基于小波的多重分形业务排队分析

Multifractal Queue Analysis Based on Wavelet

最近的研究表明 ,网络业务最为显著的统计特性是尺度特性 ,也就是分形行为 ,尤其是在小尺度下的多重分形特征 ,因而对网络业务分形特性下排队行为的研究 ,对网络容量规划和控制具有重要的意义 把小波分析方法引入网络业务的排队分析 ,由于小波固有的尺度特性 ,在小波域内能够更加精确简洁地进行排队尾分布分析 通过对真实网络业务的排队仿真 ,对小波方法和分析方法进行了比较 。

Recent research shows that one of the remarkable characteristics is the fractal nature of the investigated network traffic, especially the multifractal phenomena over small time scale Research on the queue analysis with multifractal traffic plays an important role in network dimensions and control Wavelet analysis is introduced into the analysis of network traffic queue With the scale nature of wavelet, it is more convenient and accurate to perform the tail asymptotics analysis in wavelet domain These results are validated by queueing simulation of measured network traffic with comparison of the analysis results