There are two types of Router

Routers have data-link layer destination addresses. They look at network layer address, which are in the data-link layer frame.

Most advanced routers include bridging functionality. These are called Brouters.

The destination address on the Ethernet or token ring packet is the MAC address of the router. They discard this MAC sub layer "envelope" and look inside it at the contents of the data field of the packet. What it is looking at is the Network layer destination address and it reads this and consults its routing tables to determine the best path to forward on.

The routing table can contain

After reading information in envelope it repackages as required for the delivery route chosen. The Network layer address does not change but a fresh data-link layer frame is created. In this is the MAC address of the next router and the best path to the ultimate destination. It also includes as the source address its own address to tell where it’s coming from, and not that of the original source. This functionality is referred to as "end-to-end" addressing. Also before sending it confirms the existence of the destination address. This is what is known as "forward-if-proven-remote logic"

Routers only understand Network numbers therefore they can only communicate with other routers and local network interface cards.

Another function of a router is that it listens to and identifies busy routes thus cutting down on congestion.

Look at Fig. 9.3 and 9.4 in handout

Like a bridge routers filter and isolate traffic and connect segment. They have access to more of the information in packets than bridges do. They also provide better traffic management and don’t pass on broadcast traffic.

They know the address of each segment and are thus able to make better decisions on which path to choose. They will only pass on information if the address is known and they wont let bad data through thus reducing traffic.

Routers are good at making use of bandwidth on large networks containing redundant paths – load balancing. They choose the best route by the number of hops, the cost, and line congestion. They can also prioritise processing of packets due to session time-out restrictions or time sensitivity of embedded data.

Routers are slower than bridges, but this is because they do more. They are more complicated to configure and manage than bridges. With more routers it leads to proportionately more complex management of the network. They also must have all supported protocol stacks installed and properly configured and are more costly than bridges.

Routers can route over most protocols. This gives them the ability to process multiple network layer protocols simultaneously. Protocols that are routable are:

Some routers can handle the non-routable protocols even though they don’t have any network layer address scheme. Protocols that are non-routable are:

Because of the different protocols there are several paths available for the router to use.

There are two different routing protocols – routing information protocols (RIP) and Open Shortest Path First (OSPF). The advantages of OSPF are that in handles larger internetworks as well as having a smaller impact on network traffic.

RIP uses distance vector algorithm. This measures the number of hops to the destination router with a max of 16 hops. OSPF uses the link state algorithm. This decides between multiple paths to a given router on

Most importantly it uses less bandwidth in its efforts to keep routing tables up-to-date.

Another advantage of OSPF over RIP is that while RIP uses information supplied by neighbouring routers only, OSPF receives information from all routers on a given internetwork.