主动队列管理中增强型自适应虚拟队列管理算法之设计及稳定性分析

Design and stability analysis of enhanced adaptive virtual queue algorithm for active queue management

提出了一种基于速率的增强自适应虚拟队列管理算法(EAVQ).在该算法中引入主从拥塞尺度和期望链路利用比的概念;以输入速率为主要拥塞尺度,以便保留原有自适应虚拟队列管理算法(AVQ)中响应速度快、队列时延短、链路利用率高等优点.同时,以期望链路利用比为辅助拥塞准则,设计了一种基于速率的期望链路利用比自适应机制,解决了AVQ中参数设定困难、队列抗干扰能力弱,及存在链路损失等缺点;在改善系统动态性能的同时保证了链路容量的充分利用.在线性化基础上给出了一般网络结构下TCP/EAVQ系统的局部稳定条件.通过仿真验证了EAVQ的有效性.

A rate-based enhanced adaptive virtual queue（EAVQ） algorithm is proposed, in which the concepts of the principal and subordinate measures of congestion, as well as the desired link-utilization ratio are introduced. The input-rate of network link is considered the principal measure of congestion to hold the advantages of fast dynamic response, short queue delay time, and high link-utilization in adaptive virtual queue（AVQ）. Meanwhile, the desired link-utilization ratio is used as the subordinate measure, for which a rate-based adaptive system is designed to solve the AVQ-related problems, such as parameter adjusting, disturbance rejection, and link-capacity loss. The EAVQ algorithm improves the transient performance of the system and ensures the entire utilization of link capacity. Based on the linearization, the local stability conditions of the TCP/EAVQ system in the general network structure are also given. Simulation results show the excellent performances of EAVQ.