Probability admission control in class-based Video-on-Demand system

Abstract

Admission control is applied to maintain a predefined Quality of Service in online streaming services such as Video-on-Demand system. In common admission control, a new request to a specific class will be rejected when the ports dedicated to such a class are fully-occupied. This proposed system is different from previous works as follows: (1) we experimented on the advent of probability for this fully-occupied class so that it might use other class' ports which are still available. (2) Every higher priority class (popular class) has higher probability to get admitted in lower priority class (less-popular class). Conversely, lower priority class has lower probability to get admitted in higher priority class. It is expected that our proposed system will not only increase the performance of popular class but also the performance of overall system. The proposed admission control policy is validated through simulation using NS-2 simulator. In general, with such experiments, it is shown that blockage has significantly been reduced for popular class and also for overall system.