| Summary: | 碩士 === 國立交通大學 === 電信工程系所 === 94 === The objective of this thesis is to propose a stall avoidance scheduler for an N-channel Stop ¡ And ¡ Wait (SAW) Hybrid Automatic Request or Retransmission Query
(HARQ) mechanism to improve the end-to-end throughput and reduce total system delay in the WCDMA systems with High Speed Downlink Packet Access (HSDPA). An increasing demand for both larger system capacity and higher data rates
has launched the evolution of the mobile communication market. HSDPA appears as an umbrella of features using key technologies such as Fast Link Adaptation, Fast Scheduling, Adaptive Modulation and Coding (AMC) and Fast Retransmission
Strategy (HARQ) whose combination improves the network apacity, increases the peak data rates up to 10 Mbps for downlink packet tra±c and decreases the system
throughput delay. HSDPA can allow users to enable high peak data rate services with a lower cost per delivered data bit, improve the QoS of already existing services.
In order to satisfy the growing aggressive demand for the wireless packet data services, scheduling algorithms are widely used in WCDMA wireless multimedia networks to provide the guaranteed quality of service. The design of such such algorithms is a challenge due to the mobility and connectivity problems that encountered in Wireless Networks. Among many protocol enhancements designed to maximize downlink throughput, the N-Channel SAW-HARQ is one of the required schemes that can boost the user throughput. The proposed Stall Avoidance (SA+QER) scheduler is to
be implemented in MAC layer and its role is to improve the end-to-end throughput and reduce the overall end-to-end delay in the presence of packet stall (or gap) in the
receiver reordering bu®er in the WCDMA systems with HSDPA.
The overall performance comparison of conventional cheduling policies with timer and window mechanisms is presented through simulations. Simulations results show that in terms of received goodput in upper layer and head-of-line delay, the proposed stall avoidance scheduler outperforms all the other existing schedulers, including maximum C/I, round robin, proportional fair, queue-based exponential rule. Furthermore, the overall system delay is signi‾cantly reduced by the stall avoidance scheduler.
|
|---|
