A fast rerouting scheme for OSPF/IS-IS networks

Most current backbone networks use link-state protocol, OSPF or IS-IS, as their intra-domain routing protocol. Link-state protocols perform global routing table update to route around the failures. It usually takes seconds. As real-time applications like VoIP emerge in recent years, there is a requirement for a fast rerouting mechanism to route around failures before all routers on the network update their routing tables. In addition, fast rerouting is more appropriate than global routing table update when failures are transient. We propose such a fast rerouting extension for link-state protocols. In our approach, when a link fails, the affected traffic is rerouted along a pre-computed rerouting path. In case rerouting cannot be done locally, the local router signals the minimal number of upstream routers to setup the rerouting path for rerouting. We propose algorithms that simplify the rerouting operation and the rerouting path setup. With a simple extension to the current link state protocols, our scheme can route around failures faster and involves minimal number of routers for rerouting.