant

IPv6 will coexist with IPv4 for a number of years after which all computers will
run only IPv6. During this period it is suggested to have a dual protocol stack
consisting of IPv6 and IPv4 (Figure 31) running on every host that uses both sets
of protocols.
Application
TCP
IPv4 IPv6
Ethernet
Figure 31: Typical Dual Stack Configuration
8.7.1 IPv6 Tunnelling

IPv6/IPv4 dual protocol stack will also be implemented in routers thus facilitat-
ing communication between two IPv6 compliant computers at the two ends of
1
H Factor is defined by Huitema [19] as the ratio between log(number of addresses) and the
number of bits in the address
the IPv4 based Internet [24]. This is known as IPv6 tunnelling (see Figure 32).
Tunnel
Dual stack Dual stack

router router
IPv6 / IPv4 IPv4 /

IPv4 Network IPv6
IPv6 hosts IPv6 hosts
Figure 32: IPv6 Tunnelling over a IPv4 Network
9 Data Communication in Real Time

Packet switching techniques based on ITU-T X.25 and IP have been tradition-
ally used for non Real Time data transfer. Since they do not guarantee packet
sequence integrity and consistent latency times in delivery, they are inherently
unsuitable for Real Time applications.
The following are essentially required for carrying Real Time data on existing
networks.
1. Enough bandwidth for extremely dense audio and video traffic.

2. A transport protocol appropriate for the streaming requirements of real time
data (RTP).
3. A protocol to reserve network bandwidth and assigning priorities for various
types of traffic (RSVP).
9.1 RTP Data Transfer Protocol

A Real time Transport Protocol is therefore needed to provide end to end network
transport functions suitable for applications communicating in real time. Such
applications include transmission of interactive audio and video data or real time
simulation data over multicast or unicast network services.
The largest (and the oldest) network which supports real time data communi-
cation is the telephone network. This falls into the category of circuit switched
networks. They have been designed to support Plain Old Telephone Services
(POTS). However in terms of a network protocol there is not much. Once the
35
network connection is established the communication process is largely in the

hands of the two persons communicating with each other.
In view of this a Real time Transport Protocol (RTP) along with a profile for
carrying audio and video over RTP were defined by the IETF in January 1996
[31, 32].
9.1.1 Characteristics of RTP

The Realtime Transport Protocol has the following characteristics.
• Payload type identification

• Sequence numbering
• Time stamping
• Delivery monitoring.
Real Time applications typically run RTP on top of UDP to make use of its mul-
tiplexing and checksum services (see Figure 33). Tailoring RTP to the application
is accomplished through auxiliary profile and payload format specifications. A
payload format defines the manner in which a particular payload, such as an
audio or video encoding, is to be carried in RTP. A profile assigns payload type
numbers for the set of payload formats that may be used in the application.
Real time Server
Real time Application
RTP and RTCP
UDP
Router Router
RTP RTP RTP
RTCP
RSVP RSVP
RTCP control RSVP control
information information
RSVP
RTP and RSVP

RSVP - Resource Reservation Protocol
RTCP - Real time Transport Control Protocol
Application
RTP - Real time Transport Protocol
UDP - User Datagram Protocol Recipient
Figure 33: Real Time Application running RTP on top of UDP
However RTP is not limited to be used with UDP/IP. It can equally be used
with other underlying network or transport protocols such as ATM or IPX. More-
over, RTP supports data transfer to multiple destinations using multicast distri-
bution if provided by the underlying network, a feature which makes RTP ideal
for multi party multimedia conferencing.
RTP is designed to work in conjunction with RTCP (Real time Transport Con-
trol Protocol) to monitor the quality of service. RTP delivers real time traffic with
timing information for reconstruction as well as feedback on reception quality.
The Resource Reservation Protocol (RSVP) is used to reserve network bandwidth
and assign priority for various traffic types.
9.1.2 Definitions in RTP

The following definitions are extracts from the relevant RFCs [31, 32].
• RTP Payload
The data transported by RTP in a packet, for example audio samples or
compressed video data.
• RTP Packet
A data packet consisting of the fixed RTP header, a possibly empty list of
contributing sources, and the payload data. Typically one packet of the
underlying protocol contains a single RTP packet, but several RTP packets
may be contained if permitted by the encapsulation method.
• RTCP Packet
A control packet consisting of a fixed header part similar to that of RTP
data packets, followed by structured elements that vary depending upon
the RTCP packet type. Typically multiple RTCP packets are sent together
as a compound RTCP packet in a single packet of the underlying protocol.
This is enabled by the length field of the fixed header of each RTPC packet.
• Port
The “Abstraction” that transport protocols use to distinguish among mul-
tiple destinations within a given host computer (TCP/IP protocols identify
ports using small positive integers. The transport selectors (TSEL) used by
the OSI Transport layer are equivalent to ports). RTP depends on the lower
layer protocol to provide some mechanism such as ports to multiplex the
RTP and RTCP packets of a session.
• Transport Address
The combination of a network address and the port that identifies a trans-
port level end point, for example an IP address and a UDP port. Packets
are transported from a source transport address to a destination transport
address.
• RTP Session
The association among a set of participants communicating using RTP. For
each participant, the session is defined by a particular pair of destination
transport addresses consisting of one network address and a port pair for
RTP and RTCP. The destination transport address pair may be common
for all participants, as in the case of IP multicast, or may be different for
each, as in the case of individual unicast network addresses plus a common
37
port pair. In a multimedia session, each medium is carried in a separate

RTP session with its own RTCP packets. The multiple RTP sessions are
distinguished by different port number pairs and/or different multicast ad-
dresses.
• Synchronisation Source (SSRC)
The source of a stream of RTP packets, identified by a 32 bit numeric SSRC
identifier carried in the RTP header so as not to be dependent upon the
network address. Examples of synchronisation sources include the sender
of a stream of packets derived from a signal source such as a microphone,
a camera or an RTP mixer. If a participant generates multiple streams in
one RTP session, for example from separate video cameras, each must be
identified as a different SSRC.
• Contributing Source (CSRC)
A source of a stream of RTP packets that has contributed to the combined
stream produced by an RTP mixer. The mixer inserts a list of the SSRC
identifiers of the sources that contribute to the generation of a particular
packet into the RTP header of that packet. This is called the CSRC list.
An example application is audio conferencing where a mixer indicates all
the talkers whose speech was combined to produce the outgoing packet,
allowing the receiver to indicate the current talker, even though all the audio
packets contain the same SSRC identifier (that of the mixer).
• End System
An application that generates the content to be sent in RTP packets and/or
consumes the content of received RTP packets.
• Mixer
An intermediate system that receives RTP packets from one or more sources,
possibly changes the data format, combines the packets in some manner
and then forwards a new RTP packet. Since the timing among multiple
input sources will not generally be synchronised, the mixer will make tim-
ing adjustments among the streams and generate its own timing for the
combined stream. Thus all data packets originating from a mixer will be
identified as having the mixer as their synchronisation source.
• Monitor
An application that receives RTCP packets sent by participants in an RTP
session, in particular the reception reports, and estimates the current qual-
ity of service, fault diagnosis and long term statistics.
9.1.3 RTP Fixed Header Fields

The RTP header format is shown in Figure 34. The first twelve octets are present
in every RTP packet, while the list of CSRC identifiers is present only when
inserted by a mixer.
0 1 2 3 4 8 15 31
V=2 P X CC M PT sequence number
time stamp
synchronisation source (SSRC) identifier
contributing source (CSRC) identifier
Figure 34: RTP Header Format
The RTP header provides the timing information necessary to synchronise

and display audio and video data and to determine whether packets have been
lost or arrive out of order. In addition, the header specifies the payload type,
thus allowing multiple data and compression types. This is a key advantage over
most proprietary solutions, which specify a particular type of compression and
thus limit users’ choice of compression schemes.
9.1.4 Multiplexing RTP Sessions
In RTP, multiplexing is provided by the destination transport address (network

address and a port number) which define an RTP session.
9.1.5 Real time Transport Control Protocol (RTCP)
The operation of RTCP is based on the periodic transmission of control packets

to all participants in the session, using the same distribution mechanism as the
data packets. The underlying protocol must provide multiplexing of the data and
control packets, for example using separate port numbers with UDP.
The primary function of RTCP is to furnish information on the quality of data
distribution. This feedback is a critical part of RTP’s use as a transport proto-
col, since applications can use it to control how they behave. The feedback is
also important for diagnosing distribution faults. For instance, by monitoring re-
ports from all data recipients, network managers can determine the spread of a
problem. When used in conjunction with IP multicast, RTCP enables the remote
monitoring and diagnosis.
In addition RTCP controls the rate at which participants in an RTP session
transmit RTCP packets. In a session with a few participants, RTCP packets are
sent at the maximum rate of one every five seconds whereas for a larger group,
RTCP packets may be sent only once every 30 seconds. In other words, the more
participants there are in a conference, the less frequently each participant sends
RTCP packets. This makes RTCP scalable to accommodate tens of thousands of
users.
39
9.2 Real Time Data Transfer using ATM

