An extended discrete autoregressive model for variable bit rate video traffic encoded by scalable video codec

Designing an effective and high performance network requires accurate characterization and modeling of network traffic. This paper provides a study of the transmission, modeling, and analysis of variable bit rate (VBR) video, which is fundamental for designing protocols and for effective utilization of networks in video transmission. To meet the specified requirements of future networks, the scalable video codec (SVC) was chosen as the video compression standard. The main objective of this work was to propose a statistical model that will allow better coordination between generated SVC video traffic and original video traffic. The performance of the proposed model was evaluated by comparing the model with three existing low-complex statistical models using graphical and statistical measurements as well as cross-validation. For all evaluation techniques, the accuracy of the proposed model was evaluated, and the frame size distribution of the modeled traffic was found to match that of the original video traffic better than other existing models. Copyright (c) 2015 John Wiley & Sons, Ltd.