A DISCRETE-TIME QUEUING MODEL TO STUDY THE CELL DELAY VARIATION IN ANATM NETWORK

This paper studies two different physical situations that are importan t to dimension User Parameter Function mechanisms and the buffers need ed at the ATM Adaptation Layer: (i) the cell delay variations a Consta nt Bit Rate traffic is subject to when mixed with Variable Bit Rate tr affic in an ATM network, (ii) the cell loss rate when a Worst Case Tra ffic is multiplexed with Variable Bit Rate sources. The model we use i s a queueing system in which the arrival process is a special class of discrete time batch Markovian arrival process (D-BMAP): the superposi tion of periodic and Variable Bit Rate traffic. An exact solution of t his model in a realistic situation is not possible due to computing ti me and memory requirements. An approximate solution is obtained throug h spectral decomposition and using asymptotic analysis.