Abstract
One of the key challenges of next-generation wireless networks (such as enhanced general packet radio service) is the ability to provide different grades of quality of service to users. Weighted fair queueing (WFQ) is a popular scheduling algorithm to guarantee bounded delay, guaranteed throughput, and fairness. However, adapting the WFQ scheduler to wireless networks is not trivial due to location and weather-dependent variable data rates. In this paper, optimal schemes that are based on combining the WFQ with compensation techniques are proposed, which swap channel accesses among traffic flows that experience variable data rates, so that the variable data rates are transparent to the flows. It is also proposed here three periodic compensation methods and two event-driven methods. Two of the periodic methods provide maximum achievable throughput and maximum fairness. The two event-driven methods are completely distributed and are executed much less frequently than the periodic methods. © 2007 IEEE.