Audio and video applications generate lots of bits, and the traffic has to be
streamed or transmitted continuously rather than in bursts. This is in con-
trast to conventional data types such as text, files and graphics, which are able
to withstand short and inconsistent periods of delay between packet transmis-
sions. What is needed then is a network capable of transporting both streaming
and bursty data. ATM (Asynchronous Transfer Mode) is a transmission tech-
nique which has been designed primarily to achieve this [10, 38].
ATM is a connection oriented protocol designed to support high bandwidth,
low delay (even services with predictable delay), packet like switching and mul-
tiplexing. The design of ATM ensures the capability to carry both stream traffic
(such as voice and video) and bursty traffic (such as interactive data) with guar-
anteed QOS. It uses a fixed cell size for all types of traffic. In the case of stream
traffic ATM guarantees the integrity of cell sequence which is essential for the
successful delivery of such traffic.
ATM has grown out of the need for a worldwide standard to allow interchange
of information regardless of the “end system” or type of information. Histori-
cally there have been separate methods used for the transmission of information
among users on LANs and the users on WANs. This situation has been made
more complex by the user’s need for connectivity expanding from the LAN to
MAN to WAN. ATM is a method to unify the communication of information on
LANs and WANs.
ATM is the only technology based on standards, and has been designed from
the beginning to accommodate the simultaneous transmission of data, voice and
video. It is an easily scalable backbone which can be upgraded merely by adding
more switches or links. ATM is switched instead of routed and therefore it
is faster since not every IP packet at every node is examined to determine its
destination. ATM LAN Emulation (LANE) allows transparent interconnection of
“legacy” LANs based on Ethernet or FDDI technology, making the ATM backbone
look like a fast Ethernet or FDDI to workstation applications. As more and more
ATM nodes are deployed, the differences between local and wide area networks
will disappear to form a seamless network based on one standard.
To use the limited bandwidth more efficiently, ATM uses circuit switching
principles to give the users a full channel to themselves. Since these users do
not use the full channel all the time, ATM uses statistical analysis to time division
multiplex several users onto the same line. This allows each user to have all of
the channel’s bandwidth for the period of time in which it is needed.
To achieve this ATM uses two connection concepts; the Virtual Channel (VC)
and the Virtual Path (VP).
A virtual channel (also known as a virtual circuit) provides a logical connec-
tion between end users and is identified by a VCI (Virtual Channel Identifier) in
the ATM header (see Figure 35). A virtual path defines a collection of virtual cir-
cuits traversing the same path in the network and is identified by a VPI (Virtual
Path Identifier). The VPI emulates the functions of the trunk concept in circuit
switching. Thus virtual paths define the cross connection functions across the
at the UNI at the NNI

GFC VPI VPI
VPI
VCI VCI
PTI CLP PTI CLP
HEC HEC
Payload Payload
GFC - Generic Flow Control HEC - Header Erorr Control

VPI - Virtual Path Identifier CLP - Cell Loss Priority
VCI - Virtual Channel Identifier UNI- User Network Interface
PTI - Payload Type Identifier NNI - Network Node Interface
Figure 35: ATM Cell Formats
Higher Layers Higher Layers
ATM Adaptation ATM Switch ATM Adaptation

Layer (AAL) Layer (AAL)
ATM Layer ATM Layer ATM Layer
Physical Layer Physical Layer Physical Layer
Physical Medium
Figure 36: ATM Protocol Architecture
network, whereas virtual channels are concerned with switching and connec-
tion establishment functions. Virtual paths are statistically multiplexed on the
physical link on a cell multiplexing basis.
GFC (Generic Flow Control) is used to control the amount of traffic entering
the network.
VPI and VCI are used for routing. VPI will change from one node to the next
when it travels through the ATM layer. VCI is predefined and usually remains
the same throughout the duration of the transmission. PTI (Payload Type Iden-
tifier) is used to distinguish between cells that are carrying user data and those
carrying control information. CLP is a single control bit which provides selective
discard during network congestion and HEC is used to check header errors.
The ATM cell formats used at the UNI (User-Network Interface) and NNI
(Network-Node Interface) are shown in Figure 35.
9.2.1 ATM Protocol Structure
Figure 36 shows the ATM layered architecture as described in ITU-T recommen-

dation I.321 (1992). This is the basis on which the B-ISDN Protocol Reference
Model has been defined.
41
• ATM Physical Layer

The physical layer accepts or delivers payload cells at its point of access to
the ATM layer. It provides for cell delineation which enables the receiver
to recover cell boundaries. It generates and verifies the HEC field. If the
HEC cannot be verified or corrected, then the physical layer will discard the
errored cell. Idle cells are inserted in the transmit direction and removed in
the receiving direction.
For the physical transmission of bits, 5 types of transmission frame adap-
tations are specified (by the ITU and the ATM Forum). Each one of them
has its own lower bound or upper bound for the amount of bits it can carry
(from 12.5 Mbps to 10 Gbps so far).
1. Synchronous Digital Hierarchy (SDH) ≥ 155 Mbps;

2. Plesiochronous Digital Hierarchy (PDH) ≤ 34 Mbps;
3. Cell Based ≥ 155 Mbps;
4. Fibre Distributed Data Interface (FDDI) = 100 Mbps;
5. Synchronous Optical Network (SONET) ≥ 51 Mbps.
The actual physical link could be either optical or coaxial with the possibility
of Unshielded Twisted Pair (UTP Category 3/5) and Shielded Twisted Pair
(STP Category 5) in the mid range (12.5 to 51 Mbps).
• ATM Layer
ATM layer mainly performs switching, routing and multiplexing. The char-
acteristic features of the ATM layer are independent of the physical medium.
Four functions of this layer have been identified.
1. cell multiplexing (in the transmit direction)

2. cell demultiplexing (at the receiving end)
3. VPI/VCI translation
4. cell header generation/extraction.
This layer accepts or delivers cell payloads. It adds appropriate ATM cell
headers when transmitting and removes cell headers in the receiving direc-
tion so that only the cell information field is delivered to the ATM Adaptation
Layer.
At the ATM switching/cross connect nodes VPI and VCI translation occurs.
At a VC switch new values of VPI and VCI are obtained whereas at a VP
switch only new values for the VPI field are obtained (see Figure 37). De-
pending on the direction, either the individual VPs and VCs are multiplexed
into a single cell or the single cell is demultiplexed to get the individual VPs
and VCs.
Virtual
channel
switches
VCIn VCIm VCIp
VCIm
Virtual
path
switches
VPIa VPIb VPIc VPId
VCIn VCIm VCIp

Virtual channel z connection
VPIa VPIb VPIc VPId

Virtual Virtual
Virtual path x connection
path w connection path y connection
Figure 37: VC/VP Switching in ATM
• ATM Adaptation Layer (AAL)

The ATM Adaptation Layer (AAL) is between ATM layer and the higher lay-
ers. Its basic function is the enhanced adaptation of services provided by
the ATM layer to the requirements of the higher layers.
This layer accepts and delivers data streams that are structured for use with
user’s own communication protocol. It changes these protocol data struc-
tures into ATM cell payloads when receiving and does the reverse when
transmitting. It inserts timing information required by users into cell pay-
loads or extracts from them. This is done in accordance with five AAL ser-
vice classes defined as follows.
1. AAL1 - Adaptation for Constant Bit Rate (CBR) services (connection

oriented, 47 byte payload);
2. AAL2 - Adaptation for Variable Bit Rate (VBR) services (connection ori-
ented, 45 byte payload);
3. AAL3 - Adaptation for Variable Bit Rate data services (connection ori-
ented, 44 byte payload);
4. AAL4 - Adaptation for Variable Bit Rate data services (connection less,
44 byte payload);
5. AAL5 - Adaptation for signalling and data services (48 byte payload).
In the case of transfer of information in real time, AAL1 and AAL2 which
support connection oriented services are important. AAL4 which supports
a connection less service was originally meant for data which is sensitive
to loss but not to delay. However, the introduction of AAL5 which uses a
48 byte payload with no overheads, has made AAL3/4 redundant. Frame
Relay and MPEG–2 (Moving Pictures Expert Group) video streaming are two
services which will specifically use AAL5.
43
9.2.2 ATM Services

• CBR Service
This supports the transfer of information between the source and destina-
tion at a constant bit rate. CBR service uses AAL1. A typical example is the
transfer of voice at 64 Kbps over ATM. Another usage is for the transport of
fixed rate video.
This type of service over an ATM network is sometimes called circuit emu-
lation (similar to a voice circuit on a telephone network).
• VBR Service
This service is useful for sources with variable bit rates. Typical examples
are variable bit rate audio and video.
• ABR and UBR Services
The definition of CBR and VBR has resulted in two other service types called
Available Bit Rate (ABR) services and Unspecified Bit Rate (UBR) services.
ABR services use the instantaneous bandwidth available after allocating
bandwidths for CBR and VBR services. This makes the bandwidth of the
ABR service to be variable. Although there is no guaranteed time of delivery
for the data transported using ABR services, the integrity of data is guaran-
teed. This is ideal to carry time insensitive (but loss sensitive) data such as
in LAN-LAN interconnect and IP over ATM.
UBR service, as the name implies, has an unspecified bit rate which the
network can use to transport information relating to network management,
monitoring, etc.
9.2.3 IP over ATM

The transmission of classical IP traffic over ATM can be accomplished as shown
in figure 38.
Application
TCP Layer
IP Layer
AAL
ATM Layer
Physical Layer
Figure 38: Transmission of IP over ATM

9.3 IP/TV – A Real life Example

IP/TVTM is a client server application that multicasts live or pre-recorded digital
video and audio streams in real time to an unlimited number of users over any
IP based local or wide area network including the global Internet, using fully
compliant TCP/IP protocol stacks supporting real time protocol components. It
uses state-of-the-art Internet standards such as IP multicasting, RTP, RTCP and
RSVP in its FlashwareTM software suite to provide high quality, synchronised au-
dio/video information over existing packet switched networks simultaneous with
current network data traffic.
Server Viewer 1 Viewer 2
IP/TV IP/TV IP/TV

or Viewer Viewer
MBONE
TCP/IP TCP/IP TCP/IP

with Audio / Video with with
Flashware Streams Flashware Flashware
IP based LAN/WAN
World Wide Web

