IEEE TRANSACTIONS ON COMPUTERS, TC -2007-09-0492R2, APRIL 8, 200 torrent

Proactive Content Poisoning To Prevent
Collusive Piracy in P2P File Sharing
Xiaosong Lou, Student Member IEEE and Kai Hwang, Fellow IEEE Computer Society
Abstract: Today’s peer-to-peer (P2P) networks are grossly abused by Illegal distributions of music, games, video streams, and
popular software. These abuses have resulted in heavy financial loss in media and content industry. Collusive piracy is the main
source of intellectual property violations within the boundary of P2P networks. This problem is resulted from paid clients
(colluders) illegally sharing copyrighted content files with unpaid clients (pirates). Such an on-line piracy has hindered the use of
open P2P networks for commercial content delivery. We propose a proactive poisoning scheme to stop colluders and pirates
from working together in alleged copyright infringements in P2P file sharing. The basic idea is to detect pirates with identity-
based signatures and time-stamped tokens. Then we stop collusive piracy without hurting legitimate P2P clients.
We developed a new peer authorization protocol (PAP) to distinguish pirates from legitimate clients. Detected pirates will
receive poisoned chunks in repeated attempts. A reputation-based mechanism is developed to detect colluders. The system
does not slow down legal download from paid clients. The pirates are severely penalized with no chance to download
successfully in finite time. Based on simulation results, we find 99.9% success rate in preventing piracy on file-level hashing
networks like Gnutella, KaZaA,Area, LimeWire, etc. Our protection scheme achieved 85-98% prevention rate on part-level
hashing networks like eMuel, Shareaz, eDonkey, Morpheus, etc. Our new scheme enables P2P technology for building a new
generation of content delivery networks (CDNs). These P2P-based CDNs provide faster delivery speed, higher content
availability, and cost-effectiveness than using conventional CDNs built with huge network of surrogate servers.