Résumé: This paper presents a fuzzy logic controller (FLC) approach to control traffic sources in asynchronous transfer mode (ATM) networks. The proposed traffic controller consists of two main parts: fuzzy connection admission controller (FCAC) and intelligent usage parameter controller (IUPC). The FCAC, for all incoming connections, estimates the required bandwidth and then decides whether to accept or reject them. Simulation results show that the proposed FCAC increases the accuracy of required capacity estimation in comparison with other fuzzy and non-fuzzy bandwidth estimation methods. The IUPC is responsible for monitoring and regulating the input traffic stream after the FCAC accepts the connection. This is achieved by employing a fuzzy controller with the main goal of dynamically tuning the leaky rate based on the state of the traffic source. The necessary control actions are taken based on accurate estimation of the mean cell rate in real time. To exploit the benefits of statistical multiplexing gain, the IUPC in our FLC approach uses a network congestion internal feedback to make a decision as to whether to discard or tag the violating cells. To highlight the merit of the proposed method, different experimental simulations are performed. It is observed that the proposed FLC traffic controller not only has a high statistical multiplexing gain but also can protect quality of services (QoS) of well behaving connections much better than traditional traffic controllers

Langue: Anglais