AFRICAN VIRTUAL UNIVERSITY

BY L. CHINYAMA
Table of Contents

Preface 1
Chapter 1 2
Computer Networks 2
Advantages of networking 2
The concept of networking 2
Network Types 3
Local Area Network 3
Metropolitan Area Network 3
Wide Area Network 3
Network Categories 3
Peer-to-peer 3
Server – based 3

Chapter 2 6
Network Topologies 6
The Bus Topology 6
Star Topology 8
Ring Topology 9
Mesh Topology 10
Hybrid Topologies 11
The Star bus Topology 11
The Star ring Topology 11

Chapter 3 12
Media Access Methods 12
CSMA/CD (Carrier-Sense Multiple Access With Collision Detection) 12
CSMA/CA (Carrier-Sense Multiple Access With Collision Avoidance) 13
Token Passing 13
Demand Priority 13

Chapter 4 14
Network Architecture 14
Ethernet and IEEE 802.3 15
Fast Ethernet (IEEE 802.3u 16
Gigabit Ethernet (IEEE 802.3z and 802.3ab) 18
Wireless (IEEE 802.11) 19
100VG-AnyLAN (IEEE 802.12) 19
Token Ring (IEEE 802.5) 20
Other Network Standards 20
Fibre Distributed Data Interface (FDDI) 20
Apple LocalTalk 20
Chapter 5 21
Transmission Media 211
Thinnet (Thin Ethernet) 222
Thicknet (Thick Ethernet) 24
Unshielded Twisted Pair (UTP) 25
Shielded Twisted Pair (STP) 25
Wiring Standards for Twisted Pair cables 28
T-568A Standard 30
T-568B Standard 31
Fiber Optic Cable 36

Chapter 6 43
Signal Transmission 43
Baseband Transmission 43
Broadband Transmission 43

Chapter 7 44
Wireless Media 44
Infrared 45
Laser 45
Radio 45

Chapter 8 46
Network Adapter Cards 46
NIC Performance 49
NIC speed 49
Network Adapter Card addresses 50
How ARP resolves media access control addresses for local traffic 51
How ARP resolves media access control addresses for remote traffic 52
ARP Cache 53
Plug and Play (PnP) Technology 53

Chapter 9 54
Network Operating Systems 54
Interoperatibility 54

Chapter 10 55
The OSI (Open System interconnection) model 55
How the layers interact 60

Chapter 11 61
Protocols 61
Protocol Suite or stack 61
TCP Well-Known Ports 65
Registered Ports 65

Data Communications and Networking Copyright © 2005, African Virtual University ii

Chapter 12 69
Network Connectivity Devices 69
Repeaters 69
Hubs 70
Bridges 72
Router 75
Brouter 81
Switch 82
Gateways 85

Chapter 13 86
Ethernet Frame 86

Chapter 14 88
Switching Methods 88
Packet switching 88
Circuit switching 89
Message switching 90

Chapter 15 91
Internet Protocol (IP) Addresses 91
Subnet Mask 92
IP Address Classes 92
IP Address Rules 93
Finding the Network ID 94
Determining the Maximum number of hosts per subnet 95
Subnetting 96

Chapter 16 105
DHCP Server 105
Using multiple DHCP servers 106
How a Client Computer obtains an IP Address from A DHCP server 107
DHCP Lease Renewal 108

Chapter 17 109
Network Management 109
Account Management 109
Performance Management 110
Security Management 111
Fault Management 111
Configuration Management 113
Network Management Tools 114
Hardware 114
Digital Voltmeter (DVM 114

Data Communications and Networking Copyright © 2005, African Virtual University iii
 Oscilloscope 114
Advanced Cable Testers 114
Time Domain Reflectometer (TDR) 114
Protocol Analyser 115
Software 116
Simple Network Management Protocol (SNMP) 116

Chapter 18 117
WAN Transmission 117
T1 118
Switched 56 118
Packet- switched Networks 118
X.25 119
ISDN (Integrated Service Digital Network) 122
Frame Relay 123
ATM (Asynchronous Transfer Mode) 123

Glossary 125

Data Communications and Networking Copyright © 2005, African Virtual University iv

Preface

There is a lot of literature on Data Communications and Networking in the form of books
and on the Internet. A search on the Internet will present to you a lot of sites and tutorials,
some with indispensable information. However, the numerous literatures on broad a
subject, coupled with newer technologies being introduced by some vendors, tend to
create a lot of confusion; especially for beginners. This book clears all the confusion and
presents the subject in a straightforward and explicit way, giving valuable information
and detailed explanations. It also dispels the fears that most people have concerning this
subject.

It is my hope that you will enjoy the subject.

Data Communications and Networking Copyright © 2005, African Virtual University 1

1
Computer Networks

A computer network is a system of interconnected computers that have software

components installed on them to facilitate communication and sharing of resources.

Advantages of Networking

• It allows the sharing of resources. Resources include data, applications and

peripherals. Some resources are very expensive such that it is difficult to buy or
provide for each user in an organization. These resources can be shared on a
network, and this effectively cut costs. For example, instead of buying a high
quality printer for each member of an organization, only one high quality printer
can be shared on the network. Jobs are sent for printing on the shared printer,
where they are queued and printed according to the set priories.

• It allows standardization of applications and software so that everyone on the

network uses the same application and the same version of the application.
Standardisation of applications greatly simplifies support for the applications as
well.

• It allows online communication (sending messages back and forth), e.g. email
communication, and video conferencing where virtual meetings can take place
between users in different locations

• It also allows scheduling. Scheduling is the act of setting dates for events,
activities and appointments so that every member of the organization is made
aware of upcoming events. Most scheduling utilities use an electronic calendar
and flash messages to users as a reminder of upcoming events.

The Concept of Networking

At its most elementary level, a network consists of two computers connected to each
other by some communication medium so that they could share resources. All networking
no matter how sophisticated arose from this simple system.

Data Communications and Networking Copyright © 2005, African Virtual University 2

Network Types

There are three basic types of networks:

• Local Area Network (LAN)

• Metropolitan Area Network (MAN)
• Wide Area Network (WAN)

Local Area Network

This is a network confined to a small geographical area such as an office, building or a

single floor of a building.

Metropolitan Area Network

This is a network that covers an entire city by interconnecting multiple LANs within a
city. It is larger than a LAN but smaller than a WAN.

Wide Area Network

This is a network covering a large geographical area such as across cities, states and
continents. This is accomplished by connecting different LANs using services such as
dedicated leased phone lines, dial-up phone lines (both synchronous and asynchronous),
satellite links, and data packet carrier services.

Network categories

Networks can be divided into two main categories:

Peer-to-Peer

Server – based

Peer-to-peer networks require no dedicated servers and no administrators. There is no

hierarchy among the computers. All of the computers are equal and therefore known as
peers. Normally each computer functions as both a client and a server by accessing
shared resources from other computers on the network as well as sharing its resources
with other computers on the network. Peer–to-peer networks are also called workgroups.
The term workgroup implies a small group of people. Peer-to-peer networks typically
have a practical limit of 10 computers. In a peer-to-peer network there is no central
administration since there are no dedicated servers.

Data Communications and Networking Copyright © 2005, African Virtual University 3

 Peer-to-peer

Fig 1-1 Peer-to-Peer Network

Each user has control over all their resources and decides how these resources get shared.
Security in a peer to peer is share-level. This means that a password is set on each shared
resource. Any user knowing the password can access the resource.

Share-level security is also called password-protected shares. The lack of dedicated

servers in a peer-to-peer network makes the network simple and easy to setup, and lowers
the cost of setting up the network. Servers are one of the most expensive pieces of
equipment on a network.

Peer-to-peer networks are good choices for the needs of small organizations where
the users are located in the same general area such as a small office, and where
security is not a major concern and where the projected future growth of the
network is limited.

Operating systems such as Microsoft Windows NT workstation, Windows 9x, etc, have
inbuilt peer-to-peer networking capabilities (no additional software is required to setup a
peer-to-peer networks when using them).

Server-based networks offer greater security as well as centralized, coherent

administration of resources. The network administrator centrally administers shared
resources and network security. The central administration of security creates consistent
security policies that are applied to each user on the network. shared resources and
network security Server-based networks can accommodate a large number of computers
and users, unlike peer-to-peer networks that have a practical limit of 10 computers. A
server-based network consists of dedicated servers that hold various resources. Client
computers access the shared resources from the servers. The client computers usually
have the ability to carry out some of the processing locally. A dedicated server is a
computer that only functions as a server (making its resources available to the network
and responds to requests from client computers) and will not function as a client
computer at any time. Dedicated servers quickly service the requests from the client
computers and ensure the security of the resources they hold.

Data Communications and Networking Copyright © 2005, African Virtual University 4

server
 sse
server

Fig 1-2 Server-Based Network

The servers are usually located in a separate room away from the client computers. Only
the Administrator and some authorized users have access to this location. This ensures
physical security of the servers and keeps the curious users from tempering with the
servers. Each user should give authentication information (identify yourself and verify
that you are authorized to use the network) by supplying a valid username and password
combination to connect to the network, and this logon determines access to shared
resources. This type of security where a username and password combination is required
to log on is called user-level security or Access permissions. User-level security is a more
extensive security model and provides a higher level of control over access to resources.

Specialised Servers
As the network increases in size and traffic, more than one server will be required on the
network to accommodate the expanding needs of users. Some tasks such as e-mail,
application serving and database can be resource-intensive. Each of these tasks often
requires its own server to provide acceptable performance. Spreading the load among
many servers ensures that each task is performed in the most efficient manner possible.
The different types of dedicated servers include the following:

• File and print server. It manages user access and the use of file and printer
resources.
• Application server. It provides user access to application programs.
• Mail server. It manages electronic messaging among network users.

Combination Networks

A combination network is one that combines peer-to-peer networking and server based
networking. This gives the users the benefit of both peer-to-peer and server based
networks in one network. In this type of network there are dedicated servers running
server operating systems, allowing client computers to access resources from them. The
workstations or client computers are running the client operating systems, and can still
share resources among themselves as in a peer-to-peer network.

Data Communications and Networking Copyright © 2005, African Virtual University 5

2

Network Topologies

Network topology refers to the geometric arrangement or physical layout of computers,

peripherals and cable links in a network. There are three primary topologies:

• Bus
• Star
• Ring

The Bus Topology

The bus topology is the simplest and most commonly used topology for networking
computers. It is also known as a linear bus because computers are attached or connected
to a single linear cable called a trunk or backbone. Computers on the bus network
communicate by sending packets of data to each other. A computer will send a packet of
data to all computers on the network (broadcast). Each computer listens and examines
every packet on the wire to determine if the packet is intended for it, and accepts only
packets addressed to it. The bus topology is a passive topology because computers on the
network are not actively involved in the movement of data. On an active topology
computers regenerate the data signals and pass the signals to the next computer on the
network. The number of computers on the network affects network performance.
Performance degrades as more computers are added to the bus network.

Fig 2-1

Data Communications and Networking Copyright © 2005, African Virtual University 6

Signal Bouncing

Because the electrical signal representing the data is broadcast to all computers on a bus
network, it will travel from one end of the cable to the other. If the signal is allowed to
continue uninterrupted, it would keep bouncing back and forth along the cable and
prevents other computers from putting data on the transmission medium. A terminator,
which is a resistor placed at each end of the cable provides electrical resistance to ensure
that network traffic does not echo back through network and be received again by the
computers when it reaches the end of the cable. One of the terminators should be
grounded by attaching a conductor such as a wire to a position of zero electrical potential.

Advantages

• It uses the least amount of cable and is easy to expand.

• It is relatively simple, inexpensive and reliable.
• Failure of one computer does not bring the whole network down.

Disadvantages

• A cable break can bring the entire network down.

• Network becomes slower and slower as more and more computers are added to
the network.
• It is difficult to troubleshoot and isolate problems.

Data Communications and Networking Copyright © 2005, African Virtual University 7

Star Topology

On a star topology, each computer is connected by a cable segment to a central device,

which can be a hub, a switch or a repeater. Computers communicate with each other by
passing data through the hub. The network requires more cables because each computer
requires a cable segments to connect to the central device. If a computer fails or the cable
segment that connects the computer to the hub is broken, only that link is affected the rest
of the network continues to function normally.

Fig 2-2

Advantages

• Network expansion and reconfiguration are relatively simple. For instance, new
computers can be easily added to the network.
• Failure of one computer does not affect the rest of the network.
• Each computer is attached by its own cable segment to the central hub, which
makes it easy to isolate individual computers from the network by disconnecting
them from the wiring hub.
• Network management and monitoring can be centralized because all data goes
through a central device, which can be equipped with diagnostic features that
make it easy to troubleshoot network problems. For instance, a hub or switch may
also have extra features like LEDs that indicate activity and errors on each port,
making it even easier to isolate problems. Intelligent hubs can track network
activity such as the number of packets transferred and the occurrence of errors
within the packet.
• Can accommodate different wiring.

Disadvantages

• Uses a large amount of cable because each computer must have a separate cable
run from the central device.
• If the central device fails, the entire network fails.

Data Communications and Networking Copyright © 2005, African Virtual University 8

Ring Topology

On a ring topology, computers are connected one node to the other by a single cable
forming a circle. Computers on the ring network use a signal called a token to carry the
data from one computer to the other. The access method used by the computers to put
data on the communication medium is called token passing. Unlike the bus topology, the
ring topology is an active topology. Each computer on the ring acts as a repeater to boost
the signal and pass it to the next computer until it reaches the destination computer. The
receiving computer strips the data from the token and returns the token to the sending
computer with an acknowledgment. After verification, the token is regenerated and
release for use by any other computer requiring data transfer. While the token is in use by
one computer, other computers on the ring cannot transmit, because only the computer
with the token can transmit data. This ensures that there are no collisions and hence no
times spend waiting for computers to resend data after a collision.

Fig 2-3

Advantages

• All computers have equal access to data. During peak usage performance is even
for all users. Ring networks perform well with heavy network traffic.
• Each computer acts as a repeater and as a result there is little signal degradation.

Disadvantages

• Network expansion or reconfiguration will disrupt network operations.

• If one computer fails, the whole network is goes down.

Data Communications and Networking Copyright © 2005, African Virtual University 9

Mesh Topology

Apart from the three primary topologies, we also have the mesh topology. The mesh
topology connects each computer to every other computer on the network (each computer
has a point-to-point connection to every other computer on the network). Meshes use a
significantly larger amount of network cabling than the other network topologies, which
makes it more expensive. Every computer has multiple possible connection paths to the
other computers on the network for redundancy. As a result, a single cable break will not
stop network communications between any two computers.

Fig 2-4

The total number of connections required can be calculated using the following formula:

Where n is the number of stations.

Advantage

• Reliability is increased because there are alternative routes by which data can be
transmitted in the event of one of the links failing.

Disadvantages

• Additional hardware can make mesh topologies expensive.

• When a new computer or device is added, a connection to all existing devices or
computers must be made.

Data Communications and Networking Copyright © 2005, African Virtual University 10

Hybrid Topologies

These are topologies made by combining the three primary topologies. Examples are the
Star-Bus and the Star-Ring topologies.

The Star-Bus

It is a combination of the bus and the star topologies. It consists of several star topology
networks linked together by a linear bus trunk.

The Star-Ring

The Star-Ring is a combination of the ring and star topologies. It is sometimes called a
star-wired ring. It contains a main hub that has a logical ring configuration internally. The
main hub is connected to other hubs that can have computers connected to them.

A network’s topology affects its capabilities. The choice of topology determines the
following:

• The type of equipment required on the network

• The network growth
• The way a network is managed and ease of troubleshooting
• The type of cabling

Data Communications and Networking Copyright © 2005, African Virtual University 11

3

Media Access Methods

Media access method refers to the manner in which a computer gains access to the
physical transmission medium. There are three access methods that are used by
computers to put data on the network cable:

• Carrier Sense Multiple Access

With Collision Detection (CD)

With Collision Avoidance (CA)

• Token Passing
• Demand Priority

Carrier Sense Multiple Access with Collision Detection (CSMA/CD)

With Carrier Sense Multiple Access with Collision Detection (CSMA/CD), each
computer intending to send data to another computer on the network must first ‘sense’ or
check the network cable for traffic. If there is data on the cable, no computer can transmit
until the data has reached its destination and the cable is free of data. If there is no traffic
a computer can put data on the network cable. But there is a possibility that two or more
computers might have checked the cable at the same time and found it to be free. These
computers will go ahead and put data on the network cable, and a data collision occurs.
The computers involved in the collision enter a back-off phase in which they are given
random wait periods before they can attempt to retransmit again. This is necessary to
clear the collision or resolve the contention and makes sure that the computers will
attempt to access the network cable at different times when they start retransmission.
Although chances of collisions occurring are relatively small, collisions do take place and
cause network interruption each time they occur. CSMA/CD is a probabilistic method
because it is not always certain that data packets are delivered to their destination with no
collisions taking place. This method is also known as a contention method because
computers compete for the opportunity to send data.

Data Communications and Networking Copyright © 2005, African Virtual University 12

Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA)

With Carrier-Sense Multiple Access with Collision Avoidance (CSMA/CA), each

computer signals its intention to transmit data to the other computers on the network
before it actually transmits the data. In this way, the other computers will know when a
collision might occur and avoid transmission collisions. Because a computer has to first
broadcast its intention to send data before it actually sends the data, time is wasted during
the broadcasting. This makes CSMA/CA slower than CSMA/CD. CSMA/CA is not as
popular as CSMA/CD. CSMA/CA is a deterministic method because it is always certain
that data packets are delivered to their destination with no collisions taking place.

Token Passing

It is the access method used on the ring topology. With Token Passing, a computer that
intends to send data has to get hold of the token first. When the computer is in possession
of the token, it adds the data to be send in the appropriate field on the token. It then
passes the token to the next computer, which will also pass the token on until it reaches
the destination computer. The destination computer strips the data from the token and
returns the token to the sending computer with an acknowledgement. The sending
computer releases the token for use by other computers after receiving the
acknowledgement. While the token is in use by one computer, no other computer can
transmit data, because a computer has to be in possession of the token for it to be able to
send data.

Demand Priority

It is a relatively new access method designed for a specific network type called 100VG-
AnyLAN that uses a smart hub to handle multiple transmission requests from computers
on the network. Stations must request permission from the hub before they can transmit.
Transmissions can be prioritized (e.g. time sensitive traffic such as video traffic is given
first priority). If a computer intends to transmit, it signals the hub. Such a signal is called
a demand. The hub returns a signal giving permission to transmit a packet of data.

When multiple computers intend to transmit, the hub cycles through each of the
requesting computers. The highest priority request is serviced first. If the requests are of
the same priority, the requests are serviced by alternating among them in a round-robin
fashion. Computers that have not signaled to the hub that they intend to transmit are
skipped and do not take part in the transmission cycle.

Data Communications and Networking Copyright © 2005, African Virtual University 13

4

Network Architecture

It is a set of rules and design principles that specify the way in which media, hardware
and software are integrated to form a network.

The Institute for Electronic and Electrical Engineers (IEEE) created the 802 committee in
1980 to draft standards for local area networks (LANs). Within the 802 committee are
many sub-committees, each responsible for developing standards for specific LAN types
or managing a specific aspect of a LAN type. Some of the sub-committees and their
associated LAN types are as follows:

802.1: Spanning Tree standard for bridges used to interconnect similar LANs based on
the OSI Model.

802.1a: Network Management architecture.

802.1b: Network management protocols.

802.1D: Bridging standard.

802.1d: Source route standard.

802.1g: Remote bridge standard for wide area networks (WANS).

802.2: Logical Link Control (LLC).

802.2i and j: Acknowledge connectionless LLC service.

802.3: Carrier Sense Multiple Access with Collision Detection (CSMA/CD). Bus
Network (Ethernet).

802.3i: CSMA/CD over twisted-pair and fibre optics.

802.3u: Fast Ethernet.

802.3x: Gigabit Ethernet flow control.

802.3z: Gigabit Ethernet.

802.3ab: Gigabit on category 5 UTP cabling.

802.4: Token Bus Network (e.g. ArcNet)

Data Communications and Networking Copyright © 2005, African Virtual University 14

802.5: Token Ring Network.

802.5j: Fibre optic token ring.

802.6: Metropolitan Area Network (MAN).

802.7: Broadband Networks.

802.8: Fibre optic network.

802.9: Integrated voice and data service (ISDN).

802.9a: Isochronous Ethernet (ISONET).

802.11: Wireless LAN.

802.14: Integrated service over private cable.

802.16: Broadband Wireless Access.

