Stochastic processes for computer network traffic modeling

Computer networks such as local area and wide area networks possess complex characteristics due to the heterogeneous nature of the supported traffic. The network traffic exhibits highly irregular fractal-like structure and long term correlations. Various stochastic processes such as fractional Gaussian noise, multiplicative cascades, linear fractional stable motion have been proposed to model network traffic. These stochastic processes are relatively unheard of in the networking community, until recently. This paper provides a thorough review of these stochastic processes and their application to wireless traffic modeling.