Data Communications and Computer Networks

Data communications has an ancient history, as people have always had an interest in communicating with each other. Different methods have been used and associated with each method are various advantages and disadvantages. A major problem with communications is ensuring that the receiver gets the message sent by the transmitter. In every form of communication there are common elements: 1. transmitter (sender, source) 2. receiver (destination) 3. message to be communicated 4. medium (how message is carried) Examples of medium: Medium Smoke signals Tomtom drum Pony express Carrier pigeon Post Telegraph Telephone Computer Cable Problem (Noise) Fog, Darkness Thunder Bandits Hunter Strike, Loss Broken wires Electrical Electrical

Anything that interferes with the message is technically called Noise. Convergence of Computing and Communications Communication facilities have an ancient history, but we tend to think of the advent of the telegraph and later the telephone as the beginning of modern communications. Extensive telegraph and telephone networks were established all over the world, decades before the emergence of computers.

Comp 1001 Networks

The first public telephone exchange was opened in the U.S. in 1878 and operators were used to connect subscribers. Strowger invented the automatic exchange (switch) in 1891 and this system remained in use until the 1960s when crossbar switches were introduced. The connection between two exchanges is called a trunk and trunk switches in each exchange route calls. The connection between a subscriber and the local exchange is called the local loop. In the 1970s computer controlled switches were introduced and digital switching began. Here, voice signals are converted to digital signals. In the telegraph network text was transmitted using codes, beginning with Morse, and then Baudot codes. These were predecessors of the modern ASCII code, which is frequently used in computers to represent text. The concept of a start-stop code system was developed to tell a receiver that a character was being transmitted. Seven pulses were transmitted on the line. Five were used to represent the character. A start pulse indicated that a character was to be transmitted and a stop pulse that the character was finished. This idea is the basis of the RS232 serial interface. The serial interface is the communications interface used between computers and devices such as modems, printers and computer terminals. RS232 is the name of the standard that defines the interface (e.g. how many wires are used, what each wire is used for and so on). This type of transmission is asynchronous. Characters are transmitted independently of each other as opposed to synchronous transmission where blocks of characters are transmitted and precise timing is critical.

Computer Networks
Definition A computer network is an interconnected collection of autonomous computers. The goals of a computer network include: Resource sharing: programs (O.S., applications), data, equipment (printers, disks) are available to all users of the network regardless of location. High reliability: By replicating files on different machines and having spare cpus, users are more immune from hardware/software failure. Less cost: Small machines have about 1/10 the power of a mainframe but 1/1000 the cost. By using such machines with file server machine(s), a local area network LAN can be cheaply installed. It is easy to increase the capacity by adding new machines. Communications medium: Users have access to email and the Internet

Comp 1001 Networks

Wide Area Networks (WANs)

When the computers in a networks are separated by long distances (from a few Kms to global distances) the network is called a WAN. LANs can carry data at varying speeds depending on the communications medium used. If they are based on the telephone network this speed may range from 56 Kbps to less than 10 Mbps. However, if they are based on optical fibre, much higher speeds are possible. WANs are widely used in banking. The financial markets are also heavily dependent on wide area computer networks. They provide for remote database access, which is the basis for airline reservations and home banking. They also provide electronic mail or e-mail.

Local Area Networks (LANs)

When the computers in a network are located close together (usually less than 1or 2 Kms), the network is called a LAN. LANs are used in many offices, schools, colleges and hospitals. They are suitable for networking within a building or campus area. Because of the short distances involved they are typically (though not necessarily) faster than WANs in that it is cheap to use high quality communications media such as coaxial cable and optical fibre. They can carry data at very high speeds e.g. from 10 Mbps to 100 Mbps being very common speeds, while speeds of up to 1000 Mbps are possible.

Client-Server Model of Computer Systems

Networks often provide services on one or more machines for all users of the network. The machines providing the service are called servers. The machines (users) using the service are called clients. Several servers may be used so that if one crashes, users can still access their data and the network services. This is a decentralised computing system whereby users can avail of local processing on their own machines as well as the shared services of the network. A good example is a LAN with a file server. The users store files access software on the file server, but use their own PCs for processing. Dedicated servers may also be used for email or for printing: mail servers and print servers.

Mainframe-Terminal Model of Computer Systems

This is the traditional centralised computing system, whereby a single central mainframe provides all the processing for users who are connected to it via computer terminals. A computer terminal is simply a screen and keyboard although a PC may be used to act as a terminal via terminal emulation software.

Comp 1001 Networks

The terminals may be local (on same site) or remote and connected via a modem. Response time varies according to the number of users logged on and so is not readily predictable. A major disadvantage is that all processing terminates in the event of a mainframe crash.