Ethernet and IEEE 802.3

Ethernet is the de facto standard for Local Area Networks. It was originally developed by
Xerox Corporation in conjunction with Intel and Digital Equipment Corporation in the
1970’s. In 1980, when the IEEE Local Area Network committee was established to
standardize Local Area Networks, the IEEE 802.3 subcommittee adopted Ethernet and is
now responsible for the administration of Ethernet and IEEE 802.3. Ethernet and IEEE
802.3 are slightly different although they are supported together and used
interchangeably, Ethernet does not totally comply with the 802.2 standard.

The major difference between them exists in the data link layer, which is defined by
IEEE 802.2. IEEE 802.2 divides the data link layer into the Media Access Control
(MAC) sub-layer and the Logical Link Control (LLC) sub-layer. Ethernet does not divide
the data link layer or offer LLC services such as flow control.

The differences between IEEE 802.3 and Ethernet do not hinder hardware manufactures
because IEEE 802.3 and Ethernet both support MAC addresses and the same physical
layer. All networks that use Ethernet/IEEE 802.3(including IEEE802.3u, 802.3z and
802.3ab) use CSMA/CD as the access method.

Data Communications and Networking Copyright © 2005, African Virtual University 15

Table 4-1 Ethernet / IEEE 802.3 Standards

Designation Supported Maximum Transfer Topology

Media Segment Speed
Length (meters) (Mbps)
10Base-5 Coaxial 500 10 Bus

10Base-2 Thin Coaxial 185 10 Bus

(RG-58 A/U)

10Base-T Category 3 or 100 10 Star, using either

above simple repeater
unshielded hubs or Ethernet
twisted-pair switches
(UTP)

1Base-5 Category 3 100 1 Star, using

UTP, or above simple repeater
hubs

10Broad-36 Coaxial (RG- 3600 10 Bus (often only

58 A/U CATV point-to-point)
type)

10Base-FL Multimode 2000 10 Star (often only

optical fibre point-to-point)

The Ethernet/IEEE 802.3 standard defines a LAN that operates at 10Mbps using
baseband or broadband signal. This standard uses coaxial, twisted pair, or fibre optic, as
shown in the table above

Data Communications and Networking Copyright © 2005, African Virtual University 16

Fast Ethernet (IEEE 802.3u)

It is a faster version of the IEEE 802.3. It was originally developed by vendors such as
3Com, Cabletron, Digital and Intel. The IEEE 802.3u is responsible for fast Ethernet.
Fast Ethernet defines a LAN that operates at 100Mbps using baseband signal. Vendors
support Ethernet 100BaseX cards that use both 10Mbps and 100Mbps data rates. This
standard uses either twisted pair or fibre optic cable as shown in table 4-2.

Table 4-2 IEEE 802.3u

Designation Supported Maximum Transfer Topology

Media Segment Speed
Length (meters) (Mbps)
100Base-TX Category 5 100 100 Star, using either
UTP simple repeater
hubs or Ethernet
switches

100Base-FX Multimode 2000 100 Bus Star (often

optical fibre only point-to-
point)

Data Communications and Networking Copyright © 2005, African Virtual University 17

Gigabit Ethernet (IEEE 802.3z and 802.3ab)

Gigabit Ethernet is the fastest 802.3 LAN technology and is used primarily for network
backbones. The gigabit Ethernet standard transfers data at 1000 Mbps using CSMA/CD
as the access method. The two types of gigabit Ethernet are IEEE 802.3z and 802.3ab.
The 802.3z standard is specified for copper wire and fibre optic cables. The 802.3ab
standard specifies gigabit Ethernet over Category 5 UTP cable.

Table 4-3 IEEE 802.3z and 802.3ab

Designation Supported Maximum Transfer Topology

Media Segment Speed
Length (meters) (Gbps)
1000Base-SX Fiber-optic - 260 1 Star, using
two strands of buffered
multimode distributor hub
62.5/125 fiber (or point-to-
point)
1000Base-LX Fiber-optic - 440 (multimode) 1 Star, using
two strands of 5000 single buffered
multimode mode) distributor hub
62.5/125 fiber (or point-to-
or monomode point)
fiber
1000Base-CX Twinax(a cable 25 1 Star (or point-to-
that consists of point)
two
individually
insulated coax
cables wrapped
in a single
cover) 150-
Ohm-balanced,
shielded,
specialty cable
1000Base-T Category 5 100 1 Star

Data Communications and Networking Copyright © 2005, African Virtual University 18

Wireless (IEEE 802.11)

The original version of the IEEE 802.11 released in 1997 specifies two data rates of
1Mbps and 2Mbps.The other versions offer better data transmission rates and
communication distances. For instance the 802.11b operates at 11Mbps across a
maximum range of 50-300 feet from each other. A wireless network uses a device called
a Wireless Access Point (sometimes referred to as a transceiver) to connect computers on
a cabled network to computer that are not connected by cables within a defined network
area. The Wireless Access Point allows the sending and receiving of signals between
cabled computers and those that are not connected by cables. The computers that are not
connected by cables are each equipped with a wireless Network Interface Card (NIC) and
an aerial or antenna for sending and receiving signals. A Wireless Access point has at
least one fixed Ethernet port to allow the wireless network to be bridged to a traditional
wired Ethernet network, allowing wireless and wired devices to coexist on the same
network.

100VG-AnyLAN (IEEE 802.12)

The 100VG-AnyLAN proposal was originally developed by AT&T and HP, and is
managed by the IEEE 802.12 committee. It was developed to transport both Ethernet and
token ring frames. IEEE 802.12 is usually referred to as 100VG-AnyLAN. IT uses the
demand priority access method.100VG-AnyLAN can use many different cables (hence
the term “AnyLAN”), each with its own maximum segment length as shown in table 4-4.

Table 4-4 IEEE 802.12

100VG-AnyLAN Media

Cat 3 or Cat 4 UTP (must be voice grade)

Cat 5 UTP
STP (Type1)
Fibre optic (multimode)

Data Communications and Networking Copyright © 2005, African Virtual University 19

Token Ring (IEEE 802.5)

The token Ring network is specified in the IEEE 802.5 definition. IBM initially
developed token ring for its mainframe environment. The IEEE 802.5 standard complies
with the original IBM development. The standard uses data rates of 4Mbps and
16Mbps.It uses the token passing access method.

Table 4-5 IEEE 802.5

Media MAC Signal Propagation Speed Topologies Maximum

Method Method Connections
Twisted- Token Forwarded from 4 Ring 255 nodes per
pair passing device to device (or Mbps Star-using segment
(various port to port on a hub) 16 Token Ring
types) in a closed loop Mbps repeater hubs

Other Network Standards

These are standards that are not included with the IEEE LAN standards. These standards
are either proprietary or standardized by other organizations. They are:

Fibre Distributed Data Interface (FDDI)

FDDI was developed by American National Standards Institute (ANSI) and is

classified under the International Standards Organization (ISO) 9314. Fiber
Distributed Data Interface, shares many features with token ring, such as token
passing access method, and the continuous network loop configuration. It is not
a LAN standard since it can span or cover a distance of 200 kilometers,
connecting up to 1000 stations. FDDI has better fault tolerance because of its
use of a dual, counter-rotating ring that enables the ring to reconfigure itself in
the event of a link failure. FDDI also has higher transfer speed (100 Mbps),
compared to 4 or 16 Mbps for Token Ring. FDDI uses a physical ring topology.
Each device in the ring attaches to the adjacent device using a two stranded fiber
optic cable. Data travels in one direction on the outer strand and in the other
direction on the inner strand. When all devices attached to the dual ring are
functioning properly, data travels on only one ring. FDDI transmits data on the
second ring only in the event of a link failure.

Apple LocalTalk

Apple Company developed LocalTalk network for its Macintosh computers. It

uses CSMA/CA as the access method. The computers are connected using twisted-
pair cables in a bus topology. A Local Talk network is limited to a 300-meter span with a
maximum of 32 active nodes. Multiple LocalTalk networks can be interconnected by
using routers or other similar intermediate devices.

Data Communications and Networking Copyright © 2005, African Virtual University 20

5
Transmission Media

Media is the means by which signals travel from one network device to another. The
main categories of transmission media used in data communications networks are listed
below:

• Coaxial
• Twisted pair
• Fibre optics
• Wireless

Media can generally be classified as guided (e.g. copper and fibre) and unguided
(wireless) media. The media is responsible for carrying the signal in a network.

In guided communication media, the signal(s) are confined to the medium and do not
leave it (except for small leakage amounts), whereas in unguided communication media
the signal moves freely in the atmosphere.

Coaxial cable

Fig 5-1 Configuration Diagrams for Coaxial Cables

It consists of a core made of solid copper or stranded copper surrounded by an insulator.

A braided metal shield or wire mesh of aluminium or copper covers the insulator on top.
Sometimes a layer of foil is added to the braided metal shield to form a dual shield. Quad
shielding using 2 layers of foil and 2 layers of the braided metal may also be used and is
suitable for environments that are subjected to higher interference.

The shielding protects transmitted data by absorbing stray electromagnetic signals or

noise, so that the noise does not get to the core and distort the data. The conducting core
is responsible for carrying the data. It is separated or isolated from the wire mesh by an
insulator or dielectric.

Data Communications and Networking Copyright © 2005, African Virtual University 21

The core and the wire mesh must always be separated from each other. If they come into
contact, the cable will experience a short and noise or stray signals may flow to the core
and distort the data.

Coaxial cable is more resistant to interference and attenuation than twisted pair cables.
Attenuation is the loss of signal strength, which takes place as the signal propagates along
the cable due to the resistance of the wire to signal flow. After traveling a certain distance
(which is different for each type of cable), the signal gets so weak that it can be wrongly
interpreted at the receiver end.

There are two types of coaxial cables:

Thinnet (Thin Ethernet)

It is a flexible and easy to bend cable, being 0.25 inches in diameter. The IEEE
specification refers to this type of cable as 10Base2, referring to its main specification of
10 Mbps data rate, using baseband transmission and 185m (nearly 200m) maximum
segment length. It is also known as cheapernet because it is cheaper than the other types
of Ethernet cables. The cable between computers must be at least 0.5 metres (20 inches)
long. Networks that use this type of cable can support a maximum of 30 nodes per 185m-
cable segment. The 5-4-3 rule of thinnet allows for 5 cable segments of 185metres each,
connected using 4 repeaters, but only 3 of the segments can have some nodes attached.
The 5 cable segments allow the network to cover a total distance of 925m with a total of
90 nodes on the three segments that can take some nodes. Networks that use thinnet have
the cable connected to network adapter cards, hubs and other networking devices using
BNC connectors (British Naval connector or Bayonet –Concelman connector). Thinnet
coaxial cable can carry a signal to a distance of 185m before the signal suffers from
attenuation. Thinnet cable is classified in a group referred to as the RG-58 family and has
50-ohm impedance. Impedance is the resistance in ohms that the wire presents to
alternating current flow.

Fig 5-2 The 5-4-3 rule of Thinnet

Data Communications and Networking Copyright © 2005, African Virtual University 22

Table 5-1 Coaxial Cable Types and Characteristics

Cable Description

RG-58/U Solid copper wire core, 50 ohms. It is a thinnet cable used in 10Base2
Ethernet networks

RG-58A/U Stranded wire core, 50 ohms. It is thinnet cable used in 10Base2

Ethernet networks

RG-58C/U Military specification of the RG-58A/U.It has a wide temperature

range (-300 C to +1000 C). Used in 10Base2 Ethernet networks

RG-8, RG-11 . Thicknet, 50 ohms. Used in 10Base5 Ethernet network

RG-59 CATV coaxial cable. It is thicknet cable used for Cable TV

broadband transmissions. It looks very much like thinnet but they are
not interchangeable. It is a 75-ohm cable.

RG-6 Larger in diameter and rated for higher frequencies than RG-59,

but used for broadband transmission as well. It can handle the full

range of frequencies for cable, satellite and high definition TV, 75

ohm

RG-62 ARCnet network cable, 93 ohms. It is Commonly known as ARCnet

Data Communications and Networking Copyright © 2005, African Virtual University 23

Thicknet (Thick Ethernet)

Thicknet cable is also known as Standard Ethernet. It is a relatively rigid type of coaxial
cable, being 0.5 inches in diameter. IEEE specification refers to this type of cable as
10Base5, referring to its main specification of 10 Mbps data rate, using baseband
transmission, and 500m maximum segment length. Thicknet is generally used to provide
the network backbone. It has a thicker core than that of thinnet. The thicker the copper
core, the further the cable can carry the signal. This means that thicknet can carry signals
further than thinnet. Thicknet can carry a signal to a distance of 500m before the signal
suffers from attenuation. Therefore each thicknet cable segment should be up to 500m.To
transmit the signal further than 500m, a device known as a repeater is required.

When using thicknet, connection to the network adapter card is made using a device
known as an Attachment Unit Interface (AUI). The AUI connects to the thicknet cable
using a vampire tap, so called because it clamps onto the cable. The minimum cable
length between connections is 2.5 meters. It can support up to 100 nodes per segment
.The 5-4-3 rule applies to thicknet as well, giving a total length of 2500m and a total of
300 nodes supported. The Attachement Unit Interface connector is also referred to as a
DB15 connector. Because thicknet is a thicker cable, it is more difficult to work with. It
does not bend easily and is more expensive than thinnet cable. The major advantage of
thicknet over thinnet is that it can carry the signal further than what thinnet can do. The
outer covering or jacket for both thicknet and thinnet is made of pvc (polyvinyl chloride)
or Teflon.

Combining Thicknet and thinnet

Large networks often combine thicknet and thinnet cables. Thicknet cable is used for the
backbone to connect smaller thinnet-based networks because of its ability to support data
transfer over longer distances than thinnet. For instance on a star bus topology thicknet
forms the backbone while thinnet forms the branching segments that plug into repeaters
or hubs and connect to the computers.

Data Communications and Networking Copyright © 2005, African Virtual University 24

Twisted pair cables

In its simplest form, a twisted pair cable consists of a pair of insulated copper wires that
are twisted around each other and enclosed in a protective sheath called the outer jacket.
A number of twisted pairs are often grouped together and enclosed by the outer jacket.
The actual number of wire pairs in a twisted pair cable varies. Each pair is colour-coded
to differentiate it from the other pairs. There are two types of twisted pair cables:

Unshielded Twisted Pair (UTP)

Fig 5-3 UTP Cable Showing the Wire Pairs

The wires are twisted around each other to cancel out electrical noise (crosstalk) between
one pair and the other and to reduce electromagnetic interference (EMI) from external
sources. There are standards governing how many twists are required per foot of cable.

Shielded Twisted Pair (STP)

Fig 5-4 STP Cable Showing the Wire Pairs and the Shielding

Data Communications and Networking Copyright © 2005, African Virtual University 25

Shielded Twisted Pair cable (STP) is similar to UTP, except that a shielding of lead or
aluminium foil covers each pair of wires. The shielding increases the cable’s immunity to
EMI. Thus STP has more resistance to EMI compared to UTP. It is often used in
AppleTalk and IBM Token Ring networks. It is more expensive than UTP.

When LANS were initially developed, STP cabling was primarily used. However, a
large degree of twists in UTP cables has the effect of canceling out EMI. This makes
UTP cable suitable for LAN connections despite the lack of a shielding. UTP cable is
also preferred for LAN connections because it is considerably less expensive than STP.
Twisted pair cable is now replacing coaxial cable.

The Electronic Industry Association and Telecommunications Industry Association

(EIA/TIA) came up with a standard for commercial building wiring using UTP cables
called the EIA/TIA 568. The standard specifies five categories for UTP cables. Two
additional categories, category 6 and category 7 are also used commercially, but are not
standardized.

UTP cable is rated by category according to its use and data transmission speed as shown
in table 5-2.

Data Communications and Networking Copyright © 2005, African Virtual University 26

Table 5-2 Categories 0f UTP cables

Category Description

1 Traditional telephone cable, supports voice only, not data

2 Transmits data up to 4Mbps. Contains 4 pairs of wires.

Used for some token ring networks

3 Transmits data up to 10Mbps. Contains 4 pairs of wires.

Used for Ethernet Network.

4 Transmits data up to 16Mbps. contains 4 pairs of wires.

Used for some token ring networks.

5 Transmits data up to 100Mbps. Contains 4 pairs of wires.

Used for Ethernet and fast Ethernet. CAT 5 is the de
facto standard for LAN connections. CAT5E
(Enhanced) is an enhanced version of CAT5 used in
1000BaseT networks (Gigabit Ethernet). It has more
twists than traditional CAT5.The maximum cable
segment for CAT5E is 350m

6 Transmits data up to 155Mbps.Contains 4 pairs of wires.

Used for fast Ethernet

7 Transmits up to 1000Mbps.Used for gigabit Ethernet

Data Communications and Networking Copyright © 2005, African Virtual University 27

The maximum cable segment length for UTP and STP is 100m.UTP using the 10BaseT
specification is the most popular standard. 10BaseT means 10Megabits per second using
Baseband transmission over Twisted pair cable.

The main limiting factor of twisted pair cables is the distance the signal can travel before
attenuation. This limitation is caused by a phenomenon known as the skin effect:

Current flowing in a wire tends to flow only on the outside surface of the wire, thus
using less of the available cross sectional area. This has the effect of increasing the
electrical resistance of the wire, which in turn causes the attenuation of the signal
after a short distance.

To increase the cable length beyond 100m, the 100m cable segments can be joined
together using repeaters.

Twisted pair cables use an RJ-45 telephone connector at each end of the cable to connect
to computers and Network devices. The RJ-45 connector takes 8 wires. It looks like the
RJ-11 telephone connector, but the RJ-45 is slightly larger than the RJ-11 connector and
it will not fit into the RJ-11 telephone jack (slot). The RJ-11 connector only takes 4 wires
whereas the RJ-45 takes 8 wires.

Wiring Standards for Twisted Pair cables

The EIA / TIA define two standards for wiring: The T-568A and the T-568B standards.
Cat 5 UTP cable is the de facto standard for LAN connections. It has 4 pairs of wires but
only two pairs are used. The orange pair is used for transmitting data (balanced
transmitter) and the green pair is used for receiving data (balanced receiver). The other
pairs are unused. A balanced line driver or transmitter is at one end of one of these lines
and a line receiver is at the other end.

Data Communications and Networking Copyright © 2005, African Virtual University 28

A simplified schematic for one of these lines and its transmitter and receiver is as
follows:

Fig 5-5 Balanced Transmitter and Receiver

Twisted-pair Ethernet employs two principle means for combating noise. The first is the
use of balanced transmitters and receivers. A signal pulse actually consists of two
simultaneous pulses relative to ground: A negative pulse on one line and a positive pulse
on the other. The receiver detects the total difference between these two pulses. Since a
pulse of noise (shown in red in the diagram) affects the two wires in a similar way,
effectively, it will not distort the desired signal. In other words, it will not affect the
difference in potential between the two wires. Also, the magnetic field surrounding one
wire from a signal pulse is a mirror of the one on the other wire. At a very short distance
from the two wires the magnetic fields are opposite and have a tendency to cancel the
effect of each other out.

Data Communications and Networking Copyright © 2005, African Virtual University 29

EIA/TIA 568A and 568B STANDARDS

The T-568A and T-568B standards only differ in pin assignment for the two pairs that
handle data. The pin assignments for the unused pairs are exactly the same for the two
standards.

Table 5-3 T-568A STANDARD

Pin Color Signal

White/Green
1 RX data +

Green
2 RX data -

White/Orange
3 TX data +

Blue
4 Unused

White/Blue
5 Unused
Fig 5-6
Orange
6 TX data - RJ-45 Connector pin-Numbering
And The T-568A wiring Standard
White/Brown
7 Unused

Brown
8 Unused

Data Communications and Networking Copyright © 2005, African Virtual University 30

Table 5-3 T-568B STANDARD

Pin Color Signal

1 White/Orange TX data +

2 Orange TX data -

3 White/Green RX data +

4 Blue Unused

5 White/Blue Unused

6 Green RX data -
Fig 5-7
7 White/Brown Unused
RJ-45 Connector Pin-Numbering
8 Brown Unused And The T-568B wiring Standard
(Viewed with clip down and pins facing up)

NB. Odd pin numbers are striped while even pin numbers have solid coloured wires.

Data Communications and Networking Copyright © 2005, African Virtual University 31

Straight-Through Cables

The T-568B standard is the most widely used for straight-through cables. A straight
through cable has identical ends. There is no functional difference which standard you
use for straight through cable, although the T-568B seems to be popular. Straight-through
cables are used for the following connections:

• Switch to router cabling

• Switch to hub
• Switch to PC or Server Cabling
• Hub to PC

Fig 5-8 Computers Connected To a Hub Using Straight-Through Cables

Crossover Cables

To make a crossover cable, wire one end using the T-568A and the other end using the T-
568B. You should use a crossover cable to connect units with identical interfaces. If you
use a straight through cable, one of the two units must, in effect perform the crossover
function. Crossover cables are used for the following connections:

• Switch to switch
• Hub to Hub
• Router to Router
• PC to PC

Fig 5-9 Computers Connected Using Crossover Cable

Data Communications and Networking Copyright © 2005, African Virtual University 32

Table 5-4 Making Your Own Patch Cables

Patch Cable Assembly Instructions

1) Skin off the 8) Insert the
cable jacket wires into
approximately 1" or the
slightly more. connector
(pins facing
2) Un-twist each up).
pair, and straighten
each wire between
the fingers.

