Cluster router: Homogeneous OSPF for distributed control plane

The exponential growth of Internet traffic requires that routers should be scalable. Cluster is an easy and flexible scalable way to improve the performance of the router. A critical issue in the design of cluster router is to provide a global and consistent routing table in such distributed architectures. In order to shorten the time of OSPF routing table computation, accelerate the network convergence, improve the reliability of the router and avoid the overload of the single control element CPU, distributed homogeneous OSPF is proposed in this paper. Each routing node of cluster router runs the OSPF protocol process on its own control element respectively, and keeps up the identical link-state database (LSDB). The shortest expected delay (SED) task scheduling algorithm is designed for the cluster router to compute the routing table on demanding. Compared with some improved OSPF schemes, our distributed homogenous OSPF can calculate routing table in real time and speed up LSAs processing in parallel. It decreases the internal communication overhead and provides a redundant backup scheme among these routing nodes to improve the reliability of cluster routers.