Packet Scheduling Algorithm Based On High Speed Downlink Packet Access

The ever increasing traffic generated by mobile internet browsing, file downloading, and wireless multimedia services have generated continuous demand for system capacity, communications quality, and coverage area for the existing cellular mobile communications system. The performance of the downlink channels has become the key aspect for the complete wireless communications system. To meet this need, 3GPP specifications introduce a new concept denoted high speed downlink packet access(HSDPA).Schedule policy based on data services plays a key role on the performance of the HSDPA. Considering the priority and delay requirement of service, packet scheduler concentrates on the trade-off between the cell throughput and user fairness. The main content of this paper is packet scheduling algorithm based on HSDPA technology. In packet switching network the essential assignment of packet scheduler is to determine the order and time slot of packets service and satisfy the QoS requirement.The research work in this paper is implemented through computer simulations. The whole paper is organized as follows. Firstly, the main techniques of HSDPA are introduced. Secondly, two different types of packet scheduling algorithms are proposed. One suits for Non Real Time traffic and the other for Real Time traffic. In this paper, we focus on M-LWDF which is an interesting algorithm for providing streaming services on HSDPA because it includes in the priority computation not only the instantaneous channel quality experienced by the user, but also the packet queuing delay. Finally, the scheme of simulation model is designed and simulation experiments are performed to obtain data results by OPNET. The performance and pros and cons of the M-LWDF and modified M-LWDF algorithms are analyzed and compared at last.Simulation results show that: i) the M-LWDF algorithm is a viable solution for scheduling streaming flows in HSDPA; ii) this algorithm is unfair because it provides a better QoS to users under favourable propagation conditions; and iii)a maximum Node B queuing of 2~3 seconds represents a good compromise between cell throughput and delay jitter introduced in the Node B for streanming...