基于小波技术的网络流量特性刻画

CHARACTERIZING NETWORK TRAFFIC BASED ON THE WAVELET TECHNIQUE

网络流量特性是我们深刻理解整个网络体系的架构及网络控制机制和行为特征的切入点 ,所以当务之急是寻求一种全面切实的方法来刻画现代的网络流量特征 .本文以此为契机 ,通过对流量特性与尺度特性的研究 ,利用小波技术对网络流量特性如局域网的自相似特性及广域网的多分特性进行刻画 ,并对我们从国内收集的有代表性的网络流量进行分析 ,力求通过对真实的网络流量的小波变换系数 (小波系数 ,尺度系数 ) ,能量谱的分析 ,来进一步翔实 。

The understanding nature of network traffic is start point that we deeply research the network architecture, network control schemes, network behaviors. It is necessary to explore the efficient and comprehensive method to characterize the traffic. The paper just build on the turning point, and adopt the wavelet technique to capture the self similar of LAN and multifractal of WAN. The paper full and accurate characterize the feature of the network traffic through analyzing the scaling coefficient and wavelet coefficient and energy spectrum of the collected typical real trace by us.