A scalable video-on-demand system using future reservation of resources and multicast communications

In a video-on-demand (VoD) system the user can select and play movies according to his/her own quality of service (QoS) requirements. Upon receipt of the user request, a typical VoD system checks whether there are enough available resources to deliver the requested movie to the user's host. If the response is yes, the movie presentation can start; otherwise, a rejection is sent back to the user. This means that the response is based only on the system's load at the time the request is made and assumes that the service duration (movie length) is infinite. In this paper we propose a scalable VoD (SVoD) system which decouples the starting time of the service from the time the service request is made and requires that the duration of the requested service must be specified. In response to a user request, SVoD determines the QoS with which the movie can be presented at the time the service request is made, and at certain later times carefully chosen. As an example if the requested QoS cannot be supported at the time the service request is made, SVoD allows to compute the earliest time, when the user can play the movie with the desired QoS. This time can also be determined in a way to use multicast communication to deliver the same movie to several users. The scalability achieved by SVoD is quantified and compared with that of typical VoD. The performance quantification, performed through the use of simulations, shows that SVoD achieves (in average) high resource utilization and decreases (in average) the blocking probability of user requests. In the worst case (very unlikely to happen) SVoD performs similarly to VoD. However, at any case SVoD provides a more useful feedback to the user helping him/her to make appropriate decisions. (C) 1998 Elsevier Science B.V.