Modeling MPEG coded video traffic by Markov-modulated self-similar processes

H. Liu, N. Ansari, Y. Shi

Research output: Contribution to journalArticlepeer-review

3 Scopus citations


Markov modulated self-similar processes are proposed to model MPEG video sequences that can capture the LRD (Long Range Dependency) characteristics of video ACF (Auto-Correlation Function). The basic idea is to decompose an MPEG compressed video sequence into three parts according to different motion/content complexity such that each part can individually be described by a self-similar process. Beta distribution is used to characterize the marginal cumulative distribution (CDF) of the self-similar process. To model the whole data set, Markov chain is used to govern the transitions among these three self-similar processes. In addition to the analytical derivation, initial simulations have demonstrated that our new model can capture the LRD of ACF and the marginal CDF very well. Network cell loss rate using our proposed synthesized traffic is found to be comparable with that using empirical data as the source traffic.

Original languageEnglish (US)
Pages (from-to)101-113
Number of pages13
JournalJournal of VLSI Signal Processing Systems for Signal, Image, and Video Technology
Issue number1-2
StatePublished - Aug 2001

All Science Journal Classification (ASJC) codes

  • Signal Processing
  • Information Systems
  • Electrical and Electronic Engineering


  • MPEG
  • Markov chain
  • Markov modulated process
  • Self-similar process
  • Video traffic modeling


Dive into the research topics of 'Modeling MPEG coded video traffic by Markov-modulated self-similar processes'. Together they form a unique fingerprint.

Cite this