On the power of cooperation in multimedia caching

Abstract

Real time multimedia applications such as Internet TV, Video On Demand, Distance Learning, and Video Conferencing are becoming more and more popular over the Internet. Streaming media caching is a critical ingredient in the ability to provide scalable real-time service over the best effort Internet. In many cases, bandwidth becomes the system bottleneck and the cache cannot provide the required quality for all streams simultaneously. In this paper we study new algorithms, based on cooperation, which can improve the cache ability to provide service to all of its clients. The main idea is based on the willingness of streams to reduce their used bandwidth and allow other streams that may need it more, to use it. Our extensive simulation study indicates that our algorithms can reduce the pre-caching time by a factor of 3, or increase the probability for adequate service level by 30% using the same pre-caching time. © IFIP International Federation for Information Processing 2006.