Networks – General Information - Part1

Centralised versus Distributed Systems – a brief history

In the 1960-70’s large mainframe computers had been developed and these were the majority of computers found around the country. These computers were very expensive to buy and costly to maintain. To a large extent they were limited in what they could do. The computers were mainly located at the headquarters of organizations and therefore the use of them was limited to only those who had access to them.

As they became smaller computers became cheaper and more widespread. As this happened they were used more at smaller locations and the needs of these local groups was developed. At about the same time software became cheaper and easier to use. Whereas before people require a large amount of computer skills to operate even basic software, with the newer software being developed it was made more user friendly, requiring a lot less training to be able to use and install it. Much of the software being developed also met the very different needs of these smaller groups.

The next major step in networks was the linking of these computers and making them more usable. This had the advantage of combining the local needs of users to centralized information and resource sharing.

Local Area Networks (LANs)

A definition of a LAN is a collection of personal computers (PCs) that are cabled together so that they can share resources and communicate with each other.

Simple network

Each computer is connected via a Network Interface Card (NIC)

It is now possible to connect users anywhere in the world and to the users they appear and behave like they are part of a small LAN when in reality they are part of wider network called a WAN (Wide Area Network).

The standards used for these connections was developed by the Institute of Electrical and Electronic Engineers and these standards are abbreviated IEEE (number).

The Evolution of the LAN

The demand for LANs has increased over last few years to a point where over 50% of computers in the world are now part of a LAN. With large organizations the trend has gone even further with the linking together of smaller LANs into larger networks. There has also been a great demand that different types and makes of LAN’s should be able to communicate together.

It started :-

1. Simple resource sharing such as printers.
2. File and document sharing was introduced. At this stage the network operating systems were the biggest weakness
3. As better operating systems were developed, software companies made the user interfaces, menu systems, network functions and control/monitor systems easier to understand and operate.
4. Software began to develop for applications to use the advantages of networking. Email programs and other applications were tailored especially for networks.
5. The last step was in the interlinking of LANs to mini computers and mainframes. This final step brought the benefits of both worlds together and enabled the establishment of versatile networks that could cater for most types of commercial requirements in the workplace.

Characteristics of a LAN

1. Geographically localised – either a building or building complex.
2. Uses intelligent PCs as workstations and not dumb terminals as used in mainframes.
3. Each PC performs the users own processing.
4. Cable of transferring data a high speed.
5. All network devices connect to a trunk cable or hub device.
6. Share hardware resources such as printers.
7. Share software resources such as application packages.
8. Share data resources such as database records.
9. Privately owned and managed by individual companies.
10. They can be expanded to add other users.
11. They can connect equipment from different manufacturers.
12. Cheaper than WANs.

Each PC on a LAN usually has limited recourses of their own. LANs sometimes use additional computers called servers

Servers:

1. Holds the application software run by the users – such as Word for Windows
2. Stores data files needed by users.
3. Controls all printing by using print queues.

The Use of LANs

Typical office layout in which a LAN could be installed.

Needs

• Print server
• Network software
• Company data files
• Disk space for users
• CDROM access

Software

When installing application software you can

1. Load each computer with its own version of the application
2. Can configure the network OS to distribute the software automatically to users
3. Can install a multi-licence of the software on a network file server.

Case number 1 is very time consuming even with the initial installation, not to mention upgrades.

The second to cases are far more efficient. These can be menu driven and easily used by people on the network. The greatest advantage comes from the fact that the installation need only be done once at the server and that once it is on everyone has access to same software.

Data Files

It is import that there is only version of a data file as each update happens. This one file should be the same across all users on the network. The advantage of centralized data is that the information it contains is current and available to all. Also being on the server it is more likely to be backed up on a regular basis.

Share Peripherals

Some peripherals such as laser printers, CDROM stackers are expensive. Networking allows these resources to be shared and cost savings to be achieved.

Shared Communication

One of the greatest uses for networks is electronic mail (email). Servers also handle switching of other digital links such as voice and video.

Common Topologies

There are three broad categories

Bus Topology

1. Until recently this was the most commonly used of the three.

• It consists of common trunk cable
• Each workstation is directly connected to the cable.
• Each workstation has its own NIC and this card is an active component of the network.
• Each trunk cable has terminators.
• Bus topology uses coax cable as the trunk
• There are two sizes – thick and thin
• The distance between workstations is governed by the size of the cable.

Questions (handout of blanks)

• How many servers would you need
• Where would you place the servers and printers?
• Where would the cabling actually be laid?

Answer diagram below:

2. Tree Networks are a variation on the above system. Each branch becomes a sub-network where traffic from one branch to another can be controlled to optimise the network performance.

Ring Topology

They consist of a series of point-to-point links with controlled data flow in ONE DIRECTION. All workstations are connected to one trunk cable laid out in a loop. See diagram below.