Server
Instructions for IP/TV Programs

Program Guide
Multicast Programs
Figure 39: Real Time Audio/Video Over IP based LAN/WANs
IP/TV1 consists of a Viewer, a Program Guide and a Server (Figure 39). The
program guide shows a schedule of multicasts and can be accessed via a Web
browser with HTTP (Hyper Text Transport Protocol). MBONE session information
can be accessed with the Program Guide which controls the number of streams
allowed on the network and the format of those streams, ie. audio only, audio
and video or some other combination. The server delivers pre-recorded or live
multimedia streams based on the Program Guide schedule and parameters such
as start time and file name.
The IP/TV viewer, a tool for signing up for scheduled multicasts, is designed to
provide VCR like controls as well as “channel changing” controls. With a software
based codec (compliant with ITU–T video conferencing standard H.320/H.261)
colour video running at a rate of 30 frames per second uses about 500 Kbps
bandwidth. The use of IP multicasting helps to conserve network bandwidth by
transmitting over the network a single data stream that can be picked up by any
interested user. The use of RSVP provides the ability to reserve bandwidth on
RSVP compliant routers, thereby giving priority to time dependent audio/video
streams over less critical network traffic thus ensuring the desired Quality of
Service (QOS).
1
available with Flashware from Precept Software Inc.
45
9.4 Delivering Real Time Data to the Desk Top

9.4.1 Ethernet Based Solutions
Ethernet was invented by Dr Robert M Metcalfe in the 1970s at the Xerox PARC
(Palo Alto Research Centre). This was at a time when a network bandwidth of 3
Mbps was found to be adequate to serve the needs of the users whose computers
were connected to such networks. Formal specifications for Ethernet were pub-
lished in 1980 by a multi vendor consortium that created the DEC-Intel-Xerox
(DIX) standard and turned the 3 Mbps Ethernet into an open production quality
Ethernet operating at 10 Mbps. The popularity of Ethernet made the LAN Stan-
dards Committee of the Institute of Electrical and Electronic Engineers (IEEE
802) to adopt and publish this as an IEEE standard in 1985. The IEEE specifi-
cation which is based on the original DIX technology and provides an “Ethernet
like” system, is called the IEEE 802.3 Carrier Sense Multiple Access/ Collision
Detect (CSMA/CD) standard.
The original specification of 1985 specifies copper based transmission media,
specially, thick (10base5) and thin (10base2) coaxial cables. Since then it has
grown to include twisted pair (10/100baseT) and optical fibre (10/100baseFL). It
is estimated that there are about 500 million Ethernet connections in the world
today, making it the most widely deployed LAN technology.
However, the majority of today’s computing is based on client/server and peer-
to-peer (p2p) technologies. As more applications are moved from desktops to
severs, and the number of desktop clients increases, the demand placed on a
typical Ethernet operating at 10 Mbps tends to affect the application response
time and impair the users’ ability to effectively access, manipulate and transmit
information. Moreover the high performance servers of today need maximum
bandwidth and highly reliable network connections to make the most of their
powerful processing capabilities.
Clients, on the other hand, require high speed, low cost connections to the
high bandwidth networks. This asymmetric approach to client/server network
design has led to the development of Fast Ethernet (IEEE 802.3u – 100base..
family) and Gigabit Ethernet (IEEE 802.3z – 1000base.. family) specifications
giving a speed advantage of 10 and 100 times respectively, compared to the
original Ethernet. Associated with the high bandwidth is also the switching
technique, which makes the Ethernet and its variants to be appropriate for the
delivery of real time data.
9.4.2 Signal Topology and Timing
The signal topology of the Ethernet is also known as the logical topology to distin-
guish it from the actual physical layout of the media cabling. The logical topology
of an Ethernet provides a single channel (bus) that carries signals to all stations
attached to it. Repeaters which amplify and re-time the signals can be used to
link multiple Ethernet segments together to form large Ethernets.
The signal timing is based on the amount of time it takes to traverse the
complete media system from one end to the other and back. This is also called
the round trip time, the maximum of which is strictly limited to ensure that
every interface can hear all network signals within the specified amount of time
provided in the Ethernet Medium Access Control (MAC) mechanism.
The longer a given network segment is the more time it takes for a signal to
traverse it. Therefore configuration guidelines (rules) are provided to make sure
that the round trip timing limits are met, no matter what combination of media
segments are used in the network configuration.
However the expansion of Ethernet using repeaters is limited because of the
signal timing restrictions. In order to meet the expansion needs of today, two
kinds of hubs known as repeater hubs and packet switching hubs are available.
The advantage of using a switching hub is that each port of the hub operates on
full duplex mode and provides a connection to an Ethernet media system that
operates as a separate Ethernet LAN, and the round trip timing rules for each
LAN stop at the switching hub port. This in effect allows several individual LANs
each supporting hundreds of computers to be linked together forming a much
larger LAN, but still meeting the requirements of signal timing.
The use of switching hubs allows, in addition, to mix Ethernet links operating
at 10 Mbps with links operating at higher link speeds and to operate either some
or all of the links at higher bandwidths, namely 100 Mbps or 1000 Mbps.
9.4.3 High Performance Ethernets
Two approaches have been considered in making the Ethernet to operate at

higher bandwidths. In the first approach, the original Ethernet system with the
CSMA/CD MAC mechanism has been speeded up to 100 Mbps. This approach
is therefore called 100baseT Fast Ethernet (IEEE 802.3u).
The second approach has been to create an entirely new MAC mechanism
using hubs that control access to the medium using a demand priority mecha-
nism. The design of the medium access control based on demand priority mech-
anism has allowed the transportation of token ring frames in addition to the
standard Ethernet frames. Therefore it is called 100VG-AnyLAN (IEEE 802.12).
Switching invariably becomes central to the operation of high performance
Ethernets. The backplane speed (in packets per second – PPS) which describes
the capacity of the switch to move the data from the incoming ports to the outgo-
ing ports and is an important parameter. Another factor which is important for
the efficient functioning of a switch is the amount of buffer memory provided. In
general, higher the number of ports the more buffer memory is needed.
9.4.4 Switched Ethernets
An Ethernet switch at its basic level can be thought of as a bridge with many
ports and low latency. A network segment connected to each of the ports phys-
ically represent a separate Ethernet with its own repeater count and timing re-
strictions. Switches are useful in segmenting large networks for improved perfor-
47
mance and/or easy manageability. However, the many segments of the network
operate as one single logical network.
The design of Ethernet switches gives rise to the following classification based
on the type of packet forwarding technique used in the switch architecture. The
packet switching type employed has an effect on latency which describes the
delay in the switch. The latency has the greatest impact in environments where
real time video and audio applications are supported.
• Store and Forward Switches
A store and forward switch stores each incoming frame in a buffer, checks
for errors, and if the frame is good then forwards the frame to its destination
port. The store and forward technique has the advantage that it prevents wast-
ing network bandwidth by effectively blocking the damaged frames. However
the disadvantage is that it increases the latency and therefore results in lower
throughput in networks with few errors. Store and forward switches are useful
in networks which may experience high error rates.
In general Store and Forward switching is likely to be the best choice when a
network needs efficiency and stability.
• Cut Through Switches
In a cut through switch a frame is forwarded immediately upon receiving its

destination address thus resulting in a very low latency in the switch. The dis-
advantage is that it propagates errors and therefore is only suitable for networks
which experience a few occasional errors.
• Hybrid Switches
Hybrid switch is the result of an attempt to achieve the best of both, the Store
& Forward and Cut Through techniques. In its normal operating mode a hybrid
switch operates as a Cut Through switch constantly monitoring the rate at which
damaged or invalid frames are forwarded. When the error rate is above a certain
threshold then the switch reverts to Store and Forward mode and continues to
operate in that mode until the error rate has fallen to an acceptable level before
reverting to Cut Through mode.
The Hybrid switch has the performance advantage of a Cut Through switch
when the error rate is low, and the error trapping ability of a Store and Forward
switch when the error rates are high.
The above types are only applicable when the source and destination ports are
all running at the same speed. If the switch has to perform a speed conversion,
which is the case in many new network installations especially if a standard
Ethernet is migrated to high performance Ethernet, then the switch must operate
in the Store and Forward mode to cater to the differing link speeds.
9.4.5 Fast Ethernet

Fast Ethernet (IEEE 802.3u) operating at 100 Mbps is designed as the most
direct and simple extension of 10baseT Ethernet operating at 10 Mbps. The
Medium Access Control mechanism used in the Fast Ethernet is simply a scaled
up version of the MAC technique used by 10baseT Ethernet. This makes Fast
Ethernet (100baseT) to be similar to the conventional 10baseT, only faster. It
uses the same reliable, robust and economical technology of 10baseT. The seam-
less compatibility between 10baseT and 100baseT allows easy migration to high
speed network connections. The Fast Ethernet specification includes a mech-
anism for auto negotiation of the media speed. This allows the installation of
a dual speed Ethernet interface which automatically detects and sets its link
speed.
Although Fast Ethernet preserves the critical 100 m maximum UTP cable
length from the hub to the desktop, the rules applicable to the 100 Mbps tech-
nology are different because of the scaling of the MAC interface.
100baseT Fast Ethernet is a natural evolution from the standard 10baseT
Ethernet. Figure 40 shows the two in comparison.
Feature 10baseT Ethernet 100baseT Fast Ethernet
Speed 10Mbps 100Mbps
IEEE Standard 802.3 802.3
Media Access Protocol CSMA/CD CSMA/CD
Topology Bus or Star Star
Media support Coax, UTP, Optical Fiber UTP, STP, Optical Fiber
Hub to node distance 100 meters 100 meters

