Chapter 4 Computer Networks and Communication

Computer Network

A computer network is a group of computers that are connected together so that they can share
resources. Resources shared can be processor, storage devices, printers, scanners or fax
machines. The network size differs widely from as few as two computers connected together in
an office, up to millions of computers on the internet. The variations in numbers has accounted
for different technologies to be used in interconnecting these computers. Popular computer
networks were constructed using cables (wired connections) whilst new technology has brought
up wireless connections. Computers need a network interface card (NIC) to be able to get
connections from other computers. Cables for wired networks or wireless adapters are slot on
these cards. The card also determines the speed of data transmitted on the network.

Advantages of computer networks

- Allows sharing of peripherals such as printers or fax

- Control of other computers (workstations) by the administrator is easier. For example, a
bank teller’s transactions can be monitored by management.
- Allows sharing of information e.g. the internet or concurrent database access.
- Users can log into the system from workstations that might be geographical dispersed.
- The server can upgrade software in use by workstations for the whole organization
without any travelling. Antivirus can be updated and new versions of Adobe Reader can
be downloaded.
- Multiple computers can share one software license e.g. Pastel Accounting license for
multiple users.
- Enables communication between organisations and individuals by use of emails

Disadvantages of computer networks

- They employ complicated operating systems (e.g. Linux and Windows Server 2003)
which are difficult to use.
- Failure of the main server or its cabling results in failure of the whole organisation due to
unavailability of network.
- Virus spread is faster due to file sharing.
- Complicated operating systems used require employing network administrators which
increases costs.
- Computer networks require more hardware cabling e.g. hubs, switches which also
increase costs.
- Hacking is easier then on a standalone computer.
Local Area Networks (LAN)

This refers to computers that are connected to share resources over a small area. They can be
networked in a building or a small area of not more than one kilometre. LANs are common at
home, offices of an organisation, schools and colleges. A LAN is made up of workstations and
servers. A server is a computer that is responsible for managing workstations by providing
required resources in the network. Resources required may be in the form of storage or
peripheral devices.

Network topology

This refers to the layout of computers connected on the network. The following are classes of
network topologies that can be uses on LANs.

 Star network

In this topology, all computers are connected to each other through a central computer that is
usually the server. This server contains a network operating system that manages resource
requests from all computers on the network. Each workstation is connected to the central
computer by cables. The server must have high processing power and large memory so that it
can manage transactions. The server might be connected to the internet. All software required
for use by the network may be installed on the server since they are accessible to all
workstations. The server provides security checks e.g. virus scanning as well as managing
print queue. Fig 4.1 shows star topology.

Fig 4.1 Star Topology

Workstation A Workstation A

File Server Internet connection

Workstation A
Workstation A

Printer
Advantages of star topology

- It is easy to trace faults (either of a workstation or faults of cables) since each workstation
is connected directly to the server.
- Since each workstation is uniquely identified by the server, security is easier to enforce.
- Failure of one of the workstations or failure of a cable to a workstation affects only that
part of the network not the entire network.
- The increasing number of workstations does not reduce its speed since each workstation
is directly connected to a high speed server.

Disadvantages of star topology

- There are many cables which increases cost

- It is difficult to install due to the large numbers of cables especially when the network
covers more than one building.
- Failure of a cable means no communication to that workstation
- The network operating system required for this topology is more expensive and
complicated as compared to other network topologies
- Since all communication must pass through the server, it may become congested
- Failure of the server means no network at all

 Bus Network

The bus network has a server also, but the server is connected to workstations through a main
bus cable. Each computer is connected to the network by cables branching off this main
cable. This topology is more practical and it is widely used by organisations today.

Fig 4.2 Bus Topology

Workstation File Server
Internet connection

Printer

Workstation Workstation Workstation

Advantages of bus networks

- Cheaper to install since there is less cabling

- Easier to install
- Easier to add other workstations
- Allows more than one servers so that backup will be available when one of the servers
stops functioning
- Coaxial cables may be used which increases speed

Disadvantages of bus topology

- Failure of the main cable means failure of the whole network

- Fault finding is difficult
- It becomes slower when more users enter into the network
- Less reliable due to its reliance on the main cable
- Not as secure as star network due to the transmission of all data on one main cable

 Ring Network

Fig 4.3 Ring Network

Workstation
Printer

Workstation

Workstation

Workstation

Workstation
Printer

Internet
connection

Serial data transmission

In serial data transmission, data is send and received as a series of bits, that is one bit at a time
and in sequence. The data travels in a queue. Data that travels over long distances is usually
transmitted by serial mode. The long distance in computer studies means any distance that is
more than one metre. Therefore, most of the cabling used on networks uses serial data
transmission. Also that is why the mouse and keyboard use serial transmission.

