This training is intended to provide the basics of the Routing Information Protocol (RIP). RIP is a distance vector routing protocol based on the Bellman-Ford algorithm. RIP is mainly used as an interior gateway protocol but, in some cases, it can be found in external networks. The distance vector algorithm passes complete routing tables to its neighboring routers. RIP uses hop count to determine the best path. If the count is the same it will automatically perform a load balancing round robin.

A problem arises when the hop count is the same but the link bandwidth is different. For example, if one link is 1Gig and the other is 100Mbps. Normally, one would want to use primarily the 1Gig link but because the hop count is the same, the paths will have equal-cost links (pinhole congestion).

For the basic information about RIP, please read:

• RFC 1058
• RFC 2453

As described in RFC2453, RIP version 1 does not consider autonomous systems and IGP/EGP interactions, subnetting, and authentication. The lack of subnet masks is a problem since they use subnet masks to determine a route. There are also limitations to the protocol such as its limit to 15 hops for longest path, the protocols' dependence upon "counting to infinity" to resolve certain situations (max hop count), and the protocol's use of fixed metrics to compare routes. RIP version 2 makes improvements in the way the protocol deals with changes in topology by making provisions for timing out routes, preventing instability usually by assigning a large metric value to represent "infinity." Another feature added to improve RIP are "trigerred updates," or the setting of a timer for the exchange of updates. Also, Split Horizon states that information cannot be sent back in the direction from which it was received.

Route poisoning and Holddowns are also solutions to RIP convergence issue. By Route poisoning, inconsistent updates are avoided. A route is poisoned by making it "unreachable" or "infinite." A Holddown prevents routes from changing too rapidly by allowing extra time to come back up. Both of these solutions improve convergence a bit by preventing updates to be transmitted to other routers.

RIP protocol includes request and response commands. Upon reception of a response message, the router will update its routing table with:

• The address of the destination
• The metric associated with that destination
• The address of the next router
• A recently updated flag
• Several timers

A response mesage will be sent at a regular interval of 30 seconds or triggered by an update to the routing tables