有目标的随机早期检测算法  

Targeted Random Early Detection Algorithm

作　　者：马利[1] 陈杰[1] 张波[1] 

机构地区：[1]南京信息工程大学计算机与软件学院

出　　处：《武汉理工大学学报》2010年第16期152-154,158,共4页Journal of Wuhan University of Technology

基　　金：公益性行业科研专项(GYHY(QX)2007-6-14)

摘　　要：提出了一种基于随机早期检测算法RED的改进算法Targeted RED,该算法能够在不同的网络流量下保持路由器队列的稳定性:不管是在低数据流量还是在高数据流量的情况下,队列长度都保持在一定的范围内。进而保证了网络时延的稳定,提高了网络的服务质量。TRED算法同ARED算法的目标是相同的,但是ARED引入了更多的参数,带来了更多的参数设置问题,而TRED却未引入更多的参数,这是TRED优于ARED的地方。使用网络模拟软件NS-2进行算法性能验证,实验效果良好。This paper propose an improved random early detection algorithm which is based on RED algorithm： Targeted RED（TRED）.The algorithm can maintain the stability of the router queue in different network traffic.No matter in low data traffic or in the case of high data traffic,the queue length is maintained at a certain range.TRED ensures the stability of the network latency and improve the network quality of service.TRED algorithm has the same objective of ARED algorithm.But ARED uses more parameters which bring more parameter setting problem.TRED do not have the same problem.So TRED advanced than ARED in some aspects.Related experiments＇ effect is good by using the network simulator NS-2 to validate algorithm performance.

关 键 词：RED ARED TargetedRED NS-2 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]