Adaptive Content Poisoning To Prevent Illegal File Distribution in P2P Networks *

Digital content owners have attempted to use content poisoning to disrupt illegal distribution of copyrighted files in peer-to-peer (P2P) systems. This paper provides an analytical model to quantify the impact of content poisoning. Tradeoffs between content poisoning and download efficiency are revealed. In particular, we apply this poisoning model to transform and evaluate three popular P2P networks: BitTorrent, eDonkey, and eMule, for copyright-protected content delivery. We find that chunking protocol, hashing scheme, peer selection, decoy number, and decoy blacklisting play crucial roles to prevent illegal file distribution over P2P networks. We discover that BitTorrent is most resistant to content poisoning. Index poisoning could be a viable alternative to cope with copyright violation on BitTorrent. The eDonkey is less resistant to content poisoning than eMule. A new challenge-response protocol (CRP) is proposed using majority-vote on multiple chunk copies received. To control the level of poisoning, we adapt repeated peer selection (RPS) and decoy blacklisting techniques. Both eDonkey and eMule can apply CRP, RPS, and blacklisting to perform fine tuning of the content poisoning process. These mechanisms can transform eMule to become less resistant than eDonkey, which will serve the best interest of content owners. Finally, we elaborate on integrated research tasks needed to achieve seamless copyright protection in P2P networks. Keywords: Peer-to-Peer systems, content delivery, copyright protection, content poisoning, chunking protocols, decoy deployment, distributed file sharing, P2P reputation systems,