Split horizon

From Wikipedia, the free encyclopedia

In computer networks, distance-vector routing protocols employ the split horizon rule which prohibits a router from advertising a route back out the interface from which it was learned. Split horizon is one of the methods used to prevent routing loops due to the slow convergence times of distance-vector routing protocols.

In this example A uses B to reach C.

A will not advertise its route for C (A to B to C) back to B. On the surface, this seems redundant since B will never use A's route because it costs more than B's route to C. However, if B's route to C goes down, B could end up using A's route, which goes through B; A would send the packet right back to B, creating a loop. With split horizon, this particular loop scenario cannot happen which improves convergence time in complex, highly-redundant environments.

An additional variation of split horizon does advertise the route back to the router that is used to reach the destination, but marks the advertisement as unreachable. This is called split horizon with poison reverse.

With poison reverse, when a routing update indicates that a network is unreachable, routes are immediately removed from the routing table. This breaks erroneous, looping routes before they can propagate through the network. This approach differs from the basic split horizon rule where routes are eliminated through timeouts. Poison reverse has no benefit in networks with no redundancy (single path networks). One disadvantage to poison reverse is that it might significantly increase the size of routing annoucements exchanged between neighbors. This is because all routes in the distance vector table are included in each announcement. Although this is generally not an issue on local area networks, it can cause periods of increased utilization on lower-capacity WAN connections.