重尾ON/OFF源模型生成自相似业务流研究  

Study of the Self-similar Traffic Stream Generated by Heavy-tailed ON/OFF Source Model

作　　者：胡严[1] 张光昭[1] 

机构地区：[1]中山大学电子与通信工程系,广东广州510275

出　　处：《电路与系统学报》2001年第3期72-76,共5页Journal of Circuits and Systems

基　　金：广东省自然科学基金资助项目

摘　　要：因为传统模型没有考虑网络业务流量各种时间尺度都具有突发性,因此是不完善的.一系列的测量结果表明,网络业务流量显示自相似性.重尾分布ON/OFF模型能比较好地解释自相似业务流的产生原因,可以把LAN上的业务分解为多个活动(active)主机对之间的业务流.本文通过系统的仿真实验研究了重尾ON/OFF源模型生成自相似流的机理,并与理论结果进行了对比,而且补充了理论结果.It is widely accepted that conventional models do not adequately model network traffic since they do not take into account its inherently bursty nature. A series of measurements have showed that network traffic is self-similar. The heavy-tailed ON/OFF Source model can make clear the reason why the network traffic is self-similar. It is possible to separate the aggregate LAN traffic into the individual components representing traffic flows between each active pairs of host computers. In this paper, we study mechanism of heavy-tailed ON/OFF source model producing self-similar traffic stream is studied by systematic simulations, and the simulation results are compared with theoretical results as to complement the latter.

关 键 词：自相似业务流 ON/OFF源模型 重尾分布 电信 

分 类 号：TN91[电子电信—通信与信息系统]