There is a common misconception that serial transmission is slow. This is not always the case.
The use of Universal Serial Bus (USB) has seen a very high speed serial connection that can be
used even for processes that formerly required parallel transmission like printing. Therefore,
transmission speeds depend on the type of the cables used. Serial network cables in use on the
internet can be as fast as 100Mbps, which is one million bits per second. This is fast enough to
sustain communication. Fig 4.4 shows serial data transmission of the word DADDY.

Fig 4.4 Serial Data Transmission

Step 1 ADDY D
Step 2 DDY DA
Step 3 DY DAD
Step 4 Y DADD
Step 5 DADDY

Parallel data transmission

Unlike serial transmission of data, parallel communication is done by sending of data

simultaneously using different routes. This is the case with data bus cables in the computer. The
more the wires used; the faster the communication. These buses are used inside the computer as
32-bits connections that connect the processor and memory. Parallel communication is used for
printing functions which require large volumes of data.

Fig 4.5 Parallel transmission

D D
A A
One Step D D
D D
Y Y
Parallel transmission uses more cables than serial transmission therefore; it is more expensive
than serial transmission. There is also need for timing data transmitted so that the data arrives in
the correct sequence. This is called synchronisation. The signal also degrades as distance
increases because of the interference between the signals and other signals from the environment.
Therefore, parallel transmission is not very suitable for long distances, but more efficient over
short distances.

Handshaking

This term describes of agreeing to communicate by computers. This is done before any data has
been send. Handshaking is similar to the shaking of hands by human beings which shows
agreement. Handshaking in the context of computer science means that the devices notify each
other that they are both ready to communicate, they agree on maximum speeds both parties can
sustain, and the protocols that they are going to use when communicating. Computers have
different processing speeds, so they have to notify each other so that data transfer rate matches
speeds on both sides otherwise, the receiver might be overloaded if its processing speed is less
than the sender’s speed.

Bit rate

This is the speed on which data transmitted travels on a network medium. It is the actual speed of
the transfer of data. This speed is measured in bits per second (bps). Bit rate and bandwidth are
different in the sense that that bandwidth describes the amount of data that can be transmitted on
the medium whilst bit rate specifies its speed. A 32Kb network card can transmit 32 K bits of
data per second, so that’s its speed.

Bandwidth

Bandwidth describes the volume of data that can be transmitted on a transmission medium. This
is the term that describes the range frequencies supported by the communicating channel. Higher
frequency means more data can be transmitted over the communication media at one time. This
expresses how many bits the media can transmit at a time. The data transmitted is carried in the
form of a wave. It is measured in bits per second. For example, a 56kb modem can transmit
56000 bits per second. Bandwidth ranges from as low as bits per second (bps) to Megabits per
second (Mbps). Bandwidth has two classes that are baseband and broadband.

Baseband

This term describes communication media that use single frequency band to transmit data.
Baseband is usually used in LANs. This can achieve great speeds over short distances.

Broadband
This describes communication media that support more than one band of frequencies to support
communication at one time. Broadband is more useful when there are larger volumes of data that
need to be transmitted. Internet use requires broadband technology since the many users have
different data requirements. One user may want to download a PDF document, another one
downloading music or videos from the internet. To support these transactions concurrently, there
is need for division of frequencies that support these data requests so that they run
simultaneously.

Synchronous transmission

This is a technique whereby the two communicating devices match their communicating speeds.
The speed at which the communicating devices can transmit data might be different; they need a
timer that manages the transmission speed to suit the receiving speed. This timer is called a
system clock. If there is no system clock then there is risk for losing some data as the receiver
will be overloaded. This timer will send requests for communication before actually sending the
data and it waits for a response from the receiver. If the receiver acknowledges, then the sender
sends, otherwise it waits until the receiving device is ready to receive.

