Adaptive IEEE 802.11 DCF scheme with knowledge-based backoff

The IEEE 802.11 distributed coordination function (DCF) does not require the presence of a central controller (access point). It is based on carrier sense multiple access with collision avoidance and binary exponential backoff. Under saturation conditions, DCF operates very far from the theoretical limit. Enhancements to the DCF scheme to achieve the theoretical limit (Cali, F. et al., 1998; Qiang Ni et al., 2003) have drawbacks. We present an adaptive DCF (ADCF) scheme that adapts its backoff procedure based on such knowledge. Through simulation experiments, we compare the performance of ADCF with that of DCF and the slow contention window decrease (SD) scheme of Qiang Ni et al. Under saturation conditions, we observe that both DCF and SD operate far from the theoretical throughput limit due to their inability to negotiate contention for the medium quickly. By making use of the additional knowledge, the proposed ADCF scheme operates closer to the theoretical limit. For example, when DCF and SD operate 35% and 20%, respectively, below the theoretical throughput limit, ADCF operates only 10% below the theoretical limit. Moreover, by recording the channel state information observed during the recent past, ADCF is able to respond better to variations in the offered traffic.