3) Place the wires

in the order of one
of the two standards
shown above (568B
or 568A). Bring all
of the wires
together, until they
touch.

4) At this point,
recheck the wiring
sequence with the
diagram.

5) Optional: Make a
mark on the wires
at 1/2" from the end
of the cable jacket.
6) Hold the grouped 9) Push
(and sorted) wires moderately
together tightly, hard to
between the thumb, assure that
and the forefinger. all of the
wires have
7) Cut all of the reached the
wires at a perfect end of the
90-degree angle connector.
from the cable at Be sure that
1/2" from the end the cable
of the cable jacket. jacket goes
This is a very into the
critical step. If the back of the

Data Communications and Networking Copyright © 2005, African Virtual University 33

 wires are not cut connector
straight, they may by about
not all make 3/16".
contact. We suggest
using the cutter on
the RJ-45 or RJ-11
gripping tools for
this purpose.
7B) Conductors 9) Place the
should be at a connector
straight 90-degree into a crimp
angle, and be 1/2" tool, and
long, prior to squeeze
insertion into the hard so that
connector. the handle
reaches its
full swing.
10) Repeat the process on the other end. For a straight through cable, use the same
wiring.
For a "crossover" cable, wire one end 568A, and the other end 568B.

11) Use a cable tester to test for proper continuity.

Data Communications and Networking Copyright © 2005, African Virtual University 34

Table 5-5 Notes Regarding Making Category 5 Patch Cable

The RJ-45 plugs are normally made for either solid conductors or stranded
conductors. It is very important to be sure that the plug that you use matches
the conductor type. It is extremely difficult to tell the difference between the
two by looking at them. When you buy these plugs, be sure to categorize, and
store them carefully. Using the wrong type can cause intermittent problems.
Some RJ-45, 8 Conductor Plugs that are sold are rated for both Solid and
Stranded cable.
Ordinarily, it would be taboo to untwist the pairs of any category 5 cable. The
one exception to this rule is when crimping on RJ-45 plugs. It would be
impossible to insert the wires into the channels without first untwisting and
straightening them. Be sure not to extend the un-twisting, past the skin point.
If you do it properly, you will wind up with no more than 1/2" of untwisted
conductors (up to 1/2" of untwist meets the CAT5 specification)
If the completed assembly does not pass continuity, you may have a problem
in one, or both ends. First try giving each end another crimp. If that does not
work, then carefully examine each end. Are the wires in the proper order? Do
all of the wires fully extend to the end of the connector? Are all of the pins
pushed down fully? Cut off the suspected bad connector, and re-terminate it.
If you still have a problem, then repeat the process, this time giving more
scrutiny to the end that was not replaced.
It is good to be prepared to make your own patch cables. There may be many
instances where you may fall short on supply, and making a cable will surely
get you out of a jam. However, there comes a point where the practicality
curve will lead you to factory made cables. Making several cables can be very
labour intense. Factory made cables typically have better tolerances, and
consequently have better quality than field made cables.

Data Communications and Networking Copyright © 2005, African Virtual University 35

Fiber Optic Cable (also called optical fibre)

b)

Fig 5-10 Structure of a Fibre Optic cable (a), (b) shows different type of Fibre
Cable Available

Fibre optic cable consists of a thin cylinder (the core) of high-purity glass (pure silicon
dioxide) or plastic surrounded by a cladding or insulator. The core is responsible for
carrying the light. Plastic core is easy to install but cannot carry the signal pulse as far as
glass core can do. The purpose of the cladding or insulator is to trap the light inside the
core using a technique known as Total Internal Reflection. The insulator is covered on
top by a protective outer jacket of PVC or Teflon.

In fibre optic cables, digital data is converted to light pulses, usually by light emitting
diodes and appropriate circuit. Data transmission rates vary from 100 Mbps up to
2Gbps.The maximum distance that a signal can be transmitted without a repeater is 2000
metres.

Data Communications and Networking Copyright © 2005, African Virtual University 36

Fibre optic cable has a number of advantages over copper cables:

• Data travels in the form of light, which is immune to crosstalk and

electromagnetic interference (EMI) from sources such as electric motors,
lightning strikes, etc. Thus can span long distances -2000 metres.
• It is more secure than copper cables because data travels in the form of light
instead of electrical signals that can be tapped and stolen.
• It has a large bandwidth than copper cables, giving it the ability to handle higher
data rates. Bandwidth refers to the data carrying capacity of the cable. The fastest
fibre can carry information up to 2 Gigbits per second, which is enough to carry
40 000 telephone conversations or 250 television channels.
• It has lower losses than copper cable, which increases the allowable distance that
the signal can travel before it suffers from attenuation.
• Fibre optic cable posses no threat in hazardous or flammable environments such
as petrochemical plants where a spark could trigger an explosion.
• Modern fibre optic cables are at least one quarter the weight of copper cables of
similar capacity.
• Supports voice, video and data.

Fibre optic cable has the following disadvantages:

• It is more expensive than copper cables.

• Requires expertise to install (Difficult to install): The Connectors for fibre optic
cables are more complex to attach to the cable and require precise physical
alignment. For example if several fibre cable segments need to be joined end to
end, the couplers must ensure that the fibres are correctly aligned and butted to
reduce any joining losses to a minimum.
• Not very flexible, difficult to work with.
• If the cable is bend sharply (beyond the recommended bend radius) signal losses
occur in the fibre

Fibre optic cable acts as a wave-guide allowing light to propagate through it. It uses the
principle of reflection (Total Internal Reflection) for light to propagate through.

Data Communications and Networking Copyright © 2005, African Virtual University 37

To understand how light propagates along a fibre optic cable, we need to look at two
important principles: Refraction and Reflection.

Refraction

Fig 5-11 A Light Beam Being Refracted

When a light beam strikes the surface of a material and passes through the material with
its path bend as shown in the figure above, it is said to be refracted. The angle of
incidence ( 1) is the angle at which the light is projected onto the surface of the material.
It is measured from a line drawn perpendicular to the surface of the material. The angle
( 2) at which the light leaves the material is known as the angle of refraction.

When light is launched in the fibre optic core at a small angle of incidence ( 1) it is lost
(refracted) through the cladding at an angle ( 2). The relationship between the angle of
incidence ( 1) and the angle of refraction ( 2) is given by Snell’s Law:

(1)

Where n1 is the refractive index of the fibre core and n2 is the refractive index of the
cladding. By analyzing Snell’s Law, conditions could be established under which light
will not be refracted, but reflected:

If the angle of incidence is gradually increased, a point will be reached where light no
longer escape from the core of the cable, but remains in the core and propagates inside
the core. The smallest angle of incidence, at which light is first retained inside the core, is
called the critical angle ( c). At the critical angle, the light will travel horizontally inside
the core (θ2 = 90o). The value of the critical angle is found by rearranging equation (1)

For 1 = c and θ2 = 90o, equation (1) can be written as:

n1sin c = n2 sin 90o, but sin 90o = 1

Thus n1sin c = n2, which can be written as:

Sin c = n2/n1

Data Communications and Networking Copyright © 2005, African Virtual University 38

Therefore (2)

Any angle of incidence greater than the critical angle will allow total internal reflection.
All angles are measured with respect to the normal (a line drawn perpendicular to the
core and cladding)

Thus two conditions are required for Total Internal Reflection to take place in the fibre
core:

i) The refractive index of the core (n1) should be greater than that of the cladding (n2).
Typical values are1.47 for the cladding and 1.5 for the core.

ii) The angle of incidence ( 1) should be greater the critical angle ( c)

The refractive index is a measure of the ease with which light travels through a material
and is different for different materials.

Fig 5-12 Light Propagating Along the Fibre Core by Total Internal Reflection

Fibre-Optic cable Light Sources

There are two major sources of light for fibre-optic cables:

LED (Light Emitting Diodes) -LED is commonly used with mono-mode fibre. It is
relatively weak.

ILD (Injection Laser Diodes) – ILD emits a strong, intense, narrowly focused light beam.
It is commonly used with multi-mode fibre, which helps to counteract multi-mode’s
lower performance.

Data Communications and Networking Copyright © 2005, African Virtual University 39

Mode of Propagation

In fibre optic terminology, mode simply means the path that light will follow along the
fibre cable. If there is only one path for light to travel along the fibre, the cable is referred
to as a mono-mode fibre. If there is more than one path, the cable is referred to as multi-
mode fibre.

In mono-mode fibre, the core has a very small diameter typically 8-12µm. With this
exceedingly small core diameter, light entering the fibre will be restricted to travel along
one path and in a straight line. Mono-mode fibre has the advantage of lower modal
dispersion and can be used over longer distances without repeaters. Mono-mode fibre is
faster than multi-mode fibre because there is no separation of the light beam, which
causes the light pulses to arrive at the end of the cable at different times in multi-mode
fibre. However the exceedingly small core diameter makes the fibre very expensive and
difficult to work with, as a result, the majority of LAN applications use multi-mode fibre.

Fig 5-13 Light Traveling a Straight Path in Mono-Mode Fibre

Multi-mode fibre has a relatively larger core diameter (typically 50µm) than mono-mode
fibre. This allows light to take more than one path when propagating along the core.
Multi-mode fibre has the disadvantage that the beam of light traveling in each mode will
travel a different distance. This is because the higher angle beam bounces from side to
side many more times than the lower angle beam. As a result two beams of light from the
same pulse, entering the fibre at the same time arrive at the end of the fibre separated in
time. This spreading of a pulse due to multiple rays traveling using different paths of
different distances and speed through a fibre is called modal dispersion (sometimes
referred to as Differential Mode Delay DMD)

Data Communications and Networking Copyright © 2005, African Virtual University 40

Fig 5-14 A Comparison of the Core Diameters for Mono-Mode and Multi-
Mode Fibres

Index Profile

The index profile is how the refractive index of a fibre optic cable behaves throughout the
cable. There are two types of index profiles:

Step Index

Step index fibre has a central core of uniform refractive index. The core is surrounded by
an insulator or cladding with a uniform refractive index which is less than that of the
central core. Step Index is so called because there is an abrupt change in refractive index
as we move from the cladding to the core of the fibre (the refractive index ‘steps’ up) as
shown below:

Fig 5-15

Data Communications and Networking Copyright © 2005, African Virtual University 41

 Fig 5-16 Light Beam moving in a zig-zag manner in Multi-Mode Step Index
Fibre

Graded Index

Graded Index fibre has a core of non-uniform refractive index. In this type of fibre, the
refractive index of the core is highest at the center of the core and gradually decreases
until it matches that of the cladding as we move away from the centre of the core. One
way of improving the operation of multi-mode fibre is to use graded index fibre instead
of step index fibre. The varying refractive index of the core across the diameter forces the
rays to travel a curved path as they move along the fibre. The rays that travel near the
outer edge of the fibre travel faster than those that stay near the middle, so that although
they travel a longer distance, they arrive at the same time. This means that light pulses are
not distorted in the same way that they would be in a step index fibre.

Fig 5-17

Fibre optic cables are unidirectional (light only travels in one direction) so two cables are
used, one for sending and another for receiving.

Data Communications and Networking Copyright © 2005, African Virtual University 42

6

Signal Transmission

Two techniques are used to transmit data over a cable. These are baseband and broadband
transmissions.

Baseband Transmission

Baseband transmission is used to transmit a digital signal over a single digital

transmission line or channel. All the available frequencies of the channel (bandwidth) are
used to transmit a single digital signal. Baseband systems use repeaters to strengthen
attenuated signals

Broadband Transmission

Broadband transmission is used to transmit analog signals. It divides the channel

bandwidth into multiple sub-channels. Each sub-channel will have its own range of
frequencies and is capable of transmitting its own analog signal. This allows multiple
analog signals to be transmitted simultaneously over a single cable in the same direction.

Broadband media is unidirectional, signals move in one direction. To enable

communication in both directions, there must be two data paths, one for transmitting data
and the other for receiving data. There are two ways in which two-way communication
can be achieved:

• Using dual cable (Dual cable Broadband)

• Split cable (Mid split broadband)

In dual cable broadband, each computer is attached to two cables, one for transmitting
data and the other for receiving data.

The split system uses a single cable, but divides the bandwidth into two channels each
using different frequencies. One channel is used to transmit signals and the other is used
to receive signals. Broadband systems use amplifiers to regenerate the analogue signal at
its original strength.

Data Communications and Networking Copyright © 2005, African Virtual University 43

7

Wireless Media

A wireless network is not completely free of cabling. It consists of wired devices that
communicate with devices on a traditional cabled network. A device called a wireless
access point (also called a Transceiver) is used to transmit and receive data between the
wireless devices and devices on the cabled network. Each computer in a wireless network
has a network card and an aerial. Wireless network adapter cards do not need to be
physically connected to a network cable, they use the aerial (antenna), which is omni
directional to transmit or receive data. The Wireless Access Point connector has at least
one fixed Ethernet port that connects it to a traditional (cabled Ethernet network) Ethernet
network at one end while receiving signals from wireless devices at the other end as
shown in fig 7-1.

Fig 7-1 Wireless Ethernet Network

Wireless networks are suitable for people or employees who move around frequently
within their working environments, such as sales people and doctors. For instance,
doctors and medical personnel could use notebook computers connected to a wireless
LAN as they visit patients it their rooms. In other words, a wireless network increases
mobility of the users.

Data Communications and Networking Copyright © 2005, African Virtual University 44

Wireless LANS could also be used in buildings were cabling would be difficult or
inconvenient to implement. For example it can be used in Laboratories where excessive
cables are a hindrance to those who need to move around frequently in the laboratory.
Wireless networks are also used for temporary installations within networks.

Wireless networks provide a certain degree of portability allowing equipment to be

moved easily without worrying about disconnecting and reconnecting wires.

Wireless networks allow users to extend their networks beyond the physical limitations of
cabling

The drawback of wireless LANS is that they do not have the data rate capabilities of
wired networks.

There are four techniques that are used to transmit data in wireless LANS:

Infrared (IR)

Infrared beam is used to transmit data between Devices. Infrared transmissions are
limited to 30m (100 feet). Within this range however, infrared is relatively fast. IR
technology is familiar to many people because of its use in TV and Radio remote control
units. IR is normally a line-of-sight technology, which means there must be no object
obstructing the path of the infrared light between the transmitting and receiving devices.
The need for this unobstructed pathway is a drawback for IR-based wireless
communications. It supports transmission speeds of up to 10Mbps.

Laser

Laser is an acronym for “Light amplification by stimulated emission of radiation”. Like

IR technology, it also requires a line-of-sight between communicating devices.

Radio

Data transmission over radio can be implemented using broadcast in two ways:
• Narrowband radio
• Spread Spectrum radio

Narrowband radio is also called single frequency radio. Narrowband radio transmits data
in a similar way to radio stations. A transmitter sends a signal on a specified frequency
and a receiver tuned to that frequency picks up the signal. It can support data
transmissions up to 4.8 Mbps, but can broadcast a signal up to 5 kilometers. It does not
require a line-of-sight. The only important requirement is that both the sender and
receiver must be tuned to the same frequency for them to communicate. When data is
transmitted over narrowband radio, it is easy for an unauthorized listener to intercept the
signals. A more secure and more reliable radio technology originally developed by the
military is called spread spectrum.

Data Communications and Networking Copyright © 2005, African Virtual University 45

There are two types of spread spectrum radio:

Frequency Hopping Spread Spectrum (FHSS) – With FHSS, the transmitter hops from
one frequency to another, and the receiver must know the frequencies, the pattern and the
timing of the hops. This makes it difficult for an unauthorized person to intercept the
signal.

Direct Sequence Spread Spectrum (DSSS) – This technique uses special encoding
method that creates a redundant bit pattern for each bit of transmitted data. This provides
fault tolerance because if some bits are damaged during transmission, the original data
can still be recovered without need for retransmission

Spread Spectrum usually transmits at 250Kbps, but some implementations offer outdoor
transmission rates of 4Mbps.

Data Communications and Networking Copyright © 2005, African Virtual University 46

8

Network Adapter cards

A network adapter card is an add-in card that contains the circuitry required for
communication over a network. It acts as the physical interface between the computer
and the network cable. It is also called a Network Interface Card (NIC)

It also Control the flow of data between the computer and the cabling system to ensure
that the receiving computer is not overwhelmed by too much incoming data. A
component called a transceiver (Transmitter/Receiver) on the network adapter card
handles incoming data for the computer from the network cable, and output data from the
computer to the network cable.

a) b)
Fig 8-1 Network Adapter Card Types

The network adapter card fits into one of the slots on the computer’s motherboard called
a bus .The network adapter card is attached to the network cable by a connector. The
number of bits that the bus can transfer at a time gives a measure of the bus speed. For
example the older 8-bit buses could move 8-bits of data at a time. Buses were later
designed to handle 16-bits and 32-bits of data. The types of buses in the computer
determine the types of network adapter cards that can be used to connect to the network.
There are several types of buses, as shown in Table 8-1.

Data Communications and Networking Copyright © 2005, African Virtual University 47

Table 8-1 Bus Types

BUS Abbreviation BUS Architecture Description

ISA Industry standard 16-bit system developed in 1983.The

Architecture majority of computers use this type of
bus. It is less expensive than EISA
EISA Extended Industry 32-bit design. It is compatible with ISA
Standard standard (i.e. EISA slots will accept ISA
Architecture cards)
MCA Micro Channel IBM’s proprietary bus design for its
Architecture PS/2 computers Can handle 16 or 32
bits of data at a time. A very fast
architecture. It is rarely seen today.
NUBUS Macintosh NUBUS Apple’s 32-bit design that is used only
on Macintosh computers
VESA Video Electronics 32-bit design. It was designed to
Standard accelerate video displays
Association
PCI Peripheral It is the bus that is used in modern
Component computers. It can operate at 32Mbps or
Interconnect 64Mbps.It supports plug and play (PnP)
devices. The slots on the motherboard
have a beige or white –coloured plastic
and are shorter than ISA/EISA slots It is
the most popular and fastest bus design.
PC card PCMCIA (Personal An input/ output bus that interfaces
Computer Memory several different devices, which are the
Card International size of credit cards in laptop and
Association) notebook computer. It has the ability to
‘hot swap’ or change the cards without
restarting the computer.

Data Communications and Networking Copyright © 2005, African Virtual University 48

NIC Performance

Although each network adapter card is designed to meet certain standards, some network
adapter cards provide features that increase the data transmission speed. Table8-2 shows
the features that can improve network adapter card performance.

Table 8-2 Some Features Of Network Adapter Cards

Feature Description
Shared Adapter Memory Network Adapter cards contain RAM that is shared with
the computer
Shared system Memory The Network Adapter card selects a section of computer
memory used to process data
Bus Mastering The Network Adapter bypasses the CPU and assumes
control of the system bus and loads the data directly into
the system memory without interrupting the CPU. This
method is available only with certain buses that support
it (e.g. PCI)
RAM buffering RAM chips on the Network Adapter card create a buffer
to hold the data overflow. Data overflow occurs when
the Network Adapter card receives more data than it can
transmit to the computer across the bus at one time.
Onboard Microprocessor The Network Adapter card contains a microprocessor so
that it does not rely on the cpu to help data processing

Boot ROMs

In some environments, security is such an important consideration such that workstations

normally do not have disk drives. Without disk drives, users are not able to copy
information to either floppy or hard disks, and therefore cannot take any data from the
company or organization.

Diskless computers boot from a network adapter card that has a special chip called a boot
Rom. This chip contains the boot program that connects the computer to the network
when the computer is switched on.

NIC Speed

Some Network adapter cards can operate at either 10 or 100Mbps and includes an auto-
negotiation capability that lets the adapter sense the speed of the network and adjust to
that speed without human intervention.

