Efficient simulation of rare jitter probabilities in ATM switches

Abstract

The ATM Forum has defined two different types of cell delay variation (CDV) measures: the 1-Point CDV which is a measure of jitter, is typically specified for constant bit rate (CBR) sources and the 2-Point CDV, which can be specified for both CBR and variable bit rate (VBR) sources. We model the traffic using the ATM Forum standardized traffic descriptors and consider dispersion as a measure of jitter. The events associated with excessive jitter in ATM networks are typically rare (<10/sup -6/), hence Monte Carlo simulation is not feasible. We extend our previous work on remote delay quantiles for heterogeneous systems to generate efficient simulation techniques using importance sampling to estimate the 2-Point CDV for mixed CBR and VBR sources. Subsequently, we present a novel importance sampling simulation methodology to estimate the 1-Point CDV for CBR sources in the presence of background VBR traffic. For both cases, we observe from experimental results that the improvement in simulation efficiency is inversely proportional to the probability being estimated.