自相似流量关键参数分析  

Analysis of Key Parameters of Self-similar Traffic

作　　者：谭献海[1] 黎燕敏[1] 潘启敬[1] 金炜东[1] 

机构地区：[1]西南交通大学信息科学与技术学院,成都610031

出　　处：《计算机科学》2008年第3期28-30,48,共4页Computer Science

基　　金：西南交通大学科学研究基金项目(2005A03);国家自然科学资金项目(No.90104002)

摘　　要：大量的研究结果表明,网络流量过程普遍存在着自相似和长相关特性,自相似和长相关特性对网络性能具有重要的影响。目前绝大部分研究都集中在Hurst系数的估计及其性能影响上,这是不全面的。本文深入研究影响网络性能的自相似流量关键参数,通过仿真分析Hurst系数和方差系数对网络性能的影响,表明Hurst系数和方差系数对网络性能均有重要的影响。分析了方差对网络性能影响的原因,研究了Cγ与方差之间的关系及其计算方法,给出了基于IDC的复合分形更新过程参数的估计算法,分析了分形开始时间对网络性能的影响。There is mounting experimental evidence that network traffic processes exhibit ubiquitous properties of self-similarity and long-range dependence （LRD）, and self-similarity and long-range dependence have great impact on network performances. However, most current researches on self-similar traffic mainly focus on the estimation of Hurst index and its impact on network performances, which is not overall. In this paper, the key parameters impacting the network performances of self-similar traffic are investigated. The impact of Hurst index and variance coefficient on network performance is studied by mean of simulation. Analytical results demonstrate that both Hurst index and variance coefficient have great impact on performances. The reason for the impact of variance on performances is analyzed. The relationship and its calculation of Cy and variance are studied. The estimation algorithm of parameters in Superposition of Fractal Renew Process （Sup_FRP） based on Index of Dispersion for Counts （IDC） is proposed. Finally, the impact of fractal onset time on performances is analyzed.

关 键 词：网络流量 自相似 分形布朗运动 复合分形更新过程 计数离差系数 

分 类 号：TP393[自动化与计算机技术—计算机应用技术] U463.6[自动化与计算机技术—计算机科学与技术]