Uplink scheduling algorithms for the rtps traffic class for IEEE 802.16 networks

IEEE 802.16 MAC provides extensive bandwidth allocation and QoS mechanisms for various types of applications. However, the scheduling mechanisms for the uplink and downlink are unspecified by the IEEE 802.16 standard and are thus left open for vendors' own implementations. Ensuring QoS requirements at the MAC level for different users with different QoS requirements and traffic profiles is also another challenging problem in the area. The standard defines five different scheduling services one of them being the real-time Polling Service (rtPS). In this thesis, we propose an uplink scheduler to be implemented on the WiMAX Base Station (BS) for rtPS type connections. We propose that the base station maintains a leaky bucket for each rtPS connection to police and schedule rtPS traffic for uplink traffic management. There are two scheduling algorithms defined in this study: one is based on a simpler round robin scheme using leaky buckets for QoS management, whereas the other one uses again leaky buckets for QoS management but also a proportional fair scheme for potential throughput improvement in case of varying channel conditions. The proposed two schedulers are studied via simulations using MATLAB to demonstrate their performance in terms of throughput, fairness and delay. We show that the leaky bucket based scheduler ensures the QoS commitments of each user in terms of a minimum bandwidth guarantee whereas the proportional fair algorithm is shown to opportunistically take advantage of varying channel conditions.