Poisson approximation of input traffic sources in asynchronous transfer mode (ATM) networks

Abstract

The future telecommunication networks are expected to support a wide variety of multimedia services with very different bit-rate and performance requirements. Under these circumstances, it has been recognized that the conventional simple source modeling, namely, the Poisson arrival process, is no longer adequate to describe various kinds of traffic sources accurately. Other source modelings, which give more accurate results but greatly complicate the performance analyses, have been proposed. However, all these models share a common property, that is, they can be decomposed into a superposition of independent identical interrupted Poisson processes (IPP). In order to obtain a balance between the accuracy and the complexity of analysis, we propose to approximate an IPP, a basic component of various models, by a Poisson distribution and use the Chen-Stein method to find an upper bound of this approximation error.