Business Library

The first two articles introduced some basic LAN concepts:

1. The Ethernet LAN
2. Cabling of Ethernet LAN

As these articles discussed, Ethernet local area networks (LANs) can be quickly and cheaply put together using both copper and fibre cabling. Strictly speaking, however, the networks described so far are "LAN segments" - they are complete networks in their own right.

However, a need quickly arises which requires two or more LAN segments to communicate with each other. This might be to share data or physical devices, so this article outlines the three most common ways of achieving this:

1. Repeating

In simple terms, Repeaters are devices which re-time and regenerate the signal on the LAN to enable it to extend over longer distances than cabling by itself permits. You met a device which behaves as a repeater in Article 2 - the Hub.

Note that as you are just "repeating" signals between the two LAN segments to effectively extend the cabling, the result is a single bigger LAN segment.

2. Bridging

A bridge is a more complex device which enables two LAN segments to remain autonomous whilst at the same time allowing data to pass between them. Clearly, this requires the information on one network to be selected for insertion onto the other. Another less obvious advantage is that the autonomy of the two LAN segments removes any critical timing relationship between them - hence bridges can be used to join two LAN segments over telephone lines at relatively low speeds. It also provides a mechanism by which different LAN technologies can be linked. So the LAN segments could be in the same building or opposite ends of the Earth!

The way this is achieved is by the bridge examining the information flowing across one LAN segment and checking for those addressed to devices on the other LAN segment. Obviously this means that it is an "active" device on both LAN segments as it must capture, examine, and if necessary forward information to the other network. Upon arrival at the second LAN it must also observe the same rules as everyone else - which means it must have "buffers" in which to store information as the media might be busy doing other things.

"Yeh but wait a minute, where did these addresses come from all of a sudden?" Every network adapter (the card which goes into the PC or device) has a globally unique LAN address "burnt" onto it (more precisely this is a Media Access Control (MAC) address). The address of the source and destination are transported across the LAN with the data - so all the bridge has to do is "listen" on either side and make a note of who is where! An example of a MAC address is "00-60-0C-34-C3-20".

However the drawback with bridging is that the two LANs must use compatible MAC addresses and communications protocols. Media Access Control gives you the clue - we are talking compatible (not necessary identical) physical media here.

3. Routing

The third way of joining two or more LAN segments is considerably more complex. It relies on the use of communications protocols which work above the physical media - so it doesn't matter what type of network is at either end. The MAC functions are still there but now all they do is transport information complying with this higher network protocol across a link of some sort. The boxes which provide this type of network are known as Routers.

Network protocols make use of a second type of address - a network address - which unlike MAC addresses are assigned to devices by the network administrator.

The most common type of network address is the Internet Protocol (IP) address, an example of which is "223.112.42.17". There is a worldwide authority for allocating unique ranges of IP addresses - which are essential for any networks which need to communicate outside their enterprise to or via someone else's network. These are structured in a similar way to telephone numbers so that a portion of the address gives an indication of where you are - in IP terms which "subnet" you are in. The routers themselves work out how to get to and from all the subnets they know about or can discover.

To view this page, for example, your workstation has to have such a globally unique IP address in order to be able to use the Internet (your Internet Service Provider might dynamically allocate one temporarily when you connect). That's a bit closer to home than you might have imagined!

To communicate across the Internet you use a suite of protocols known as Transmission Control Protocol / Internet Protocol (TCP/IP).

In simple terms, TCP takes care of the end to end transfer of information and that's the problem of your workstation and the system you are communicating with to look after.

Your workstation uses IP to specify where the information is to go to (or identify where it came from in the other direction), and the network uses IP to "route" information around the network based on the subnet to which the information is addressed.

Business Implications?

• LAN segments can be linked together to enable intra and inter enterprise communications.
• Using repeaters is cheap but is only possible on very limited distances and between identical LAN types. It also forms a single larger LAN segment on which all cable length restrictions still apply.
• Bridges provide a degree of isolation between LAN segments which use compatible LAN protocols.
• Routers enable a wide variety of totally different network types to exchange information provided they use a common network protocol such as IP.
• Repeating is by far the simplest to set up, bridging is more complex, and routing more so. If you want to link two LAN segments consider them in this order and choose the first one which satisfies your requirements.