A Markov-based model of MPEG-2 audio/video traffic

MPEG-2 audio/video standard is the internationally accepted encoding standard for digital audio and video transmission. Accurate performance evaluation models of the related traffic are therefore required to design and manage future networks efficiently. In this paper a Markov-based model of MPEG-2 audio/video traffic is proposed. Each monomedia source is modeled with a switched batch Bernoulli process (SBBP), whose statistical changes are considered as correlated events driven by the user behavior. Applying the master/slave model, the MPEG-2 multimedia source as a whole is modeled with an SBBP model. Finally, the MPEG-2 traffic stream, put out by the MPEG-2 multiplexer, is modeled as an on-off source and the distributions of the the on and off state durations are calculated