Network Configuration.
The user machines in a network are called hosts. The hosts are connected by a subnet which carries messages between hosts. The subnet is made up of transmission lines (trunks, channels, circuits) and switching elements (computers). Transmission lines carry bits and the switching elements connect the transmission lines. The shape of a computer network can vary conceptually from a single straight line, usually referred to as a bus, to a many sided polygon with each node connected to all the others. There are two types of subnet design: 1. Broadcast subnets: In this system a message is broadcast over the network and all machines have the possibility of receiving the message. LANs usually use broadcast subnets. Each machine has its own unique address and typically will only "listen" to messages that are sent to this address. 2. Point to Point subnets: Here, a message is transmitted from one computer to another computer and so until the destination computer is reached. This is analogous to the postal system where a letter is transferred from post-office to post-office. WANs usually use point to point subnets

Broadcast Sub-networks These are typically configured as either a bus or a ring network. They can be further classified as Static or Dynamic. In a static broadcast subnet each computer gets a chance to transmit and can only broadcast a message when it's turn comes around. This is a rather inefficient use of network time, since if the computer, whose turn it is to transmit, has nothing to transmit, then the network is left idle. It has the advantage that two stations can never transmit a message simultaneously. A simultaneous transmission of messages causes a collision where the messages get corrupted and so are not received. They must be re-transmitted. A dynamic broadcast subnet makes more efficient use of the network. This system allows any station to transmit at any time the network is free of traffic. In this case, when a computer wishes to transmit, it follows the following protocol:

Comp 1001 Networks

1. Listen to see if the network is free (Carrier Sense) 2. If the network is free, transmit the message otherwise wait for a small amount of time and repeat from step 1. 3. Check to see if the message is still on the network (Collision Detect). Two machines could have carried out step 1 at same time, found the network free and proceeded to transmit their messages, thus causing a collision. If a collision is detected then wait for a random but small amount of time and repeat from step 1. In the event of a collision, both computers will wait for random time periods so that it is unlikely that they will cause another collision. The likelihood of collisions is directly related to the number of active users on the network i.e. the network traffic. The above protocol is referred to as CSMA/CD, which stands for Carrier Sense Multiple Access/Collision Detect. Multiple Access means that many users can access the network at any time. Point to Point Networks The second type of subnet, the point to point subnet, is mainly found in Wide Area Networks (WANs). If possible, the point to point subnet transmits directly to the relevant station. If no direct route is available, it will send the message to a "switch " which re-transmits the message to the destination. The best known example of this type of network is the telephone network (Public Switched Telephone Network or PSTN. also referred to as the Plain Old Telephone System or POTS)

Network Topologies.
Computer networks can be configured in a number of ways as shown below. The ring and bus topology are used in broadcast subnets. Messages are broken into smaller units called packets for transmission on a network. With the ring configuration each packet of information is sent off around the ring on its own. The same applies to a bus . The packets travel out to the ends of the line where they will encounter a device called a terminator. It serves the network as a refuse collector, intercepting and absorbing any signal that reaches the end of the bus and prevents it from being reflected back. The complete network, which is a point to point subnet, differs from the others. In this configuration each station is connected directly to every other station on the network. Its biggest advantage is speed, however it is not very practical because of the number of connections that would be required.

Comp 1001 Networks

The loop is the next configuration design for a point to point network. Each packet is transmitted along the line until it encounters a computer. The computer at this point in the network then holds the packet, and any other that is transmitted, until it has accumulated the entire message. Only when the intercepting computer has collected and reassembled the entire message is that message forwarded on to the next computer. This process is continued station by computer until the entire message reaches its final destination. This is called store and forward.

The principles for the tree, intersecting loop and star configurations are the same as those discussed in the loop. The telephone network is based on the tree topology. A terminalmainframe model has a star topology.

Comp 1001 Networks

Network Topology

a) Ring

b) Bus

c) Complete

d) Loop

e) Tree

f) Intersecting Loop

g) Star

Comp 1001 Networks

Data Transmission
There is a maximum limit to the amount of data that can be transmitted using any transmission medium (Shannons Law). Each medium has its own maximum. Data is transmitted as signals. The number of signals per second is the frequency of the signal. This is measured in Hertz (Hz). One cycle per second is 1 Hz. One million cycles per second is 1 MHz. Bandwidth The maximum range of frequencies that can be transmitted is the bandwidth of the medium. This determines the maximum amount of data that can be transmitted. A coaxial cable might have a bandwidth of 10 to 100 MHz. An optic fibre has a bandwidth of up to 108 MHz. The higher the bandwidth the more data that can be transmitted per second, hence the popularity of optic fibres. In fact, computers cannot yet utilise the full capacity of optic fibres i.e. they cannot transmit at the speed which the fibre is capable of carrying.

