论文部分内容阅读
基于在广域网（ＷＡＮ）、局域网（ＬＡＮ）、综合业务数字网（ＩＳＤＮ）中传输的业务以及可变速率（ＶＢＲ）业务等多种通信业务中普遍存在着自相似特性（长相关），叙述了自相似流量的数学定义和一些基本性质，描述了异步传输模式（ＡＴＭ）缓存的排队模型和一个具有自相似特性的业务量模型，给出了在自相似业务量下的ＡＴＭ 缓存溢出概率的下界，并与传统模型作了比较．比较结果表明，传统模型对ＡＴＭ 缓存性能的估计过于乐观．因此，在分析ＡＴＭ 网络性能的时候必需充分考虑业务量的自相似特性．
Self-similar features (long-term correlation) are commonly found in a variety of communication services based on services transmitted over wide area networks (WANs), local area networks (LANs), integrated services digital networks (ISDNs) and variable rate (VBR) services, The mathematical definition and some basic properties of self-similar traffic are described, and the queuing model of asynchronous cache (ATM) cache and a traffic model with self-similar characteristics are described. The lower bound of ATM buffer overflow probability under self-similar traffic is given , And compared with the traditional model. The comparison shows that the traditional model is too optimistic about the performance of ATM caching. Therefore, when analyzing the performance of ATM network, we must fully consider the self-similar characteristics of traffic.