数据业务自相似特性研究

Study on self-similarity of data service

合适的业务源模型能够在建立理论分析模型或进行系统性能仿真时,使网络系统性能评价更加准确。自相似特性是数据业务的重要特征。本文分析了自相似业务源的建模现状,深入讨论了如何将自相似模型应用于GPRS业务源建模。通过时间序列自相似特性分析工具SELFIS,对中国移动GPRS网络中从一个网关支持节点GGSN处采集的在单位时间内收到的字节数进行分析,得出GPRS数据业务具有较强自相似性的结论。

When building the model for theoretical analysis or having a simulation for system performance, a fit model for source traffic will make the performance evaluation more accurate. Self-similarity is one of the important characteristics of the data service. In this paper, the current status and recent development of self-similar traffic modeling is analyzed. And we focus on applying it to GPRS source traffic modeling. Through SELFIS, which is a software tool for self-similarity analysis of time series, the number of received bytes every time unit at one of GGSN selected from China Mobile GPRS network is analyzed. It is concluded that the self-similarity of GPRS traffic is strong.