(Maximum)
Media Interface AUI MII
Figure 40: Comparison between Ethernet and Fast Ethernet
Fast Ethernet also supports multiple media types. For 100baseT the same
cabling installed for a 10baseT network can be used. Figure 41 shows the three
media specifications supported by Fast Ethernet, namely 100baseTX (2 pairs of
high quality cables), 100baseT4 (4 pairs of normal quality cables) and 100baseFX
(optical).
• 100baseTX
The 100baseTX specification supports 100 Mbps transmission speed over
two pairs of UTP Category 5 or STP Category 5. The RJ45 connector used for
49
Fast Ethernet Media Access Control (MAC)
Media Independent Interface (MII)
(AUI Equivalent)
100BaseTX 100BaseT4 100BaseFX
2 pair Cat 5 UTP 4 pair Cat 3, 4 and 5 62.5/125 micron

2 pair STP UTP fibre
Figure 41: Fast Ethernet Media Support
100baseTX UTP is exactly the same as that used by 10baseT UTP. For STP wiring,
100baseTX also specifies the traditional DB-9 connector.
• 100baseT4
100baseT4 media uses four pairs of Category 3, 4 or 5 UTP wiring to carry

data at 100 Mbps. The signalling scheme in 100baseT4 uses three pairs of
wires for data and the fourth for collision detection. Because 100baseT4 can use
Category 3 it enables migration to 100baseT4 without having to rewire.
100baseT4 also uses RJ45 connector.
• 100baseFX
100baseFX media specification defines 100 Mbps operation over two strands
of 62.5/125 micron fibre and allows transmission over greater distances than
UTP. The fibre optic connectors are the same as those defined for 10baseFX
networks.
Fast Ethernet specification includes a Media Independent Interface (MII) which
defines a standard interface between the CSMA/CD MAC layer and any of the
three media specifications supported. This is much like the AUI connector for
standard Ethernet. The MII defines a 40 pin connector to connect to the external
transceivers.
The Fast Ethernet however does not support coaxial cabling largely due to its
inability to support high data rates over the distances of interest. The ISO 11801
cabling standard is applicable to Fast Ethernet implementation.
9.4.6 Gigabit Ethernet

The ever increasing processing power in the computers and the deployment of
power hungry applications on the networks must be matched by developing high
speed network connections to reduce traffic bottlenecks, improve overall perfor-

mance and ultimately enhance the productivity of the users who use the net-
work. The development of Gigabit Ethernet is clearly a solution to match the
network infrastructure with the capability of todays desktop computers.
Gigabit Ethernet (IEEE 802.3z) is an extension of the highly successful 10
Mbps Ethernet and 100 Mbps Fast Ethernet standards and therefore is fully
compatible with the huge installed Ethernet base. Gigabit Ethernet employs
all the features of Ethernet specification including frame format, support for
CSMA/CD protocol, full duplex transmission, flow control and management ob-
jects. This compatibility preserves investment in network administrator expertise
and support staff training.
The enhancements in Gigabit Ethernet includes the support for fast optical
fibre connections at the physical layer of the network and a MAC layer specifica-
tion which sustains a tenfold increase in the MAC layer data rates. This is a key
to its ability to support data intensive applications such as imaging and video
conferencing.
The proliferation of Gigabit Ethernet will take place in phases. Initially Gigabit
Ethernet will be used as backbone switch-to-switch connections. The next phase
will be to deploy switch-to-server connectivity to boost access to critical server
resources. This evolution will be driven by the increasing installation of PCs with
100 Mbps network interfaces. Finally as the desk top costs come down and user
network demand increases, Gigabit Ethernet switches will enter the backbone
and will take over the switch fabric.
The original IEEE 802.3z specification for 1000baseX comes in three flavours,
namely:
• 1000baseSX – 850nm laser on multi mode fibre

• 1000baseLX – 1300nm laser on single and multi mode fibre
• 1000baseCX – short haul copper STP.
The 1000baseT (IEEE 802.3ab) standard specifies CAT-5 balanced copper

transmission media in order to take advantage of the already installed base of
UTP Category 5 cabling up to 100 m distances.
Recently the IEEE has setup the 802.3ae committee to draw up standards for
CSMA/CD that is capable of operating at speeds up to 10 Gbps.
9.4.7 Wireless LANs

A wireless local area network (WLAN) uses radio frequency (RF) technology to
transmit and receive data over the air. The Institute of Electrical and Electronics
Engineers (IEEE) have established the IEEE 802.11 standard, which is the pre-
dominant standard for wireless LANs. Any LAN application, network operating
system, or protocol including TCP/IP, will run on 802.11-compliant WLANs as
they would over Ethernet.
The Table 1 on page 51 shows the IEEE 802.11 family of standards for Wire-
less LANs (WLANs).
51
802.11 2.4 GHz band, supports data rates of 1-2 Mbps

802.11a High speed WLAN standard for 5 GHz band.
Supports up to 54 Mbps
802.11b WLAN standard for 2.4 GHz band.
Supports up to 11 Mbps. Also known as WiFi.
802.11e QoS requirements for all WLAN services
802.11f Defines inter-access point communication
to facilitate multi-vendor distributed WLANs.
802.11g An additional modulation technique for 2.4 GHz band,
intended to provide rates up to 54 Mbps
802.11h Defines the spectrum management of the 5 GHz band
for use in Europe and Asia Pacific.
802.11i Addresses security weaknesses of both
authentication and encryption protocols.
Table 1: IEEE 802.11 Wireless LAN Standards
The indoor operating ranges for Wireless LANs are typically 30 m at 11 Mbps
and 90 m at 1 Mbps [13]. The typical outdoor operating ranges are much larger
(120 m at 11 Mbps and 460 m at 1 Mbps).
WLANs can be configured in a variety of ways, but the configuration which

truly illustrates the benefits of WLAN technology is when it is used to provide a
networking Hot Spot.
A hotspot provides wireless LAN service, for free or for a fee, from a wide va-
riety of public meeting areas, including coffee shops and airport lounges. There
are currently thousands of hotspots worldwide and new access points are being
added daily. To use hotspots, your notebook must be configured with Wi-Fi Cer-
tified (Wi-Fi Alliance is a non profit international organisation set up to certify
interoperability of wireless local area network products based on IEEE 802.11
specification) technology so you can connect with other products. Wi-Fi Certified
notebooks can send and receive data anywhere within the range of a wireless
LAN base station.
The IEEE has two more committees developing standards for Broadband
Wireless communication systems. IEEE 802.16 committee is working on Broad-
band Wireless Access Standards where as the IEEE 802.20 committee works on
Mobile Broadband Wireless Access (MBWA) standards.
9.4.8 Bluetooth Technology

Bluetooth technology1 is defined as the chip technology enabling seamless voice
and data connections between a wide range of devices through short-range digi-
tal two-way radio. It is an open specification for short-range communications of
data and voice between both mobile and stationary devices. For instance, it spec-
ifies how mobile phones, Wireless Internet Devices (WIDs), computers and PDAs
interconnect with each other, with computers, and with office or home phones.
It also provides a Service Discovery Protocol (SDP) for discovering Bluetooth de-
vices. Bluetooth is thus used for deploying wireless personal area networks in
homes and offices, the so called Home Area Networks (HANs).
Bluetooth is an industry standard initially developed by Ericsson in the 1990s
and backed by a group of electronics manufacturers including 3Com, IBM, Intel,
Lucent Technologies, Microsoft, Motorola, Nokia and Toshiba. The development
is now led by Bluetooth SIG (Special Interest Group) which published Bluetooth
version 1.0 specification in July 1999 that allows any sort of electronic equip-
ment – from computers and cell phones to keyboards and headphones – to make
its own connections, without wires, cables or any direct action from a user. Blue-
tooth is intended to be a standard that works at two levels:
1. It provides agreement at the physical level – Bluetooth is a radio-frequency

standard.
2. It also provides agreement at the next level up, where products have to
agree on when bits are sent, how many will be sent at a time and how the
parties in a conversation can be sure that the message received is the same
as the message sent.
The companies belonging to the Bluetooth SIG, and there are more than 1,000
of them, want to let Bluetooth’s radio communications take the place of wires for
connecting peripherals, telephones and computers.
Bluetooth communicates at a frequency of 2.4 gigahertz, which has been set
aside by international agreement for the use in Industrial, Scientific and Medical
(ISM) devices. A number of devices that are already in use take advantage of
this same radio-frequency band. Baby monitors, garage-door openers and the
newest generation of cordless phones all make use of frequencies in the ISM
band. Making sure that Bluetooth and these other devices do not interfere with
one another has been a crucial part of the design process. In order to limit
spectral emissions and interference, it provides three different power classes
corresponding to 10, 20 and 100 m distance operation.
Originally Bluetooth aimed to provide low cost, low power connections be-
tween a variety of consumer products. As an example, a Bluetooth link between
a laptop and a 3G cellular phone could enable the laptop computer to gain access
1
Named after Harald Bluetooth who was king of Denmark in the late 900s. He managed to
unite Denmark and part of Norway into a single kingdom and then introduced Christianity into
Denmark. He left a large monument, the rune stone in Jelling, in memory of his parents. He was
killed in 986 during a battle with his son, Svend Forkbeard.
53
to the Internet by means of the 3G packet data transmission. The work of IEEE
802.15 committee aims to standardise the physical and datalink layers of Wire-
less Personal Area Networks (WPANs) based on Bluetooth specification which
specifies a complete system from the physical layer to the application layer. Al-
though the Bluetooth SIG and IEEE 802.15.1 versions are not identical, it is
hoped that they will soon converge to a single standard.
9.4.9 Future of Wireless Technology
Wireless Internet will drive four wireless applications namely, messaging, brows-
ing, interacting and conversing. The emerging trend is for mobile networks to
migrate to fully IP based 4G systems. These will use a new spectrum and packet
switching technology to support data rates of 10+ Mbps. The 3G and 4G systems
will provide multimedia services to users everywhere, while WLANs will provide
broadband services in hotspots, and WPANs will connect personal devices to-
gether at very short distances.
9.5 ADSL – delivering RT multimedia to the home and small