Transmission Media Copper Wires

o Twisted Pair (tp) o Coaxial Cable (coax)

Fibre Optic Wireless

o Microwave o Infrared o Radio

Twisted Pairs (tp) They are used by telephones for the local loop (connection between your home phone and the local telephone exchange). They carry electrical signals. A tp consists of two insulated copper wires (1mm diameter) twisted to reduce electrical interference. Capacity: dependent on the distances involved but can be up to several Mbps over a few Kms. For example ISDN (Integrated Services Digital Network) lines offer speeds from 64Kbps to over 1 Mbps and have been available to home users for Internet access, for several years. More recently (2003), DSL (Digital Subscriber Line) and in particular ADSL (Asymmetric DSL) lines are available to home users with speeds of 1.5 to 6 Mbps. ISDN and ADSL both use digital

Comp 1001 Networks

transmission and so must use a digital line unlike the standard analog telephone line where a modem is used. You must install an ISDN card or an ADSL card into your PC to use an ISDN or ADSL line. TPs may be shielded (stp ) or unshielded (utp ) with the shielded having extra insulation. However, it is the rate of twisting (number of twists per inch) that is the most important characteristic. They are also classified into Category-5 (CAT-5) and Category-6 (CAT-6). CAT-5 can carry 10 or 100 Mbps (10/100Mbps) over short distances e.g. up to 100 metres approx. The communications standard used in this case is called 10/100-Base-T. This is the type of cable that is often used in building to connect PCs to a LAN. Usually, the CAT-5 cable connects to a device know as a hub which is less than 100 metres from each PC. There may be a hub for each floor/laboratory in a building. CAT-6 cable operates at 100/1000Mbps (Gigabit Ethernet) and is typically used to interconnect hubs. It is more expensive than CAT-5 cable. Large organisations frequently have a so-called "backbone" network that interconnects separate LANs in different buildings/rooms as in the diagram below. Over short distances CAT-6 cable may be used but optic fibre is also often used as it can cover longer distances.

Third Floor

PCs to hub via CAT-5 Hub

PCs

Hub

CAT-6 Backbone
Hub

Duct to carry cable

First Floor

Servers
UPS

Three floor Building with CAT-6 Cable Backbone connecting LANs on each floor

Comp 1001 Networks

The UPS is an Uninterruptible Power Supply that usually consists of some form of battery backup power system. This will keep the system powered on, in the event of a mains electricity failure. Depending on the UPS, it may be able to provide power for periods of several minutes to several hours and at least long enough for the system to be shut down in a controlled fashion so that files get saved and so on.

Coaxial (Coax) Cable

Carry electrical signals. It consists of a copper core surrounded by 3 outer layers of insulation. It has a high bandwidth and good noise immunity. The original Ethernet standard was based on 10 Mbps coaxial cable. Ethernet is the most popular LAN standard and was developed at Rank Xerox (who also developed the mouse, laser printer and Graphical User Interface (GUI) software. Ethernet LANs can be based on tp, coax or optic fibre. There are two forms of coax: Thick Coaxial: operates over distances of up to 500m based on 10-Base-5 standard. Thin Coaxial: operates over distances of up to 200m based on 10-Base-3 standard. Capacity 10 to 100 Mbps for distances of up to 1 km. Frequently used in LANs but is being replaced by utp/stp in most LANs.

Optic Fibre
Uses light to carry data and has a huge bandwidth. Very thin glass fibres used. To date capacity of 1000 Mbps over 1 km is feasible. It is used in WANs, LANs for interconnecting hubs and also for linking telephone exchanges. Excellent noise immunity as it does not suffer from electrical interference and is therefore suitable for harsh environments such as factory floor.

Wireless
Line of Sight: Infrared and Microwave
Physical cables have a major problem if you have to cross private or public property where it may be difficult or very expensive to get permission, in addition to the costs of laying the cable. Using line of sight transmitters avoids this problem. Microwaves can be used over long distances e.g. A 100m tower can transmit data for distances over 100 km. Cheaper than digging a trench. Relatively high speeds of 10 Mbps upwards are possible. Satellite : operate in same fashion as microwaves where the satellite operates as a Big microwave repeater in the sky!! Satellite communication has a high bandwidth giving up 50 Mbps speeds and a given satellite may be able to have many "channels" at this speed.

