A comparative study of numerical and heuristic approaches to calculation of sustainable cell rate

Abstract

This paper highlights the comparative advantages of a novel numerical method for the calculation of the sustainable cell rate (SCR) parameter in ATM networks. The proposed algorithm is suited to the performance analysis of a leaky bucket (LB) fed with a wide range of discrete traffic sources, including Markov modulated Bernoulli processes (MMBP) and general on-off source models. The alternative approach, is a heuristic framework, based on the time /spl epsi/-quantile function associated with a traffic source. This has also been used for finding the SCR value of the source. The effectiveness and capabilities of these two approaches are investigated, and comments on the suitability of application for each approach is given.