Enhanced Algorithm for Logical Topology-Based Fault Link Recovery in Crossbar Networks

Abstract

Conventionally, virtual topology nodes are mapped with actual physical topology nodes and links mapped with lightpaths (LPs). But this is computationally unwieldy becoming huge and more complex with increasing network size. This work proposes a scheme for optimal utilization of the network, minimizing network congestion by allocating efficient LPs in existing physical systems. When network size increases, the degree of the sub-trees also increases but without any increase of carrier wavelengths; the number of hops (and carrier wavelengths) decreases as it generates direct paths between the longest path nodes. Structural metrics of the physical and logical topologies of this method are more efficient. It supports multicast connections and minimizes data replication. The outer iteration of the proposed algorithm is executed twice; the first iteration yields the network’s logical topology, while the second gives “complete bipartite graph.” Using the logical topology, the path length of the longest distance node is reduced, and direct connections between nodes are created. The reconfiguration is used only for the nearest node of logical topology, and this reduces the number of hops between the nearest nodes to three. The maximum path length is 6 for any n × n crossbar network. This will cause reduction in the number of protection cycle.