A discrete time queueing model to study the cell delay variation in an ATM network

Abstract

This paper studies two different physical situations that are important to dimension User Parameter Function mechanisms and the buffers needed at the ATM Adaptation Layer: 1. (i)the cell delay variations a Constant Bit Rate traffic is subject to when mixed with Variable Bit Rate traffic in an ATM network, 2. (ii)the cell loss rate when a Worst Case Traffic is multiplexed with Variable Bit Rate sources. The model we use is a queueing system in which the arrival process is a special class of discrete time batch Markovian arrival process (D-BMAP): the superposition of periodic and Variable Bit Rate traffic. An exact solution of this model in a realistic situation is not possible due to computing time and memory requirements. An approximate solution is obtained through spectral decomposition and using asymptotic analysis.