Data Communications and Networking Copyright © 2005, African Virtual University 49

Network Adapter Card Addresses (Hardware Addresses or Physical Addresses)

Each Network Adapter card has a unique physical address that is hard wired or burned
into its read-only memory (ROM) on the adapter card during the manufacture process.
This address is also called the Media Access Control (MAC) address. It is a 6-byte
address. An example of a MAC address is shown below:

40:8E:0A: 12:34:56

Each byte is separated from the other by a colon. The first three bytes identify the
manufacturer of the adapter card, and the last three bytes represent the unique address of
the adapter card. Authority and registration of media access control addresses are
overseen by the Institute of Electrical and Electronics Engineers (IEEE). Currently, the
IEEE registers and assigns unique numbers for the first three bytes of the media access
control address to individual manufacturers. Each manufacturer can then assign the last
three bytes of the media access control address to individual network adapters. MAC
addresses are called physical addresses as opposed to logical addresses (IP Addresses).
Logical addresses are found at the network layer (layer 3). They are used to send data
between hosts on different networks (remote network). Physical addresses are found at
the data link layer (layer2), and are used to send data between hosts on the same physical
network. In order to use MAC addresses to send data a protocol called Address
Resolution Protocol (ARP) resolves the IP Address of a machine to its MAC address.

Data Communications and Networking Copyright © 2005, African Virtual University 50

The following illustration shows how ARP resolves IP addresses to hardware addresses
for hosts on the same local network.

Fig 8-2 How ARP Resolves Media Access Control Addresses For Local Traffic

In this example, two TCP/IP hosts, Hosts A and B, are both located on the same physical
network. Host A is assigned the IP address of 10.0.0.99 and Host B is assigned the IP
address of 10.0.0.100.

When Host A tries to communicate with Host B, the following steps resolve Host B's
IP address (10.0.0.100) to Host B's MAC address:

a) Host A checks its own local ARP cache for a matching hardware address (MAC
address) for Host B.
b) If Host A finds no mapping in the cache, it broadcasts an ARP request frame to all
hosts on the local network with the question "What is the MAC address for
10.0.0.100?" Both hardware and software addresses for the source (Host A), are
included in the ARP request.

Each host on the local network receives the ARP request and checks for a match
to its own IP address. If a host does not find a match, it discards the ARP request.

c) Host B determines that the IP address in the ARP request matches its own IP
address and adds a MAC address/IP address mapping for Host A to its local ARP
cache.
d) Host B sends an ARP reply message containing its MAC address directly back to
Host A.
e) When Host A receives the ARP reply message from Host B, it updates its ARP
cache with a MAC/ IP address mapping for Host B.

Data Communications and Networking Copyright © 2005, African Virtual University 51

Once the media access control address for Host B has been determined, Host A
can send IP traffic to Host B by addressing it to Host B's media access control
address.

ARP is also used to forward IP datagrams to local routers for destinations that are not on
the local network. In this situation, ARP resolves the media access control address of a
router interface on the local network. The following illustration shows how ARP resolves
IP addresses to hardware addresses for two hosts on different physical networks
connected by a common router.

Fig 8-3 How ARP Resolves Media Access Control Addresses For Remote Traffic

In this example, Host A is assigned an IP address of 10.0.0.99 and Host B uses an IP

address of 192.168.0.99. Router interface 1 is on the same physical network as Host A
and uses the IP address 10.0.0.1. Router interface 2 is on the same physical network as
Host B and uses the IP address 192.168.0.1.

When Host A tries to communicate with Host B, the following steps resolve Router
interface 1's IP address (10.0.0.1) to its MAC address:

a) Because the data being send by Host A is intended for a computer on a remote
network, IP (Internet Protocol) determines that the forwarding IP address to be
used to reach host B is 10.0.0.1 (the IP address of its default gateway). Host A
then checks its own local ARP cache for a matching hardware address for
10.0.0.1.

b) If Host A finds no mapping in the cache, it broadcasts an ARP request frame to all
hosts on the local network with the question "What is the MAC address for

Data Communications and Networking Copyright © 2005, African Virtual University 52

 10.0.0.1?" Both hardware and software addresses for the source (Host A), are
included in the ARP request.

Each host on the local network receives the ARP request and checks for a match to its
own IP address. If a host does not find a match, it discards the ARP request.

c) The router determines that the IP address in the ARP request matches its own IP
address and adds a hardware/software address mapping for Host A to its local
ARP cache.
d) The router then sends an ARP reply message containing its MAC address directly
back to Host A.
e) When Host A receives the ARP reply message from the router, it updates its ARP
cache with a MAC/IP address mapping for 10.0.0.1.

Once the media access control address for Router interface 1 has been determined,
Host A can send IP traffic to Router interface 1 by addressing it to the Router interface 1
MAC address. The router then forwards the traffic to Host B through the same ARP
process.

ARP Cache

To reduce the number of address requests, a client normally caches resolved addresses for
a short period of time and store them in a part of the computer memory known as a cache.
Machines with ARP caches search their caches before sending a request. By using a
cache, a machine can accommodate new additions to the network dynamically. Without
the ARP cache, the resulting ARP request and replies would generate additional network
traffic, which would impact on network performance.

Plug and Play (PnP) Technology

Most computer hardware (including network adapter cards) can now be automatically
configured on a system using Plug and Play technology. Plug and Play is a mechanism by
which a computer automatically configures hardware devices added to it. It uses the
computer BIOS, Operating System, device drivers and system resources for the automatic
configuration. Each device manufacture is responsible for assigning the PnP ID for each
product and storing it in the hardware. Each Plug and Play device must:

• Be uniquely identified by a Plug and Play device ID number in order to allow the
operating system to recognize the device.
• State the services it provides and resources it requires.
• Identify the driver that support it so that the operating system can load the
appropriate driver and configure it.

After installing a PnP device, it should work when you start up your computer.
Plug and Play eliminates the need to set jumpers, and it also makes
installation of devices quick and easy.

Data Communications and Networking Copyright © 2005, African Virtual University 53

9
Network Operating systems

A network operating system (NOS) is system software that controls the network and
allows multiple users to access network resources. A network operating system has all the
features of a stand-alone operating system but offers many more facilities in a complex
environment. Some of the functions of a network operating system include managing
multiple users on the network, providing access to file and print servers and
implementing network security. Examples of Network operating systems are Novell
Netware, Microsoft Windows NT /2000/xp Linux and UNIX.

A network operating system is made up of two components:

• Client software
• Server software

The client software is the software that is installed on the client computers. Examples of
client operating systems are Windows 9x, Windows NT workstation, Windows 2000
professional, etc. The client software consists of a component called a redirector that does
the following:

It intercepts requests at a client computer. After intercepting a request, it determines if the

request can be satisfied on the local computer or on another computer on the network or
the server. If the request cannot be satisfied locally, the redirector redirects the request to
the appropriate computer where the request can be met.

An example of a local request is a request for a directory listing on the local computer
hard drive. A request that cannot be satisfied locally (remote request) can be a request for
a resource on a mapped network drive.

The sever software is the software that is installed on the server. It enables the server to
share its resources such as data, printers, fax machines and disk drives with other
computers on the network. It also allows the level of sharing to be determined to allow
different users different levels of access to shared resources. It also provides the means
for the Network Manager or system administrator to manage users. Some Network
operating systems provide network management tools that enable the administrator to
track network behaviour and performance.

Interoperatibility

It is the integration of computers running different operating systems in a single network

in a manner that allows them to share resources. For example a network might include a
Novell server, Unix server and an NT sever with client computers running operating
systems such as windows 9x, Macintosh client and OS/2 client. While such a network can
offer many challenges, it will work quite well if it is properly planned and implemented.

Data Communications and Networking Copyright © 2005, African Virtual University 54

10
The Open System Interconnection (OSI) Reference model

It is a conceptual model developed by the International Standards organization (ISO) that

describes how data communications takes place using seven layers as well as
standardizing the data communication process. It is called any open system because it is a
published standard available for use by any vendor, thereby facilitating the integration of
hardware and software from different vendors. It removes the restrictions that are found
in closed networks. A closed network is one that is restricted to one manufacture’s
equipment and standards. Such networks are not attractive to the user because they
restrict the choice of hardware and software that can be used.

It makes it possible for different manufactures to come up with networking products

(hardware and software) that are compatible (allows mix and match of networking
products). This gives the user a wide range of manufactures’ products to choose from.

It also breaks down the networking process into seven fundamentally different layers.
The breaking down of the networking process into the seven layers makes it easy to
understand the networking process. Each of the seven layers represents a different stage
and type of processing that occurs when data is to be transmitted from one computer to
the other. In other words, each of the seven layers tries to deal with specific aspects of the
networking process.

Data Communications and Networking Copyright © 2005, African Virtual University 55

Table 10-1 Seven Layers of the OSI Reference Model

Layer Layer Description

Number
Application 7 This is where the user interacts with the computer. It
provides a set of application programming interfaces (APIs)
for application programs to access network resources such
as files and services such as file transfer, e-mail messaging.
An API is a set of functions and commands that an
application needs to create links with various services on
the network.
Presentation 6 The presentation layer transforms data into a format suitable
for communication. This includes compression (e.g. MPEG
for video compression and coding), encryption, encoding
(e.g. ASCII, EBCDIC) and decryption of data.
Session 5 Initiates, manages and terminates a communication session
between two communicating stations. A communication
session consists of service requests and service responses
between the stations. These requests and responses are
coordinated by protocols implemented at the session layer.
It also determines the form of communication to be used
(simplex, half duplex or full duplex), and regulates which
station transmits, when and for how long.
Transport 4 It accepts packets from the session layer and breaks down
large packets for transport across the network. It ensures
end-to-end integrity of the data packets during transmission
(ensures that packets are delivered and arrive at their
destination error free, in sequence and with no losses or
duplications). This is achieved through various means of
error control, error recovery and verification of sequence
numbers for the packets.
Network 3 It provides routing of data packets to their destination by
determining the best route/path that data can take to the
destination. The network layer may also split large packets
into smaller packets if the packet size is larger than the
largest frame size that the data link layer will accept.

Data Communications and Networking Copyright © 2005, African Virtual University 56

Data Link 2 It accepts Data packets from the network layer and
repackages them into units called frames that are passed to
the physical layer for transmission. Each frame is
transmitted with the necessary synchronization, error
control, and flow control information to provide reliable
transfer across the physical link.

The IEEE differentiates the OSI Data Link layer into to

separate sub-layers. These are the Media Access Control
(MAC) sub-layer and the Logical Link Control (LLC) sub
layer.

The MAC sub-layer is responsible for putting data on the

transmission medium. It provides a system through which
stations can share the communication medium. This
function is called Media Access Control .It permits media
access methods such as CSMA/CD, Token Passing, etc. It
also handles physical addressing to uniquely identify each
computer on the network by a pre-coded address that is
hard-wired in to the network interface card.

LLC sub-layer establishes and maintains the connection

(logical connection) between one station and the other while
transmission is taking place. It is also responsible for
framing and flow control. Flow control is synchronizing the
sending and receiving stations so that the receiving station
is able to handle the data at the rate at which it is being sent.
This keeps rapidly transmitting stations from ‘drowning’ or
‘overwhelming’ slower receivers. The LLC is defined in the
IEEE 802.2 specification.

Physical 1 It is the bottom most layer of the OSI reference model. It

layer specifies the type of communication media (coaxial, twisted
pair, or fibre optic cable), the type of connectors to be used
and the timing signals for data. It also defines the electrical
specifications (maximum and minimum voltage levels for
signals, e.g. +5v for logic 0 and – 5v for logic 1 in the
RS232 communication standard)

Data Communications and Networking Copyright © 2005, African Virtual University 57

The following mnemonics help to remember the seven layers in their exact order:

From top to bottom: All People Seem To Need Data Processing

From bottom to top: Please Do Not Take Sales Persons’ Advice

Movement of Data from the sending station to the Receiving Station

Data from the sending station Descends through the OSI layers from the Application
layer to the Physical Layer. At each layer the data is formatted using some protocols that
work at that particular layer so that it can be successfully passed on to the next layer.
When the data reaches the target station, it moves up the layers in reverse order until it
reaches the physical layer of the target station. At each layer the formatting information
that was added at the sending station is removed, when the data reaches the application
layer of the receiving station, all the formatting information has been removed and the
data is now in human readable form.

Data Communications and Networking Copyright © 2005, African Virtual University 58

The following diagram illustrates the movement of data from the sending station to the
target or receiving station.

Sending Station Receiving Station

Application Application

Presentation Presentation

Session Session

Transport Transport

Network Network

Data Link Data Link

Physical Physical

Transmission Medium

Fig 10-1 Movement of Data Form the Sending Computer to the Receiving
Computer

Data Communications and Networking Copyright © 2005, African Virtual University 59

How the Layers Interact

Each layer communicates and works with layers immediately above and below it and its
peer layer in the other station as shown in the following diagram:

Fig 10-2 Interaction of the OSI Layers

For example, the data link layer in computer A, communicates with the network and
physical layers of computer A, and the data link layer of computer B

Data Communications and Networking Copyright © 2005, African Virtual University 60

11

Protocols

A protocol is a set of rules and procedures governing the format of data and the orderly
exchange of data over a network.

The protocol defines how computers identify one another on a network, the form that
data should take in transit, and how the data is processed once it reaches its final
destination.

There are many different protocols that work at the different layers of the OSI reference
model. Each protocol implements the functions of the OSI layer(s) at which it works.

Protocol Suite or Stack

It is a set of protocols that are typically used together at some or all the seven layers of
the OSI reference Model. Examples of protocol stacks are TCP/IP, AppleTalk and IBM’s
Systems Network Architecture (SNA)

TCP/IP (Transmission Control Protocol/ Internet Protocol)

TCP/IP is a protocol suite that comprises more than 100 different protocols that work
together to provide a set of network functions at the different layers of the OSI reference
Model. The name TCP/IP is derived from two of its protocols (TCP and IP) although it
comprises many protocols.

The US Department of Defense (DOD) developed TCP/IP as part of its Defense

Advanced Research Project Agency (DARPA), which was aimed at interconnecting
various defense department computer networks.

In order to come up with this network, there was need for a protocol that was efficient,
reliable, that has low overheads and that is routable. This led to the development of
TCP/IP. Some top minds worked in collaboration in the development of TCP/IP. These
were affiliated with major Universities in the USA.

Data Communications and Networking Copyright © 2005, African Virtual University 61

Advantages of TCP/IP

Reliable.TCP/IP has been in use for more than 20 years and has proven to be a stable and
robust protocol suite. It guarantees the transfer of data packets from one host to
the other by offering failure recovery and the ability to handle high error rates.

Interoperatibility.TCP/IP is a cross-platform protocol suite that allows communication

among different computers (computers using different operating
systems) over different types of networks.

Scalable. It can be used in small to large networks

Routability. TCP/IP is a routable protocol that allows data to be moved from one
network to the other through different communication paths (routes)

Low Overhead. TCP/IP has minimal amount of extra bits that function as packaging
bits for the data being transmitted.

Useful Utilities. It has many utilities available for troubleshooting.

Internet Connectivity. It enables computers to be connected to the Internet.

TCP/IP and OSI Model

Each TCP/IP layer corresponds roughly to the functionality of the equivalent layer in the
OSI model. Table 11-1 identifies the relationship between the two. The TCP/IP model is
also referred to as the Department of Defense (DoD) reference model.

Table 11-1 A Comparison of the OSI Layers and TCP/IP Layers

OSI TCP/IP
7 Application 4 Application
6 Presentation
5 Session
4 Transport 3 Transport
3 Network 2 Internet
2 Data Link 1 Network Interface
1 Physical

Data Communications and Networking Copyright © 2005, African Virtual University 62

The Application layer is broadly equivalent to the Application, presentation and session
layers of the OSI reference model. Examples of protocols found at this layer are Telnet,
FTP, HTTP and SMTP.

The Transport layer is similar to the OSI transport layer, but with elements of the OSI
session layer functionality. Two protocols found at this layer are TCP and UDP.

The Internet layer is responsible for the routing of data packets across networks. It
corresponds to the Network layer of the OSI model. IP and ARP are both found at the
Internet layer.

The Network Interface layer combines the data link and physical layers of the OSI
reference model. It deals with physical and mechanical specifications for the media and
connectors, as well as electrical specifications for signals and access methods.

The following are some of the core protocols that make up TCP/IP:

Telnet (Terminal Emulation over a Network). It is a protocol that is used for remote login
(allowing remote terminal sessions to be created on one computer from another)

FTP (File Transfer Protocol). It enables the sending and receiving of files from one
TCP/IP host to another.

SMTP (Simple Mail Transfer Protocol). It is used for the transfer of mail messages
(email messages) between Internet hosts.

TCP (Transmission Control Protocol). A connection oriented protocol that guarantees the
delivery of data packets from one station to the other without errors.

UDP (User Datagram Protocol). It is a connectionless protocol (i.e. it does not guarantee
the delivery of data packets). There is no logical connection to be established
during communication when using this protocol. This means no
acknowledgement or handshaking takes place between stations when using this
protocol. As a result, data can be transmitted with some errors. The advantages to
be realized from UDP are that data transmission is faster since no time is wasted
in handshaking and the UDP header is smaller compared to the TCP header. This
implies that UDP has very little overhead compared to TCP

The decision whether to use TCP or UDP lies in the application. Some
applications are designed to use TCP and others are designed to use UDP

IP (Internet Protocol). It is a connectionless protocol that is used for IP addressing and

routing of data packets on the network. IP addressing is assigning a 32-bit address
to a host or node on an IP network so as to uniquely identify it.

Data Communications and Networking Copyright © 2005, African Virtual University 63

ICMP (Internet Control Message Protocol). It is a protocol that is used for reporting error
messages regarding the delivery of data packets. An example of an error message
that can be reported is the “Host unreachable” or “Destination unreachable” error
message. ICMP is also used to perform diagnostics and to verify connections
using the PING utility. The PING utility sends an echo request data packet asking
for acknowledgment that a particular IP Address exists on the Network. If a
computer with the requested IP Address exists and receives the request, it notifies
the sending computer of its existence. The ping utility will then report the
existence of the IP address and how long it took to get to that station (the echo
reply). If the station does not exist, the sending station receives a time-out error
message and the request packet is dropped after the TTL has expired.

IGMP (Internet Group Management Protocol). It is a protocol that manages multicasting

to IP groups. All hosts that belong to a particular multicast group “listen” for
packets of data that are sent to their multicast group.

Port Numbers

TCP and UDP protocols communicate using ports. A port is a number that an application
uses as a sending and receiving address for data. An applications listens to its appropriate
port so that it can access and process data appropriately. Each port is used by a specific
application or service. Each port number is a 16-bit integer value. A port is assigned a
number between 0 and 65535, which is coded into the TCP or UDP protocol header. TCP
or UDP port numbers between 0 and 1024(ports ranging from 1 to 1023) are the well-
known ports (also called reserved ports). They are the ports to which standard
applications listen for services that a host can provide.

Data Communications and Networking Copyright © 2005, African Virtual University 64

Some of the well-known TCP ports are listed in table 11-2

Table 11-2 TCP Well-Known Ports

Port Number Associated Application/Service

21 FTP (File Transfer Protocol)

22 Secure shell (SSH) for secure login

23 Telnet

25 SMTP (Simple Mail Transfer Protocol)

53 DNS (Domain Name Server)

79 Finger

80 HTTP (Hyper Text Transfer Protocol)

110 POP3 (Post Office Protocol Version 3)

The IP address, type of port (UDP or TCP) and the application port number are usually
combined into one functional unit called a socket. An example of a socket is as follows:

131.107.2.200:TCP:80

Registered Ports (user ports)

The ICANN does not control registered port numbers, which range from 1024 to 65535
and are considered non-privileged. Therefore, any process can use them.

Data Communications and Networking Copyright © 2005, African Virtual University 65

IPX/SPX IPX/SPX stands for Internetwork Packet Exchange/ Sequenced Packet
Exchange. It is a protocol suite that was developed by Novell. It is used primarily
by Novell NetWare. As its name implies, it has two parts:

IPX It performs the functions of the network layer on the OSI reference model, that of
forwarding and routing data packets.

SPX It is a transport layer protocol that is responsible for numbering data packets and
guaranteeing the delivery of the packets to their destination and in sequence.
Although IPX/SPX is owned by Novell, it is also used and supplied with
Microsoft products. Nwlink IPX/SPX is Microsoft’s version of Novell Netware
IPX/SPX.

SNA (Systems Network Architecture) It is IBM’s protocol suite. It was one of the
pioneering protocols. It was developed in the 1960s for the mainframe computers
and the AS400 computers.

AppleTalk It is a set of protocols that perform most of the functions of the seven
layers of the OSI reference model. It was developed by Apple Computer
Company to network its Macintosh computers in the Apple
Network Architecture.

Many additional networking protocols exist. Some of them are as follows:

NETBIOS (Network Basic Input/Output System) It is a session layer protocol. It has the
disadvantage that it is non-routable. However, it is very easy to work with.

NETBEUI (NETBIOS Extended User Interface) It is a smaller, faster and efficient

transport layer protocol. NETBEUI is non-routable. An alternative protocol to
NETBEUI is NETBIOS over TCP/IP (NBT). NBT is routable because it uses
TCP/IP as its transport.

