PWFQ:一种基于优先级共享带宽的加权公平排队算法  

PWFQ: A Priority-Based Weighted Fair Queueing Algorithm

作　　者：许孙娟[1] 叶家骏[1] 邹君妮[1] 

机构地区：[1]上海大学通信与信息工程学院,上海200072

出　　处：《上海大学学报（自然科学版）》2006年第1期1-4,共4页Journal of Shanghai University:Natural Science Edition

基　　金：上海市重点学科建设资助项目(T0102);上海市科委科学技术发展基金资助项目(04dz12045)

摘　　要：对WFQ算法无法保证实时业务QoS的原因进行了深入分析,指出了WFQ算法的公平调度原则存在的不足,提出了一种基于优先级分配带宽的PWFQ算法.这种算法根据业务的延时性能对带宽的敏感程度划分业务的优先级,在链路带宽不足的条件下,能够有效保证高优先级业务的QoS要求.仿真结果显示,PWFQ算法不仅改善了实时业务的延时和延时抖动性能,并且能够将它们的最大延时控制在允许的范围之内.Having analyzed the reason why WFQ algorithm cannot ensure the quality of real-time service, the paper points out the shortcomings existing in the fair queueing principle of WFQ algorithm and proposes the PWFQ algorithm which distributes bandwidth based on priority. Under the condition that the link bandwidth is inadequate, the proposed algorithm can guarantee the quality of service with high priority by assigning priority based on the degree of sensitivity of traffic delay to the bandwidth. Simulation results show that the PWFQ algorithm not only can improve delay performance and delay jitter of real-time traffic, but can also meet the worst-case delay bound requirements.

关 键 词：调度算法 公平性 优先级 延时 

分 类 号：TN911.254[电子电信—通信与信息系统]