Connection Admission Control In ATM Networks Based On The Foreground and Background Neural Networks

摘要

Connection admission control (CAC) in ATM networks is the set of actions taken by the network to decide whether to accept connection requests during the phase of call establishment or call re- negotiation.CAC is an integral part of the preventive congestion control in ATM networks whose aim is to ensure network performance.The CAC algorithm has the characteristics of the multitude of control parameters,high degree of computation complexity and strong time restrictions.In this paper we present a CAC mechanism featured by combination of foreground control and background learning which is based on neural networks having the capabilities of self-learning and high-speed processing.A case study is given,after which we discuss the practicability of the proposed algorithm.