A conceptually simple framework for simulating hierarchical MPEG video traffic

This paper presents a methodology to synthesize MPEG-4 video traffic traces with hierarchical levels of compression. In real-time video streaming applications, scalable video transmission is employed to adjust quality of service (QoS) requirements to the available channel bandwidth. The proposed model is a very simple tool to simulate scalable video traffic transmission, in order to evaluate the performance of computer networks under different traffic loads and transport protocols. The MPEG-4 streaming is classically modeled as a self-similar random process, due to the long range dependence of the autocorrelation function. The presented model is based on statistical analysis collected from real video traffic data for high, medium and low resolutions. The statistical properties for the synthesized traffic are generated using finite impulse response (FIR) digital filters. The proposed simulation method approximates the autocorrelation function of the real MPEG-4 traffic, while presenting low computational complexity.