Wavelet-based modeling and smoothing for call admission control of VBR video traffic

Standard video compression techniques (e.g., MPEG) produce variable bit rate video stream for constant quality. The bursty nature of video traffic may cause significant inefficiency in bandwidth allocation. A number of works have been done to enhance bandwidth efficiency by respectively focusing on variable bit rate (VBR) traffic modeling [1] [2], smoothing [3] [4] [5] [6] and call admission control [7] [8] [9]. In this paper, we propose a unified modeling and smoothing scheme for call admission control based on the computationally efficient wavelet model in [1]. The wavelet-based model, which captures the multifractal nature of VBR traffic, not only tracks the autocorrelation of the real traffic well, but also gives an accurate description of the heavy tail distribution VBR traffic exhibits. Simulation results show that this method outperforms other existing ones in estimating necessary bandwidth for video traffic transmission. In addition, based on this tree-structured wavelet model, call admission control scheme can be implemented in conjunction with tree-structured wavelet-based traffic smoothing techniques (e.g., PCRTT [5] and WTS [6]). Consequently, additional bandwidth efficiency scan be gained via traffic smoothing.