摘要
自相似性是网络通信量的普遍性质并且对网络性能有很大的影响。论文利用蒙特卡罗方法研究了自相似排队系统的性能问题。研究表明长程相关和短程相关对于排队系统性能具有非常不同的影响,尤其是在缓存较大的情况下。同时,还发现通信量中长程相关发生作用的尺度与通信量以及排队系统本身的参数都有关,这对于实际的网络设计具有较强的参考意义。
Self-similarity is a ubiquitous character in traffic and has a great impact on network performance.In this pa-per,the performance of self-similar queueing system is studied with Monte-Carlo method.It is showed that long-range dependence has a quite different impact on queueing performance from that of short-rang dependence,especially on the condition of large buffer size.The onset time scale of long-range dependence is related to the parameters of traffic and queueing system,this is a good reference for the design of network in practice.
出处
《计算机工程与应用》
CSCD
北大核心
2004年第6期30-33,共4页
Computer Engineering and Applications
基金
国家863高技术研究发展计划资助(编号:2002AA303260)
关键词
蒙特卡罗
排队长度丢包率
长程相关
Monte-Carlo,Queueing length,Loss rate,Long-range dependence
