Scalable Scheduling with Burst Mapping in IEEE 802.16e (Mobile) WiMAX Networks

            ne olumn tripping with non-increasing rea first mapping algorithm (OCSA) was proposed by Chakchai So-In et al to schedule bursts on downlinks of base station by giving priority to the largest bursts to OFDMA frame. However, size-based scheduling that favor large items are known to exhibit poor average delay performance especially under workload distributions that are highly skewed (i.e., heavy tailed workloads). In this paper, we first study OCSA and use numerical results to show that it starves short bursts at high loads. We then propose improvement to OCSA (iOCSA) and a new algorithm called ne olumn tripping with ncreasing rea first mapping algorithm (OCSIA). In contrast to OCSA, OCSIA gives priority to short bursts. Our detailed numerical results to compare OCSIA to OCSA under varying workload distributions clearly show iOCSA improves the performance of OCSA, and OCSIA significantly outperforms OCSA under heavy tailed workloads without starving large bursts.