Resiliency in Networks Through Meshed Trees

Network control protocols for switched networks utilize shortest path problem algorithms to build a shared shortest path tree to allow for traffic to be flooded. These algorithms were not built for the dynamic nature of computer networks and require additional latency and computation for protocols to reactively repair the tree in the event of a component failure. We argue that tree recovery downtime can be improved through an algorithm that proactively plans for any change in the network. Proactively is achieved through the construction of a Meshed Tree, a structure created for resiliency through the storage of a preferred forwarding path and backup paths in the event of a failure. Meshed Trees are built through a proposed Meshed Tree Algorithm (MTA), which can be applied to protocols. We compare MTA against implementations of the Rapid Spanning Tree Algorithm (RSTA) and a generic Link State Algorithm in failure recovery simulations. Results show that MTA better handles failures by reducing the amount of overhead and computation required to repair the tree.