Multicommodity flow optimization in support for packet-switched network traffic engineering

Abstract

Many contemporary IP routing protocols are based on algorithms that are non-traffic-aware. The chief problem is that very often many comparable traffic flows are misled to converge on bottleneck links since these algorithms are oblivious to traffic characteristics and link loading. One possible remedy to this is to reconstitute them with a better algorithm. We recommend the maximum concurrent multicommodity flow (MCMF) algorithm. It considers traffic demands, network topology and resources in its computation to generate an optimal solution that satisfies all traffic demands as much and as fairly as possible while not overusing the resources. It does this by spreading traffic load across all feasible links in the network. In simpler terms it means placing traffic where the capacity exists. We develop a basic traffic engineering framework that shows how the MCMF algorithm can be applied. There exists numerous algorithms for solving this kind of problem. But this paper focuses particularly on a fully polynomial time approximation scheme (FPTAS), namely, the Garg-Konemann (1997) algorithm. However, its convergence rate is inconsistent due to its dependency on initial guess of a crucial parameter. We extend this algorithm in order to eliminate this issue and include empirical results that demonstrate its superiority over the original version. We conclude our paper with our opinion whether if the algorithm is fast enough to qualify as an on-line IP routing engine.