Data Communications and Networking Copyright © 2005, African Virtual University 66

High Level Data Link Control (HDLC)

It was developed by the International Standards Organization (ISO). It has become the
most widely accepted data link protocol. It offers a high level of flexibility, adaptability,
reliability and efficiency of operation for data communication needs. HDLC operates at
the Data link layer of the OSI reference model. It supports both half duplex and full
duplex communication modes in point-to-point networks.

HDLC Frame Format

Flag Address Control Data CRC Flag

8-bits 8-bits 16-bits Variable 16-bit 8-bits

Fig 11-1

The beginning and end of an HDLC frame are marked by flag character sequence
(01111110 in binary or 7E in Hexadecimal delimits the frame)

The header of the HDLC frame consists of the address and control fields. The address
field always contains the address of a secondary station. When the primary station
transmits, it is the receiving secondary station's address, if it is a secondary station
responding, it is its own address. The control field identifies the type of frame and is also
used to control the communication process using commands and responses and sequence
numbers to maintain the flow of data.

The trailer consists of a 16-bit Cyclic Redundancy Check (CRC), which detects any
errors during transmission by allowing integrity of the transmitted frame data to be
checked.

HDLC Frame Types

There are three types of frames, which are as follows:

• Information Frame (I-Frame), this carries data and message acknowledgements

• Supervisory Frames (S-Frame), this carries commands and responses
• Unnumbered Frame (U-Frame), this carries additional commands, and responses.

HDLC has two operational modes:

Normal Response Mode (NRM), also known as Synchronous Data Link Control (SDLC)

Link Access Procedure Balanced (LAPB)

Of the two, LABP tends to be more popular than SDLC.

Data Communications and Networking Copyright © 2005, African Virtual University 67

HDLC specifies the following: types of stations on a link

• Primary station
• Secondary station
• Combined station

A primary station is a station that controls all the other connected stations on the link. It is
also responsible for the organization of data flow on the link

A secondary is a station under the control of the primary station. It has no ability or direct
responsibility for controlling the link. It only sends response frames when requested by
the primary station.

A combined station is a station that incorporates the functions of the primary and
secondary stations. All combined stations are able to send and receive commands and
responses without any permission from any other station on the link.

LAPB is a bit-oriented synchronous protocol that provides transparency when

transmitting data. To prevent the flag information from being interpreted as data, HDLC
uses a technique known as bit stuffing or zero insertion. When the flag sequence is used
to represent data, a ‘0’ bit is inserted after any sequence of five consecutive 1’s in the
payload (actual data bits) as illustrated below:

Data before zero insertion 011101111111

Data after zero insertion 0111011111011 (zero bit added after five consecutive ones)

At the receiver end, upon seeing five consecutive 1s followed by a zero, the receiver
assumes that the zero was stuffed and removes it. The actual flag pattern is not subject to
bit stuffing so as to distinguish when the flag sequence represents flag information and
when it represents data. Bit stuffing only occurs when the flag sequence is used to
represent data.

The time between frame transmissions is called the inter-frame fill. During the inter-
frame fill, continuous flags are transmitted on the link to keep the link active.

Data Communications and Networking Copyright © 2005, African Virtual University 68

12

Network Connectivity Devices

Connectivity device is a general term that refers to the simple and complex devices that
are used to connect cable segments and network segments together.

Simple connectivity devices include the following types of connectors:

BNC T– Connector. It connects two cable segments and also offers a connection to the
NIC. If only one cable is to be connected to the BNC T-connector, a terminator must be
connected to the other side of the T-connector.

BNC Barrel Connector. This is a straight cylindrical unit to which a cable attaches at
each end, thus enabling you to join two pieces of cable to extend the total cable length.

Warning. The use of barrel connectors should be kept to a minimum because of signal
loss that can occur at each connection point.

RJ Connectors

RJ (Registered Jack) connectors are so called because they are registered with the Federal
Communication Commission (FCC). The RJ connectors consist of a plug and a
receptacle (also called the Jack).
RJ-11 connectors are used on telephone cables while RJ-45 connectors are used on
computer networks.

Complex Connectivity Devices

Simple connectivity devices merely connect cables and devices, but complex
connectivity devices do a lot more on top of that. Some of the complex connectivity
devices are as follows:

Repeaters

A Repeater works at the physical layer of the OSI reference model. It is the least
expensive and most common device used to extend the length of the network cable by
joining two segments of cables and regenerating or boosting the signal from one segment
before passing it on to the next segment. The two segments to be joined must use the
same media access method, protocol, and transmission technique. For example an 802.3
CSMA/CD (Ethernet) LAN segment cannot be joined to an 802.5 (Token Ring) LAN
segment

Data Communications and Networking Copyright © 2005, African Virtual University 69

Fig 12-1 A Repeater Joining Some Network Segments

However, a repeater does not give any more bandwidth or allow data to be transmitted
faster. In other words, a repeater cannot be used to enlarge a network beyond the
capabilities of its underlying architecture.

Repeaters can connect different physical media, e.g. thinnet and fibre optic. Repeaters
offer no filtering; they send every packet of data from one cable segment to the other
even if the data consists of malformed packets. Broadcast storms also pass through the
repeater from one segment to the other. A broadcast storm is a condition or phenomenon
that occurs in networks when some of the computers on the network send broadcast
packets that flood the whole network causing network overload by saturating the capacity
of the network or bandwidth.

Repeaters also allow isolation of segments in the event of failures or fault conditions.
Disconnecting one side of a repeater effectively isolates the associated segments from the
network. Using repeaters simply allows you to extend your network distance limitations
and to regenerate signals.

Hubs

A hub is a central device through which all network traffic must pass through. There are
two types of hubs:

• Passive Hub
• Active hub

A passive hub only acts as connection points; it does not regenerate the signal. Passive
hubs do not require electrical power to run.

An active hub regenerates and retransmits the signal the same way a repeater does Thus
an active hub can be used to extend the distance that a signal can be transmitted. An
active hub is sometimes referred to as a multiport repeater. An active hub requires
electrical power to run.

Data Communications and Networking Copyright © 2005, African Virtual University 70

NB signals become weak after traveling some distance. Regeneration is the process by
which the signal is given more power to travel further.

Hubs create a physical star network while maintaining a logical bus or ring configuration
internally.

Benefits of using Hubs

• Many hubs have different ports to accommodate a variety of cabling types, e.g. a
single hub can take coaxial and twisted pair cables.
• Hubs also offer centralized monitoring of network activity and traffic. Many
active hubs contain diagnostic capabilities to indicate whether or not a connection
is working.

Data Communications and Networking Copyright © 2005, African Virtual University 71

Bridges

Bridges are devices that operate at layer 2 of the OSI reference model. That is why they
are widely referred to as data-link devices. When bridges were introduced in the early
1980’s, they were used to connect two homogeneous LAN segments, that is LAN
segments with identical network architecture (i.e. 802.3 to 802.3 or 802.5 to 802.5).
Nowadays, bridges are available that can connect different networks or LAN segments
using different access methods (for example joining an Ethernet segment to a token ring
segment), and different types of cables. Bridges can perform the same functions as
repeaters, extending the length of a network, but they can also reduce traffic by
segmenting the network based on media access control addresses.

Fig 12-2 A bridge connecting some network segments

Because bridges operate at the layer 2 of the OSI, they use hardware addresses. A bridge
builds a routing table of computer MAC addresses. It listens to all traffic. Each time a
bridge is presented with a packet, it captures the source address together with the port and
segment on which the computer is connected. In other words, it “learns” which
computers are on which segment and which computer is associated with a particular
address.

NB The address table built by a bridge is called a routing table because it is used to
determine to which side of the bridge the packet should be routed. Don’t confuse this
with the routing table used by a router. A bridge’s routing table uses hardware addresses,
but a router’s routing table is based on higher-level IP addresses.

A bridge is intelligent enough to do some routing as follows:

A bridge filters network traffic based on the media access control (MAC) address of a
packet to determine whether a specific packet belongs on a local segment or another
segment. The action taken by a bridge when it receives a packet depends on which
segments the source and destination nodes are on. When a bridge receives a packet from
a computer on one of the connected segments, it examines the destination address for that
packet. It then determines whether the destination computer resides on the same network
segment by verifying the MAC address from the MAC address table.

Data Communications and Networking Copyright © 2005, African Virtual University 72

 • If the source and destination addresses of a data packet are located on the
same segment, the packet will not be forwarded across the bridge to another
segment.

• If the destination address is in the routing table and on the other segment, the
bridge forwards the packet to the other segment. In other words, the bridge passes
the packet across the bridge to the other segment.
• If the destination address is not in the routing table, the bridge forwards the packet
to all segments except the source segment (segment on which the source computer
resides).

Bridges can also reduce network traffic by segmenting the network (dividing one network
into segments). This isolates traffic or problems and reduces traffic on both segments,
making more bandwidth available.

Although a bridge can reduce network traffic by segmenting the network traffic based on
media access control addresses, it cannot filter broadcasts. Bridges simply pass on all
broadcast packets (packets with no entry in the MAC address table). If the destination
address of a packet is not in the bridge’s MAC address table, the bridge forwards the
packet to all of the connected nodes, thereby contributing to broadcast storms.

Bridges are slow in operation. There can be some data loss in bridges, and if a data loss
occurs, the lost data cannot be recovered. However, new technology has been developed
in the form of switches and routers that replaced bridges.

Data Communications and Networking Copyright © 2005, African Virtual University 73

There are three types of bridges:

i) Transparent Bridge

The traditional bridge (also called transparent bridge) requires little or no setup. The
bridge will automatically initialize itself and configure its own routing information after it
has been enabled. A transparent bridge is intelligent, because it makes all routing
decisions:

• It stores the known MAC addresses of the end stations in a table.

• It examines the destination MAC address and automatically sends the packet to
the destination.

Transparent bridges are also known as spanning tree, IEEE 802.1 D. The bridge is said
to be transparent because computers on the network are not aware of its presence (it is
invisible to the computers). Computers on the network just see one large network
without realizing that the network is made up of segments connected by the bridge.
Transparent bridging is found primarily in Ethernet networks.

ii) Source Routing Bridge

It is a special type of bridge used on Token Ring networks. It does not maintain a routing
table of its own, but depends on host computer to make routing decisions. Each computer
or device on the network must maintain its own table of all the routes (using hardware
addresses) required for communication with other network computers. To send data, the
source computer must incorporate the routing information in the header of the packet to
be sent. Packets are forwarded using information specified in the packet’s MAC header
as opposed to a path determined by the bridge. The routing information in the packet’s
MAC header lists the hardware addresses needed for the packet to reach its destination.
The source-routing bridge examines this routing information and forwards the packet
according to this information.

iii) Source-Routing Transparent (SRT) Bridges

This occurs when transparent and source routing bridges coexist. It combines the
algorithms of transparent bridging and source route bridging to enable communication in
mixed Ethernet and Token ring environments.

Data Communications and Networking Copyright © 2005, African Virtual University 74

A network can have more than one bridge. This provides fault tolerance, but can lead to
bridging loop problem, which occurs when there are multiple paths between two points
and packets end up going around in circles. This creates unnecessary traffic (broadcasts).
The Spanning-Tree Algorithm (STA) was developed to solve the bridging loop problem.
The STA continually explores the network so that a link failure is discovered to avoid
loss of connectivity or creation of new loops.

Bridges can also be classified as:

• Local bridges
• Remote bridges

A Local bridge provides direct connection between multiple LAN segments that are in
the same area (close together).

A remote bridge is used to connect multiple LAN segments that are far apart using
communication links such as leased telephone line/public dial-up lines.

Router

A router is a networking device that provides inter-network connectivity, i.e. it

interconnects different networks allowing data to be moved from one network to the
other. Without routers and routing protocols, communication would be limited to those
nodes or stations that are on the same physical network. Conceptually, routers are similar
to bridges, except that they operate at the network layer (layer3) of the OSI reference
model, and gather information about the networks to which they are connected. Like
bridges, Routers provide filtering, but does so using network addresses to route data
packets across multiple networks instead of MAC addresses. Routers also do not pass on
broadcast traffic. This reduces network traffic significantly and prevents propagation of
broadcast storms. Routers are more intelligent than bridges; they make complex routing
decisions to select the optimal path to a given destination from among multiple paths.
Routers can be configured to support single or multiple protocols. A router creates and
maintains a table, known as a routing table, of the routes available to get to the different
networks to which it is connected. The router uses this information, together with
information about prevailing network conditions, to determine the best route for a given
packet.

Data Communications and Networking Copyright © 2005, African Virtual University 75

Routing tables can be configured manually using static routes or dynamically using some
routing protocols. Whether a router is configured statically or dynamically or a
combination of both, the objective is the same, to facilitate communication between hosts
or nodes on different networks.

Static routing requires the system Administrator to manually setup and configure the
routing table and to specify each route between networks. Routing information entered
this way does not change. The router does not have the capability of adjusting to changes
in the network (it is not possible to learn new routes). As a result, if a defined route fails
or becomes unavailable the destination becomes unreachable. This technique is simple to
implement and simple to understand, but it becomes difficult to keep routing tables
accurate when the network is growing rapidly and too much work if the network is very
large. It has the advantage of eliminating all traffic related to routing information updates:
Static routing conserves bandwidth because the routers do not generate route update
traffic, which consumes a lot of network bandwidth.

In dynamic routing, routers have the ability to discover the best routes using some
protocols. In a network a link may fail unexpectedly, or a new link may be established. A
dynamic routing protocol discovers these changes and automatically adjusts its routing
table and informs other routers of these changes through the transmission of various
messages, such as the routing update message which consists of all or part of a routing
table. The routers use this information to make packet-by-packet decisions about how to
send data across the network. The protocols facilitate the exchange of routing information
between networks and upgrade routing tables automatically, thus requiring a minimal
amount of setup and configuration. The process of updating routing tables to reflect
changes that have occurred in the network is called convergence.

Some of the protocols that are used for dynamic routing are:

• Routing Information Protocol (RIP)

• Interior Gateway Routing Protocol (IGRP)
• Enhanced Interior Gateway Routing Protocol (EIGRP)
• Open Shortest Path First (OSPF)
• Exterior Gateway Protocol (EGP)
• Border Gateway Protocol (BGP)
• DEC’s DNA Phase IV
• DEC’s DNA Phase V
• Novell’s Netware Link Service Protocol (NLSP)

Data Communications and Networking Copyright © 2005, African Virtual University 76

One way of classifying routing protocols is according to whether they are Interior
Gateway protocols (IGPs) or Exterior Gateway Protocols (EGPs).

Interior Gateway protocols are those protocols that are used to exchange routing
information within a single Autonomous System (AS). An Autonomous System is a
collection of networks under a single administrative authority (i.e. a single company or
organization) that use a common routing strategy. An autonomous system is also known
as a routing domain. Examples of Interior gateway protocols are: OSPF, RIP, IGRP/
EIGRP.

Exterior gateway protocols (EGPs) are those protocols that are used for inter AS routing
(routing between separate autonomous systems/routing domains), so that each AS may be
aware of how to reach others through the Internet. Examples of exterior gateway
protocols are EGP and BGP.

Routing protocols are also classified as Distance Vector or Link-State according to the
algorithms they use to route data.

Distance Vector routing protocols use distance vector algorithm to calculate the routes for
network traffic. The name distance vector is derived from the fact that routes are
advertised as vectors of distance and direction, where distance is defined in terms of a
metric and direction is defined in terms of the next-hop router. The primary distance
routing algorithm is the Bellman-Ford algorithm Examples of distance vector protocols
are:

• RIP
• IGRP
• Apple Talk’s Routing Table Maintenance Protocol (RTMP)
• DEC’s DNA Phase IV

The major disadvantage of distance-vector protocols is slow convergence.

When an RIP router is initialized, it broadcasts a request for routing table information
from all the neighbouring routers it can see. The routing information that it receives from
the neighouring routers includes, routing metric, and the hop count for each known route.
The information received is then entered into the local routing table and is used by the
router to determine the shortest path to each destination. Once the table is built, the router
sends a RIP broadcast every 30 seconds indicating its known destinations and the cost (in
hops) to get to each.

RIP has the following disadvantages:

RIP is used in small to medium-sized networks. A destination is considered unreachable

if the hop count is greater than 15. The maximum number of hops allowed with RIP is
15.

Data Communications and Networking Copyright © 2005, African Virtual University 77

RIP determines the best path by looking at the number of hops between the two end
nodes. Assuming all hop counts take the same amount of time, the optimum path is the
one with the smallest hop count. This technique ignores differences in line speed, the line
utlisation and all other metrics, many of which can be important factors in choosing the
best path between two nodes.

RIP has slow convergence: It takes a relatively long time for network changes/updates to
be propagated to all routers. This is because RIP sends full table updates instead of
sending only the changes across the network at 30 seconds intervals. This process takes
several minutes for a distant router to be notified of a change across the network.

This is so because the RIP broadcasts or announcements consume a lot of bandwidth.

This lead to the development of more efficient and complex Link-State routing protocols.

Because of time delays in transmitting updated routing table information, routing loops
can occur in large networks that use RIP.

Link-State routing protocols use Link-state routing algorithms to route network traffic.
The primary Link-state routing algorithm is the Dijkstra’s algorithm. The major
disadvantage of Link-State protocols is that they require more processing power to
process the routing information (It is compute and memory intensive). They also require
a lot of memory to store Link State Advertisements (LSA) and topological database.
They maintain a complex database on the topology of the network Examples of Link
State protocols are:

• OSPF
• BGP
• EGP
• DEC’s DNA Phase V
• Novell’s Netware Link Service Protocol (NLSP)

OSPF is a link-state routing protocol that uses some algorithm to calculate routes to
destinations. Link state protocols do not just take a route with the least number of hops to
get to the destination. They calculate routes based on line speed, traffic or utilization, and
cost. Link state routing protocols reduce broadcast traffic because the do not send out
periodic full table updates as with distance vector protocols. The only time when link
state protocols exchange full tables or complete copies of their routing tables is upon
initialization. Thereafter route updates contain only the changes not the entire table, and
are only send to routers within the same logical area. If no changes occur, they do not
generate an update. OSPF was developed by the Internet Engineering Task Force (IETF)
as a replacement for RIP. Link-state algorithms are more efficient and create less network
traffic than distance-vector algorithms. OSPF is more sophisticated than RIP and requires
more knowledge and experience to set up and administer.

Data Communications and Networking Copyright © 2005, African Virtual University 78

It is designed for use in medium to large and complex networks. It generally response to
path changes more quickly than RIP.

OSPF can operate within a hierarchy. The largest entity within the hierarchy is the
Autonomous System. An autonomous system can be divided into a number of areas,
or groups of contiguous networks and attached hosts. Routers with multiple
interfaces can participate in multiple areas. These routers, which are called area
border routers, maintain separate topological databases for each area.

A topological database represents an overall picture of networks in relationship to

routers, and contains the collection of link state advertisements received from all
routers in the same area. Because these routers share the same information, they
have identical topological databases.

OSPF employs two different types of routing, depending on whether the source and
destination are in the same or different areas. Intra-area routing occurs when the
source and destination hosts are in the same area. Inter-area routing occurs when
they are in different areas. An OSPF backbone is responsible for distributing routing
information between areas. It consists of all area border routers, networks not wholly
contained in any area, and their attached routers. Fig 12-3 illustrates this.

Fig 12-3 An Autonomous System and the Associated Areas

In the above diagram, routers 4, 5, 6, 10, 11, and 12 make up the backbone. If
Host H1 in Area 3 wants to send a packet to Host H2 in area 2, the packet is
sent to Router 13, which forwards the packet to Router 12, which sends the
packet to Router 11. Router 11 then forwards the packet along the backbone to
area border Router 10, which sends the packet through two intra-area routers
(Router 9 and Router 7) to be forwarded to Host H2.

Data Communications and Networking Copyright © 2005, African Virtual University 79

Because routers perform more complex functions on each packet, they are slower than
bridges in operation. Similar to a bridge, a router provides filtering, but offers more
embedded intelligence, sophisticated network and traffic management than bridges.
Routes have the potential for flow control and comprehensive error protection.