Comp 1001 Networks

10

Wireless
Radio LANs or wireless (wi-fi) LANs are becoming common in offices, universities, hotels, restaurants and airports. A wireless LAN enables users to connect to the Internet from a laptop computer with a wireless network card. In UCD, Commerce students use such laptops with wireless cards to connect to the college network, for course work and email.

LANs and WANs continued

As noted above it, LANs in the same building (or different buildings) may be connected by hubs. It is also possible to connect LANs together using either repeaters or bridges. When two LANs are connected using a repeater, all the traffic on one LAN is simply copied to the second LAN even though it may not addressed to a machine on the second LAN. So, I the figure below, all traffic on LAN A in Building 1 is repeated on LAN B. This can lead to increased message collision frequency. A bridge on the other hand only passes traffic from one LAN to another LAN if the traffic is addressed to a machine on the second LAN. This leads to more efficient usage of the network. So, traffic from LAN B below will only be routed to LAN C via the bridge, if that traffic is destined for a machine on LAN C. Another possible configuration is to connect LANs in different buildings to a site-wide or backbone LAN using bridges. This enables communication between all buildings. The bridges ensure that only traffic addressed to a LAN in a given building will enter that building. This means that the LANs in each building will not be swamped with network traffic that is intended for other buildings. The site wide backbone would typically be a fibre optic connection.

LAN A R LAN B B LAN C

LAN X R B LAN Y Building 4

Building 1

Site-wide Backbone

Stations ( Mainframes or PCs or servers)

B Bridge R Repeater

Comp 1001 Networks

11

Wide Area Networks

The type of data communications facility used is a function of the nature of the application, the number of computers involved and the distance between them. Two computers in the same room can be simply linked with a point-to-point wire link:

Computer A

Computer B

Interface
If the computers are different locations (cities, countries) then public carrier facilities may be used. One common technique is to use the public switched telephone network (PSTN) and a modem which enables you connect computers to the PSTN:

Computer A PSTN Modem Modem

Computer B

The model above is the typical method that a home user employs to connect to the Internet, where Computer A represents the user's computer and Computer B represents the Internet server of the user's ISP (Internet Service Provider e.g. Eircom in Ireland). The ISP server would have a high-speed connection to the Internet which is shared among the users. In the case of a number of computers to be connected which are separated by long distances then two possibilities for networking present themselves: the use of private data networks and the use of public data networks. Both e public and private networks are connected to the Internet. The earliest computer networks made use of the PSTN (telephone network) for data transmission between user equipment located at different locations as shown in an earlier diagram. The data rates using the PSTN are typically less than 56,000 bps and the cost is based on connection time. Together these factors make the PSTN unattractive for large-scale data communications and led large organisations to develop private data networks. Private data networks were set up using leased lines from the telephone authorities to connect a connect a number of switching nodes or multiplexers. ISDN lines allows speeds from 64 kbps to several Mbps. Since it very expensive to set up such networks, they are used by very large organisations such as banks. They are known as private enterprise-wide networks. Such organisations would install their own switching system which can handle both voice and data communications.

Comp 1001 Networks

12

A device known as a PBX (Private Branch Exchange) or a PABX (Private Automatic Branch Exchange) may be used to connect telephones, terminals and computers in a building or a campus. The exchange routes internal calls directly and is connected to the PSTN for external calls. The IMUX is an intelligent Multiplexer (see later) that allows a single high-speed line to be shared by a number of users at the same time. A DSE (Data Switch Exchange) allows computers exchange data over the lines.
Enterprise-wide Network

Site A PBX IMUX DSE

Site B PBX

Leased Lines

IMUX DSE

IMUX

Phones Computers

Data Switching DSE Exchange

PBX
Site C

Private Network

DSE

PBX

Private Branch Exchange

IMUX Intelligent Multiplexer

Comp 1001 Networks

13

Data Transmission: Analog and Digital

Analog Transmission: Dominated the last 100 years and is here for a while yet. Network designers made use of the existing telephone network which was aimed at voice transmission. This is actually very poor for computer networking. For example 2 computers connected by a direct cable can achieve a data rate of up to 100 Mbps with very low error rate. Using phone lines, 56 K bps is the maximum transmission speed with a relatively high error rate. It is approximately 10 orders of magnitude worse: the cost of bus ticket to town versus a moon landing is same order of magnitude. Modems Phone lines deal with frequencies of 300 to 3000 Hz. A computer outputs a serial stream of bits (1s, 0s). A modem is a device that accepts such a bit stream and converts it to an analog signal, using modulation. It also performs the inverse conversion. Thus two computers can be connected using two modems and phone line. Using a modem, a continuous signal (tone) is sent in the range 1000 to 2000 Hz. To transmit information, this carrier signal is modulated. Its amplitude, frequency, phase or a combination can be modulated .

