Improving Data Availability in P2P Streaming Systems Using Distributed Virtual Cache

Viewing multimedia contents through P2P streaming systems is gaining to be an emerging practice. In such systems, the steaming mode for multimedia content is used rather than the downloading mode. The advantages of this mode are numerous, especially in terms of time latency. Nonetheless, P2P streaming systems suffer from data availability problem resulting from the unpredictable peer´s departure. In this paper, we present a novel approach based on the use of an embedded distributed virtual cache amongst the peers in theP2P network. An efficient allocation policy, that balances the streaming load over peers, is developed to distribute the cached data over the resulting distributed virtual cache. Additionally, we present an adaptive cache management policy, based on a closed-loop feed back control system that automatically adjusts the size of the cached data to the system workload. The simulation results we obtained show clearly the effectiveness of our approach.