Unicast IP Routing
Unicast IP Routing |
|
While simple and well supported in the industry, RIP for IP suffers from some problems inherent to its original LAN-based design. The combination of the these problems makes RIP a desirable solution only in small to medium-sized IP internetworks.
RIP uses a hop count as the metric for the route stored in the IP routing table. The hop count is the number of routers that must be crossed to reach the desired network. RIP has a maximum hop count of 15; therefore, there can only be 15 routers between any two hosts. Networks 16 hops and greater away are considered unreachable. Hop counts can be customized so that slow links are set to multiple hops; however, the accumulated hop count between any two networks must not exceed 15.
The RIP hop count is independent of the Time-to-Live (TTL) field in the IP header. On an internetwork, a network 16 hops away would normally be reachable for an IP packet with an adequate TTL; however, to the RIP router, the network is unreachable and attempts to forward packets to hosts on the network result in ICMP Destination Unreachable–Network Unreachable messages from the RIP router.
RIP allows for multiple entries in the routing table for a network if there are multiple paths. The IP routing process chooses the route with the lowest metric (lowest hop count) as the best route. However, typical RIP for IP router implementations, including Windows 2000, only store a single lowest metric route for any network. If multiple lowest hop count routes are received by RIP, the first lowest metric route received is stored in the routing table.
If the RIP router is storing a complete list of all the networks and all of the possible ways to reach each network, the routing table can have hundreds or even thousands of entries in a large IP internetwork with multiple paths. Because only 25 routes can be sent in a single RIP packet, large routing tables have to be sent as multiple RIP packets.
RIP routers advertise the contents of their routing tables every 30 seconds on all attached networks through an IP subnet and MAC-level broadcast. (RIP v2 routers can be configured to multicast RIP announcements.) Large IP internetworks carry the broadcasted RIP overhead of large routing tables. This can be especially problematic on WAN links where significant portions of the WAN link bandwidth are devoted to the passing of RIP traffic. As a result, RIP-based routing does not scale well to large internetworks or WAN implementations.
By default, each routing table entry learned through RIP is given a timeout value of three minutes past the last time it was received in a RIP announcement from a neighboring RIP router. When a router goes down due to a hardware or software failure, it can take several minutes for the topology change to be propagated throughout the internetwork. This is known as the slow convergence problem.