论文部分内容阅读
文中介绍了９０年代高速网络性能评价领域中一个重大发现，即真实的网络业务具有自相似性．传统的基于Ｍａｒｋｏｖ模型的性能评价结果对自相似业务已不再适用，需要研制新的模型与工具．文中介绍了自相似过程的数学定义及其性质，讨论了重尾分布和诺亚效应及它们与自相似过程的关系，还介绍了几种生成自相似业务的方法和自相似业务模型下队列系统性能评价方面的研究成果．
This paper presents a major discovery in the field of high-speed network performance evaluation in 1990s, that is, the real network business has self-similarity. The traditional performance evaluation results based on Markov model are no longer applicable to self-similar services, and new models and tools need to be developed. This paper introduces the mathematical definition of self-similar process and its properties, discusses the heavy-tailed distribution and Noah’s effect and their relationship with self-similar process, and also introduces several methods to generate self-similar business and self-similar business model queue system Performance evaluation of the research results.