Rate-adaptive weighted fair queueing for energy-aware scheduling

In this paper we study energy-aware scheduling that trades energy consumption against a traditional performance measure of delay. We use the power-rate function f(x)=c+xαf(x)=c+xα for x>0x>0 and f(0)=0f(0)=0 to model the power consumption, where c>0c>0 represents the base power. We give a definition of a rate-adaptive version of the Weighted Fair Queueing scheduling algorithm, and prove its energy consumption is within a bounded factor of the best possible when the algorithm guarantees the classic end-to-end delay for every connection.