Segmentation of fractal network traffic with wavelets and log-likelihood statistics

Network traffic exhibits long-range dependence (LRD). The estimators of LRD parameters based on the discrete wavelet transform (DWT) are the best in terms of efficiency and accuracy. These estimators do not consider non-stationarity or changes in the LRD parameters. We propose using the Schwarz information criterion (SIC) to detect changes in the variance structure of the wavelet decomposition, and segmenting the traffic into pieces with homogeneous characteristics for the Hurst parameter. The procedure can be extended to the stationary wavelet transform (SWT), a non-orthogonal transform that provides higher accuracy in the estimation of the change points. The SIC analysis can be performed on-line (in the sense of progressively). The algorithm has been tested against synthetic and real traffic traces with promising preliminary results.