Abstract:

Wireless packet scheduling has been a popular paradigm to provide packet-level quality of services (QoS), in terms of throughput, delay and fair sharing, over error-prone channels. However, the state-of-the-art scheduling solutions are designed for single-rate environments. They cannot exploit the multi-rate capacility offered by the physical layer of the current WLANs. In this paper, we propose Adaptive Wireless Fair Scheduling (AWFS), which oppotunistically exploits the multirate feature and provides packet-level QoS in the presence of channel errors. AWFS departs from the throughput-oriented fairness and compensation, and adopts temporal fair sharing and fair compensation defined in virtual time. It offers a packet-level solution that ensures virtual temporal share, works with variable packet size and occaasionally idle flows, and operates in the infrastructure mode. AWFS improves overall channel utilization while nsuring fair sharing. It can operate with current 802.11b/e MAC design. through both analysis and simulations, we evaluate the effectiveness of AWFS. The simulations show that AWFS can achieve signficant throughput gain by improving overall throughput up to 159% compared with state-to-the-art scheduling algorithms in the simulated scenarios.