Component failures in high-speed computer networks can result in significant topological changes. In such cases, a network reconfiguration algorithm must be executed to restore the connectivity between the network nodes. Most contemporary networks use either static reconfiguration algorithms or stop the user traffic in order to prevent cyclic dependencies in the routing tables. The goal is to present NetRec, a dynamic network reconfiguration algorithm for tolerating multiple node and link failures in high-speed networks with arbitrary topology. The algorithm updates the routing tables asynchronously and does not require any global knowledge about the network topology. Certain phases of NetRec are executed in parallel, which reduces the reconfiguration time. The algorithm suspends the application traffic in small regions of the network only while the routing tables are being updated. The message complexity of NetRec is analyzed and the termination, liveness, and safety of the proposed algorithm are proven. Additionally, results from validation of the algorithm in a distributed network-validation testbed Distant, based on the MPI 1.2 features for building arbitrary virtual topologies, are presented.

Summary form only given. Component failures in high-speed local and system area networks can result in significant topological changes. In such cases, a network reconfiguration algorithm is executed to restore the connectivity. Most of the current networks use either static reconfiguration algorithms or stop the user traffic to prevent cyclic dependencies in the routing tables. The goal is to validate an extension of the dynamic reconfiguration algorithm NetRec, which was previously published by the authors. The extensions is designed to increase the network availability in the presence of multiple link and node failures. It updates the routing tables asynchronously and doesn't require any global knowledge about the network topology. Certain phases of NetRec are executed in parallel, thus reducing the reconfiguration time. We present results from validation of the algorithm in a distributed network testbed, based on the MPI 1.2 features for building virtual topologies.

High-speed local and system area networks may change their topology due to switching on/off of routers and hosts or due to component failures. In such cases, a reconfiguration algorithm must be executed to restore the network connectivity and thus achieve high system reliability. However, most of the solutions are based either on redundant network paths or on regular network topologies.The purpose of this paper is to specify NetRec, a novel algorithm for dynamically reconfiguring an arbitrary network topology when a permanent node fault occurs. Unlike other reconfiguration algorithms, NetRec is applicable for all high-speed computer networks and is compatible with all modern routing techniques, including wormhole-based system area networks. It restores the network connectivity by building a tree that spans all immediate neighbors of the faulty node that are still connected to the network. The algorithm is distributed and does not require any global knowledge.

Focuses on the utilization of alternative communication paths in local and system area networks with static routing. A lot of research work has been devoted to employing such paths for fault tolerance, but the issue of utilizing them for performance enhancement has been largely neglected, especially for static routing networks. This work formally proves that the throughput of multiple paths is maximal if the traffic is uniformly distributed over them. Based on this, a procedure for destination partitioning in static routing networks is introduced. It is applicable to arbitrary multi-path topologies and traffic patterns that lend themselves to partitioning. The procedure is applied to several topologies with different degree of equivalent paths coverage and their performance is evaluated through simulations. The results demonstrate that the network performance is significantly improved when the proposed partitioning procedure is applied.