Multifractal Based Network Traffic Modeling

This text provides an overview of existing broadband traffic modelling based on the Poisson process and its variants like the MM1 models. It also provides coverage of models based on self-similar processes. Throughout the book, the authors have focused on the problem of broadband traffic modelling keeping in mind long range dependencies in broadband traffic. A new model is being proposed for modeling broadband traffic based on the theory of multiplicative cascade processes. This is referred to as the Variance Gaussian Multiplier (V.V.G.M) multiplicative cascade process. The proposed model incorporates variations of the multiplier distribution used to generate the process. Information theoretic principles are used to explain increased burstiness of network traffic. Kalman filter based techniques are developed for predicting traffic burstiness. The question of complexity of broadband traffic is also examined by defining a volatility parameter.