摘要
近期真实网络流量分析表明很多信息源会产生在多时间尺度下具有自相似特性的信息流，据此人们采用模拟及分析的方法对此种业务流对 Ａ Ｔ Ｍ 交换机／复用器性能的影响进行了研究，发现了与传统短相关模型完全不同的排队特性。但是到目前为止对于自相关信息流的复用问题尚未发现有人进行研究。而这个问题的解决对自相似业务流条件下的 Ａ Ｔ Ｍ 交换机／复用器性能分析、接入控制（ Ｃ Ａ Ｃ）等都具有很重要的意义。本文基于临界时间尺度的概念对这个问题的解决进行了有益的探索，得出了较好的结论，并采用模拟的方法对得出的结论进行了验证，模拟结果表明估计的参数较准确地描述了自相似业务合成流的特征。
Recent traffic analyses of a lot of networks have shown that many traffic sources produce traffic streams that are self similar over several time scales,and people have studied the influences of self similar traffic streams on ATM switch/multiplexer using simulation and analysis methods,and found that the queuing properties of self similar traffic streams are different from those of traffic streams with short range dependence.But there is no multiplexing analysis of self similar traffic streams until now.The solution of this problem is helpful to the performance analysis of ATM switch/multiplexer and call access control(CAC).In this paper,we explore this problem using critical time Scale and obtain better conclusion,in addition we justified the conclusion using simulation method,It has shown that the estimated parameters correctly describe characteristic of self similar multiplexing traffic stream.
出处
《通信学报》
EI
CSCD
北大核心
1999年第8期7-12,共6页
Journal on Communications