business
ADSL technology is the result of the recent advances in modem technology that
converts existing twisted pair telephone subscriber lines into access paths for
the transfer of multimedia and high speed digital data. Present day ADSL can
sustain downstream transmission speeds up to 6 Mbps to a subscriber and in
excess of 800 Kbps in both directions (duplex).
These transmission rates expand the existing access capacities by a factor of
100 or more (V.90 modem supports 56 Kbps nominally) without new cabling,
thus transforming the existing public telephone infrastructure which was only
used to deliver voice, text and low resolution graphics in the past to a powerful
system capable of bringing multimedia, including full motion video, in quality
assured real time, to the home and small business.
ADSL is seen as an interim solution which will play a crucial role in the next
decade for the delivery of information in video and other multimedia formats.
This is expected to pave the way for a full broad band service, which will take
decades to reach all prospective customers. In this interim period ADSL will
bring movies, television, video catalogues, remote CDROMs, corporate LANs and
the global Internet into homes and small businesses.
An ADSL circuit consists of an ADSL modem at each end of a twisted pair
telephone subscriber line creating three information channels as follows (see
Figure 42):
• high speed downstream channel (up to 6 Mbps)

• medium speed duplex channel (∼800 Kbps)
• POTS (Plain Old Telephone Service) or an ISDN channel (64 – 128 Kbps).
Information
Server
Telephone
Network Subscriber line
ADSL ADSL
Modem Subscriber
Equipment Modem 1.5 Mbps - 6 Mbps
(5.5 km) (2.7 km)
16 to 800 kbps
Internet
Figure 42: ADSL Channels
It is this asymmetric nature in the data transmission rates in the channels

which gives rise to its name Asymmetric Digital Subscriber Line.
The downstream data rates of an ADSL channel depends on:
• length of copper line

• wire gauge
• presence of bridged taps
• cross coupled interference.
In transmitting digital compressed video as a real time signal, ADSL cannot use
link level or network level error control procedures commonly found in other
data communication systems due to the fact that they are generally based on
error recovery by re–transmission best suited for non-real time services. ADLS
modems therefore incorporate forward error correction.
Multiple channels are created by the ADSL modem by dividing available band-
width of a telephone line in one of two ways (see Figure 43):
• FDM – Frequency Division Multiplexing

• Echo Cancellation.
In FDM one frequency band is assigned for upstream data and another for
downstream data. The downstream path is then divided by FDM into one or
more channels at the desired data rates.
Echo cancellation, a technology used in V.32 and V.34 modems, assigns the
upstream band such that it overlaps the downstream band and separates the
two by local echo cancellation.
With either technique ADSL splits a 4 KHz region for POTS at the DC end
of the frequency spectrum and guarantees the availability of telephone channel
even if ADSL fails.
55
Upstream Downstream Downstream
Upstream
POTS
POTS
0 4 kHz 1 MHz 0 4 kHz 1 MHz
Frequency Frequency
Figure 43: Creating ADSL Channels by (i) FDM and (ii) Echo Cancellation
9.5.1 ADSL Standardisation

At the present moment the following standards are applicable to ADSL:
1. ANSI T1.413 Issue I (1995) – up to 6.1 Mbps

2. ETSI Annex to T1.413 – to reflect European requirements
3. ANSI T1.413i2 (Issue II 1998) – includes multiplexed interfaces at user end
and protocols for configuration and network management.
4. ITU–T G.dmt (G.992.1) and G.lite (G.992.2) (1999) – the latter allows the
splitting to be done at the telephone network equipment end, at the expense
of line speed.
The final step has helped moving towards vendor interoperability and service
provider acceptance, further increasing ADSL deployment.
ATM Forum has recognised ADSL as a physical layer transmission protocol
for UTP medium.
10 WAP – Wireless Application Protocol

WAP is designed as an application environment based on a set of communica-
tion protocols for wireless hand-held devices to enable manufacturer, vendor and
technology independent access to the Internet and advanced telephony services.
It bridges the gap between the mobile world and the Internet as well as corporate
intranets and offers the ability to deliver an unlimited range of services to sub-
scribers independent of their network service provider or type of terminal device.
It is developed by mobile and wireless communication companies and includes
a mini-browser, scripting language, access function and layered communication
specification. This enables the mobile user to access the same wealth of infor-
mation from a wireless device as they can from a desktop.
WAP is based on a secure specification that allows users to access information
instantly via hand-held wireless devices such as:
• mobile phones
• pagers
• two way communication radios
• smart phones
• communicators, etc.
WAP supports most wireless network transmission technologies in use today,
most notable being CDMA (Code Division Multiple Access), GSM (Global Sys-
tem for Mobile communications), TDMA (Time Division Multiple Access), HSCSD
(High Speed Circuit Switched cellular Data), CDPD (Cellular Digital Packet Data),
SMS (Short Messaging Service), GPRS (General Packet Radio Service), UMTS
(Universal Mobile Telecommunication System), etc.
The operating systems that support WAP are many. Those specifically engi-
neered for hand-held devices include:
• PalmOS
• Windows CE – Compact Edition
• J2ME – Java 2 Mobile Edition, etc.
Although WAP supports HTML and XML, the WML (Wireless Mark-up Lan-
guage) is specifically designed for small screens of typically 150 x 150 pixel
resolution and one hand navigation without a keyboard as opposed to using
point-and-click action of a mouse. Like XML, WML is scalable and extensible as
it allows users to define new markup tags.
WAP supports WMLScript (similar to JavaScript) but makes minimal demands
on limited memory and CPU power of hand-held devices. Browsers that are
specially designed to run on WAP devices and called micro-browsers are of small
file size.
10.1 WAP Specification

The WAP specifications define a set of protocols to be used in application, ses-
sion, transaction, security and transport layers. These are designed to enable
operators, manufacturers, and application and content providers to meet the
challenges in advanced wireless service provision.
WAP utilises standard Internet components such as XML, UDP (User Data-
gram Protocol), TLS (Transport Layer Security) and IP. These are optimised for
the mobile hand-held device environment with its own unique constraints; low
bandwidth, high latency and less connection stability. WAP utilises binary trans-
mission for greater compression of data and WAP sessions are designed to cope
with intermittent transmission coverage.
The WAP content is produced using WML and WMLScript and is scalable from
two line display on a basic hand-held device to a full graphic screen on smart
phones and communicators.
As WAP is based on a scalable layered architecture each layer can develop
independently of others. This will also enable content providers to customise
57
SNMP
JAVA Logging Subscriber DB Applications
M Billing Subscriber Compiler

A Data Data & Encoder
N
A Context Manager
G
WSP
E HTTP
M WTP
E
N WTLS
T
WDP/Adaption TCP/IP
Bearers Internet
Firewall etc.
Figure 44: Layered Architecture of the WAP Gateway
content to match client expectations and differentiate themselves from their com-
petitors with new, enhanced information services.
While WAP users benefit from easy, secure access to Internet based infor-
mation such as unified messaging, banking, and entertainment, they will also
enjoy considerable freedom of choice when selecting mobile devices and network
operators.
10.2 Architecture of the WAP Gateway

Figure 44 shows the layered architecture of the WAP Gateway.
• HTTP Interface – This serves to retrieve from the Internet, the WAP content
requested by the mobile device. WAP content (WML, WMLScript) is then
converted into a compact binary byte code for efficient transmission over
the air channel.
The WAP microbrowser software resident within the mobile device interprets
this byte code and displays the interactive WAP content.
• WSP – The WAP Session Protocol layer provides a light weight session to al-
low efficient exchange of data between applications. It supports connection-
oriented for two way communication between the device and the network
and connection-less for broadcasting or streaming the data to the device.
• WTP – The WAP Transport Protocol layer provides transaction support adding
reliability to the datagram services provided by the WDP layer.
Technology Flow Rate Operating Mode Status

GSM 9.6 Kbps Circuit in use
HSCSD 56 Kbps High flow in use
packets
GPRS 112 Kbps Packets extension
of GSM
UMTS 2 Mbps High flow not based on
packets GSM technology.
Changes required.
Table 2: Wireless Bearer Technologies
• WTLS – Wireless Transport Layer Security checks data integrity, performs

client and server authentication and has encryption facilities to provide se-
cure transport service required by secure applications such as e-commerce,
e-banking, etc.
• WDP – The WAP Datagram Protocol is for the transport layer that sends and
receives messages via any available bearer network.
The WAP gateway server can operate under two possible scenarios:
1. If the web server provides content in WML, the WAP gateway transmits this
data directly to the WAP client.
2. If the web server delivers content in HTML, the WAP gateway server encodes
HTML into WML before transmitting to the WAP client.
The present day wireless bearer technologies include GSM, HSCSD, SMS,
GPRS and UMTS among others. Table 2 compares some of their important char-
acteristics.
WAP was primarily developed for 2G mobile systems that provide slow data
rates between 9.6 – 14.4 Kbps. The development of 3G wireless systems such
as UMTS with data rates of 2 – 4 Mbps is expected to resolve the problem of
limited bandwidth. With such systems it is natural that WAP will not be needed.
However the WAP Forum argues that, even in 3G systems, bandwidth will play
a crucial role as new applications will continue to demand higher bandwidths
and data rates. In addition they can benefit from the ability to deal with small
screen size, low power consumption, carrier independence, multi-device support
and intermittent coverage.
Figure 45 shows the ’wapalised’ ICTP homepage as it would appear on the
screen of a WAP enabled mobile phone.
10.3 Mobile Web Services

