Enhancing Peer-to-Peer Parallel Data Access with PeerFecT

摘要

Data access performances over a Peer-to-Peer (P2P) network is a critical issue since downloading time remains the longest phase of P2P usage by end-user. In the present effort, we propose an innovative approach called PeerFecT (Peer-to-peer with Fec dissemination Technique) based on an erasure code, often called Forward Error Correction (FEC) code, to speed-up data access into P2P networks. The main idea is to use the erasure code to dilute the information over the peers. This dilution allows a greater and flexible choice among all the disseminated encoded data and then statistically enhances the overall throughput of the transfer. This paper focus on the performance gain this technique can provide. A performance evaluation based on an original model using the results of a measurement campaign of sequential and parallel downloads in a real P2P network over Internet is presented. The results show that for the same amount of information disseminated into the P2P network, our solution is more efficient compared to classical file replication approaches.