Asynchronous transmission

Asynchronous transmission starts by dividing the data one character at a time. Each individual
bit is attached a start bit and a stop bit. These bits are used for error detection and flow control.
The bits are reassembled at the receiver using the bit sequence numbers as specified by the bit
sequence. Asynchronous transmission uses parity bits to check and correct errors.

Parity bits

These are error control bits. There are two types of parity that is even and odd. Odd parity
assumes that a bit stream is should have an odd number of ones (1). Even parity also assumes
that the total number of ones (1s) must be even.

For example, transmitting a bit stream 10001100, even parity will do the following.

Fig 4.6 Even Parity (with odd number of ones)

1 0 0 0 1 1 0 0 1

A parity bit added

1 0 0 0 1 1 0 0 1

From the above diagram, the initial bit stream was 10001100 which consisted of three ones (1).
Even parity emphasises that there must be an even number of ones so it adds another one at the
end of the bit stream to make it 100011001. This time the number of ones becomes four, which is
even. All the conversions that are carried out at the receiver are based on that all bit streams are
supposed to be even. Therefore any bit pattern that is not even is assumed to be corrupted.
Investigations are then made from this bit pattern that has errors.

Alternatively, we might have a bit pattern like 10110100 that is supposed to be identified by
even parity. Since the bit patterns have an even number of ones (1), even parity will add a zero
so that the number of ones is not affected. Fig 4.7 illustrates this.

Fig 4.7 Even Parity (with even number of ones)

1 0 1 1 0 1 0 0 0

1 0 1 1 0 1 0 0 0 Parity bit added

Modems

The term modem stands for Modulator / Demodulator. A modem is a piece of hardware that is
used to connect a computer to a telephone line. The telephone line will then connect to another
computer thereby allowing two computers to communicate using the telephone line. There are
differences in the type of data that is processed by computers and the type of data handled by
telephone lines. Computers process data in binary form (ones and zeros). This form of data is
called digital. On the other hand, telephone lines can only handle data that is analogue. Analogue
data is data in the form of waves. Therefore, the modem modulates data (converts binary digits
from computers into electrical signals that are in wave form). Also the modem demodulates
electrical waves into digital form. Demodulation is the transformation of electrical waves from
the telephone line to digital form so that they are usable by computers.

A modem can be external or internal. External modems are slot on the computer like other
computer peripherals. Internal modems are either in-built on the motherboard or they are
expansion cards that are slot on the motherboard so that they are cased in the same box as the
motherboard.

Packet switching

This is a communication method whereby data to be transmitted is broken down into packets. A
packet is a chunk of bits. A packet is also called a datagram. Each packet takes its own route that
is free and faster. Each packet has a sequence number that is used when reassembling packets.
The packet has a data portion as well as a source and destination address attached to it. The
destination address of the packet is used by other nodes when they forward the packet to its
destination. Also, a checksum is attached to the packet so that error handling is done. The
checksum will be responsible for the identification of errors in the data transmitted.

This mode of transmission is faster and secure since each packet takes a different route. This
method is suitable for non-voice data which can be split are up but still retaining its logic and
meaning. This is different from voice data which does not tolerate this mechanism. If words are
interchanged over a medium, the meaning of the message will be distorted; otherwise the
receiver may not tolerate the delays of reassembling the message.

Fig 4.4 Packet Switching

Sender Receiver

Key

Packet 1
Packet 2
Packet 3

Circuit switching

Communications between computers on the network involves other intermediate computers on

the network. These are used as nodes which redirect the requests of two communicating devices.
Therefore, the sender and receiver must be provided with a dedicated link that lasts for the
duration of the communicating period. This is analogous to telephone calls whereby the link is
created by dialling numbers of the receiver. This communication in networks is especially
applicable to online transaction processing in distributed database systems.

The nodes are other computers on the network which might also be communicating with each
other. These nodes will keep the established link open as long as the circuit has not been
terminated either by the sender or the receiver. Fig 4.4 shows how a circuit switch looks like.

