Deployment of wireless local area networks (WLAN) has undergone dramatic growth in recent years. Multiple WLAN standards have been developed, including several versions of IEEE 802.11. The widespread deployment of IEEE 802.11 is based on the Distributed Coordination Function (DCF) component of the standard, which incorporates channel contention resolution using CSMA (Carrier Sense Multiple Access). However, an important shortcoming of past deployments of CSMA-based protocols has been their inability to achieve optimal performance. In recent years, researchers have developed new throughput-optimal CSMA algorithms. These results have opened up the possibility of designing CSMA-based protocols that can perform close to the optimal. However, the prior work on throughput-optimal CSMA makes several simplifying assumptions, including assumptions of perfect carrier sensing, symmetric interference relationship between links, and fixed rate transmissions.
The thrust of this project is on developing CSMA-based scheduling mechanisms that can work well in practical networks. Towards this goal, the project investigates systematic relaxation of unrealistic assumptions and constraints, and attempts to develop CSMA algorithms that can perform well under the relaxed conditions. The scope of the project includes implementation of selected protocols on an experimental testbed. The project explores CSMA in the context of wireless multi-hop mesh networks, as well as infrastructure-based wireless LANs. Expected results from the project include new theoretical algorithms as well as efficient CSMA-based protocols that can perform well in practice.