Impact of Partial Demand Increase on the Performance of IP Networks and Re-optimization Approaches.
Eueung Mulyana, Ulrich Killat · 2004
An important aspect in management of IP networks is to control traffic traversing links on the network, while optimizing performance. Using offline approaches, this means that recomputation should be carried out on a periodical basis to adjust network configuration to the most actual traffic situation obtained from measurements or forecasts. In this work we investigate the effect of partial demand increase on the performance of the network and propose a simple policy scheme to decide whether re-optimization should be performed. Two re-optimization approaches based on plain localsearch and simulated-annealing are presented. We apply our method for metric based traffic engineering scheme to the German scientific network (G-WiN) for which a traffic matrix and several traffic-increase patterns were randomly generated. Several computational results are provided.