EIGRP uses DUAL to achieve route-loop freedom. By design DUAL ensure no route-loop in the network.
- Then why does EIGRP use poison reverse UPDATE when there is a topology change? Is there any other reason to send poison reverse other than route-loop avoidance?
- Furthermore, why do we use poison reverse instead of split horizon technique? ie not to send an UPDATE (poisoned) back on the interface on which it learnt about a destination.
split horizon
andpoison reverse
two different mechanism to avoid routing loops?Split horizon
says that an update should not be sent on the intf on which it learnt about a destination. Inpoison reverse
, we send an update on the same intf on which we learnt about the destination. Only that we send the update with infinite metric.