Table 11-3 Information Contained in a Routing Table

Column Description
Network Address (the destination This column represents all the networks
network IP address) that the router knows. These are the
addresses to which a packet arriving at
this router can be send

Net Mask (The destination network It is a 32-bit number associated with a

subnet mask) particular IP Address entry in the routing
table
Gateway Address It is the address of the neighbouring
router that connects to the network to
which packets can be send. The gateway
will then forward the packet to the
appropriate network
Metric (sometimes called the hop count) This indicates the cost or hops associated
with a particular network route (the
distance to the target). The router must
find the path representing the least cost to
get a packet to its destination. The lower
the cost or hop count, the better or more

Data Communications and Networking Copyright © 2005, African Virtual University 80

 efficient a particular route. A metric
value of 1 means that the number of
routers between the host’s network and
the destination network is one.
Time-To-Live (TTL) Is a value specifying how much longer
(in seconds) a particular entry will remain
in the routing table. Each packet of data
to be routed will also have a TTL value
that decrements with each hop. When the
TTL reaches zero, the packet is
destroyed.
Source It indicates which protocol announced the
routing entry to the router. Routers use a
number of protocols to learn IP routing
information. The most common protocol
for the LAN environment is RIP

A default route with address 0.0.0.0 is often called the 'route of last resort'. It is the last
route where a packet of data is send to when a router gets a packet and cannot establish
the network address where the packet should be forwarded. If no default route is defined,
a data packet whose destination network cannot be established is discarded and an ICMP
error message is send to the originator. The Administrator inserts the default route
manually, it is not learned.

If no default route is available, the router drops the packet, and then sends a message to
the sending station indicating that it was unable to forward the packet because the
destination was unreachable.

Brouter

A brouter operates at layer 2 and layer 3 of the OSI reference model. It can function as
either a bridge or a router (it incorporates the functionality of both bridges and routers)
depending on the network transport protocol in use. For instance, a brouter can act like a
router with routable protocols such as TCP/IP and acts as a bridge with non-routable
protocols such as NetBEUI.

Data Communications and Networking Copyright © 2005, African Virtual University 81

Switch

A switch is a high-speed packet routing device that enables multiple physical LAN
segments to be connected to form a large single network. Switches operate at much
higher speeds than bridges and can support new functionality such as virtual LANs.

Switches can be categorized according to the OSI layer at which they operate. According
to this, we have layer 2 and layer 3 switches. The difference between layer 2 and layer 3
switching is the type of information inside the frame that is used to determine the output
interface. Layer 2 switches use MAC addresses to determine the output port.

Layer 3 switches use IP addresses to determine the output interface (in layer 3 switching,
frames are switched based on IP address information/ network layer information). Layer
3 switches also support router functionality

Technically speaking, hubs operate using a broadcast model and switches use a virtual
circuit model.

For example, when four computers are connected to a hub, and two of these computers
are communicating with each other, the hub simply pass through all network traffic to
each of the four computers

Unlike hubs, which broadcast messages to all ports regardless of the destination address,
switches use internal address tables to selectively forward frames to only the port
associated with the receiving node. Each port on a switch will receive only packets that
are meant for it.

Fig 12-4 The difference between a hub and a switch in a situation where station A
Transmits data to station B.

Data Communications and Networking Copyright © 2005, African Virtual University 82

With the hub the data is transmitted to all stations since it is broadcasted to all ports on
the hub. With the switch the data is only send to the port to which station B is attached.

There are two kinds of switches-the workgroup switch and the enterprise switch.

The workgroup switch works as described above. It is the direct replacement of the
hub, allowing computers to be directly connected to it and allocating full bandwidth to
the connected computers.

An enterprise switch connects to the network backbone and has no stations directly
connected to it. Hubs, bridges and routers are directly connected to it.

The diagram below illustrates the concept:

Fig 12-5 Connecting Hubs, Switches and Routers

The bandwidth of the enterprise switch should be greater than the combined bandwidth of
the entire network to which it is connected.

Data Communications and Networking Copyright © 2005, African Virtual University 83

Switch Architecture

At present there are three basic architectures for Data Link Switches:

Cut-through. The frame is forwarded to its destination as soon as the switch has
copied the first 6-bytes following the preamble that make up the destination address
information into the buffer. It then looks up the destination address in its switching
table and forward the frame to the appropriate port. Cut-through provides reduced
latency because it begins to forward the frame after reading the destination address
of the frame. This has the advantage of being very fast, but there is no error
checking - too many frame re-transmissions may slow the network down.

Store-and-forwrad. The entire frame is copied into a buffer and Cyclic Redundancy
Check (FCS) is performed to check for errors before the frame is transmitted to its
destination. If the frame has errors, it is discarded and retransmission is carried out. This
has the advantage of ensuring that bad frames are not transmitted, but is much slower
than the cut-through architecture. If the frame does not contain any errors, the switch
looks up the destination address in its switching table and forwards the frame to the
appropriate port.
Fragment-free. It combines the best of both cut-through and store-and-forward by acting
as a cut-through switch while monitoring the traffic for errors. The switch uses cut-
through until a certain error threshold is reached. It will change over to store and forward
switching when the defined error threshold is reached. The error threshold is normally
configurable. The main disadvantages are that of cost - this type of switch is more
expensive - and the reliance on management setting the correct error threshold.

Access to Full Bandwidth

A switch can allow computers to access full bandwidth on a network: A key problem
associated with LANs is the fact that the bandwidth of the connecting cable is shared
among all workstations connected to the LAN. For instance, in a 10-station 100BaseTX
network, each station (on average) obtains 10Mbps. This means that the available
bandwidth is divided among the stations.

If computers are directly connected to a switch (micro-segmentation) each station

attached will access the full bandwidth. For example, on a 10BaseT network, each
workstation will access a data transmission rate of 10Mbps. Each station receives instant
access to the full bandwidth and does not have to contend for the available bandwidth
with other stations.

Servers can be connected directly to a switch, allowing network users to utilize the
network’s bandwidth when accessing server resources.

Data Communications and Networking Copyright © 2005, African Virtual University 84

Gateways

Gateways are more complex than bridges and routers. They are used to connect two
heterogeneous networks (connecting networks using different protocols, data formats
language and architecture). An example is connecting an Apple network to an IBM SNA
network.

A gateway is called a protocol converter because it changes the format of data from one
network to the format used on another network. It stripes old protocol stacks that were
used to format the data on one network and repackages the data with the protocol stack of
the destination network. Gateways can operate at any of the OSI layers.

Data Communications and Networking Copyright © 2005, African Virtual University 85

13

Ethernet Frame Format

Often you find the words frame and packet used synonymously, although frames are only
defined at the data link layer of the OSI reference model and packets at the network layer
and above.

The frame structure varies according to the network technology. For example several
flavours of Ethernet, Token ring and FDDI frames exist. The following figure illustrates
the format of an Ethernet frame as defined in the original IEEE 802.3 standard

Preamble Start Destination Source Data Data (- Pad (0- Frame

(7-bytes) byte (1 address (2 address field x bytes y bytes) check
byte) or6 bytes) (2 or 6 length sequence
bytes) (2- (FCS)
bytes) (4-bytes)

Fig 13-1 Ethernet Frame

Preamble

It is a seven-byte pattern consisting of alternating 0s and 1s that allows the transmitting

and the receiving stations to synchronise their communication.

Start Byte

It is a special pattern 10101011 that indicates the start of a frame. It also acts as a signal
to other nodes on the network so that they are made aware that another node has started
transmitting

Destination Address

It identifies the physical address (MAC Address) of the station or stations that are to
receive the frame. If the first bit is a 0, the field specifies a particular station (a single
station) as the destination. If the first bit is a 1, the destination address is a group address
(multicast address) and the frame is send to all stations in the predefined group specified
by the address. If all bits are 1s, the frame is broadcast to all stations on the network.

Data Communications and Networking Copyright © 2005, African Virtual University 86

Source Address

It identifies the station that send the frame and therefore who to respond to or ask for
more information.

Data Length

It specifies the number of bytes in the combined data and pad fields.

Data field

It contains the actual data being transmitted (the actual data being transmitted is also
referred to as the payload)

The Pad field

The Pad field contains some extra bytes that are added (padded) to the actual data bytes
if the data bytes are less than 46-bytes. This is because the data field must be at least 46
bytes as specified by the IEEE 802.3 standard, and a maximum of 1500 bytes. If it is less
than 46 bytes, then the pad field will make up the difference.

Frame Check Sequence (FCS)

It checks for errors in the transmitted data using a 32-bit Cyclic Redundancy Check
(CRC). This ensures that all data received at the destination has no errors. The receiving
node calculates the CRC value for the incoming data and compares it with the transmitted
value. Any discrepancy between the two values means that data has been corrupted
during transmission, and the receiving station would request for a resend.

Data Communications and Networking Copyright © 2005, African Virtual University 87

14

Switching Methods

Switching is the process by which packets are received, stored and transmitted to their
destination.

There are three methods that are used for switching:

Packet Switching

Packet switching involves dividing data into small manageable units called packets. Each
packet has a destination address, which makes it possible to send individual packets
separately across the network. Packets belonging to the same data may each take different
routes to their destination. When a packet reaches a node, the logic at that node
determines the next node where the packet should go in a route leading to the destination.
Because of the intelligence built at each node (routing strategy or routing algorithm),
dynamic routing of data is possible. The routing strategy looks at factors such as the cost
of a connection from one node to the other and the best route at that time. This ensures
that no routes are congested. Although each packet may travel along a different path, the
packets are ordered and sequenced by protocols when they get to their destination. The
packet size is kept very small so that if there is an error in transmission, a small packet
retransmission will be easier than the retransmission of a large packet. Also, small
packets are held at the node for very short periods of time.

The major disadvantage of packet switching is that each packet requires a copy of control
information (it requires a header field for storing source and destination addresses, etc)

In packet switched networks, packets can be addresses to stations in three ways:

i) Unicast

In unicast packets of data are addresses to a single station. For example, if station A
communicates with station B, the packets exchanged between these two stations are
unicast.

ii) Multicast

In multicast, packets are addressed to a predefined group of stations on a network, but not
necessarily all of the stations on the network.

Data Communications and Networking Copyright © 2005, African Virtual University 88

iii) Broadcast

In broadcast, packets are addressed to all stations on the network. Broadcast offers a way
to communicate the same data to every station on a network at once. All stations receive
the broadcast packets regardless of whether a particular station is interested in the packet.

Circuit Switching

Transmission of data through a circuit switched network takes the following phases of
operation:

a) Connection establishment phase

Before any data can be transmitted from one station to the other, an end-to-end (station-
to-station virtual circuit) communication path has to be established.

b) Data transfer phase

After the establishment of a communication path between the two stations, data is
transferred. The virtual circuit is maintained for the period of data transfer. Generally the
full duplex mode of communication is used.

c) Circuit disconnection phase

The link or communication path between the two stations can be disconnected only after
the two stations have finished communicating (circuit disconnection)

The most common example of a circuit switched network is the telephone network,
which is primarily used for voice communication. When a call is made, the switching
equipment in the Telephone Exchange establishes a circuit or logical connection between
the caller and the answerer. The circuit is maintained for the duration of the call. The
major advantage of circuit switching is that data from one station to the other does not
contain the additional formatting bits and flow control information. As a result data
immediately appears in the same form at the destination.

Circuit switching has two major disadvantages:

• Both stations must be available at the same time for the data transfer to take place.
• During data transfer the communication path setup between the two stations is
entirely dedicated to those two stations. It cannot be used for communication by
other stations on the network. It will only be available to other stations when the
two stations have finished communicating and have broken down the
communication path (circuit disconnection)

Data Communications and Networking Copyright © 2005, African Virtual University 89

Message Switching

In message switching no connection has to be established between the two stations.

Rather, when a station wishes to send a message, it adds a destination address to the
whole message block. The message is then passed through the network from one node to
the other until it reaches its destination. The message is temporarily stored at each node.
Message switching uses the store and forward method, where the message is stored at a
node and then forwarded to another node until it reaches its destination.

Advantages of message switching over circuit switching:

• Simultaneous availability of the sender and receiver is not required, since the
network can store the message pending the availability of the receiver.
• Higher priority messages experiences less delay than the lower priority messages

Disadvantages of message switching:

• It needs sufficiently large data buffers to hold the incoming message

• It has a large delay for the messages to reach their destination
• A failure in a connection may mean the loss of the entire message. Consequently
message switched networks are no longer common, they have given way to the
more efficient packet switched networks.

Data Communications and Networking Copyright © 2005, African Virtual University 90

15

Internet Protocol (IP) Addresses

Each TCP/IP host is identified by a logical IP address. The IP address is a network layer
address and has no dependence on the data link layer address (such as a MAC address of
a network interface card). A unique IP address is required for each host and network
component that communicates using TCP/IP.
The IP address identifies a system’s location on the network in the same way a street
address identifies a house on a city block. Just as a street address must identify a unique
residence, an IP address must be globally unique and have a uniform format.
Each IP address includes a network ID and a host ID.
• The network ID (also known as a network address) identifies the computers that are
located on the same physical network bounded by IP routers. All computers on the
same physical network must have the same network ID. The network ID must be
unique to the internetwork.
• The host ID (also known as a host address) identifies a workstation, server, router, or
other TCP/IP host within a network. The address for each host must be unique to the
network ID.

An IP address is 32 bits long. Rather than working with 32 bits at a time, it is a common
practice to segment the 32 bits of the IP address into four 8-bit fields called octets. Each
octet is converted to a decimal number (the Base 10 numbering system) in the range 0-
255 and separated by a period (a dot). This format is called the dotted decimal notation.
Table 15-1 shows an example of an IP address in binary and dotted decimal formats.
Table 15-1 Binary and Dotted Decimal Forms of an IP Address
Binary Format Dotted Decimal Notation
11000000.10101000.00000011.00011000 192.168.3.24

Data Communications and Networking Copyright © 2005, African Virtual University 91

Subnet Mask

A subnet mask also called a net mask is a 32-bit binary number (similar to an IP Address)
used to distinguish the network and the host portions of an IP address. It can also be
expressed in the dotted decimal notation as shown in table 15-2

Table 15-2 Binary and Dotted Decimal Forms of a Subnet Mask

Binary Format Dotted Decimal Notation
11111111.11111111.11111111.00000000 255.255.255.0

IP Address Classes

IP Addresses are divided into five classes: A, B, C, D and E. The following table shows
the characteristics of each IP address class.

Table 15-3 Characteristics of IP Address Classes

IP Address class Valid Address Default Subnet Description

Range Mask
A 1.0.0.0 to 126.0.0.0 255.0.0.0 or /8 The first octet is of the form
0xxxxxxxx, and ranges from
1 to 126 in decimal. Can
allow a total of 16,777,214
hosts.
B 128.0.0.0 to 255.255.0.0 or /16 The first octet is of the form
191.254.0.0 10xxxxxx, and ranges from
128 to 191 in decimal. Can
allow a total of 65,534 hosts.
C 192.0.1.0 to 255.255.255.0 or /24 The first octet is of the form
223.255.254.0 110xxxxx, and ranges from
192 to 223 in decimal. Can
allow a total of 254 hosts.
D 224.0.0.0 to N/A Multicasting network. The
239.255.255.255 first octet is of the form
1110xxxx, and ranges from
224 to 239 in decimal.
E 240.0.0.0 to N/A It is for experimental
247.255.255.255 purposes. The first octet is of
the form 1111xxxx, and
ranges from 240 to 247

Data Communications and Networking Copyright © 2005, African Virtual University 92

IP Address Rules

IP Addressing follows certain guidelines in order to function properly. Although we have

learnt about the ranges for the IP Address classes, not all addresses within these ranges
can be used. This section describes the exceptions

Loopback Addresses

Any address staring with 127 is a loop back address and should never be assigned to a
machine. It is used for testing TCP/IP configuration. The loopback address is typically
127.0.0.1

Broadcast Addresses

Limited Broadcast: 255.255.255.255 is the broadcast address. It is used for configuring

hosts when they boot up. Computers can broadcast this address to obtain IP Addresses
(for example from a DHCP server) messages to all computers on the network. It cannot
be used as a network address.

Net-Direct Broadcast: netid.255.255.255. It is used to broadcast to all hosts in a

network. For example if the network portion of your IP Address is 192.32.200 and the
host portion is 12, your computer can broadcast messages to all network hosts by using
the destination address 192.32.200.255

Reserved IP Addresses

Certain addresses have been set aside and reserved for private use only. Routers will not
route data from these addresses. They are for internal private use only. If you would like
to use TCP/IP on your internal network (intranet) and not use the Internet, you can use
the private addresses. The private addresses are as follows:

Class A: 10.0.0.0 through 10.255.255.255

Class B: 172.16.0.0 through 172.31.255.255

Class C: 192.168.0.0 through 192.168.255.255

Data Communications and Networking Copyright © 2005, African Virtual University 93

Finding the Network ID

The Network ID portion of an IP Address can be determined by performing a logical

Bitwise-AND operation between the subnet mask and the IP address. Both the IP address
and the Subnet mask are converted to binary and then the AND operation is performed.

Consider the following IP Address and subnet mask:

IP Address 192.20.16.5

Subnet mask 255.255.255.0

Converting to binary and performing the AND operation

IP 11000000 . 00010100 . 00010000 . 00000101

Subnet mask 11111111 . 11111111 . 11111111 . 00000000

ANDing Result 11000000 . 00010100 . 00010000 . 00000000

Converting the result of the ANDing back to decimal gives us 192.20.16.0

The Number 192.20.16.0 is the Network ID. A common convention in TCP/IP is to

discard the trailing zero octets in the network ID. This way the Network ID simply
becomes 192.20.16. Since an IP address is made up of the Network ID plus The Host ID,
the Host ID is simply the part that is missing from our Network ID for us to have a
complete IP Address. In this case, the Host ID is 5.

It is easy to determine the Network ID and the Host ID for a given IP address when using
the default subnet masks. All you have to remember is that the octets in the subnet mask
which have all their bits set to 1 points to the portion of the IP Address that represent the
Network ID. The zero octets (octets with all their bits set to zero, in the subnet mask),
point to the portion of the IP Address representing the Host ID.

Consider the following two IP Addresses:

192.20.1.5 and 192.20.6.8

These are class C IP Addresses, but one cannot tell whether the two addresses are on the
same subnet or not without a subnet mask. If the two IP Addresses are using a subnet
mask of 255.255.255.0, the two IP Addresses are located on different subnets (192.20.1
and 192.20.6 respectively).

Data Communications and Networking Copyright © 2005, African Virtual University 94

Therefore hosts using these IP Addresses and the common subnet mask of 255.255.255.0
should communicate with each other through a router. If the two IP Addresses are using a
subnet mask of 255.255.0.0, the two IP Addresses are said to be on the same subnet
(192.20)

Determining the Maximum Number of Hosts per Subnet

The Maximum number of Hosts per subnet is given by the formula 2n – 2

Where n = number of bits representing the host

Two host addresses are subtracted because there are two invalid addresses:

• The host bits cannot be all zeros, because when the host bits are all zeros, the IP
Address will refer to the network and not any of the hosts.
• The host bits cannot be all ones. When the host bits are all ones, the IP address
becomes a broadcast address. Hence, such an IP Address will not identify a
particular host, but send packets to all computers on the network.

Consider the following IP Address and subnet mask:

IP 15.6.100.1

Subnet mask 255.255.0.0

The Host ID is 100.1. It has two octets, which means there are 16-bits representing the
host since each octet will have 8-bits. Therefore the maximum number of hosts is:

216 – 2

n = 16

= 65536 – 2

= 65534 hosts

Data Communications and Networking Copyright © 2005, African Virtual University 95

Subnetting

It is the act of subdividing a single IP Network into multiple smaller networks

In order to create some subnets, we need to borrow bits from the host portion of an IP
Address.

Consider the following IP Address and subnet mask:

Example 1

IP Address 129.20.5.6

Subnet mask 255.255.0.0

There are 16-bits representing the host. In order to create at least 5 subnets, we should
first determine the number of bits to be borrowed from the host portion of the IP Address.
The number of subnetting bits to be borrowed is determined by the formula:

2n - 2 ≥ Number of subnets to be created

Where n is the number of subnetting bits.

To create at least 5 subnets, the value of n should be 3.This means that we need 3
subnetting bits. Therefore the subnet mask changes

From: 255.255.0.0 ( 11111111.11111111.00000000.00000000)

To: 255.255.224.0 ( 11111111.11111111.11100000.00000000)

Data Communications and Networking Copyright © 2005, African Virtual University 96

To determine the subnets that are created, you take the decimal value of the lowest order
subnetting bits. In this case it is 25 = 32. This means the subnets to be created will differ
by 32 as shown in table 15-4.

Table 15-4 Determining Network Subnets

Subnet Subnet Address in dotted Subnet address in binary

decimal

1 129.20.32 10000001.00010100.00100000

2 129.20.64 10000001.00010100.01000000

3 129.20.96 10000001.00010100.01100000

4 129.20.128 10000001.00010100.10000000

5 129.20.160 10000001.00010100.10100000

6 129.20.192 10000001.00010100.11000000

To determine the range for each subnet, we keep those octets, which are not changing as
we move from one subnet to the other (i.e. 129.20). We then consider the remaining
octets, starting from the octet that contains the subnetting bits to any other octet that
follows (if any), as we move from Right to Left.

