Wavelet-based resource allocation in ATM networks

Abstract

In recent years, numerous publications have identified a high number of traffic sources to be of self-similar nature. Such traffic can have a negative impact on traditional queuing systems. This paper presents a call admission control algorithm that uses periodic wavelet analysis applied to traffic measurements to estimate the amount of resources necessary to support a certain set of connections under quality-of-service constraints. This algorithm even exploits correlated traffic. Extensive simulation results illustrate the performance of the algorithm over a large parameter spectrum (Hurst parameter, load, and intensity), and the results consistently show that the higher the Hurst parameter, the higher the statistical gain.