Web services include well defined protocol interfaces through which services can
be provided to users over the Internet. They are built on emerging technologies
59
Figure 45: Content display on a WAP enabled mobile phone
such as XML, SOAP (Simple Object Access Protocol), WSDL (Web Services De-
scription Language) of the W3C (World Wide Web Consortium), UDDI (Universal
Description, Discovery, and Integration) and HTTP.
Mobile web services provide content delivery, location discovery, user authen-
tication, presence awareness, user profile management, terminal profile man-
agement and event notification. Initially wireless terminals are likely to access
Mobile Web Services indirectly, through application servers which will manage
the interaction with the required Web service.
11 Brief Introduction to Programming the Internet

with JavaTM
This section is intended to illustrate the important principles of TCP/IP socket
programming in Java and as such the examples are selected to illustrate them
concisely. In this respect the code examples would fall seriously short of produc-
tion quality code. Although Object Oriented design principles have been adhered
to, large amounts of error handling code have been avoided thereby sacrificing
robustness for brevity and clarity. For a more complete treatment of Java and
object oriented programming the reader is referred to the standard text books
[1, 11] and the Chapter on Introduction to Java Programming by Carlos Kavka.
The java.net class library package in Java1 provides a variety of classes, in-
terfaces and exception handling features useful in the implementation of net-
working applications. The low level APIs allow direct access to network protocols
through TCP sockets and UDP datagrams. The high level APIs such as the URL
and URLConnection, on the other hand, allow the faster development of network
applications, with much less code to write, albeit at the expense of flexibility.
The classes in the java.net package can be classified into three broad cate-
gories, namely:
1. Web Classes – to access web documents via their URLs.
2. Raw Classes – to provide stream based interfaces to network hosts.
3. Extension Classes – to extend the web classes to cater for new document
and content formats.
The following sections outline a number of relevant classes and some code
segments to illustrate the use of Java APIs in network programming [5, 18, 20].
11.1 InetAddress Class

The java.net.InetAddress class represents an Internet Protocol (IP) address
and can be used when creating DatagramPacket or Socket objects. The class
does not have public constructor functions. Instead it supports three static
methods which return one or more instances of InetAddress. The InetAddress
class contains two fields namely, hostName (of type String object) and address
(of type int) which are not public. The hostName field contains the fully quali-
fied domain name of host, for eg. www.ictp.trieste.it and the address field
contains the 32 bit IP address, for eg. 140.105.16.63 (this is IPv4. This rep-
resentation will be different for a 16 byte IPv6 address). The getLocalHost,
getByName and getAllByName methods must be used by applications to create a
new InetAddress instance (see the Method Summary at http://java.sun.com/j2se/1.5
There are other methods in the java.net package which return InetAddress
objects. In Java 2 version 1.5 (Sun now calls this Java 5.0) these methods
include:
1
http://java.sun.com/j2se/1.5/docs/api
61
DatagramPacket.getAddress(),
DatagramSocket.getInetAddress(),
DatagramSocket.getLocalAddress(),
MulticastSocket.getInterface(),
ServerSocket.getInetAddress(),
Socket.getInetAddress(), and
Socket.getLocalAddress().
The following code segments illustrate the use of some of the classes and
methods available.
import java.net.*;
public class ICTPByName {

public static void main (String [] args) {
try {
InetAddress ICTP_address =
InetAddress.getByName("www.ictp.trieste.it");
System.out.println (ICTP_address);
}
catch (UnknownHostExecption e) {
System.out.println
("Could not find ICTP Web address");
}
catch (SecurityException e) {
System.out.println
("A Security Manager exists - does not allow operation");
}
}
}
When the code above is executed successfully, it will give:
www.ictp.trieste.it/140.105.16.63
Creating InetAddress Objects
A string containing the dotted quad form of the IP address can be passed to
InetAddress.getByName().
The other two methods are:
InetAddress.getAllByName(String hostName) – returns an array con-

taining all addresses;
InetAddress.getLocalHost()
All three methods throw UnknownHostException. The methods, except

getLocalHost, also throw SecurityException.
InetAddress Class
public String getHostName() — returns the hostname of an InetAddress
object as a string;
public byte[] getAddress() — returns the address as a byte array.
Beware!!! Type byte in Java is signed! Addresses between 128 and 255 are
returned as negative numbers!! Add 256 to the negative bytes!!!
public String getHostName() — returns the address as a string.
Notice that there are no setHostName() and setAddress() methods. Java
guarantees that packages out side java.net cannot change an InetAddress ob-
ject’s fields behind its back.
The InetAddress class has seen many changes in J2SE 1.4 in order to ac-
commodate IPv6 support. Inet4Address and Inet6Address classes have been
added as two subclasses of InetAddress to represent the 32 bit IPv4 addresses
and the 128 bit IPv6 addresses respectively. However, in most common cases
the IPv6 support is transparent to the developer who will continue to manipu-
late instances of InetAddress. All static methods in InetAddress return either
an Inet4Address or Inet6Address depending on the resolution. Similarly all
socket methods returning an InetAddress, such as Socket.getInetAddress
will return an instance of Inet4Address or Inet6Address.
11.2 URL Class

This class is associated with a number of constructors which set up URL objects
with the following general format:
a protocol, a host, a directory and a file name
This constructor throws a MalformedURLException if the format of the URL

is incorrect. After successfully constructing the URL object Java language pro-
vides methods to determine its component parts such as protocol, port, host,
etc.
HTTP provides a client with the ability to retrieve information from the server
about a resource, such as its size, when last modified and content type. Ac-
cess to this information is provided by the URLConnection class, an instance of
which may be obtained from the openConnection method of an URL object. The
getDate, getExpiration and getLastModified methods return dates in the
form of the number of milliseconds since January 1, 1970.
java.net.URL Class
It is a final class hence cannot be subclassed.
It is Java’s abstraction of the Uniform Resource Locator (URL).
URL object is not stored as a string. Rather it is treated in Java as an object with
fields such as:
63
protocol
hostname
port
path
query string, etc. . .
each of which may be set independently.
Constructing URL Objects - Protocol Types
Six constructors available.

All of them throw MalformedURLException.
All VMs (Virtual Machines) support protocol types HTTP and file.
JDK supports HTTP, file, FTP, mailto.
Constructing a URL from a string
The following code segments illustrate the use of URL concepts.
try {
URL u = new URL("http://mlab-14.ictp.trieste.it");
}
catch (MalformedURLException e) {
System.out.println("Invalid String");
}
Constructing a URL from its component parts
try {
// String protocol, String hostname, String filename
URL u = new URL("http", "www.ictp.trieste.it", "/index.html");

}
// All known virtual machines should recognise http!
}
The above assumes the default port which for HTTP is port 80.
For cases where the default port is not applicable the following allows the
redefinition of a port number in the constructor:
try {
// String protocol, String hostname, int port, String file
URL u = new URl("http", "web.server.somewhere", 1221, "/myfile");

}
System.err.println (e);
}
Splitting a URL object into its components

Java provides five public methods:
getProtocol() – returns the protocol of this URL;

getHost() – returns host name;
getPort() – returns the int port number or -1 if not specified;
getFile() – returns the file portion of the URL;
getRef() – returns the named anchor part of the URL.
Receiving Data from a URL

The following code segment illustrates how to receive data contained in a URL.
try {
URL u = new URL("http://www.hamsterjam.com");
InputStream is = URL.openStream(); // returns raw data
// ASCII if text file
// binary if an image file
int c;
while ((c = is.read()) != -1)
System.out.write(c);
}
catch (IOException e) {
}
}
11.3 Socket and ServerSocket Classes

TCP is a protocol which allows reliable transport of stream data, of course at the
expense of speed (due to increased processing time in handling error detection
and correction) and complexity. Applications built using TCP/IP often use the
client-server model of communication such as web page delivery or file transfer
65
using FTP. In this model a server waits (or listens) for a client to request a ser-
vice. For standard applications such as HTTP and FTP these services are offered
through well known ports; 80 for HTTP and 21 for FTP. Other services may be
offered at any port number assigned by the server’s operating system. On Linux,
the port numbers associated with particular services are in /etc/services.
A Socket object represents the Java version of a TCP connection. The sending
and receiving paths are represented by an InputStream and an OutputStream
in the socket.
When a Socket object is created a connection is made to a specified computer.
When a ServerSocket connection is made the server on which it is made listens
to clients and establishes a connection when it hears the connection request.
There are a number of constructors for the Socket and the most popular ones
take the name of a host computer and a port number as arguments.
Socket Constructors
The following constructor creates a TCP socket to the specified port on the spec-
ified host and attempts to connect to the remote host.
try {
Socket Google = new Socket ("www.google.com", 80);
// String hostname, int portnumber
// ...........
// send and receive data
// ...........
}
catch (UnknownHostExeption e) {
}
catch (IOExeption e) {
System.err.println (‘‘An I/O error has occurred’’);
}
System.out.println
}
The following is similar to the previous constructor, but uses the InetAddress
object, rather than the hostname.
try {
InetAddress GoogleAddress = InetAddress.getByName("www.google.com");
Socket GoogleSocket = new Socket (GoogleAddress, 80);
//
}
}
}
If the socket has to be created using a particular NIC (such as a 100 Mbps
Ethernet card as opposed to a 10 Mbps card on a multihomed host), then:
try {
InetAddress FEcard = InetAddress.getByName("FE.mymachine.com");
Socket FEGoogleSocket = new Socket("www.google.com", 80, FEcard, 0);
// String host, int port, InetAddress interface, int local port
// ’0’ for local port means ’don’t care’
//
}
}
}
The following constructor uses InetAddress class:
try {
InetAddress ATMcard = InetAddress.getByName("ATM.mymachine.com");
InetAddress Google = InetAddress.getByNAme("www.google.com");
Socket ATMGoogleSocket=new Socket(Google, 80, ATMcard, 0);
}
}
}
Getting Information about a Socket

Given a socket object the getInetAddress() method returns the IP address of
the remote machine the socket is connected to or null if not connected.
Similarly the getPort() method returns the port to which socket is con-
nected on the remote machine or 0 if it is not connected yet.
There are two ends to a socket. The getLocalPort() method returns the
local port number to which the socket is bound or -1 if the socket is not bound.
11.4 DatagramSocket and DatagramPacket Classes

The reliable two way stream transport of TCP is not always required, specially
when the amount of data to be transported is relatively small (eg. credit card
verification or obtaining time from a NTP server), speed of communication is of
prime importance or an application is run over a highly reliable network segment
such as a local area network.
The two main classes in Java associated with potentially unreliable UDP com-
munication are DatagramSocket and DatagramPacket. It is the very nature of
67
DatagramSocket that it is not associated with a particular receiving endpoint.

