Performance Evaluation of a Distributed CSMA Algorithm for Throughput and Utility Maximization Internal Report- Work in progress

Maximizing network throughput and utility are two contrast-ing goals difficult to achieve simultaneously. This problem is particularly challenging in the context of wireless networks, where the communication channel is shared by multiple nodes, and an efficient mechanism must be used in order to ensure fair channel access and good performance. In this project, we evaluate the performance of a novel CSMA-based algorithm for throughput and utility maximization, leveraging backpressure computations to coordinate the access of multiple nodes to the medium. We base our evaluation approach on simulations considering a number of interesting topologies, and identifying factors that negatively affect the performance of the algorithm. We also study the more extreme case of large networks with dynamic traffic patterns, evaluating the performance of the algorithm in conditions of higher stress. The obtained results show that the algorithm is vulnerable to adverse topological features such as information asymmetry, for which specific nodes receive much of the throughput share while others starve. In other interesting cases, instead, we find that the algorithm tends to achieve a higher network utility than IEEE 802.11, at the cost of reducing the total network throughput. I