Fuzzy adaptive traffic enforcement for ATM networks

Future broadband networks like BISDN have to accommodate considerably diverse payload types on a common packet switched platform. The ATM is a fixed packet length transfer mode defined by the ITU-T as the preferred transfer standard for BISDN networks and is able to integrate different services efficiently. The problems associated with the characterisation of bursty and VBR (variable bit rate) ATM traffic has given rise to some concerns in the field of network management. While many conventional approaches like M/D/1 queues and the MMPP (Markov modulated Poisson process) do not yield totally acceptable results, self-adaptive fuzzy logic methods are particularly suitable for nonlinear and bursty environments like ATM systems. This paper presents a fuzzy logic CAC (connection admission control) and traffic policing method intended for the modelling and management of ATM networks