It is simply used as a means of obtaining a UDP port from the local operating
system in order to send and receive UDP packets.
The following code segments illustrate the use of some of the useful classes
and methods available in the Java Networking API.
Example – Use of InetAddress class

/* prints the name/address pair (note!) of local host and the name of a
* host specified and its IP address in dotted quad format.
* also prints ALL known IP addresses of a host. IP names are hard coded.
*/
import java.net.*;
class Hosts {
public static void main (String args[]) {

try {
// useful to know to what host a client is connecting
InetAddress localAddress = InetAddress.getLocalHost();

System.out.println (localAddress);
System.out.println ("==============================");
// find out the address of a given host
InetAddress hostAddress = InetAddress.getByName(

"www.ictp.trieste.it");
System.out.println ("My name is " + hostAddress.getHostName());
System.out.println ("My address is " +
hostAddress.getHostAddress());
System.out.println ("==============================");
// list all registered addresses of the host
InetAddress[] allAddresses = InetAddress.getAllByName(

"www.mirror.ac.uk");
System.out.println ("All known addresses for " +
allAddresses[0].getHostName());
for (int i=0; i<allAddresses.length; i++) {
System.out.println (allAddresses[i]);
}
}
catch (UnknownHostException e) {
System.out.println ("Error: Cannot resolve");
}
System.out.println
}
}
}
Example – Use of URL class

/*
* This example illustrates the use of URL class to fetch
* the contents of a web page pointed to by a given URL.
* The URL of the web server is passed through command line.
* MalformedURLException and IOException must be caught.
*/
import java.net.*;
import java.io.*;
class GetContents {
public static void main (String args[]) throws Exception {

if (args.length != 1) {
System.out.println ("Usage: java GetContents <url>");
System.exit(0);
} // about usage
String str, urlstring;
// check if url is entered with or without http://
if (args[0].startsWith("http://")) {
urlstring = args[0].substring(7);
}
else {
urlstring = args[0];
}
URL webURL = new URL("http://"+urlstring);
InputStream is = webURL.openStream();
BufferedReader ds = new BufferedReader(new InputStreamReader(is));
while (true) {
str = ds.readLine();
if (str == null) {
break;
}
System.out.println(str.trim());
}
}
}
69
Example – Use of UDP DatagramPackets and DatagramSockets

/*
* Illustrates the use of DatagramSocket and DatagramPacket to query the
* Time Of Day on well known port number 13.
*/
import java.net.*;
import java.io.*;
class UDPDaytimeClient {
public static void main (String args[]) throws
UnknownHostException, SocketException, SocketTimeOutException {
try {
InetAddress host = InetAddress.getByName(args[0]);
// get IP address of host
DatagramSocket mySocket = new DatagramSocket();
// create a datagram socket
// send a dummy packet to the UDP daytime service

// reply contains the response from the time server
final int DaytimePort = 13;
DatagramPacket query = new DatagramPacket(new byte[1],1,
host,DaytimePort);
mySocket.send (query); // send query packet

// get response
byte[] r_buffer = new byte[100];
DatagramPacket response = new DatagramPacket(r_buffer,
r_buffer.length);
mySocket.receive(response); // display response
String received = new String(response.getData());
System.out.println ("The time of day received from port "
+ response.getPort() + " of "
+ response.getAddress().getHostName()
+ " is: ");
System.out.println (received);
mySocket.close(); // close socket

}
System.err.println ("System IO error");
System.exit(1);
}
}
}
Example – Simple TCP Client

/*
* TCP Client
*/
import java.io.*;
import java.net.*;
public class TCPEchoClient {

public static void main (String args[]) throws IOException {
Socket echoSocket = null;

PrintWriter out = null;
BufferedReader in = null;
try { // use port number 4321

echoSocket = new Socket (args[0], 4321);
// enable automatic flush
out = new PrintWriter (echoSocket.getOutputStream(), true);

in = new BufferedReader (new InputStreamReader
(echoSocket.getInputStream()));
}
catch (UnknownHostException e) {
System.err.println ("Host unknown");
System.exit(1);
}
System.out.println ("Server welcome is: " + in.readLine());
out.println ("Hello");
System.out.println ("Server Responded: " + in.readLine());
out.println ("Hello again");
System.out.println ("Server now responded: " +in.readLine());
out.close();
in.close();
echoSocket.close();
}
}
71
Example – Simple TCP Server

/*
* TCP Server
*/
import java.io.*; import java.net.*; import java.util.*;
public class MyTCPServer {

public static void main (String args []) throws IOException {
ServerSocket tcpServerSocket = null;

try {
tcpServerSocket = new ServerSocket (4321);
}
System.err.println ("Could not listen on specified port");
System.exit(1);
}
while (true) {
Socket clientSocket = null;
try {
System.out.println ("Server listening");
clientSocket = tcpServerSocket.accept();
}
System.err.println ("Accept failed");
System.exit(1);
}
PrintWriter out = new PrintWriter (clientSocket.getOutputStream(),

true);
BufferedReader in = new BufferedReader (new InputStreamReader
(clientSocket.getInputStream()));
String inputLine, outputLine;

out.println ("Welcome to the Echo Server");
while ((inputLine = in.readLine()) != null) {
if (inputLine.equals("Bye")) break;
Date now = new Date (System.currentTimeMillis());
out.println (now+": " + inputLine);
}
out.close(); in.close(); clientSocket.close();
}
}
}
12 Programming Multitasking Applications

The examples in the previous section demonstrated the features in Java that
enable programming client server applications with sockets. These applications
however handle one client at a time. If a client connects while another is already
being served, the server will continue to serve the first client until it has finished
and then serve the second client. This type of servers are called iterative servers
because they handle clients sequentially. They work best where each client re-
quires a short connection time to the server. However, if service time becomes
large then the waiting time experienced by those clients in the queue may be
unacceptable.
Threads in Java (covered in standard texts on Java and in the Chapter by
Carlos Kavka) provides a convenient mechanism to develop server applications
to handle many clients simultaneously allowing each connection to proceed in-
dependently without interfering with other connections.
There are two approaches to developing multitasking or concurrent applica-
tions using threads:
1. thread per client – a new thread is spawned to handle each client connection.
2. thread pool – a fixed number of pre-spawned threads work together to serve
client connections.
They are independent of the type of client-server protocol (TCP or UDP) that
is used.
12.1 Thread per Client

In this type of server, a new thread is created to handle each connection from
the client. The server typically executes a loop that runs forever, listening for
connection requests on a specified port. When connection requests are made by
clients the server repeatedly accepts them by spawning a new thread to handle
each connection.
12.2 Thread Pool

The idea of using a Thread Pool is to reduce the overheads created by thread per
client approach. When a thread is created it consumes system resources such
as CPU time. Each thread has its own data structures that consume system
memory. In addition, scheduling and context switching, create extra overhead.
As more threads are created, the system resources consumed by these threads
increase causing the system to spend more and more time in thread management
rather than serving client requests.
One solution to this problem is to limit the total number of threads that is
allowed to be spawned and reuse them. Instead of spawning a new thread for
73
each connection request, the server creates a thread pool with a fixed number
of threads and client requests are assigned to a thread from this pool. When
the thread finishes servicing the client it returns to the pool so that it can be
reassigned to a new request. Connection requests from clients that arrive when
all threads in the pool are busy are queued until a thread becomes available.
13 Summary
TCP/IP is a time tested and a proven technology which is now universally avail-
able under almost all operating system platforms. In the recent times TCP/IP
has been used in transporting time critical streaming data such as video and
audio over the global Internet. A new version of Internet Protocol, IPv6, is being
formulated taking care of the shortcomings of IPv4 making it more suitable for
real time data communication applications.
Real time Transport Protocol (RTP), together with a host of other protocols
facilitate the transfer of real time data streams over existing LANs and WANs
based on the Internet Protocol (IPv4) technology.
The overall success and acceptability of Ethernet as a LAN technology has led
to its enhancement to high performance platforms operating at 100 Mbps (Fast
Ethernet) and 1 – 10+ Gbps (Gigabit Ethernet). These two offer the users the
familiar Ethernet operating environment with the ability to deliver high band-
width real time data such as video and imaging to the desktop. In the last few
years however, there has been a noticeable drive to develop wireless technology
to support computing in the personal and home environment as well as mobile
and as a replacement to the local area and wide area connectivity based on wired
technologies. As a result there are medium to high bandwidth wireless solutions
in the market today slowly revolutionising the way not only homes and offices
but entire cities, regions or even countries are interconnected.
Java programming environment provides a rich collection of features to en-
able network programmers to easily develop sophisticated data communication
applications for time critical data over distributed data communication networks.
Over the years the Internet and its TCP/IP protocol suite have been used for
such diverse applications for which types neither the Internet nor its protocols
were designed in the first place. In the last decade the Internet and its TCP/IP
protocol architecture have become the basis for the implementation of GRID
Infrastructure [12, 2], aimed at delivering computational power to applications
on demand. These applications, although belong to diverse areas of scientific
research on the one hand as well as business and commercial on the other,
have one thing in common. They both demand enormous CPU performance and
storage capacities, that cannot be satisfied by the use of a single or multiple
supercomputers of our time for their successful execution.
The concept of GRID infrastructure is to bring together CPUs, sensors, in-
struments, storage devices, and such like scattered all over the world into one
pool that can then be shared over the Internet. Already there are hundreds
of national and international GRID infrastructure development projects in mo-
tion. In the short term the Internet has been, and will further be, enhanced by
developing the middleware components required to perform GRID specific com-
putational tasks. These collaborations have already provided an opportunity
to scientists hitherto unavailable, in order to tackle some of the hardest com-
putational problems known to them including earth quake prediction, disaster
75
control and management, genome mapping, and weather prediction, to name a

