Robust routing mechanisms for intradomain traffic engineering in dynamic networks

Abstract

Internet traffic is highly dynamic and difficult to predict in current network scenarios. This makes of Traffic Engineering (TE) a very challenging task for network management and resources optimization. We study the problem of Intradomain Routing Optimization under this traffic uncertainty. Recent works have proposed robust optimization techniques to tackle the problem, conceiving the Robust Routing (RR) approach. RR copes with traffic uncertainty in an off-line preemptive fashion, computing a single static routing configuration that is optimized for traffic variations within some predefined uncertainty set. Despite achieving routing reliability with relatively low performance loss, RR presents various drawbacks and conception problems as it is currently proposed. This paper brings insight into the different Robust Routing shortcomings, introducing new mechanisms that improve previous proposals and alleviate these problems. Among others, we propose and evaluate new optimization objectives to attain better global performance from an end-to-end quality of service perspective.