Quality of Service based Retrieval Strategy for Distributed Video on Demand on Multiple Servers

Abstract

The recent advances and development of inexpensive computers and high speed networking technology have enabled the Video on Demand (VoD) application to connect to shared-computing servers, replacing the traditional computing environments where each application was having its own dedicated computing hardware. The VoD application enables the viewer to select, from a list of video files, his favorite video file and watch its reproduction at will. Early video on demand applications were based on single video server where video streams are initiated from a single server, then with the increase in the number of the clients who became interested in VoD services, the focus became on Distributed VoD architectures (DVoD) where the context of distribution may be distributed system components, distributed streaming servers, distributed media content etc.The VoD server must handle several issues in order to be able to present a successful service. It has to receive the clients’ requests and analyze them, calculate the necessary resources for each request, and decide whether a request can be admitted or not. Once the request is admitted, the server must schedule the request, retrieve the required video data and send the video data in a timely manner so that the client does not suffer data starvation in his buffer during the video reproduction. So, the overall objective of a VoD service provider is to provide a better Quality of Service (QoS). Some issues related to QoS are-efficient use of bandwidth, providing better throughput etc.One of the important issues is to retrieve the video data from the servers in minimum time and to start the playback of the video at client side with a minimum waiting time. The overall time elapsed in retrieving the video data and starting the playback is known as access time. The thesis presents an efficient retrieval strategy for a distributed VoD environment where the basic objective is to minimize the access time by maintaining the presentation continuity at the client side. We have neglected some of the network parameters which may affect the access time, by assuming a high speed network between the servers and the client. The performance of the strategy has been analyzed and is compared with the referred PAR (Play After Retrieval) strategy. Further, the strategy is also analyzed under availability condition which is a more realistic approach.