Self-similar traffic generation using a fractional ARIMA model. Application to the VBR MPEG video traffic

Abstract

Recent studies and traffic measurements have shown that variable bit rate (VBR) video traffic, as well as local area network (LAN) traffic, exhibits long-range dependence (LRD). When this kind of traffic is transmitted over ATM networks, the LRD property implies a larger requirement in buffer allocation and channel capacity. A new fractional ARIMA model for VBR MPEG video traffic at frame level is presented. The fractional integrative component of the model is based on a process called "fractional differencing". The model is obtained from three very long sequences. It is used to synthesize artificial traffic, which presents the desired LRD.