Fig 4.5 Circuit Switching

 Sender Receiver

Key

Node

Circuit established

Possible routes

A circuit switched network first establishes a link between the communicating devices before the
actual data is transmitted. This link will not be used by other nodes until it has been released.
This method assumes a continuous flow of data between the communicating devices. It is also
suitable for voice transmission where there is a dialogue for example communication over the
telephone.

Communication protocol

A protocol is a set of rules that govern communication between components of a network. These
rules identify the methods of identifying data on the network as well as identifying the sender
and receiver of the data. Protocols set standards just like in our daily life, we have rules for
communicating with our parents and this is different from the way we talk to friends. Common
protocols on the network are:

 Hypertext Transfer Protocol

This is a protocol that enables computers to handle different data formats. The data can be in
the form of text or graphics. Representation of text data and photographic data might be
different so this protocol makes it possible for both types of data to be processed and
displayed on one page. It also makes links to other pages.

 File Transfer Protocol

This protocol provides rules for transferring data between computers. Data transfers occur
when downloading or uploading documents. Downloading is a term that describes the
process of taking a copy of a document from the Internet Service Provider (ISP) to a
 workstation. Uploading is posting a document from the workstation to the server so that the
document can be accessible from the server.

 Transmission Control Protocol / Internet Protocol

This is a protocol that governs the transmission of data on the internet. Messages will be split
into packets where each packet will have to be managed from source to destination. Two
models were developed from this protocol, the TCP/IP reference model and the OSI
reference model.

OSI Reference Model

This is a model that explains the data as it moves between the sender and receiver as well as the
processes that manage the data transfer. It is a child of the internet protocol. OSI is Open
Systems Interconnect for short. The OSI reference model is more practical and is in use today. It
has five layers that are; medium access control, logical link, internet, transport and application
layers. Fig 4.4 shows these layers.

Fig 4.5 OSI Layers

Application Layer

Sequence
of stages Sequence
Transport Layer of stages
for
converting for
data to be converting
sent on a data
network Internet Layer received
from
network

Logical Link Control Layer

Medium Access Control Layer

Functions of the layers of OSI reference model.

 Medium Access Control (Layer 1)

This layer defines the following:

- Network topology (star, ring or bus)

- Transmission mode (simplex, half-duplex or full duplex)
- Synchronise the sender and receiver
- Representation of the data
- Characteristics of the transmission media
- The number of bits that can be transmitted per second, that is, the data transfer rate.
- Configurations that are made on the line to determine either point to point or multipoint

 Logical Link Access Control Layer (Layer 2)

This layer is responsible for managing packets of data as they are transmitted between two
adjacent nodes in a network. The nodes can be two workstations or a workstation and a
server. Its main duties are

- Framing

This is the determining of the bits that mark the beginning of a packet as well as the bits that
mark the end of a packet. Framing also identifies the address of the packet, that is its source
and destination address.

- Flow control

This is the control of the sequence by which data flows between the sender and the receiver.
This layer controls the release of the data from sender so that the receiver will not be
overloaded by data sent.

- Access control

This function tells which device can send data at what time. Is avoids overloading the link, as
well as avoiding data collision which might corrupt the data.

- Error control

This is a mechanism that determines the existence of errors as well as the mechanisms by
which these errors are corrected. This layer is responsible for parity checks on data receipt.

- Addressing

This layer keeps track of the packet’s destination address as the packet passes through some
nodes on the network before reaching its intended destination.
 The Internet Layer (Layer 3)

This layer is responsible for making an individual packet to travel from the source computer
to the destination computer. Its main duties are routing the packet and logical addressing of
the packet.

- Logical addressing. When there is more than one route possible for transmission, this
layer will be responsible for the identification of the best route to use.
- Logical addressing. This is the conversion of the source and destination address into
binary digits that can be understood by all nodes on the network. Logical addressing is
useful to other intermediate nodes so that they can forward the packets to their
destination.

 Transport layer (Layer 4)

This layer is responsible for managing the process of taking the whole message from the
sender to the destination. It manages reliability of the communicating process between the
two computers. Its duties are

- Flow control. This is similar to flow control in layer 2 above.

