![]() |
Whilst it is possible to link together Local Area Networks (LANs) using repeaters, bridges and routers, sometimes the distances are too great and costs do not permit the laying of cabling over great distances. In order to link LANs together it may be necessary to build use a Wide Area Network (WAN). WANs can be built using a range of different technologies, which will be discussed.
Before looking at these different technologies, we will first look at the three different ways in which packets of data can be routed, known as packet routing.
There are three ways in which packets of data can be routed to their destination:
Circuit switching involves the alteration of the actual physical circuit used to make the connection. The circuit used is clearly defined and does not change during the life of a connection. During the connection, the circuitry is (on most systems) not available for any other connection. Typically circuit switching is used for telephone calls over fixed (land) lines.
The entire message is forwarded from switch to switch until it reaches its destination. For example a telegram.
Packet switching refers to the routing of packets of data to the correct destination. The data is broken into small packets Using small packets of data means that if any of the data does not make it to the destination, having been lost on route, then only those missing packets need to be re-sent, not the entire message. The packets are created at the source and re-assembled into the whole message at the destination. This is performed by a Packet Assembler/Disassembler (PAD). The route the packets take can change every time a new message is sent and even packets that are a part of the same message may take a different route.
There are a number of different protocols for packet switching, these include X.25 and FDDI but there are others.
The X.25 protocol (pronounced as, “X 25”) defines how data is packaged and also performs the routing of the data. X.25 is a packet switching protocol. X.25 is robust, using confirmations to ensure that communication has been successful, providing reliable connections.
X.25 is widely used for long-distance data communications networks and for many national telephone systems and also in-store credit card transactions. X.25 networks are sometimes referred to as Value Added Networks (VANs) because they have the potential to provide other networking services.
In order to establish a connection using the X.25 protocol, an initial Call Request Packet is sent from the source to the host. The host replies with a Call Accept Packet. This occurs at every step along the route.
In order to end the communication a Call Clear Packet is sent and this is replied to with a Clear Confirmation Packet.
Since X.25 checks that the transmission of data has been successful, when TCP communication is made over an X.25 network the TCP confirmation process is redundant, X.25 having already done this at a lower level (in the OSI) model. This means that the additional TCP communication used to ensure the reliability of communication are actually redundant when communicating over an X.25 network. The additional overhead means that with modern networking protocols (like TCP) the reliability of X.25 is simply an overhead. This has been one of the factors leading to the increased use of other protocols.
Frame Relay is less reliable than X.25, but since the reliability of X.25 is redundant when using protocols such as TCP, this does not matter. In fact the lack of reliability built into Frame Relay means that it is now preferred over X.25 for building WANs. Essentially Frame Relay leaves it up to higher-level protocols, such as TCP, to ensure that the communication is reliable.
A disadvantage of this approach is that if the network is highly congested and packets are lost as a result, then it is not until the packets have reached the destination, is a request sent for the re-transmission of the missing packets. This Terminal Recovery means that more network traffic is generated when the network is already congested. In order to address issues of congestion, Frame Relay network use congestion control software to monitor and correct congestion.
ATM is a packet switching protocol, using packets of variable length. The variable length of packets is better suited to the bursts of transmission that occur over networks, reducing latency on the network. ATM also has the advantage that it can potentially be used from the desktop PC, through theLAN and out over the WAN, thus avoiding any change in network communication protocol as the boundaries between different networks are crossed. ATM also has the advantage that it can carry both data and voice simultaneously, something that X.25 and Frame Relay cannot do.
There is a disadvantage with ATM; the percentage of the packets taken-up with ATM routing and timing information tends to be greater than for Frame Relay or X.25. This means that the amount of band width used for ATM communications is greater than the equivalent communication over one of the other frame relay protocols.
Typically FDDI systems are used on large single sites, such as office complexes or the campus of a university. They consist of two rings of fibre optic cable up to 100km in length. Nodes attached to the rings provide access points to the fibre optic cables. The two rings provide a primary and a secondary (backup) communication.
This is similar to FDDI but works over much larger areas, such as metropolitan areas. Two rings of cable, usually fibre optic, are used as with the FDDI system. DQDB provides speeds of up to 100Mbps.
Telephone lines can be leased from a phone company. These can be used with a standard dial-up modem or a special network adapter card. The speed of connection can vary depending upon what has been agreed with the phone company.
ISDN uses standard telephone wires to carry data, so has the advantage that telephone cabling already in place can be used as well as national telephone systems. However, the local telephone network must be digital in order to provide ISDN services. The local loop (from the local exchange box to the telephone socket) must also be digital, in many cases this means that the local loop will require conversion. ISDN typically provides connection rates of up to 150kbps, that’s more than three times faster than typically achieved using a 56k modem.
T1 cables are provided by telephone companies. They typically carry data at up to 45Mbps. A large diameter T1 cable terminates at a T1 terminator, which feeds into the router for the LAN. These fast services are expensive, typically costing hundreds or thousands of pounds per month, depending upon the speed of the connection. Such connections may be required by companies needing to move very large quantities of data (e.g. share trading firms) or those providing digital telephone services (such as automated dial-up systems used in competitions, etc.).
In the UK DSL is usually marketed as ADSL (Asymmetric DSL), or more commonly as Broad Band. There other variations other than ADSL, such as HDSL (High-bit-rate DSL) and RADSL (Rate-Adaptive DSL). These services are actually provided by the local telephone company, although they may be rented through another company. In the UK BT control most of the local exchanges, consequently other companies wishing to provide ADSL services in the UK must actually rent the services from BT. The end user may be unaware that this is the case.
ADSL typically provides up to 8.1Mbps for downloading, although the rate provided will depend upon the contract with the service provider. The upload speeds are typically several times slower than for downloading.
In order to provide DSL services the local telephone exchange is upgraded with a Digital Subscriber Line Access Multiplexer (DSLAM). This provides a combination of voice and data services, allowing both services to be provided over the same local loop. Sometimes the term final mile is used, to indicate that DSL services are provided for the final part of the connection, from the telephone exchange to the phone socket.
Companies providing cable television services can provide fast Internet connections through a cable modem. These can workout cheaper for the end user than DSL services. Cable modems usually are a separate external box from the PC, connecting to the computer via an Ethernet connection.
There are two main types of cable modem:
One-way Coaxial Modems: These use the cable modem for downloads but require a separate dial-up modem for up-loads. Typical download rates are 2Mbps.
Two-way Hybrid Fibre-Coaxial: The cable modem is used for both down-loads and up-loads. Typical download speeds are 3-10Mbps.
There are two types of satellite, near-Earth orbit and geostationary. The near-Earth orbit move rapidly across the sky, and so can only provide communication services for a limited period. Geostationary satellites are stationary relative to the surface of the Earth, so are essentially fixed communication relay points in the sky. Geostationary satellites have the disadvantage that they are in a higher orbit and there is a delay for the signal to travel from the Earth to the satellite, usually around 0.27 seconds. This can be cumbersome for certain types of application.
There are a number of advantages to satellite systems, such as not having to install long cabling on the ground or under the ocean. The cost of laying cabling over long distances can make using a satellite a cheaper option.
For smaller communications the Very Small Aperture Terminal (VSAT) can be used, requiring a dish of 1.2 – 2.8 metres diameter. VSAT systems typically provide data rates of about 19kbps, which is fairly slow.