Skip to Main Content
We propose a novel class of opportunistic scheduling disciplines to handle mixes of real-time and best-effort traffic at a wireless access point. The objective is to support probabilistic service rate guarantees to real-time sessions while still achieving opportunistic throughput gains across users and traffic types. We are able to show a ldquotightrdquo stochastic lower bound on the service a real-time session would receive assuming that the users possibly heterogeneous capacity variations are known or estimated, and are fast fading across slots. Such bounds are critical to enabling predictable quality of service and thus the development of complementary resource management and admission control strategies. Idealized simulation results show that the scheme can achieve 80%-90% of the maximum system throughput capacity while satisfying the quality of service (QoS) requirements for real-time traffic, and that the degradation in system throughput is slow in the number of real-time users, i.e., inter- and intra-class opportunism are being properly exploited. We note however, that there is a tradeoff between strictness of QoS requirements and the overall system throughput one can achieve. Thus if QoS requirements on real-time traffic are very tight, one would need to simply give priority to real-time traffic, and in the process lose the throughput gains of opportunism.