mCoss: A multi-constraint scheduling strategy for WiMAX networks

Abstract

In this paper, we attempt to assemble the different pieces of the resource allocation puzzle of mobile WiMAX networks by addressing the main scheduling issues that are still open. We thus propose a novel multi-Constraints Scheduling Strategy (mCoSS) which maximizes the quality of service (QoS) degree of satisfaction for both realtime and non-real-time traffic in terms of delay and throughput. In the scheduling strategy presented in this paper, the access to the network is regulated via a traffic shaper which is inspired from the dual token bucket shaping mechanism. This technique allows traffic burstiness while bounding it. The modified dual token bucket mechanism is combined with a two-rounds scheduling algorithm reflecting the upper and lower bounds of service to be expected by each connection. The bandwidth request and grant policy adopted in these algorithms takes advantage of the different mechanisms proposed by the IEEE 802.16e standard. It adapts the choice of the appropriate technique to the service flow QoS constraints and the current availability of radio resources. Other concerns such as supporting the link adaptation capability and avoiding starvation of best effort traffic are also addressed in this solution. The performance of the proposed strategy is evaluated through simulation and compared to other scheduling solutions proposed in the literature. The obtained results show a nice tradeoff between fairness and efficiency with a high respect for the connections'QoS requirements.