Scheduling with QoS constraints over Rayleigh fading channels

We consider the problem of scheduling packets from multiple flows over a Rayleigh fading wireless channel. Recently, there has been much interest in opportunistic scheduling, i.e., scheduling packets from a user who has the highest SNR (signal-to-noise ratio), to maximize the network's throughput. In this paper, we compare the throughput achievable under fair opportunistic scheduling (i.e., a modification of opportunistic scheduling to ensure fair resource allocation) with the throughput under time-division multiplexing (TDM) scheduling. Using large deviations to characterize the probability that the QoS constraint (an upper bound on delay) is violated, we numerically compare the performance of the two scheduling algorithms under various channel conditions. We show that the opportunistic scheduler outperforms the TDM scheduler when the number of users is small but the TDM scheduler performs better when the number of users exceeds a threshold which depends on the channel parameters.