Analysis, modeling and generation of self-similar VBR video traffic
SIGCOMM '94 Proceedings of the conference on Communications architectures, protocols and applications
Modeling and simulation of self-similar variable bit rate compressed video: a unified approach
SIGCOMM '95 Proceedings of the conference on Applications, technologies, architectures, and protocols for computer communication
Source models for VBR broadcast-video traffic
IEEE/ACM Transactions on Networking (TON)
PMCCN '97 Proceedings of the IFIP TC6 / WG6.3 & WG7.3 International Conference on the Performance and Management of Complex Communication Networks
Wavelet analysis of long-range-dependent traffic
IEEE Transactions on Information Theory
The effect of multiple time scales and subexponentiality in MPEG video streams on queueing behavior
IEEE Journal on Selected Areas in Communications
Modeling video traffic using M/G/∞ input processes: a compromise between Markovian and LRD models
IEEE Journal on Selected Areas in Communications
Hi-index | 0.00 |
It is widely accepted that the video frame size process islong range dependent (LRD),which has been indicated bya number of statistical tests. This approach suggests the useof LRD traffic models for traffic engineering, which in manycases leads to subexponential queue tails in the networkbuffers.Basically, there are two concepts that are consid-ered as the origin of LRD in video modeling:self-similarityand heavy tailed level durations. In this paper we argue fora new approach, where the frame size sequence is describedby three processes representingthree explicit time scales,all of them having a physical meaning.We demonstrate thatis concept serves as an explanation for the outcome of theLRD tests. The properties that are attributed to long rangedependence in fact seem to be the result of non-stationarityof the large time scale process. We also show that the highfrequency component is short range dependent, thus ques-tioning the presence of long range dependence in video traffic.