Discrete-Time TES/D/1 Queueing Model with Bounded Delay for the Performance Analysis of Cell Rate Monitoring Mechanisms

摘要

One of the hardest problems in engineering and optimizing asynchronous transfer mode (ATM) networks is the control and monitoring of cell arrival processes. Control and monitoring mechanisms play an important role in ATM network management and have a crucial impact on the system behaviour and the quality of service (QoS). Due to the fixed length of cells and the slotted time behaviour of ATM networks, discretetime models are suited for the performance analysis of cell rate monitoring mechanisms in ATM systems. In this paper, we first present discrete-time queueing models for the analysis of a cell rate monitoring mechanisms called generic cell rate algorithm (GCRA), where the arrival process of the queueing models is a transform-expand-sample (TES) process. TES processes are versatile stochastic processes for modelling general stationary time series as well as their autocorrelation functions. Using a TES process as the cell arrival process in a queueing model, we can model the characteristics of the real cell arrivals in ATM networks appropriately. After developing analytical methods for the performance analysis of the GCRA, we discuss the dimensioning issues of the cell rate monitoring mechanism.