Simulation of an ATM queueing system with a fractional Brownian noise arrival process

Abstract

Poisson based traffic models do not accurately model the characteristics of many types of real network traffic. Several studies have shown that LAN traffic (and some other types) are well characterized by self-similar processes. Moreover, there are strong indications that traditional Markov queueing systems can overestimate the performance of real ATM systems. We investigate the performance of an ATM queueing system fed by a self-similar arrival process. The packet arrivals are generated by a fractional Brownian noise process and the service process is deterministic. There are no analytical or even numerical solutions for this non-Markov system, therefore we collected our performance measures through simulation. We are able to observe many of the characteristics found in measurement studies. The results are compared to another ATM queueing system with an autoregressive arrival process.