Is VBR video non-stationary or self-similar? Implications for ATM traffic characterisation

Stephen Bates, Steve McLaughlin

Research output: Chapter in Book/Report/Conference proceedingConference contribution


This paper attempts to determine whether VER video signals are either non-stationary or self-similar in nature as this has major implications for source characterisation on RTM networks. A quantitative test for determining the extent to which a signal can be considered wide sense stationary is developed from standard hypothesis testing techniques. This test is applied to VER video data and three stationary models. A test is developed from the Hurst exponent estimator and this is also applied to the four sets of data and an additional ARIMA model. The implications of the results for characterisation techniques on ATM networks are discussed.

Original languageEnglish
Title of host publicationModelling and Simulation 1996
Editors András Jávor, Axel Lehmann, István Molnár
Number of pages4
Publication statusPublished - 1996
Event10th European Simulation Multiconference 1996 - Budapest, Hungary
Duration: 2 Jun 19966 Jun 1996


Conference10th European Simulation Multiconference 1996
Abbreviated titleESM 1996


Dive into the research topics of 'Is VBR video non-stationary or self-similar? Implications for ATM traffic characterisation'. Together they form a unique fingerprint.

Cite this