Mathematical models for multi-service networks

Summary form only given. The author describes a simple and robust ATM call admission control, and develops the theoretical background for its analysis, Acceptance decisions are based on whether the current load is less than a precalculated threshold, and Bayesian decision theory provides the framework far the choice of thresholds. This methodology allows an explicit treatment of the tradeoff between cell loss and call rejection, and of the consequences of estimation error. Further topics discussed include the robustness of the control to departures from model assumptions, its performance relative to a control possessing precise knowledge of all unknown parameters, the relationship between leaky bucket depths and buffer requirements, and the treatment of multiple call types. The problem of call admission control has received considerable attention in the literature. The approach taken in this work differs in that we eschew any attempt to police mean rates, or to gain benefit from burst-scale queueing. For obvious statistical reasons a long-term mean (over say the life of a call) cannot be policed efficiently. Similarly the tail behaviour of a burst-scale queue is too sensitive to the characteristics of burst lengths to permit robust statistical multiplexing over time. Instead we focus on a very simple scheme, which aims to extract maximum benefit from statistical multiplexing over different calls without requiring detailed knowledge of their mean rates