Channel-and delay - aware scheduling algorithms for wimax

IEEE 802.16 introduces scheduling algorithm to manage and allocate network resource in order to achieve Quality of Service (QoS) of different services. Scheduling algorithms can be categorized into two main types: channel-unaware and channel-aware. Channel-unaware schedulers assume the wireless channel condition is always perfect and error-free. This assumption is impractical due to the nature of the wireless medium. Meanwhile channel-aware schedulers exploit the variation of the wireless channel by taking the wireless link quality into scheduler's decision making mechanism. Thus it can handle the variation of wireless medium in the better manner to achieve optimum network performance. In this thesis, the proposed channel-aware Weighted Fair Queuing (CWFQ) and Queue Duration Awareness (QDA) algorithm have been developed and investigated. CWFQ improves the goodput performance of WFQ through exploitation of wireless medium to achieve better spectral efficiency. Packet with better link quality is given higher service priority. Meanwhile QDA algorithmimproves the delay performance of the CWFQ scheduling algorithm by giving higher service priority to packet that has longer queue duration. This helps to reduce packet waiting time for service. Extensive simulations have been done using QualNet simulator to investigate the performance of CWFQ and QDA algorithms. The results show that CWFQ has achieved better goodput and packet delivery ratio (PDR) for real time polling service (rtPS) flow compare to WFQ, with improvement of 19.4% and 21.5% in static scenario, 12.4% and 13.5% in mobile scenario. This shows that CWFQ is able to exploit wireless channel and improve spectral utilization. For QDA algorithm, the results show that CWFQ+QDA has outperformed the existing Garroppo-WFQ (GWFQ) by 36.3% and 79.4% in static and mobile scenario. This is because QDA algorithm awards higher priority to packet that has long queued to avoid starvation of service. This shows that QDA is able to improve CWFQ's delay performance.

Text FK 2013 9 ir.pdf Download (7MB)

Actions (login required)

View Item