Network Behavior Analysis Based on a Computer Network Model

This paper applies a new traffic model, iterated function systems (IFS) for network traffic modelling, to explore computer network behaviour and analyse network performance. IFS model can generate various self-similar data flows according to a certain character. In the light of the experiments, we have discovered interesting phenomena. It is found that increasing the routers’ processing capacity will have quite different influence on the power spectrum from that for enlarging the routers’ buffer size. Furthermore, phase transition and n f / 1 type power law is found under mixed traffic model. 1 IFS for Fractal Traffic Modeling of Computer Networks In recent years the fractal nature of Internet traffic has been observed , and many self-similar models for network traffic (such as on/off models, FARIMA models, etc.) have been presented so far . Here we introduce a new approach, IFS, to produce the fractal network traffic. Definition 1. The two dimensional affine transform 2 2 : R R → Ω is defined as follows.       +           =     Ω f e y x d c b a y x