RIP (IP Routing Information Protocol)

RIP (and IGRP) always summarizes routing information by major network numbers. This is called classfull routing. Classless routing and prefix routing protocols allow contiguous blocks of hosts, subnets, or networks to be represented by a single route. RIP is a classfull routing protocol and does not support prefix routing.

RIP is a distance-vector routing protocol that practices classfull routing, which is used to discover the cost of a given route in terms of hops and store that information in the route table.

The router can then consult the table in selecting the least costly, most efficient route to a destination. It gathers information by watching for route table broadcasts by other routers and updating its own route table in the event that a change occurs.

RIP is specified in RFC 1058 and updated with RFC 1723. Some of the differences defined in RFC 1723 are added security features. RIP messages are now allowed to carry more information in their updates.

RIP Route Tables

At a minimum, RIP route tables provide the following information:

• IP destination address

• A metric (numbered from 1 to 15) indicative of the total cost, in hops, of a particular route to a destination

• The IP address of the next router that a datagram would reach on the path to its destination

• A marker signaling recent changes to a route

• Timers, which are used to regulate performance

• Flags, which indicate whether the information about the route has recently changed

• Hold-downs used to prevent regular update messages from reinstating a route that’s no longer functional

• Split horizon used to prevent routing loops

• A poison reverse update used to prevent larger routing loops

RIP sends out routing updates at regular intervals and whenever a network topology change occurs.

When a router that’s running RIP receives new information indicating a better route to a destination, the new information replaces the older entry in its table. For example, if a router loses a link, it will recalculate the routes in its own tables and then send the revised information out to all its neighbors. Each router will receive this information, update its table accordingly, and then send the information out to all its neighbors.

Note

Neighbors are routers with interfaces to a common network.

All of this sounds pretty cool, but there are a few drawbacks. For one thing, when a topology change takes place, it results in slow convergence. In a large network, this can very well lead to the counting to infinity problem, plus the routing loops that were mentioned earlier.

All things considered, you can see that RIP can be quite useful for routing within small- to moderate-sized internetworks. But, its small hop count limit and single metric don’t really allow much flexibility in complex environments.

Real World Scenario—Another Use for RIP

There’s another use for RIP that’s not usually published in textbooks or by Cisco. Many of my clients are small businesses, and they use ISDN dial-on- demand for connecting to their ISP. I have found a use for RIP in these scenarios because of its default updates (30 seconds). These types of installations don’t normally require a routing protocol, because they are what’s commonly referred to as a stub network; there are no other networks or routers present.

Normally I only configure a static (default) route to the ISP, but since ISDN is dial-on-demand, the link will time out after normal business hours and external connection requests by users to the internal e-mail server will fail. ISPs don’t normally listen to RIP updates from their client’s networks, so enabling RIP and allowing the default updates to propagate through the connection to the ISP keeps the link up.

Cisco’s recommendation for an ISP-facing router would be to disable RIP updates towards the ISP by using the passive-interface command. Since I configure the router to advertise updates out the ISP-facing router, I have had to consider the security ramifications. RIP can be configured to advertise any network you wish with the network network-number command.

My solution is to configure a bogus network number and rely on NAT, the firewall, and security policies to secure the client’s network. The link remains up, and RIP serves another use.