Bayesian based Admission Control (BAC) Policy for Distributed video-on-demand system

Video-on-Demand (VOD) is a system which allows users to select and watch a video on demand. In a video system, large number of user requests arrives for different video categories (popular and unpopular). There needs to be an admission control policy to decide which request to admit and, which should reject to maintain quality of service (QoS). An admission control policy needs to be designed very carefully because a new incoming call can only be accepted if it is required QoS can be achievable without degrading the QoS of accepted requests. Admission control is more demanding in video on demand because of its extra attributes such as movie popularity, movie length, movie resolution, request time, client location and link quality. Popularity of an incoming request is one of the most important attributes. Popular movies are expansive, and generate more revenue. High definition (HD) movies require more resources. Length/duration of the movie occupies the resources for longer time and time of the request is also very important, in peak/busy hours many requests are expected especially for popular movies. The purpose of this study is to investigate all possible attributes and to develop context aware Bayesian admission control (BAC) policy to take care of all these attributes before admitting the incoming request for all above attributes to maximize the revenue, efficient use of resources and guaranteeing promised QoS. The performance of the proposed system is evaluated via simulation. Number of blocked requests and the total generated revenue are calculated. It is observed that proposed admission control policy improves the performance of all the request classes there by improving the performance of the overall system.