Abstract
This paper criticises the notion that long-range dependence is an important contributor to the queuing behaviour of real Internet traffic. The idea is questioned in two different ways. Firstly, a class of models used to simulate Internet traffic is shown to have important theoretical flaws. It is shown that this behaviour is inconsistent with the behaviour of real traffic traces. Secondly, the notion that long-range correlations significantly affects the queuing performance of traffic is investigated by destroying those correlations in real traffic traces (by reordering). It is shown that the longer ranges of correlations are not important except in one case with an extremely high load.
Summary
The extended journal version of the authors’ ICCCN paper challenging the idea that long-range dependence is an important driver of real internet traffic queuing behaviour, combining theoretical criticism of the traffic models used to generate it with an empirical test that destroys long-range correlations in real traces and finds queuing performance barely changes except under extremely high load.
bibtex
@article{clegg2011criticismslrd,
author = {Richard G. Clegg and Raul Landa and Miguel Rio},
title = {Criticisms of modelling packet traffic using long-range dependence (extended version)},
journal = {Journal of Computer and System Sciences},
year = {2011},
volume = {77},
number = {5},
pages = {861--868},
doi = {10.1016/j.jcss.2010.08.004}
}
Authors
Richard G. Clegg, Raul Landa, Miguel Rio
Venue
Journal of Computer and System Sciences, 77(5), pp. 861-868