What is Poison Reverse?

Understanding Poison Reverse

Poison reverse is a loop-prevention technique in distance-vector routing protocols where a router advertises a route back to the neighbor it learned that route from, but with an infinite metric, explicitly telling that neighbor the route is unreachable through this path. It strengthens the simpler split-horizon rule to keep routers from forming routing loops.

In distance-vector protocols like RIP, routers share their full routing tables with neighbors. Plain split horizon prevents loops by simply not advertising a route back toward its source. Poison reverse goes further: it does advertise the route back, but poisons it with an unreachable metric (in RIP, a hop count of 16, which RIP treats as infinity). This active "do not use this route through me" signal removes ambiguity and, combined with route poisoning and triggered updates, helps the network agree quickly that a failed route is dead rather than slowly counting to infinity.

This matters for network availability and stability. Routing loops cause packets to circulate until their TTL expires, wasting bandwidth, creating black holes, and degrading or denying service, a reliability problem and a potential avenue for disruption. Poison reverse speeds the convergence on a consistent, loop-free view of the topology after a link fails, reducing the unstable window during which traffic can loop or be misrouted.

For example, suppose Router A reaches network X through Router B. With poison reverse, when B advertises its routing table back to A, it lists network X with a metric of 16 (infinity), clearly informing A not to route to X via B. If A's direct path to X then fails, A will not be fooled into thinking B offers an alternate route to X, preventing the two routers from bouncing packets back and forth in a loop while the protocol converges.