LB-SELA - RATE-BASED ACCESS-CONTROL FOR ATM NETWORKS

The statistical multiplexing of sources with diverse traffic character istics in ATM networks leads to serious congestion control problems. O ne of the most fundamental is the policing of sources, particularly th ose with bursty traffic characteristics. Because of the statistical na ture of burstiness, the policing of these sources is difficult and the known policing mechanisms cannot control them effectively. In this pa per, the Leaky Bucket mechanism is enhanced using a learning algorithm ill order to ''learn'' the behaviour of the source. As will be shown in;he simulation results, the tighter and faster control achieved by t he proposed methodology results in more statistical gain and better gu arantee of the QoS constraints. (C) 1998 Elsevier Science B.V. All rig hts reserved.