Ict network traffic based on a simple fractal structure

The results of measurements of ICT network traffic reveals the presence of fractal structures and long-range dependence. Accurate modeling of the offered traffic load is the first step in optimizing resource allocation algorithms and QoS requirements. In this article the fractal structure is represented by the simple random midpoint displacement method which approximates fractional Brownian motion (fBm), theoretical process that exhibits aforementioned features, is presented. Apparently, this algorithm generates inaccurate approximation of fBm for the higher degree of self-similarity. To improve this simple and fast algorithm, some modifications based on variance analysis are proposed. The results of simulation and statistical testing with comparison to the realtime measurements in computer network are presented and discussed.TRAFFIC INTENSITY IN ICT NETWORKS BASED ON A SIMPLE FRACTAL STRUCTURE
The results of traffic intensity measurements in ICT networks reveal the presence of fractal structures and long-term dependencies. Accurate modeling of the load is therefore the first step in optimizing network resource reservation and helps to meet QoS (Quality of Service) requirements. In this article, the fractal structure was obtained using the random midpoint displacement method, which provided an approximation of the fractional Brownian motion process, which constitutes the theoretical basis for the analysis of fractal structures. As it turns out, the presented method does not approximate this theoretical process very accurately, especially for higher degrees of self-similarity. To improve this simple and fast method, some modifications based on variance analysis were proposed. The results of simulations and statistical testing were presented and compared with real traffic recorded in an ICT network. (…)

Ict network traffic based on a simple fractal structure – PDF file for download