摘要
REM算法是一种典型的AQM算法。本文利用NS2网络仿真软件研究了REM的网络性能。仿真实验表明,拥塞度量与静态网络性能解耦,各种网络环境中的队列长度均能稳定至目标值处;而拥塞加剧时,队列收敛性变差,暂态网络性能降低。调节算法参数γ和Φ能有效地减小队列长度的过渡时间以改善网络性能,增强算法的环境适应性。
REM is a typical AQM algorithm. The performance of REM algorithm is investigated based on NS2 simulation platform. The main results of the simulation include： Congestion and performance measures are decoupled. REM queue length can achieve its desired value in the great changing network. But when the congestion deteriorates,the transient performance gets worsen,such as slower convergence speed. The transient performance can be improved effectively by adjusting the parameters of γ and Φ .
出处
《微计算机信息》
北大核心
2008年第36期99-101,共3页
Control & Automation
关键词
网络拥塞
主动队列管理
随机指数标记
网络性能
Network Congestion
Active Queue Management
Random Exponential Marking
Network Performance
