Optimizing IP Networks for Uncertain Demands Using Outbound Traffic Constraints

Conventional routing management approaches, which are based on given and often ”worst case” traffic matrices, are not sufficient anymore to deal with the increasing growth of the number of endpoints and diverse applications in IP networks. Traffic variations are becoming a more and more important issue, especially when making long-term network planning decisions as well as for medium-term network provisioning policies. In this paper, we address an offline metric-based traffic engineering (TE) problem in IP networks for uncertain demands, subject to several simple outbound traffic constraints. The resulting models belong to the polyhedral traffic model. They are intuitively tracktable and particularly appropriate for TE approach based on heuristic or metaheuristic frameworks. Impacts of the proposed uncertainty models on link utilization are compared. We also provide some computational results in terms of network utilization and other performance measures. keywords : routing, traffic engineering, IP networks, demand uncertainty