Self-similarity and Long Range Dependence in teletraffic

Abstract

Self-similarity plays an important role, at least over a finite number of scales, in natural phenomena as well as complex technology-related systems. One of the most surprising examples is provided by telecommunication networks, where the shift from circuit-switching to packet-switching has led to a deep change in the stochastic nature of traffic flows. The concepts of self-similarity, long range dependence and heavy tails, widely used in traffic modelling, are closely related among them and strongly influence network performance. Hence, this overview, based on the research and teaching experience of the author, presents in a simple way the main definitions, focusing on their physical interpretation, and highlights the relevance of these properties in network dimensioning.