This is a WAN that uses a privately owned or leased line, which only subscribers have access to. The Technology Colleges Trust Intranet is an example of a private WAN, which has many technology colleges subscribed to it. Colleges can contribute and access data on this WAN. Joint Academic Network (JANET) is another example of a private WAN, set up by the universities. Internet service providers such as CompuServe and AOL are types of private WAN.
3. Components
Workstations
A workstation is either a stand-alone computer or a dumb terminal (no processing power of its own) connected to a network. Each workstation has a network card which fits into an expansion slot inside its casing. The network card enables workstations to communicate across the network, and with the file server. The cards job is to convert computer generated data into a form suitable for transmission.
Data transmission media
This is the material through which data signals are sent. These include metal wire, optical fibre and air (for radio or microwave transmission).
Ethernet Cabling can be divided into three different types:
-
Thick Ethernet Cable is 10 mm in diameter with a solid copper core. A single network segment can stretch up to 500 metres and will support up to 100 nodes. The transmission rate is 2 Mbps.
-
Thin Ethernet is also 10 mm in diameter but with a stranded core. The maximum length is 180 metres, and it can support 30 nodes. Transmission rate is 2 Mbps.
-
10baseT standard Ethernet has a pair of wires twisted together. It provides the same transmission rate as a thick Ethernet but is much cheaper and easier to install.
Fibre Optic cabling
This is made up of thousands of glass fibre strands which transmit light or infra red rays instead of electrical signals. The transmission rate is much faster than Ethernet.
Workstation to media connectors
These connect the wires between terminals to the input socket situated at the back of the computer. Thin Ethernet requires connection by T piece. Which connects two wires to the Ethernet Network Card in the station. The 10baseT requires the telephone bell-connector-type connection.
Network cards
A network interface card (or just network card) is a piece of hardware needed inside a PC in order to use it as a terminal for a network. It is a small circuit board that slots into one of the main connectors on the motherboard. A network cards main purpose is to reduce the cabling between each terminal to a single wire.
File servers
This manages the files on a network. File servers may be 'dedicated' in which case they may only manage files and not be used for anything else. Other file servers may also run programs as well as store files. In a network where the computers connected have different specifications, the computer with the highest specification should be chosen as the file server since the speed of this computer will determine the speed of the whole network.
Printer servers
Peripherals such as printers can be shared of a network, so that any station on the network can print to a selected printer. For this to work a station must be set up to control incoming print jobs in the form of a queue, this station is known as a print server. Such a system has the advantage that individual workstations are no longer tied up managing their own printing.
Software
- Server operating system
- Station operating system
- Application operating system
Server operating system - This is required for any computer to function. A network's software is different to a standalone in that the server requires an operating system that enables stations to share resources. Server software operating systems include Microsoft LAN manager, Windows NT server, UNIX, Novell NetWare. This type of software offers extra facilities such as allowing users to access files from the file server.
Station operating system - The station must also have its own operating system which needs to be compatible with that running on the server like Windows NT server. Most Windows operating systems are compatible with that on the server.
Application software - Unless the application software is not able to run across a network then generally all software of this kind (Word etc.) is installed at the server.
Multiplexer
This piece of hardware is used to connect many computers to one server, which may be some distance away, by using a single high speed communications line and a single modem rather than each computer having their own communications line and modem. All computers are linked into the multiplexer, which is then fed into the modem and then into the communications line. At the server location, the single line is fed into the modem, which in turn feeds the multiplexer, which then splits the signals back up into their original number of signals.
