Power spectral analysis-based QoS evaluation of VBR video and its application to pricing model

Abstract

Since so many potentials of QoS degradation, it is difficult to predict the QoS degradation of VBR MPEG video application under high-speed network environment. The general procedure of the call admission requires the QoS contract between the users and the network. But, various statistical properties of the source traffic and its hard estimation make it difficult to satisfy the QoS contract. In this paper, we propose the measurement-based cell-rate autocorrelation function as the criterion of estimation on QoS degradation. The analysis of the power spectrum of MPEG video with P-MMBBP traffic modeling is ventured. Simulation results show that QoS degradation of the input traffic, particularly delay is closely related to the power spectrum. Furthermore, a new method of the pricing scheme based on the power spectral analysis is proposed and analyzed.