Performance analysis of adaptive video on demand using stochastic Petri nets

Studies the performance of an adaptive video-on-demand (VOD) system. In this system, channel resources are allocated more efficiently than in full VOD or near VOD, in which a separate channel is allocated for each user or a fixed set of movies is played at regular time intervals, respectively. In adaptive VOD, one channel can be allocated for several users (bunching property) and users´ requests can be rejected even though there are channel resources available (refusal by choice). In addition to these features, the effectiveness of adaptive VOD is based on delayed notification. Users are not notified about the acceptance or the rejection of their requests immediately the request arrives to the server but after a time delay. We model this VOD system using stochastic Petri nets and use discrete event simulation to determine the performance of the system, i.e. the number of requests accepted by the video server compared to the number of requests offered. This is compared to the performance of an adversary system which does not exhibit the refusal-by-choice property