A Simple, Two-Level Markovian Traffic Model for IPTV Video Sources

To facilitate network performance analysis and simulations for IPTV traffic, a two-level Markovian traffic model is proposed in the paper. The model considers both spatial and temporal correlation in MPEG encoded video sequences, so it can mimic the highly Variable data Bit Rate (VBR) behavior of IPTV sources. The model contains a Group of Pictures (GoP) - level Markov chain and a frame-level Markov chain, so it can capture both the inter-GoP and intra-GoP correlations. The proposed traffic model is simple to incorporate into network simulators, and can be used to obtain closed-form solutions of queue performance.