Search

Existing backoff scheme's optimization of IEEE 802.11 DCF MAC protocol consider only saturated networks or asymptotic conditions. In real situations, traffic is bursty or streamed at low rates so that stations do not operate usually in saturated regime. In this work, we propose and analyze a backoff enhancement for IEEE 802.11 DCF that requires information only about the network size and that is quasi-optimal under all traffic loads. We first analyze the performance of DCF multiple access scheme under general load conditions in single-hop configuration and we provide an accurate delay statistics model that consider the self-loop probability in every backoff state. We prove then the short-term unfairness of the binary exponential backoff used in IEEE 802.11 by defining channel capture probability as fairness metric. Motivated by the results on fairness, we introduce the constant-window backoff scheme and we compare its performance to IEEE 802.11 DCF with Binary exponential backoff. The quasi-optimality of the proposed scheme is proved analytically and numerical results show that it increases, both the throughput and fairness, of IEEE 802.11 DCF while remaining insensitive to traffic intensity. The analysis is then extended to consider the finite queuing capacity at nodes buffers using results from the delay analysis. NS2 simulations validate the obtained results.