自我類化網路訊務產生器之研究

Abstract

近年來，許多研究報告顯示電腦網路訊務系統具有長範圍相關的統計特性，因而較適合以自我類化程序來當作其模型，而傳統所使用的蒲松程序因為並不具有長範圍相關的統計特性，因此，並不再適合作為網路訊務系統的模型。當網路訊務在合成時，如果其長範圍相關的統計特性被忽略，將會導致不正確的網路系統性能評估。因此，產生具有長範圍相關統計特性的序列就變得很重要。 在這篇論文中，我們提出一個利用濾波器原理來合成自我類化網路訊務的方法。此一方法，改善了知名的傅立葉轉換和隨機中點置換這兩個方法的兩項缺點。首先，利用此濾波器方法所合成訊務序列的長度不再需要事前先指定，另外，利用此方法所合成的序列也將都是非負整數。雖然，此濾波器方法有一個缺點，即當合成的序列聚集到一個很大的數時，會喪失其自我類化的性質，但是這一個結果卻也正好符合真實網路所量測到的統計行為。

Recent empirical studies have shown that the modern computer network traffic is much more appropriately modeled by long range dependent self-similar processes than traditional short range dependent processes such as Poisson. Hence, if long range dependence is not considered for synthesizing experimental network traffic, it will lead to incorrect assessments of performance evaluation in network system. This arises the need of a well synthesizing trace with long range dependence. In this thesis, we present a filter-based method for synthesizing self-similar network traffic. This method improves the well-known methods of Paxson Fourier Transform and Random Midpoint Displacement in that the length of the synthesized traffic sequence does not need to pre-specify, and also the synthesized sequence is always non-negative. Although our method may have the drawback of becoming non-self-similar when the generated trace is aggregated under a very large window, this phenomenon turns out to match the measured behavior of true network traffic, where the self-similar nature only lasts beyond a practically manageable range, but disappears as the considered aggregated window is much further extended.