A New CAC Method Using Queuing Theory

Abstract

The CAC (Connection Admission Control) method plays an important role in the ATM (Asynchronous Transfer Mode) network environment. The CAC is the first step in the prevention of congested states in the network topology, and conducts to the optimal network resources utilization. The paper is aimed to propose an enhancement for a convolution method that is one of the statistical CAC methods used in ATM. The convolution method uses a buffer-less assumption in the estimation of the cell loss. Using formulas for the G/M/1 queuing system, the cell loss can be estimated as the buffer overflow probability. In this paper, the proposed CAC method is compared with other three statistical CAC methods, and conclusions regarding the exploitation of the CAC method are presented