Examining Interior Routers
As you learned earlier in this hour, interior routers operate within an autonomous network. An interior router should have complete knowledge of any network segments attached to other routers within its group, but it does not need complete knowledge of networks beyond the autonomous system.
Several interior routing protocols are available. A network administrator must choose an interior routing protocol appropriate for the conditions of the network and compatible with the network hardware. The following sections discuss the important interior routing protocols:
RIP is a distance vector protocol, and OSPF is a link state protocol. In each case, the real protocol must address details and problems that weren't discussed in the broad methodologies described earlier.
By the Way
Most routers available today support multiple routing protocols.
Routing Information Protocol (RIP)
RIP is a distance vector protocol, which means that it determines the optimum route to a destination by hop count. (See the section "Distance Vector Routing" earlier in this hour.) RIP was developed at the University of California, Berkeley, and originally gained popularity through the distribution of the Berkeley Systems Design (BSD) versions of Unix. RIP became an extremely popular routing protocol, and it is still used widely, although it is now considered somewhat outdated. The appearance of the RIP II standard cleared up some of the problems associated with RIP I. Many routers now support RIP I and RIP II.
As described earlier in this hour, RIP (as a distance vector protocol) requires routers to listen for and integrate route and hop count messages from other routers. RIP participants are classified as either active or passive. An active RIP node is typically a router participating in the normal distance vector data exchange process. The active RIP participant sends its routing table to other routers and listens for updates from other routers. A passive RIP participant listens for updates but does not propagate its own routing table. A passive RIP node is typically a host computer. (Recall that a host needs a routing table also.)
When you read the previous discussion of distance vector routing, you might have wondered what happens when a hop count received and incremented is exactly equal to the hop count already present in the routing table. That is the kind of detail that is left to the individual protocol. In the case of RIP, if two alternative paths to the same destination have the same hop count, the route that is already present in the routing table is retained. This prevents the superfluous route oscillation that would occur if a router continually changed a routing table entry whenever there was a tie in the hop count.
A RIP router broadcasts an update message every 30 seconds. It also can request an immediate update. Like other distance vector protocols, RIP works best when the network is in equilibrium. If the number of routers becomes too large, problems can occur because of the slow convergence of the routing tables. For this reason, RIP sets a limit on the maximum number of router hops from the first router to the destination. The hop count limit in RIP is 15. This threshold limits the size of a router group, but if the routers are arranged hierarchically, it is possible to encompass a very large group in 15 hops.
Although the distance vector method does not specifically provide for considerations of line speed and physical network type, RIP lets the network administrator influence route selection by manually entering artificially large hop counts for inefficient pathways.
Open Shortest Path First (OSPF)
OSPF is a more recent interior routing protocol that is gradually replacing RIP on many networks. OSPF is a link state routing protocol. OSPF first appeared in 1989 with RFC 1131. Several updates have occurred since then. RFC 2328 covers OSPF version 2, and some later RFCs add additional extensions and alternatives for the OSPF protocol.
Each router in an OSPF router group is assigned a router ID. The router ID is typically the numerically highest IP address associated with the router. (If the router uses a loopback interface, the router ID is the highest loopback address. See Hour 4 for more on loopback addresses.)
As you learned earlier in this hour, link state routers build an internal map of the network topology. Other routers use the router ID to identify a router within the topology. Each router organizes the network into a tree format with itself at the root. This network tree is known as the Shortest Path Tree (SPT). Pathways through the network correspond to branching pathways through the SPT. The router computes the cost for each route. The cost metric can include parameters for the number of router hops and other considerations, such as the speed and reliability of a link.