An efficient reconfiguration scheme for fault-tolerant meshes

A new reconfiguration scheme, including a reconfiguration algorithm, is proposed in this paper to lift up the fault tolerance and system reconfiguration abilities for the mesh topology. The scheme adds redundancies-spare nodes and links-to the mesh network for necessary node replacement. By collocating a suitable number of spare nodes located at the best site of the network and joined by some well-connected spare links, our scheme is simple and yet effective in performing system reconfiguration. To carry out reconfiguration in a more regulated way, a reconfiguration algorithm is provided. The algorithm works dynamically and individually: system reconfiguration starts instantly upon the emergence of a fault and the replacement of a new faulty node-is considered independently from previous replacements. Experimental performance evaluation shows that, with significantly reduced complexity, the proposed reconfiguration scheme is able to achieve desirable reconfiguration rates for the mesh network. (C) 2004 Elsevier Inc. All rights reserved.