Scheduling aware network flow models for multi-hop wireless networks

Abstract

Network flow models have proven to be an effective tool in the analysis, optimization and design of network protocols. Critical to the success of these models in general multi-hop wireless networks (MHWNs) is an accurate estimation of the effect of CSMA scheduling. While the existing models capture coarse grained estimates of interference, they do not account for the substantial impact of MAC scheduling. On the other hand, accurate models of throughput in CSMA networks exist. However, they are unsuitable for use as a part of a network flow formulation because of their complexity and some of their underlying assumptions. This paper contributes an efficient and constructive model to estimate the effect of scheduling on interfering links in general MHWN settings. We integrate this approach with a network flow routing model which works with aggregate estimates of capacity to improve the quality of the solution. Simulation results show that accounting for scheduling effects leads to large improvements in the quality of the solution.