DC SIGNALS FROM TERMINAL OR COMPUTER

MODEM

MODEM

DC SIGNALS TO A TERMINAL OR COMPUTER

This diagram illustrates the carrier signal always present between two modems. One common form of modulation Frequency Modulation (FM).
FREQUENCY MODULATION

Comp 1001 Networks

14

This is the most common form of modulation used for data transmission. In FM, the carrier signal is modulated between two different frequencies (say 1200Hz and 2200Hz) without affecting the amplitude in accordance with the digital signal that it must send. The serial interface between the modem and computer is governed by the RS-232 standard (also known as the CCITT V24 standard). Digital Transmission Digital transmission takes place in the form of pulses representing bits (1s and 0s). This is the type of communication used internally in computers. The high-speed trunks linking central phone exchanges use digital transmission. It has a lower error rate than analog transmission. The local loop (from phone to exchange) is still analog. This must be converted at the exchange to digital. A device called a Codec (coder/decoder) does this. It samples the analog signal 8000 times per second and encodes the signal digitally by representing each sample as a binary number. The technique used is called Pulse Coded Modulation or PCM.

Types of Transmission
There are three types of transmission. 1. Simplex: information is transmitted in one direction only and the roles of transmitter and receiver are fixed. This form is not used for conventional data transmission. 2. Half Duplex (HDX): transmission is allowed in both directions but in only one direction at a time. Data communication systems that use the telephone network usually transmit in HDX. 3. Full Duplex (FDX): sender/receiver can transmit and receive from each other at the same time. In order to transmit in FDX, the user usually has private direct lines.
SIMPLEX

HALF DUPLEX

FULL DUPLEX

Comp 1001 Networks

15

Multiplexing
With high bandwidth channels it is possible to share the channel so that a number of users can use the channel at the same time. This is called multiplexing. For example 4 users could each operate terminals at 10,000 bps connected to mainframe over a single 40,000 bps line as shown in the figure below. The 4 lines from the terminals are connected to a multiplexer which is connected to another multiplexer by the 40,000 bps line. The second multiplexer de-multiplexes the signal onto 4 separate lines for the computer. This gives us efficient line usage and saves money.

M U X

High Speed Line

M U X

C o m p u t e r

Terminals

Multiplexing is also used for voice transmission where optic fibres and microwave cable can handle from 8000 to 16000 simultaneous conversations. There are a number of types of multiplexing. Two common ones are Time Division Multplexing (TDM) and Frequency Division Multiplexing (FDM).

Standards and Protocols

Standards and protocols are required to govern the physical and logical connections between terminals, computers and other equipment. They are vital for data communications and computer networking Typically standards fall into two groups: official standards (from national standards bodies) and de facto standards established by common usage. An early official standard was the EIA RS-232-C for data transfer over wires. (Electronic Industries Association Recommended Standard)

Comp 1001 Networks

16

Another set of standards was the V series e.g. V-24 from the CCITT body which is a part of the U.N. Standards for LANs (local area networks) were proposed by the American Institute of Electrical and Electronics Engineers (IEEE), an influential organisation. A recommendation for standards called X-25 for access to and transmission methods for packet switched data networks (PSDNs) were proposed by the CCITT. The existence of different standards bodies regulating data communications is obviously a handicap for global standardisation. In addition, manufacturers have developed their own standards to maintain their market position e.g. Digitals Decnet standards. (The Digital corporation was taken over by Compaq which in turn has been taken over by Hewlett-Packard). The International Standards Organisation (ISO) took an initiative to develop universal data communication standards to unite standards bodies, computer and telecommunications manufacturers and users. The ISO Open Systems Interconnection (OSI ) reference model was put forward as a framework to develop standards for data communication products. An open system is one that is prepared to communicate with any other open system by using agreed rules or protocols on how the communication should take place. ARPANET: ARPANET was one of the first WANs and a forerunner of the Internet. It used a network protocol called IP (Internet Protocol) to handle the interconnection of WANs to LANs. It used a transport protocol call TCP (Transmission Control Protocol) to govern transmission of data. The two are often referred to as TCP/IP and the major protocols of the Internet. It also provided protocols for file transfer (FTP), remote login (TELNET) and e-mail (SMTP). These three protocols are still very important and widely used protocols Internetworking is the term used for the connection of two networks. The growth of internetworking between LANs and WANs and WANs and WANs led to what is now referred to as the Internet. A computer that provides for the interconnection of two different networks is called a gateway.

Comp 1001 Networks

17