DAT: An AP scheduler using dynamically adjusted time windows for crowded WLANs

This paper proposes a new packet scheduling algorithm for access points in a crowded 802.11 WLAN. Our goal is to improve the performance of efficiency (measured by packet response time or throughput) and fairness which often conflict with each other. Our solution aggregates both metrics and leverages the balance between them. The basic idea is to let the AP allocate different time windows for serving each client. According to the observed traffic, our algorithm dynamically shifts the weight between efficiency and fairness and strikes to improve the preferred metric without excessively degrading the other one. A valid queuing model is developed to evaluate the new algorithm´s performance. Using trace-driven simulations, we show that our algorithm successfully balances the trade off between the efficiency and the fairness in a busy WLAN.