Parallel processing algorithms and architecture for multimedia on-demand servers

Advances in networking and storage technologies have made it possible to build multimedia on-demand servers that provide service similar to those of neighborhood videotape rental stores. We propose a specialized high performance programmable multiprocessor architecture which allows continuous playback of mediastreams to large number of clients, a critical factor in multimedia servers. This architecture supports dynamic phase shifts in media distribution which is particularly useful in a distributed environment. We introduce algorithms for maximum concurrency extraction and to avoid task migration in the presence of dynamically changing demand. We show that our solution performs an order of magnitude better than other known solutions.<
>