- Connection control. This is the establishment of connections between the two
communicating devices as required.
- Segmentation and reassembly. This process is also called packetizing. This is the
breaking up of the packet so that they can take different routes. Each packet is given a
sequence number that is going to be used when these packets are reassembled at the
destination.
- Error control. This is almost similar to error control in layer 2 above, but the only
difference is that this layer manages the whole message not individual packets.
- Service point addressing. This ensures that the program that is supposed to receive the
message will the actual program that gets it. This layer gives port addresses for messages
so that the message is received properly by protocol that has been identified by the port.
Each protocol has a specific port number on the network.

 Application layer (Layer 5)

This layer defines what message to send, when to send the message as well as the format by
which the message will be sent. This layer is responsible for the following

- File transfer. This function allows the user to have access to other file on a remote
computer, so that they can retrieve the file or manage it from the remote computer.
- Mail service. It enables the communication on the network by the use of emails.
- Encryption. This enables data to be unreadable so that it might not be corrupted.
 - Translation. It makes it possible for different platforms to communicate although they use
different formats of data representation.
- Dialog control. This defines the communication mechanism of between the sender and
receiver so that they can enter into a dialog. This can be full duplex or half duplex.
- Synchronisation. This defines synchronisation points that will be used to verify data send
against data received. This reduces the labour of re-sending the message if some packets
are corrupted. Only corrupted packets will be resend.

TCP/IP reference model

Before the advent of the OSI reference model, the TCP/IP model was used to account for data
transmission. It has seven layers. Its layers are almost similar to those of the OSI reference
model, except that TCP/IP adds two more layers before the application layer. These two more
layers take some of the OSI reference models’ application layer’s functions that is; the
presentation layer implements data translation, compression and encryption; the session layer is
responsible for synchronisation and control of dialog; therefore leaving the application layer with
functions like file transfer, email services and directory services for managing distributed
database. Fig 4.5 shows the similarities and differences between the OSI and TCP/IP reference
models.

Fig 4.6 TCP/IP Reference Model versus OSI Reference Model

OSI Reference Model TCP/IP Reference Model

Application Layer
Application Layer

Presentation Layer

Session Layer

Transport Layer Transport Layer

Internet Layer Network Layer

Logical Link Control Layer Data Link Layer

Medium Access Control Layer Physical Layer

Types of network cables used

 Twisted pair wires

The twisted pair cable consists of two cables that are twisted together and covered in a layer
of insulating material. The cables are copper wires which is a good conductor. The twisted
copper cables used can be of different types. This is cheaper than the coaxial cable and fibre
optic. Twisted pairs cables are more prone to external interference like heat. Two lines can be
used in this type of cable, so that one line may be used for data transfer whilst the other one
is used for voice communication.

 Coaxial cable

The cable consists of a single copper wire that is cased in an insulator. The insulation is then
encased in a copper braid and finally covered by an insulation sleeve. This type of cable is
used between satellite dish connections and the decoder at home or in organisations. It is also
in LANs. The cable is well insulated and this makes it more efficient in resisting external
interference by other signals in the environment. A copper braid reduces external interference
and acts as a grounding cable. Signals can travel for a considerable distance before they
require boosting. This cable is more expensive than twisted pair cables.

 Optical fibre

It is also referred to as an optic cable. This cable consists of a considerable number of glass
or plastic cables or strands that are encased in an insulation material. Data transmission is
done by method of light that travels in the glass or plastic cables. The more the glass or
plastic strands we have in the cable, the higher the volume of data transmitted simultaneously
in the cable.

Optic fibres are more reliable and efficient in the sense that light travelling is more resistant
to interference as compared to copper wires. The signal may travel for long distances before
it requires boosting. The optical fibre cables are however not as flexible as coaxial cables,
they cannot easily bend, they are more bulky and require stronger insulation as compared to
coaxial cables since they are not as strong as other cables.
Test Questions

- State and explain the advantages and disadvantages of computer networks

- Define the terms
o LAN
o WAN
o Serial data transmission
o Parallel data transmission
o Handshaking
o Bit rate
- Describe the following topologies
o Star
o Bus
o Ring
- What are the advantages of ring and bus topology?
- What are the disadvantages of star topology?
- Explain the differences between the following terms
o Synchronous and asynchronous transmission
o Packet switching and circuit switching
o TCP/IP and OSI reference models