Auxiliary Storage
One of the most important tasks on a network is hard disc management. Users often think that the space is unlimited. When the server hard disc becomes full it stops operating because there is no room for it to make temporary files for day to day operation. Each user must be allocated a finite amount of disc space. However there are times when this is not enough. This is when auxiliary storage can be used. If the station has a hard disc, then it can be used as auxiliary storage. There are however other facilities available. Floppy discs can be used if the files are not too big. Zip/Jaz drives enable files to be copied onto to tape. These tapes can be very large (12 Gigabytes is not uncommon). This facility is often used for backing up server information for security. Writeable CD’s can also be used as auxiliary storage. However access for reloading work from these devices is generally slower than from the hard drive, either of the station of the server. Data can also be compressed in order to store it in compact form (e.g. floppy disc)
Bridge
These are used to connect two LAN’s in the same location. It can also be used to keep two LAN’s separate, only allowing specified data through it. The connection of two LAN’s enables sharing of data and easier management; i.e. rather than two LAN’s to manage it can be treated as one and managed from one location. It can also be used to separate segments on a single LAN. For example in a company there are many departments. Some departments require similar information but all will have their own information that only they will require. The bridge can be set so that it only allows specified information through. This speeds up the performance of the segments because they are no longer carrying all information, so of which is not required. A bridge can only connect networks that have the same topology.
Gateways
These are hardware devices that decide which way a packet of information needs to be sent, ensuring the information is sent in the most efficient way. This device is more intelligent than a bridge in that it can decide which way a packet of information needs to be sent. It can also connect networks of differing topologies. These are used generally to connect remote networks. A gateway/router will be used at each end of the communication line.
A LAN such as the network in college includes the following components:
- Workstations
- File server
- Network card
- Print server
- Fibre Optic Cabling
- Ethernet Cabling
- Auxiliary Storage
- Server Operating System
- Station Operating System
A WAN such as the Internet includes the following components:
- Work Stations
- File Server
- Auxiliary Storage
- Station Operating System
- Server Operating System
- Application Software
4. Benefits of a network
Shared
Software can be installed on one computer (the network server) which can then be made available to selected users or stations. This makes installation easier for the network manager, and helps security of the network. For example, in a school only certain people need access to the staff and pupil database. On a network this information would be installed on the server and made available only to those people or stations that require it, rather than installing it on individual machines.
Users can share peripherals such as printers and CD-ROM’s. These are potentially expensive, and when linked to individual machines can be relatively underused and so inefficient.
Information can be stored and backed up centrally. This improves security because only one hard disc needs to be managed, and the backup routine can be carried out automatically at one place, and also in a shorter time than can be done on many hard drives.
Only via a network can electronic mail be sent. A network therefore enables the computer to become a communications tool. Relatively new developments in hardware, have enabled text to be scanned in, digital photographs and digital videos to be taken, and all to be transferred across the network electronically in a matter of seconds (depending on size of file) - any where in the world.
Central Services
When upgrades need to be added to software held centrally it can be done on the file server instead of having to load many stand-alone machines. With stand-alone machines different versions of the same application may be used within an organisation, leading to non-standard output and difficulties with staff using different terminals. This is avoided when using a central copy of an application.
Security is likely to be improved as many organisations use diskless workstations, preventing viruses and prevents users from stealing data. It is also easier for an organisation to make sure employees are not using illegal software if software is kept on a central computer.
Efficiency
Networks enable central management and support. It is easier and often less time consuming, from a management point of view, to install users and new resources on a network. Users need only be installed once at the server, and in many cases they can be copied from existing users’ requirements. Resources need only to be installed once, rather than onto many machines individually. However when installing software and hardware onto a network it is different to installing it standalone, and in many cases the process takes more time, than a single installation. This is because the same application/resource needs to be shared amongst many users, whilst not losing any functionality.
Team working
Networks allow users to collaborate. That is it enables them to share information, and build on common files. For example several users can add to a common database. In this case all contributing to one is better than all users having their own individual database. This improves teamwork.
5. Network Topologies
A topology is the pattern in which the stations/nodes (CPUs, servers, printers, etc.) are connected by the communications channel, usually wires.
Ring topology
Ring topology is only used with LANs, with no controlling computer and the computers connected in a circle. For one computer to communicate with another the data must pass through other computers until it reaches its destination. This causes the data signal to degenerate if the wires are long, devices called repeaters increase the signal strength. (See also token ring on page 7)
One advantage is that it is easy to identify a problem with the cable or data carried.
Bus topology
All the computers are connected to a single pathway, with no controlling computer. A message sent from one computer to another is also sent to all the other computers. Only one computer is able to transmit data at a time.
