Modeling and queueing analysis of variable-bit-rate coded video sources in ATM networks

Abstract

Traffic models and queuing performance of variable-bit-rate (VBR) video sources in an asynchronous transfer model (ATM) network are studied. The VBR video source characteristics of four different coding schemes are examined. A discrete-time, discrete-state Markov chain, which has been successfully applied in the analysis of packetized voice traffic, is used to model the aggregate video traffic, with each VBR-coded source being modeled by a renewal process. The analysis indicates that the aggregate random arrival process at the input of the multiplexer tends to be more predictable (less random) than that of packetized voice traffic for the same normalized bit rate. It is shown that the parameters (link utilization, average queue size, average delay and cell loss rate) obtained for the Poisson process provide an upper bound for the VBR-coded video sources under consideration. >