The address range for the first subnet is from 129.20.a.b to 129.20.c.d

a.b and c.d are of the form: 001xxxxx.xxxxxxxx (x means the value can be a 1 or a 0)

a.b, minimum value is given by: 00100000.00000001 (32.1 in dotted decimal)

c.d, maximum value is given by: 00111111.11111110 (63.254 in dotted decimal)

Therefore the subnet 129.20.32 will have IP Address range from 129.20.32.1 through
129.20.63.254

Data Communications and Networking Copyright © 2005, African Virtual University 97

The address 129.20.63.255 will be the broadcast address for this subnet (address with all
host bits set to 1).

Similarly, for the second subnet (129.20.64) we keep those octets, which are not
changing (129.20).

The address range for the second subnet is from 129.20.a.b to 129.20.c.d

a.b and c.d are of the form: 010xxxxx.xxxxxxxx (x means the value can be a 1 or a 0)

a.b, minimum value is given by: 01000000.00000001 (64.1 in dotted decimal)

c.d, maximum value is given by: 01011111.11111110 (95.254 in dotted decimal)

NB. Never change the value of any of the three subnetting bits in calculating the
max. and min. values

Therefore the subnet 129.20.64 will have IP Address range from 129.20.64.1 through
129.20.95.254

The address 129.20.95.255 will be the broadcast address for this subnet (address with all
host bits set to 1).

For the third octet:

a.b and c.d are of the form: 011xxxxx.xxxxxxxx (x means the value can be a 1 or a 0)

a.b, minimum value is given by: 01100000.00000001 (96.1 in dotted decimal)

c.d, maximum value is given by: 01111111.11111110 (127.254 in dotted decimal)

Again the values of the subnetting bits are not changed in determining max. and
min. values

Therefore the subnet 129.20.96 will have IP Address range from 129.20.96.1 through
129.20.127.254

The address 129.20.127.255 will be the broadcast address for this subnet (address with all
host bits set to 1).

Data Communications and Networking Copyright © 2005, African Virtual University 98

Continuing in this way, we determine all the ranges as shown in table 15-5 that follows.

Table 15-5 Address Range for the Subnets

Subnet Subnet Address Range Broadcast Address

Address

1 129.20.32 129.20.32.1 through 129.20.63.254 129.20.63.255

2 129.20.64 129.20.64.1 through 129.20.95.254 129.20.95.255

3 129.20.96 129.20.96.1 through 129.20.127.254 129.20.127.255

4 129.20.128 129.20.128.1 through 129.20.159.255

129.20.159.254

5 129.20.160 129.20.160.1 through 129.20.191.255

129.20.191.254

6 129.20.192 129.20.192.1 through 129.20.223.255

129.20.223.254

Note that each subnet does not use all the available IP addresses for its range, because an
IP address that contains all binary zeros for the host portion would represent the network
address, and an IP address that contains all binary ones for the host portion is interpreted
as a subnet-direct broadcast.

Example 2

Consider the network 210.199.10.0. A class C network (class C default subnet mask is
255.255.255.0). In order to create six subnets we proceed as follows:

No of required subnetting bits is given by:

2n – 2 ≥ 6

The value of n is 3. This means that we require 3 subnetting bits.

Data Communications and Networking Copyright © 2005, African Virtual University 99

Therefore the subnet mask changes

From: 255.255.255.0 ( 11111111.11111111.11111111.00000000)

To: 255.255.255.224 ( 11111111.11111111.11111111.11100000)

To determine the subnets that are created, you take the decimal value of the lowest order
subnetting bits. In this case it is 25 = 32. This means that subnets to be created will differ
by 32 as follows:

Table 15-6 Determining Network Subnets

Subnet Subnet Address in dotted Subnet address in binary

decimal

1 210.199.10.32 11010100.11000111.00001010.00100000

2 210.199.10.64 11010100.11000111.00001010.01000000

3 210.199.10.96 11010100.11000111.00001010.01100000

4 210.199.10.128 11010100.11000111.00001010.10000000

5 210.199.10.160 11010100.11000111.00001010.10100000

6 210.199.10.192 11010100.11000111.00001010.11000000

Data Communications and Networking Copyright © 2005, African Virtual University 100
To determine the range for each subnet, we keep those octets, which are not changing as
we move from one subnet to the other (i.e. 210.199.10). We then consider the remaining
octets, starting from the octet that contains the subnetting bits to any other octet that
follows (if any), as we move from Right to Left.

The address range for the first subnet is from 210.199.10.a to 210.199.10.b

a and b are of the form: 001xxxxx (x means the value can be a 1 or a 0)

a minimum value is given by: 00100001 (33 in decimal)

b maximum value is given by: 00111110 (62 in decimal)

Therefore the subnet 210.199.10.32 will have IP Address range from 210.199.10.33
through 210.199.10.62

The address 210.199.10.63 will be the broadcast address for this subnet (address with all
host bits set to 1).

Similarly, for the second subnet (210.199.10.64) we keep those octets, which are not
changing (210.199.10). The subnet ranges from 210.199.10.a to 210.199.10.b

a and b are of the form: 010xxxxx (x means the value can be a 1 or a 0)

a minimum value is given by: 01000001 (65 in decimal)

b maximum value is given by: 01011110 (94 in decimal)

NB. Never change the value of any of the three subnetting bits in calculating the
max. and min. values

Therefore the subnet 210.199.10.64 will have IP Address range from 210.199.10.65
through 210.199.10.94

The address 210.199.10.95 will be the broadcast address for this subnet (address with all
host bits set to 1).

Data Communications and Networking Copyright © 2005, African Virtual University 101
For the third octet:

a and b are of the form: 011xxxxx (x means the value can be a 1 or a 0)

a minimum value is given by: 01100001 (97 in decimal)

b maximum value is given by: 01111110 (126 in decimal)

Again the values of the subnetting bits are not changed in determining max. and
min. values

Therefore the subnet 210.199.10.96 will have IP Address range from 210.199.10.97
through 210.199.10.126

The address 210.199.10.127 will be the broadcast address for this subnet (address with all
host bits set to 1).

Putting it all together we come up with the following table:

Table 15-7

Subnet Subnet Address Address Range Broadcast

Address

1 210.199.10.32 210.199.10.33 210.199.10.63

through
210.199.10.62
2 210.199.10.64 210.199.10.65 210.199.10.95
through
210.199.10.94
3 210.199.10.96 210.199.10.97 210.199.10.127
through
210.199.10.126
4 210.199.10.128 210.199.10.129 210.199.10.159
through
210.199.10.158
5 210.199.10.160 210.199.10.161 210.199.10.191
through
210.199.10.190
6 210.199.10.192 210.199.10.193 210.199.10.223
through
210.199.10.222

Data Communications and Networking Copyright © 2005, African Virtual University 102
Note that each subnet does not use all the available IP addresses for its range, because an
IP address that contains all binary zeros for the host portion would represent the network
address, and an IP address that contains all binary ones for the host portion is interpreted
as a subnet-direct broadcast.

The subnet mask used in each of the subnets is 255.255.255.224

Example 3

Consider the subnet mask 255.128.0.0 this can be written in binary as:

11111111.10000000.00000000.00000000

It cannot be used for subnetting because there is only one bit available for subnetting. We
need at least two bits for subnetting.

Example 4

Consider the following class C network:

197.129.59.0

i) How many subnet bits are required to create 62 subnets?

ii) Calculate the number of hosts per subnet.

Solution

i) 2n – 2 >= 62

2n >= 64

Therefore n = 6

Thus 6 subnetting bits are required to create 62 subnets.

Data Communications and Networking Copyright © 2005, African Virtual University 103
To calculate the number of hosts per subnet, we need to find the new subnet mask that
includes the subnetting bits (custom subnet mask). The default class C subnet mask is
255.255.255.0

In binary is: 11111111.11111111.11111111.00000000

If we now add the 6 subnetting bits the new subnet mask in binary is as follows:

New net mask: 11111111.11111111.11111111.11111100

This custom subnet mask clearly shows that there are two bits remaining for the host
(the bits which are zero)

Number of hosts per subnet is given by 2n –2. The value of n = 2, which gives us 2 hosts
per subnet.

In big networks, subnetting is generally recommended. Big networks with computers that
are geographically disperse or even Local Area networks, are usually divided into subnets
(smaller networks) that can be connected together by routers.

There are many advantages to be realized from subnetting:

• It permits a mix of network technologies e.g. the Ethernet (Bus) and Ring can be
connected.
• It reduces network congestion because broadcasts and local network traffic are
limited to the local subnets.
• It simplifies management because it is easier to identify and isolate network
problems in a group of smaller networks connected together than within one big
network.
• It allows many computers to be networked than can be put on one network. For
example, by subnetting, more computers can be networked than those that can be
put on a single Bus network

Data Communications and Networking Copyright © 2005, African Virtual University 104
16
DHCP server

DHCP stands for Dynamic Host Configuration Protocol. A DHCP server is a computer
configured with the DHCP protocol so as to automatically assign TCP/IP addresses to
client computers along with the correct subnet mask, default gateway, and DNS server.

Each TCP/IP host on a network must be allocated a valid and unique IP Address, either
automatically (using a DHCP sever) or manually (using static addresses). When
configuring a DHCP server, remember to give the server a static IP address. There are
three ways in which IP addresses are grouped for use and allocated by a DHCP server.
These are as follows:

Scope. A scope is a range of IP Addresses that a DHCP server can assign to the clients
that are on one subnet

Superscope. It is a range of IP Addresses that span several subnets. The DHCP server
can assign these addresses to clients that are on several subnets.

Multicast scope. A range of class D addresses from 224.0.0.0 to 239.255.255.255 that

can be assigned to computers when they ask for them. A multicast group is assigned one
IP Address. Multicasting can be used to send messages to a group of computers at the
same time with only one copy of the message. The Multicast Dynamic Client Allocation
Protocol (MADCAP) is used to request a multicast address from a DHCP server.

When setting up a DHCP server the lease duration should also be defined. The lease
duration is the period of time that a computer is given an IP address by the DHCP server.
When using a DHCP server, IP Addresses are leased instead of permanently assigned to
computers. As a result, in a DHCP environment, an IP Address is not a permanent
property of a computer.

Advantages of using a DHCP server

• A DHCP server does not allocate an IP Address to two or more hosts. If an IP

address is allocated to two or more hosts, IP conflicts occur on the network.
• A DHCP server makes it easy to configure IP Address information. Instead of the
Administrator moving from one machine to the other configuring IP information,
he only needs to specify the IP information on the DHCP server.
• DHCP capability is in-built in most server operating systems, so adding the
DHCP service does not cost anything extra.
• When IP information is configured centrally on a DHCP server, any problems
with IP information can be rectified only on the DHCP server, instead of moving
from one machine to the other rectifying problems to do with IP information.

Data Communications and Networking Copyright © 2005, African Virtual University 105
A DHCP client is a computer that obtains IP information automatically from a DHCP
server.

Using multiple DHCP servers

It is possible to have more than one DHCP server on a subnet. If there is more than one
DHCP server on a subnet, you can not control which DHCP server gives a client an IP
address. Any DHCP server that receives a client’s DHCP request broadcast can send a
DHCP offer to that client. It is up to the client to take-up an offer from one DHCP server
and ignores other offers.

When using more than one DHCP server, you must ensure that no IP Addresses are
duplicated on the DHCP servers. If Addresses are duplicated on the DHCP servers, an IP
Address can potentially be leased to two DHCP clients at the same, thereby creating
some conflicts.

For instance, if you have two DHCP servers on your subnet, you could split the IP
Address range as follows:

Table 16-1 Splitting an IP Address Range between DHCP Servers

Subnet DHCP Sever A DHCP Server B

1 120.50.7.10 – 120.50.7.84 120.50.7.85 – 120.50.7.110

2 120.50.8.10 – 120.50.8.34 120.50.8.35 – 120.50.8.110

Data Communications and Networking Copyright © 2005, African Virtual University 106
How a Client Computer Obtains an IP Address from A DHCP Server

There are four stages that a client computer goes through in order to obtain an IP Address
fro a DHCP server:

Stage 1

When a DHCP client is switched on, it broadcast a request for an IP Address lease from
the available DHCP server(s). The broadcast packet is referred to as a DHCPDISCOVER
packet.

The discover packet contains the hardware address or MAC address and name of the
client computer requiring IP address information.

Stage 2

All DHCP server(s) that receive the IP lease request respond to the DHCP client request
with an IP offer by sending a DHCPOFFER packet (each DHCP server sends its own
DHCPOFFER)

A DHCPOFFER packet is a proposal from the DHCP server to the DHCP client. It
contains an IP Address that can be used by the client, a subnet mask, a lease period (in
hours) and the IP Address of the DHCP server offering the proposal.

Stage 3

The DHCP client selects an offer from one of the DHCP servers and broadcast a
DHCPREQUEST packet containing the IP Address of the of the DHCP server it selected
to obtain IP Information from. The DHCPREQUEST is broadcasted so that the DHCP
servers whose offers were not accepted withdraw or pull-back their IP address offers.

Once the chosen DHCP server receives the DHCPREQUEST packet from the client, it
updates its DHCP database and marks the address it has given to the client as leased.

Stage 4

The DHCP server selected by the client responds with an acknowledgement packet
known as a DHCPACK.

Data Communications and Networking Copyright © 2005, African Virtual University 107
DHCP Lease Renewal

The DHCP client automatically attempts to renew its IP address lease when 50% of the
lease period has expired. To do that, the DHCP client sends a DHCPREQUEST packet
directly to the DHCP server from which it obtained the lease. A DHCPNACK (DHCP
Negative Acknowledgement) is send to the DHCP client if the lease can be renewed. If
the lease cannot be renewed by the DHCP server, the DHCP client should start the lease
process again with a DHCPDISCOVER packet to obtain another IP address when the
lease expires.

If the lease renewal was successful, the DHCP server responds by sending a DHCPACK
directly to the DHCP client that requires the lease renewal.

A DHCP client can be forced to renew its lease before 50% of the lease period has
expired by the ipconfig/renew command in windows.

Data Communications and Networking Copyright © 2005, African Virtual University 108
17

Network Management

Network management is the execution of a set of functions that are required to provide
high-quality and reliable communication on a network.

There are five major areas or components of network management:

• Account Management
• Performance management
• Configuration management
• Security Management
• Fault management

Account Management

Account Management involves creating and maintaining user accounts and providing
appropriate access to resources.

The Network Administrator is responsible for creating user accounts. A user account is
composed of a username, password and appropriate permissions to access resources. It is
a good idea to implement passwords with a minimum length. The system should be
configured to maintain a history of passwords already used, and force users to
periodically create new, unique passwords. Furthermore the administrator should
configure the system to lock accounts if an incorrect password is entered several times.
Users should be advised not to use obvious passwords such as the date of birth, and social
security numbers. The Administrator can put user accounts into groups. Groups are used
to logically organize users with similar resource requirements. This makes it easy for the
administrator to treat a large number of accounts as one account. For instance, if there are
100 users in a group, the administrator can simply send a message to the group and all the
members of the group would get the message. Members of a group automatically inherit
the permissions and Rights set for the group. The administrator can also disable and
delete user accounts if necessary. The administrator can also decide on the type of profile
the users would have after logging on the network. The administrator can allow users to
have a uniform profile or can allow individuals to create their own profiles.

Data Communications and Networking Copyright © 2005, African Virtual University 109
Performance Management

Performance management is the systematic process of monitoring and analyzing the

activities of the network to ensure that the network operates optimally and provides value
to the users. The process of performance management involves the following steps:

i) Gathering performance related data

ii) Analysing data to determine normal levels (baseline). A baseline is a range
of measurements that represents acceptable performance under typical
operating conditions, which is used as the basis of comparison for all
subsequent performance statistics of the network.
iii) Setting appropriate performance thresholds so that exceeding these thresholds
indicates a network problem worth of attention.

By carefully monitoring and analyzing the activities of the network, bottlenecks could
be identified and eliminated. A bottleneck is a system device that is slowing down
network performance. Some of the devices that can become bottlenecks are as
follows:

• The processor
• Network interface cards
• Physical memory
• Disk controllers
• Data transmission media

During the lifetime performance of a network, certain devices may degrade in

performance. It is obligatory for the system administrator to identify and eliminate
bottlenecks. In other words you need to pinpoint the cause of the bottleneck in order to
correct it. Consider a system that has a Pentium 166 MHz processor with 64 MB of
RAM. If the applications that are running are memory intensive, then the lack of memory
becomes a bottleneck. Upgrading the memory will eliminate the bottleneck.

By carefully monitoring and analyzing the activities of the network, it is possible to be

able to determine some trends and be able to predict what is most likely to happen to the
network in the near future. Such information can be used to proactively manage the
network and avoid disastrous consequences. For example, if the performance logs are
showing that the average CPU utilization on your server is increasing by 15% each
month, you can assume that within the next six months you are going to experience some
problems with the server. Before performance becomes so slow that your system is no
longer responding, you can upgrade the server CPU.

Data Communications and Networking Copyright © 2005, African Virtual University 110
Security Management

It is the process of controlling access to network resources according to user rights and
privileges. This may include setting up access lists in routers (creating firewalls to keep
intruders out). It is also essential to implement security within the organization to make
sure right people have access to resources. Without these security measures in place, you
might find someone destroying your valuable data, or selling your company secretes to
your competitors or someone invading the privacy of others. Primarily a security plan
must identify which user in the organization can see which data and perform which
activities on the network.

Many operating systems include some utilities that are used to ensure network security.
For example, Windows 2000 server includes tools or utilities that support the ability to
define a corporate-wide default security configuration for user accounts and audit
policies. Audit policies are policies that track the activities of users and record selected
types of events in the security log of the server or workstation.

Infrastructure components (hubs, switches and routers) should be kept in secure places
where only administrators and technical staff have access.

Fault Management

It is the duty of the system administrator to protect company data from being lost due to
site disasters or power outages. Most fault management systems are able to detect, log,
notify and automatically fix some of the problems so as to keep the network running.
Some of the failure recovery strategies that can be put in place are as follows:

• The use of Uninterruptible power supply (UPS)

A UPS is a device that keeps computers running (servers and clients) after a power
failure, providing power from batteries for a short period of time. The UPS can keep a
server running in the event of a power failure long enough for users to log off and the
administrator to perform a clean shutdown of the system. However, a UPS cannot be used
to guard against server disk failures.

• Backup and off-site storage

A backup system should be in place. For example, a tape backup system could be used so
that when the hard disk on the server fails or develops some problems, data can safely be
retrieved from the backup tape.

Backup should be planned so that it can be carried out on a daily basis, weekly basis, or
monthly basis. Never consider off-site storage as an option, it should be a mandatory part
of the backup process. Disaster strike when you least expect it. An unexpected fire could
wipe out your company data. Floods, earthquakes and other disasters can strike at any
moment.

Data Communications and Networking Copyright © 2005, African Virtual University 111
It is unlikely that a disaster could simultaneously strike two places at the same time.
Placing your data in more than one location helps to reduce the probability of data loss by
disasters.

To avoid data loss due to disk failure in servers, some fault-tolerant disk configurations
can be used. These are the RAID-1 and RAID-5 configurations. RAID stands for
Redundant Array of Inexpensive Disks.

RAID-1 (Mirroring)

Mirroring is the use of two separate disks in a single computer such as a server. One
physical disk is known as the primary drive and the other is the secondary drive. Data
written to the primary drive is mirrored (copied) to the secondary. Mirrored disks provide
fault-tolerance in that if one drive in the mirror set fails, the other continues to work
without an interruption in the service or loss of data.

Disk Duplexing

Disk Duplexing is a form of Disk mirroring where each of the mirrored drives uses a
separate controller. The use of separate controllers instead of one protects against
controller failure. Thus disk duplexing protects against controller failure as well as disk
failure. The main disadvantage of mirrored disks is high overhead. The same data is
written to two hard drives. For example to implement a mirror of a 4 GB drive, you need
two 4GB disks (a total of 8GB of storage space)

RAID-5 is a method of combining space on separate disks so as to have one large storage
space as well as providing data recovery if one of the disks in the set fails. RAID-5
requires a minimum of 3 disks in the striped set. Data and parity information are
distributed across all disks in the array. If a single drive in the striped set fails, the parity
information stored on the other drives can be used to regenerate or rebuild the data on a
new disk that replaces the failed disk.

Data Communications and Networking Copyright © 2005, African Virtual University 112
Configuration Management

It is the process of gathering and documenting network and system configuration

information for hardware and software, as well as making note of any changes or
modifications during the life of a network. Configuration information should contain the
following:

• A drawing of the entire network and location of all hardware/equipment and

details of cabling.
• Software information such as licensing and support details.
• Details of the different types of servers and the role played by each server e.g.
Primary Domain Controller (PDC), Backup Domain Controller (BDC), DHCP,
DNS, WINS, etc.
• A record of all problems experienced on the network, their symptoms, solutions
and dates of occurrence.