Mesh topology
All computers are linked, but not necessarily together. The Internet is a good example of a mesh topology. There may be several routes from one computer to another, so if one computer goes down there will be another route to the destination.
Star topology
This is where all the computers are connected to a central, controlling computer by separate links. If a link fails it will not effect the rest of the network and can easily be found. Also each computer has its own link so there is no competition for space, which makes it faster. However, the large amount of cabling required is costly and unsightly, and the management of the server is crucial.
6. Data flow control methods
Reservation
Line reservation is the process by which a terminal keeps exclusive use of a communications line while it is sending data to another computer in a network. There are various mechanisms that terminals can use to achieve this:
This type of line reservation is used for networks that can be managed as a ring, where each message can be sent from one computer to another until all the computers have been traversed. A special message called a token is passed from computer to computer around the ring. When a computer wishes to send a message it waits until the token arrives, captures it and transmits the message.
The main problem with this method is that the token can be lost along the line and it is difficult to discover where this has occurred and to restart the token. However, this method tolerates broken links and terminals which fail to respond properly, making it useful when reliability, fault tolerance and a high rate of data transfer are needed.
This when each terminal is given a time slot in which to transmit its data. There are no problems with collisions as no two time slots coincide, but vacant time slots can slow the whole network.
This is used to control communications channel access in LANs. The central computer asks each workstation to see if they want to send information.
The star topology has a central computer to control the network and so employs polling. This is to make sure the data is transferred to the central file server in an orderly manner.
All the terminals are polled in turn, but some will have no data to send which wastes time. It is more efficient if the operating system deals immediately with data as it is sent.
Contention
Contention occurs when two terminals transmit data at the same time. In order to avoid this a terminal must ‘listen’ to see if there is any traffic on the network. If data is detected it is captured, stored and the destination address of the data checked to see if it is for the terminal at which it has arrived or one of the others. If there is no traffic on the network, the terminal is able to transmit its own data. There is a chance that two terminals may think the line is free and try to transmit data simultaneously, with the resulting intermixing and corruption of the transmissions. Terminal must therefor have a procedure for dealing with these ‘collisions’, or find a method to avoid them.
Collision avoidance/detection
Carrier sense multiple access/collision detection (CSMA/CD) is the most common method of dealing with collisions. This method involves a ‘listen-and-wait’ scheme, which means when a collision occurs, each transmitting computer waits a random period of time before transmitting the signal. This makes a repeat collision unlikely, but if it does occur again the process is repeated.
7. Management Activities
Checking network activity levels and system reporting
Software manufacturers have produced a variety of products to make the job of managing a network much easier; this software is called network management software. The purpose of such software is to gather statistics about the movement of data around the network and watch for any faults.
Security
Security is very important to a network as data is widely distributed to many people, so there is a greater risk of misuse.
Networks are vulnerable to viruses because they can pass from one computer to the whole network, causing havoc in the process. There are a number of ways to prevent viruses getting on the network and minimise the damage if they do (e.g. scan disks, make back-ups, antivirus software, etc)
Storage
Networks have a certain amount of storage space, each user having an allocated amount. The server can monitor the amount of space and deny access if there is not enough.
System configuration
This is how the network is set up, in terms of users, passwords, station connection, registration with the server and the allocation of resources (e.g. printers, scanners, etc)
8. System security methods
Control of access
Access of control can be divided into two groups:
- Physical controls
These prevent the user from getting to the computer or making it work.
- Controlling access to the room (keypads, special cards)
- Controlling access to the building (guards, locks)
- Lock the computer so it can't be switched on
- Lock the computers away when not in use
Logical controls
These prevent the user from accessing or changing certain files. There are two types of network security: user level security and share level security. In user level, each user is given a user name and password, which have to be entered to gain access to the network. Share level is when each resource on the network is given a password.
Forced recognition
In order to make users aware of the importance of security, clauses can be written into their contracts of employment. In government departments and the armed forces, staff have to sign the Official Secrets Act making it a crime to pass information on to outside bodies.
