A tabu-search approach for designing computer-network topologies with unreliable components

Abstract

The topological design of computer networks is a part of network planning, and consists of finding a network topology configuration that minimizes the total communication cost, while considering some performance and reliability constraints. Given the computational complexity of techniques that allow an optimal solution to this problem, heuristic methods are often proposed to reduce the search of candidate topologies and provide suboptimal solutions. The tabu search (TS) approach in this paper is one such method. Our adaptation of TS to the topological design of computer networks applies some moves to a starting topology in order to reduce its total link-cost and/or to improve its average delay. Such an approach is an attempt to generalize a local-search method. Simulation results confirm the efficiency and robustness of this method for designing backbone networks interconnecting between 12 and 30 nodes. Our implementation of TS generally provides better solutions than CS, SA, and GA. Our implementation of TS uses a routing strategy where packets travel according to the shortest distances between nodes. This strategy is easy to implement and allows the direct evaluation of flow and delay, but has some shortcomings. The routes do not vary according to the charge of the network. In practice, it is suitable to use the dynamic routing strategy where routes vary according to the traffic supported by the network. The length of the tabu list can noticeably affect the final results, therefore the use of dynamic list is entirely suitable.