Documentation should be thorough and well organized. The information gathered is

usually stored in a database for easy access and retrieval.

Data Communications and Networking Copyright © 2005, African Virtual University 113
 Network Management Tools

Network management employs a variety of tools to assist in the management of the

network. Most tools can be classified as primarily hardware or software, although in
practice any software needs to be supported by hardware.

Hardware

Digital Voltmeter (DVM)

It is the most basic, all-purpose electronic measuring tool. In skillful hands, it can reveal
far more than just the amount of voltage or current passing through a wire. It can be used
to check for continuity in network cables and to check the resistance of terminators.
Continuity checks can reveal shorts or breaks in cables.

Oscilloscope

An oscilloscope is an electronic instrument that measures the amount of signal levels and
displays the signals on a screen. It can also detects shorts and open circuits.

Advanced Cable Testers

Advanced cable testers not only display information on the physical condition of the
cable. They can also provide information on message frame counts, excess collisions, late
collisions, error frame count, and congestion errors.

Time Domain Reflectometer (TDR)

It is used to check for breaks and shorts in cables. Just as any other components in a
system, cables can also develop problems. Yet unlike equipment that can be examined,
pulled apart, or replaced, cables often stretch for longer distances. It is usually difficult,
time consuming, and costly to locate a cable problem. Imagine pulling out meters of
cable or digging up cables that are buried underground to locate a break somewhere in
the cable. Fortunately, there is a way to identify the location of a cable problem working
from one end of the cable only, without needless exploration. This technique is called
Time Domain Reflectometry

The principle behind TDR is that any imperfection in a cable: a short between
conductors, an open or a low resistance between conductors acts as a partial mirror to
some of the energy or signal send through the cable. Some of the transmitted energy will
be reflected back to the transmitter by the discontinuity, instead of continuing to the
intended receiver. By carefully measuring the time between sending and receiving the
reflected signal, the distance to the fault can be calculated, since the propagation factor of
a given type of cable is known.

Data Communications and Networking Copyright © 2005, African Virtual University 114
The time between transmitting a signal and receiving the reflected signal is divided by
half to find the one-way time. The one-way time is then multiplied by the propagation
factor to find the distance to the fault.

Protocol Analyser

A protocol Analyser can be hardware or software tool used to analyse data packets
passing through a network and display their contents. It captures the data packets, and
once captured the data packets can be analysed in real time or in recorded time. The
actual data can be extracted from the formatting information and closely studied to
identify the cause of the problem. It can also identify bottlenecks, protocol problems and
malfunctioning network components. Most protocol analysers support several network
protocols, such as TCP/IP and IPX/SPX

The software version of this tool is known as the Network Monitor. It comes with
operating systems like Windows NT and Windows 2000. It is a scaled-down version that
can capture data between the host computer on which it is running and the other
computers on the network. Network monitor software has two components:

One component is installed on the server, and the other component is installed on the
client machines. The client component is required because it puts the network adapter
card into promiscuous mode, in which the card will accept packets not addressed to it
(obviously a requirement to monitor overall network traffic)

Network monitor basically allocates a lot of RAM to use as a buffer. It captures every
packet it encounters on a particular NIC to the buffer, gathering statistical data.

Data Communications and Networking Copyright © 2005, African Virtual University 115
Software

Simple Network Management Protocol (SNMP)

It is a TCP/IP protocol for monitoring networks. SNMP consists of three parts:

• SNMP Manager
• SNMP Agent
• Management Information Base (MIB)

The SNMP Manager is a computer or station that has the SNMP management software
installed. The management software sends queries to the device that are being managed
(devices running the SNMP Agent)

The SNMP agent is part of the SNMP protocol, which are programs that run in Key
Network components such as:

• Hubs
• NICs
• Routers and Bridges
• Other specialized equipment

When the SNMP Agent is running, it records information about the device on which it is
executing or running. On demand, it returns the information when requested by the
SNMP Manager.

The MIB is a database on which the agent records information pertaining to the device on
which it is executing.

The SNMP Manager (a central Management program) is used to poll the agents and
download the information stored in MIBs. The information is presented in the form of
graphs, maps and charts that can give information about the behaviour of the network.

Data Communications and Networking Copyright © 2005, African Virtual University 116
18

WAN Transmission

WANs are composed of geographically dispersed LANs that are interconnected. Some of
the technologies used for connecting LANs include the following:
• Analog
• Digital
• Packet switching
• ISDN (Integrated Services Digital Network)
• Frame Relay
• ATM (asynchronous Transfer Mode)

Analog connectivity uses dial-up lines or dedicated lines. A dial-up connection is

established for each communication session and released when the communication
session is complete. They are slower and not totally reliable for transmitting data. They
are normally used when connection between sites is required for a certain amount of time
each day. For example, to transfer files or update databases between sites. Dedicated lines
(also known as Leased lines) provide full-time dedicated connection. A leased analog line
is faster and more reliable than a dial-up connection.

The reason why digital lines are preferred to analog lines is that they provide reliable
transmissions. Digital lines are available in several forms including DDS, T1, T3, T4, and
switched 56.
Digital Data Services (DDS) can be used to implement digital connectivity. It is faster
and offers more secure transmission environment than analog lines. DDS provides point-
to-point synchronous communications at the following data rates:
• 2.4Kbps
• 4.8Kbps
• 56 Kbps.
DDS uses digital communication and does not require modems.

Data Communications and Networking Copyright © 2005, African Virtual University 117
T1

It is a digital leased line that uses two-wire pairs (one for sending and the other for
receiving) to transmit data using full duplex mode. It is a point-to-point transmission
technology (used for transmissions between two fixed points) A T1 line operates at
1.544Mbps. It is used to transmit digital voice, video and data signals. A T1 line can be
split into a total of 24 separate channels called Fractional T1 (FT1). Each of these
channels can operate at 64Kbps.Subscribers who do not need or cannot afford the cost of
an entire T1 can subscribe to one or more T1 channels. A T1 line is known as an E1 line
in Europe.
NB. T1 services are not available in all countries.
T3
It is similar to T1, but T3 has an even higher capacity. A T3 line can transmit at up to
45Mbps. A T3 line is made up of 672channels each at 64Kbps (Fractional T3). T3 and
Fractional T3 leased lines provide voice, video and data services from 6Mbps to
45Mbps.They are the highest capacity leased lines available today. They are used for
point-to-point transmissions.

Switched 56

It is a LAN-to-LAN digital dial-up service that transmits data at 56Kbps.The advantage

of switched 56 is that it is used on demand, thereby eliminating the cost of dedicated or
leased line.

Packet- switched Networks

Because packet technology is fast, convenient and reliable, it is used in transmitting data
over wide areas such as between cities, states, or countries. Data is broken down into
packets and each packet can be switched or routed separately since it contains all the
necessary information required to send it.

Data Communications and Networking Copyright © 2005, African Virtual University 118
X.25 Packet Switched Network (PSN)

It is a packet switching network standard developed by the International Telegraph and

Telephone Consultative committee (CCITT) in the mid 1970’s.X.25 was designed with
extensive error checking to ensure that data packets are reliably delivered in poor and
noisy telephone lines. The extensive error checking causes X.25 to be somewhat slower
than other WAN transmission methods. Although X.25 may be slower than T1 or ATM,
it is faster than a dial-up connection. The biggest advantage of X.25 is that it is an
established standard and is more reliable for International connections.

Fig 18-1 X.25 Network

X.25 devices fall into three general categories:

Data Terminal Equipment (DTE). It is a device that acts as the data source or sink (a
device that can send or receive data).DTEs are usually terminals, personal computers or
hosts that are connected as end systems.
Data Communication Equipment (DCE). It is a device that establishes, maintains and
terminates a connection between stations and also provides the necessary signal
conversion required for data transmission over a communication link. DCEs are
communication devices such as modems and packet switches.

Packet Switching Exchange (PSE). These are switches that make up the bulk of the
carrier’s network. They transfer data from one DTE device to another through the X.25
packet switched Network (PSN)

Data Communications and Networking Copyright © 2005, African Virtual University 119
The X.25 packet switched network uses the following protocols to perform the functions
of the first three layers (bottom three layers) of the OSI reference model.

Packet-Layer Protocol (PLP). PLP is the X.25 network layer protocol. It maps to the
Network layer of the OSI reference model. It manages packet exchange between DTE
devices across virtual circuits. PLP operates in 5 distinct modes:

i) Call setup. This mode is used to establish switched virtual circuits between DTE
devices. It is used only with switched virtual circuits (SVCs), not with Permanent Virtual
Circuits (PVCs). The call setup mode is executed on a per-virtual circuit basis, which
means that one virtual circuit can be in call setup mode while another is in a different
mode.

ii) Data transfer mode. This mode is used for transferring data between two DTE devices
across a virtual circuit. In this mode, PLP handles segmentation and reassembly, error
and flow control. Again, this mode is executed on a per-virtual circuit basis and is used
with PVCs and SVCs.

iii) Idle mode. It is used when a virtual circuit is established but no data transfer is taking
place. It is executed on a Per-virtual circuit basis and is used only with SVCs.

iv) Clearing mode. It is used to end communication sessions between DTE devices and to
terminate SVCs. This mode is executed on a per-virtual circuit basis and is used only
with SVCs.

v) Restarting mode. It is used to synchronise transmission between a DTE device and a

locally connected DCE. This mode is not executed on a per-virtual circuit basis.

Link Access Procedure Balanced (LAPB)

It is a data link layer protocol that manages communication and packet framing between
DTE and DCE devices. LABP is an implementation of the High level Data Link Control
(HDLC) protocol. LAPB is a bit-oriented protocol that ensures that frames are correctly
ordered and delivered error-free to their destination. LAPB maps to the data link layer of
the OSI reference model.

Data Communications and Networking Copyright © 2005, African Virtual University 120
X.21 bis

It is a physical layer protocol used in the X.25 packet switched network. It defines the
electrical and mechanical specifications that are required to transmit data on the physical
medium. It supports point-to-point connections.

Packet Assmbler/Disassembler (PAD)

The PAD is a device commonly found in the X.25 packet switched network. It is usually
located between a DTE device and a DCE device. A PAD performs three primary
functions:

• It buffers data sent to or from a DTE device.

• IT assembles outgoing data into packets and forwards them to the DCE device.
The assembly process includes adding the X.25 header to data packets.
• It disassembles incoming packets before forwarding the data to the DTE. The
disassembly process involves removing the X.25 header from data packets.

X.25 virtual circuits

A virtual circuit is a logical connection created to ensure reliable communication between

devices in a network. Two types of X.25 virtual circuits exist. These are switched virtual
circuits (SVCs) and permanent virtual circuits (PVCs)

SVCs are temporary connections that are established for the period of data transfer and
broken down after the data transfer.
PVCs are permanently established connections used for frequent and consistent data
transfer.

Data Communications and Networking Copyright © 2005, African Virtual University 121
ISDN (Integrated Service Digital Network)

It is a dial-up connection designed to provide voice, video and data transmission services
on digital telephone networks. Because it is a dial-up connection, it is used when needed
and released when communication is complete.

Channel Types

ISDN is built on two main types of communication channels:

• The bearer (B) channel carries voice, video and data information. A ‘B’ channels
operates at 64kbps.
• The delta (D) channel carries control and signaling information. The ‘D’ channel
speed is 16kbps for BRI ISDN and 64kbps for PRI ISDN.

Service Types

ISDN offers two services types:

• Basic Rate Interface (BRI). Basic Rate ISDN uses three channels. Two ‘B’
channels to carry the digital data and one D channel that manages the link using
signaling information at 16Kbps.Basic Rate ISDN is therefore referred to as 2B +
D. Basic Rate ISDN can therefore provide a maximum data rate of 128 Kbps.
This means a computer using Basic Rate ISDN can use both channels together for
a combined 128Kbps data rate. If compression is supported, much higher
throughput can be achieved. Sometimes the D channel is used to carry data as
well, though this option is only now being offered within vendors’ products.

• Primary Rate Interface (PRI). It uses the entire bandwidth of a T1 carrier by

providing 23 ‘B’ channels each at 64Kbps and one D channel at 64Kbps (23B +
D). The D channel is only used for signaling and link management.

Data Communications and Networking Copyright © 2005, African Virtual University 122
Frame Relay

Frame Relay is an advanced fast packet switching technology commonly employed in

WAN connections. Frame Relay was designed around faster fibre optic networks.
The specifications for Frame Relay address some of the limitations of X.25. Unlike X.25,
Frame Relay assumes a more reliable network. This enables Frame Relay to eliminate
much of the X.25 overhead required to provide reliable service on less reliable networks.
Frame Relay services typically are implemented at line speeds of 56Kbps up to
1.544Mbps (T1).

Frame Relay networks are gaining popularity because they are much faster than other
switching networks. This is because Frame Relay uses Permanent Virtual Circuits (PVC)
that gives permanent virtual pathways for WAN connection (an end-to-end permanent
virtual path). As a result there is no need for Frame Relay devices to perform
fragmentation and reassembly or to provide best-path routing.

In a permanent Virtual Circuit (PVC) the virtual circuit is permanently setup. It is

always available and connected. In switched Virtual Circuit (SVC) the virtual circuit is
setup during the period of communication and then destroyed after the communication
session.

Frame Relay networks can also provide subscribers with bandwidth as needed
(bandwidth on demand), which allows nearly any type of transmissions. Frame Relay
technology requires a Frame Relay capable router or bridge.

ATM (Asynchronous Transfer Mode)

It is an advanced packet switching technology that provides high-speed transmissions

over LANs or WANs. It is used to transmit voice, video and data. It uses fixed 53-byte
cells, whereas other technologies employ frames of variable length to accommodate
different amounts of data. 53-byte cell consists of 48-bytes of data and 5-bytes of header.
The 48-bytes of data was selected to suit both audio and data transmission needs. Audio
information for instance, must be delivered with little latency (delay) to maintain a
smooth flow of sound. Audio Engineers therefore preferred a small cell so that cells
would be more readily available when needed. For data however, large cells reduce the
overhead required to deliver a byte of information. Traffic that is time-critical such as
voice or video can be given priority over data traffic that can be delayed slightly with no
ill effect.

Data Communications and Networking Copyright © 2005, African Virtual University 123
ATM can theoretically offer transfer rates of up to 1.2 Gbps, but most commercial
hardware currently operate at 155Mbps.ATM can be implemented with any type of
media (coaxial, twisted pair or fibre-optic). To reach the maximum speeds, fibre media
should be used. ‘Asynchronous’ refers to the characteristic of ATM in which
transmission time slots do not occur at the same intervals, but are at irregular intervals.

The biggest disadvantage of using ATM right now is that it is still relatively new and,
therefore is much more expensive than other solutions. As the technology becomes more
accepted, the cost will decrease. The ATM network requires ATM compatible devices
such as routers and bridges.

Data Communications and Networking Copyright © 2005, African Virtual University 124
Glossary

A transmission method that allows data to be send and

received at irregular intervals, one character at a time using start and stop bits.

Access control list (Acronym: ACL). A list that contains information about the users
and groups that have predefined permissions to some resources.

Access control. The mechanism for limiting access to certain resources based on users’
identity and their membership in various predefined groups.

Backbone. It is a generic term used to refer to the main cable in a network that
interconnects a number of computer segments or subnets.

Bandwidth.

In electronic communication, bandwidth is the range (or band) of frequencies that an

electronic signal uses on a given transmission medium. In is expressed in terms of the
difference between the highest-frequency and the lowest-frequency available for the
signal.
In computer networks, bandwidth is often used to describe the data carrying capacity
(data transfer rate) of a communication channel.

Baseline. A range of measurements that represents acceptable performance under typical

operating conditions, which is used as the basis of comparison for all subsequent
performance statistics of the network.

Bit-Error Rate (BER). The fraction of data bits transmitted that are received in error.

Bottleneck. A bottleneck is a hardware resource that is slowing down system or network

Performance.

Cross talk. The unwanted coupling of electromagnetic signals from one communications
cable to another as on a telephone line.

Cut-through. Technique for examining incoming packets whereby an Ethernet switch

looks only at the first few bytes of a packet before forwarding or filtering
it. This process is faster than looking at the whole packet, but it also allows some bad
packets to be forwarded.

Daemon: A program that is usually initiated at startup and runs in the background and
can be called to start other processes or services. It waits and listens in the
background until summoned by another process.

Data Communications and Networking Copyright © 2005, African Virtual University 125
Glossary

Default route: A route that is used when no other routes for the destination are found in
the routing table. For example, if a router or end system cannot find a
Network route or host route for the destination, the default route is used.
The default route is used to simplify the configuration of end systems or
routers. For IP routing tables, the default route is the route with the
network destination of 0.0.0.0 and net mask of 0.0.0.0

gateway. The address of the router’s interface that is used to forward packets
not destined for the local network.

Demand-dial connection. A connection that is established when data needs to be send

and terminated after the data has been send.
Distance vector routing. A routing algorithm that utilizes constant broadcasts to send
routing tables to other routers.

Distributed computing A form of computing where all the processing is shared by

many different computers.

Enterprise Network: A network comprising all the networks within a single

organization.

Fibre core. The centre of the optical fibre through which light is transmitted.

Firewall. A combination of hardware and software that provides a security system,

usually to prevent unauthorized access from outside to an internal network or intranet. A
firewall prevents direct communication between network and external computers by
routing communication through a proxy server outside of the network. The proxy server
determines whether it is safe to let a file pass through to the network. A firewall is also
called a security-edge gateway.

Flow control. Synchronizing the sender and receiver so that data is transmitted only as
fast as the receiver can handle.

Frequency Division Multiplexing. A technique whereby the bandwidth of a single

communication channel is divided into a number of sub channels of different frequency
ranges to allow several signals to be transmitted simultaneously over the same
communication channel.

Data Communications and Networking Copyright © 2005, African Virtual University 126
Glossary

Internet. It is a system of linked networks that are worldwide in scope and facilitate data
communication services such as remote login, file transfer, electronic mail, the World
Wide Web and newsgroups. The Internet was initially restricted to military and academic
institutions, but now it is a full-fledged conduit for any and all forms of information and
commerce. Internet websites now provide personal, educational, political and economic
resources to every corner of the planet.

Internetwork. A collection of several networks that are connected by bridges, switches,

or routers, so that all users and devices can communicate, regardless of the network
segment to which they are attached.

Intranet. A private network inside a company or organization that uses the same kinds
of software that you would find on the public Internet, but that is only for internal use.

IP address. A 32-bit address used to uniquely identify a node or a host on an IP

network.
Ipconfig. Is a command-line utility that displays the current TCP/IP configuration
information on a networked computer

Learning bridge (Also known as transparent bridge). This type of bridge is transparent
to the device sending the packet. At the same time, this bridge will learn over time what
devices exist on each side of it.

Load balancing. The process of distributing some load across multiple links, servers,
processors or other devices in order to improve performance and overcome deficiencies
in existing equipment.

Logical topology. The logical path a signal follows as it passes among the network
nodes.

Network Infrastructure. It refers to the cabling system, network device, and

architecture of the network.

Network segment. A portion of a computer network separated by a computer network

device such as a hub, bridge or switch.

Node. It is any device connected to a network, such as a computer, networking

equipment or a printer.

Link-state routing A routing algorithm that only sends out route table changes.

Logical Topology. The logical pathway a signal follows as it passes among the network
nodes.

Data Communications and Networking Copyright © 2005, African Virtual University 127
Glossary
On-demand service. A type of telecommunication service in which the communication
path is established almost immediately in response to a user request.

Routing. The sending of data packets end systems (stations or nodes) located on different
networks.

Routing table. A database of routes containing information on network IDs, forwarding

addresses, and metrics for reachable network segments on an internetwork.
Server farm. A collection of server computers maintained by an organization to meet its
server needs.

Spanning Tree. An algorithm used by bridges to create a logical topology that connects
all network segments, and ensures that only one path exists between any two stations.

Synchronous Transmission. A transmission method that allows data to be

transmitted as a group or block at regular intervals without the need for start and
stop bits. The receiver is synchronized with the transmitter

Subnet. A subdivision of an IP network. Each subnet has its own unique subnet
network ID.

Time-Division multiplexing Is a technique whereby signals from multiple sources

are transmitted at different time intervals over the same channel.

Throughput. The amount of actual user data transmitted per second without the
overhead of protocol information such as Start and Stop bits of frame headers and
trailers.

Network Traffic. Data flowing through a network

Vampire tap. A device used to connect drop cables to thicknet cables.

Virtual circuit. It is a logical connection created to ensure reliable transfer of data

between stations on a network.

Well-Known Ports: Ports in the range from 0 – 1023

Workstation. A computer attached to the network.

Data Communications and Networking Copyright © 2005, African Virtual University 128