Performance analysis of CAR: centralized adaptive reservation

A centralized adaptive reservation (CAR) protocol has been proposed by Chan and Leung (see Proc. IEEE CCEC'00, Halifax, NS, 2000) for supporting constant bit rate (CBR). Variable bit rate (VBR) and unspecified bit rate (UBR) services in a star-configured network. This protocol has two main features. First, both the access probability and the bandwidth for reservation requests are varied dynamically to enhance efficiency. Second, by using a request-mini-slot method, UBR traffic can only utilize the residual bandwidth of the other traffic. We derive a Markov chain model for analyzing the performance of the CAR protocol by means of a bi-state traffic source. The focus is to analyze the performance of the VBR stations when both the access probability and the bandwidth for reservation requests are changed dynamically.