Minimal feedback optimal algorithms for traffic engineering in computer networks

This paper addresses the problem of traffic engineering in computer networks. More precisely, optimal data rate adaptation laws are provided, which are applicable to networks where both multiple paths are available between any pair of source/destination nodes and multiple classes of service are to be provided. In particular, it is shown that the algorithms presented only need a minimal amount of information to achieve the optimal operating point. More precisely, they only require knowledge of whether a path is congested or not. Hence, the control laws provided in this paper require much less feedback than currently available ones. The proposed approach is applicable to utility functions of a very general form and endows the network with the very important property of robustness with respect to node/link failures; i.e., upon the occurrence of such failure, the presented control laws reroute traffic away from the inoperative node/link and converge to the optimal allocation for the "reduced" network.