Threshold-based admission control for multi-class video-on-demand systems

The next generation of distributed digital video-on-demand (VoD) systems will use admission control tests to ensure that users receive predictable and guaranteed service. If the system cannot support a new request then that request is blocked This paper presents an analytical model to evaluate unified admission control strategies for distributed VoD systems with multiple video classes. We prove that there exists a computationally efficient technique to determine the blocking probabilities for newly arriving requests. Through numerical analysis we show the effect of different admission control policies on overall system blocking rates