Valuation promotes cooperation in peer-to-peer file-sharing
Dingding Guo, Yu‐Kwong Ricky Kwok · 2013
Existing incentive schemes for peer-to-peer (P2P) file-sharing are rate-based, giving room for strategic peers to benefit from manipulative behaviors so as to treat honest peers unfairly. Specifically, strategic peers can achieve high performance by providing high upload rates which are useless to the system. On the other hand, honest peers suffer from getting low download rates even if they devote chunks with very high values which do great help to the system and other peers. In this paper, we first show that whether to upload high value chunks or low value ones in BitTorrent is a prisoners' dilemma game. We then propose a novel value-based metric, through which peers are rewarded for uploading high value chunks. We prove that by adopting the value-based metric, chunk exchange becomes a repeated game. Our simulation results indicate that our value-based approach can effectively motivate peers to contribute high value chunks to the system, which, in turn, also benefits from value-based metric by achieving a higher propagation speed of the very first copy of the file.