few.
One salient point that makes the GRID effort different from most other high
performance computing projects is that the GRID is available to anyone who
wants to use it. In the true spirit of the Internet all the software, documentation
and literature are available in the public domain as open source. There is no
doubt that this brings a new and exciting computing paradigm to the scientists
in the developing world [21].
References
[1] Barnes D J, Object-Oriented Programming with Java; An Introduction,
Prentice Hall, 2000.
[2] Berman F, Fox G C and Hey A J G, (Eds), Grid Computing – Making the
Global Infrastructure a Reality, Wiley, 2003.
[3] Black U, Computer Networks, Protocols, Standards and Interface (2nd

Edition), Prentice Hall, 1993.
[4] Black U, TCP/IP and Related Protocols (2nd Edition), McGraw Hill, 1995.
[5] Calvert K L and Donahoo M J, TCP/IP Sockets in Java: Practical Guide

for Programmers, Morgan Kaufmann, 2002.
[6] Comer D, Internetworking with TCP/IP: Principles, Protocols and Ar-

chitecture, Prentice Hall, 1988.
[7] Comer D, Internetworking with TCP/IP - Vol I: Principles, Protocols and

Architecture (2nd Edition), Prentice Hall, 1991.
[8] Comer D and Stevens D L, Internetworking with TCP/IP - Vol II: Design,
Implementation, and Internals, Prentice Hall, 1991.
[9] Comer D, Computer Networks and Internets with Internet Applications

(4th Edition), Prentice hall, 2004.
[10] De Prycker M, ATM Solutions for Broadband ISDN (3rd Edition), Prentice
Hall, 1995.
[11] Deitel H M and Deitel P J, Java: How to Program (5th Edition), Prentice
Hall, 2003.
[12] Foster I and Kesselman C, (Eds),The GRID 2: Blueprint for a New Com-
puting Infrastructure (2nd Edition), Morgan Kaufmann, 2004.
[13] Furht B and Ilyas M, (Eds), Wireless Internet Handbook – Technologies,

Standards and Applications, CRC Press, 2003.
[14] Goncalves M and Niles K, IPv6 Networks, McGraw-Hill, 1998.
[15] Goncalves M, Voice Over IP Networks, McGraw-Hill, 1999.
[16] Hagen S, IPv6 Essentials, O’Reilly, 2002.
[17] Handel R, Huber M N, and Schroder S, ATM Networks - Concepts, Proto-

cols, Applications, Addison-Wesley, 1994.
[18] Harold E R, Java Network Programming (2nd Edition), O’Reilly, 2000.

77
[19] Huitema C, IPv6: The New Internet Protocol, Prentice Hall, 1996.
[20] Ince D & Freeman A, Programming the Internet with Java, Addison-
Wesley, 1997.
[21] Induruwa A S and Induruwa S D, eScience - A 21st century vision for

eSri Lanka, pp151-158, Proc. 22nd National IT Conference, Colombo, Sri
Lanka, 2003.
[22] Jain R, FDDI Handbook - High Speed Networking using Fibre and Other
Media, Addison-Wesley, 1994.
[23] Keshav S, An Engineering Approach to Computer Networking - ATM

Networks, the Internet, and the Telephone Network, Addison-Wesley,
1997.
[24] Miller P E & Miller M A, Implementing IPv6 (2nd Edition), Hungry Minds,
2000.
[25] Partridge C, Gigabit Networking, Addison Wesley, 1994.
[26] RFC1 1075, Distance Vector Multicasting Routing Protocol, November

1988.
[27] RFC 1112, Host Extensions for IP Multicasting, August 1989.
[28] RFC 1585, MOSPF – Analysis and Experience, March 1994.
[29] RFC 1700, Assigned Numbers, October 1994.
[30] RFC 1883, Internet Protocol, Version 6 (IPv6) Specification, April 1996.
[31] RFC 1889, RTP: A Transport Protocol for Real Time Applications, Jan-
uary 1996.
[32] RFC 1890, RTP Profile for Audio and Video Conferences with Minimal
Control, January 1996.
[33] RFC 2205, Resource Reservation Protocol, September 1997.
[34] RFC 2236, IGMP – Internet Group Management Protocol (Version 2),
November 1997.
[35] RFC 2362, Protocol Independent Multicast, June 1998.
[36] RFC 3232, Assigned Numbers: RFC 1700 is Replaced by an On-line

Database, January 2002.
1
all RFCs are available from ftp://rs.internic.net/rfc
[37] Smith P, Frame Relay - Principles and Applications, Addison Wesley,

1993.
[38] Stallings W, ISDN and Broadband ISDN with Frame Relay and ATM (4th
Edition), Prentice Hall, 1998.
[39] Stallings W, Local and Metropolitan Area Networks (6th Edition), Pren-
tice Hall, 2000.
[40] Stallings W, Wireless Communications and Networks, Prentice Hall,

2001.
[41] Stallings W, High Speed Networks and Internets: Performance and

Quality of Service (2nd Edition), Prentice Hall, 2002.
[42] Stallings W, Data and Computer Communications (7th Edition), Prentice

Hall, 2003.
[43] Stallings W, Computer Networking with Internet Protocols, Prentice Hall,

2003.
[44] Tanenbaum A S, Computer Networks (4th Edition), Prentice Hall, 2003.
[45] Wilder F, A Guide to the TCP/IP Protocol Suite, Artech House, 1993.

Internet Protocol Architectures and Network Programming

Uploaded by

Copyright:

Available Formats

Internet Protocol Architectures and Network Programming

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

Internet Protocol Architectures and Network Programming

Uploaded by

Copyright:

Available Formats

310/1779-6

Fourth Workshop on Distributed Laboratory Instrumentation

Internet Protocol Architectures

Supporting material for the

6 A word about the Internet 12

7 Internet Protocol Architecture 14

8 IPv6 – The New Generation Internet Protocol 29

9 Data Communication in Real Time 34

10 WAP – Wireless Application Protocol 55

11 Brief Introduction to Programming the Internet with JavaTM 60

12 Programming Multitasking Applications 72

3.1 Geographical Coverage

There is a significant level of deployment of all or some of the above even in

3.2 Network Topology

• Bus Topology (eg. CSMA/CD; Ethernet)

Figure 1: Token Passing Ring

If however networks are categorised according to their transfer mechanism,

Figure 2: Broadcast Network

Figure 3: Switched Network

Although initially switched networks were mainly used in telecommunica-

Figure 4: Hybrid Network

4.1 What is a Network Protocol?

• ISO - OSI (Open Systems Interconnect) Reference Model (OSI-RM)

• IBM’s SNA (Systems Network Architecture)

which are proprietary and hence are not truly interoperable.

ISO−OSI TCP/IP ITU−T IEEE802.X

Physical Link X.25−1 Physical

Figure 5: Layered Architectures of different Protocols

In Figure 6 is shown the architecture of the IEEE 802 family of standards

802.1 Network Control and Management

802.2 Logical Link Control (LLC) LLC

Figure 6: The Architecture of IEEE 802 LAN/MAN Standards

FDDI is a standard developed by ANSI-ASC X3T9 (Accredited Standards Com-

Nodes in a DQDB subnetwork are connected to a pair of buses flowing in

Figure 7: Open Bus DQDB Network

Figure 8: Looped Bus DQDB Network– DQDB Ring

SMDS (Switched Multi-megabit Data Service) which provides a connectionless

4.2 Transmission Mechanism

4.2.1 Packet Switching

Transmission mechanisms based on packet switching allow the multiplexing of

4.2.2 Frame Relay

4.2.3 Cell Switching

4.3 Physical Media

• Optical fibre cables

• It provides the highest degree of interoperability.

5.1 Building Internets and Intranets

Application Gateway Application

Network Router Network

Data link Data link

Physical Repeater Physical

Figure 9: Interconnection Components

5.3 Bridges and Switches

when protocol translation is necessary. A typical example is interconnecting two

5.6 Multiport-Multiprotocol Devices

6 A word about the Internet

Figure 10: Worldwide Internet Connectivity

Figure 11: Growth of the number of Web hosts on the Internet

to security, address space, quality of service, etc, in order to make it sustainable

Figure 12: Growth of the number of Internet hosts