There are two types of systems

1. IBM token ring, which uses either, unshielded or shielded twisted pair. They operate logically as a ring but are wired as a star. The difference being that each station is connected to the hub by two wires. One in and other out.
2. FDDI (fibre distributed data interface), which uses fibre optic cable. These are by large companies and on large LANs. They are durable, high-speed networks running at 100Mbps.

Questions (handout of blanks)

• Where would the cabling actually be laid?

Star Topology

Star networks are a series of point-to-point links connected via a hub located at the centre. Hubs at as high-speed ports and switches. Prior to 1993 star weren’t used much and were usually slow and small. When the standard was changed to IEEE 802.3 these became faster, more reliable and cheaper even though more cable was needed in an installation.

Questions (handout of blanks)

• How many and where would you place the hubs?
• Where would the cabling actually be laid?

LAN Configurations

We need to look at what type of operating system the networks will use to determine what it will support and how it will relate the users to data and to each other. Networks operating systems can be classified in two broad categories, peer to peer and server based. Each of these has its own characteristics and supports different operational environments.

Peer to Peer

These include NOS (network operating systems) such as Apple Share, Lantastic, Netware Lite and Windows for Workgroups. All these NOS support resource sharing and e-mail.

• Any workstation can share any of its resources with other users.
• Any workstation can be used to process local applications while at the same time providing services for other users.
• Each workstation is capable of providing services to others (being the server) and requesting services from other workstations (being the client). In a server based system a PC is either a server or a client not both.

In the above diagram station B acts as a server to station A who is a client requesting print services.

Server based

Server based networks are used to provide

• File serving
• Print serving
• Data serving

The limitations of the servers are :-

• It is there to serve the workstations therefore it cannot be used as another workstation.
• When workstations run application software it is downloaded from the server hard drive to memory in the workstation.
• Data may come from the server or be provided locally from the workstation.
• Servers handle the printing but the process is performed locally in the workstation.

The good points of the servers are :-

• Resource sharing
• Email
• On-line database applications. (applications run on the workstation but server provides sharing and storage. This includes file locking, access control, restrictions and update control)

Hardware Requirements for a Small LAN

1. When a server is used it must be high performance, high speed and with a large storage capacity.
2. PC’s and servers must connect to networks via a NIC, which must conform to standard being used on that network.
3. Cabling and connectors.

Metropolitan Area Networks (MAN)

1. Provide communication covering a city but uses LAN technology.
2. Conforms to the IEEE 802.6 standard.
3. Carry voice data and images at 200 Mbps.
4. Radius of up to 75 Km.
5. Can consist of a number of different LANs using different communication media.
6. It may consist of common carrier, satellite, microwave and other media.

Wide Area Network (WAN)

Links computers over a large geographical distance. This is usually greater than 10Km.

An example of a WAN would be the TAB, which covers computers all over Australia.

1. No geographical limitations
2. Connects via a terminal to the host (central computer)
3. Usually work at a lower speed than LANs.
4. Uses Telstra or other common carriers to make the connections.
5. Uses special equipment to link to the carrier network.

WANs allow users an ‘entry point’ into the network system. Just as a telephone is connect to common equipment each time a call is placed, so WAN are usually switched at the time of connection. These nodes are sometimes called exchanges for this reason. Permanent connections can be leased through providers and these are far more expensive but are permanently available to the users.

Network Options

Private Networks – these are leased lines and made permanently available to the hirer. They are high cost but come with a guaranteed performance margin, which is higher than switched lines.

Public Switch Networks – These are temporary links, which are established via whichever path is available at the time. They can switch by many varied routes depending on traffic conditions at the time.

Advantages of Each

Leased lines – high quality data transfer with fewer errors and a higher degree of monitoring by the provider.

Switch lines – provide flexibility and can switch via alternate routes when a bearer is down.

Disadvantages of Each

Leased lines - are more expensive and with any major failures require the provider to restore services.

Switched lines – require several seconds of dialling time to connect and because they use a switched network are more susceptible to noise, which can corrupt data.

WAN Topologies

Terminal/host model

1. Requires a host computer top do the processing.
2. Terminals for users to enter data (usually dumb terminal or PC emulating dumb terminals)
3. Communication equipment to connect and transmit data to the network.

There are several topologies applied to WANs. Point to point, multipoint, star, ring and mesh

Point to Point

Provide direct connection between two pieces of data terminal equipment (DTE). These links may be permanent or temporary.

Multipoint Networks (or Multi-drop)

Provides access to the remote host for several terminals by means of ONE communication line. Because access is shared it must be controlled by line control procedures, via a MUX (multiplexer).

Star Networks

Consists of a number of point-to-point links and a central hub, which is usually a computer centre. Most of the communication is between the DTE’s. The central computer provides the message switching service.

Ring Networks

Consists of a number of point-to-point links, but without a central hub. Each DTE communicates directly with the station either side and acts like a repeaters to the next.

Mesh Networks

Mesh networks require no message switch because each DTE can connect to ANY other

DTE directly.