Simulation and Analysis of Self-Similar Network Traffic Using WBFGN

In this paper we implemented the method wavelet based fractional Gaussian noise WBFGN for the generation of self-similar network traffic, which is an increment process of fractional Brownian motion (Abry and Sellan, 1996). A large collection of self-similar sequences were generated and a big Bank of data SSTDB was created and uploaded for the reference of communication scientists and other IT experts. We concentrated over two aspects firstly how accurately a self-similar signal can be generated, secondly how quickly the method can simulate the network traffic signal. A comparative analysis was also done between different types of wavelet transforms for the different values of Hurst parameter. The statistical inaccuracy and time required to produce sequences of a given length are experimentally studied, analyzed and compared.