Blacklisting Polluters and Determining Polluted Content in P2P File Sharing Systems

Many P2P file sharing systems are subject to the pollution attack, whereby corrupted copies of popular titles are aggressively introduced into the system. In this paper we explore defenses to the pollution attack. Specifically, we develop efficient mechanisms to determine (i) the IP ranges responsible for introducing polluted content (that is, the blacklist set); and (ii) the pollution levels of titles (songs, movies, etc.) that have been targeted for attack. The methodology is efficient in that it does not involve downloading files from file-sharing nodes. Instead it determines the blacklist set and the targeted content by crawling the file-sharing system, harvesting metadata, and analyzing the metadata. We illustrate the technique by harvesting metadata from the FastTrack/Kazaa network. We apply our methodology to the harvested metadata to determine a blacklist set and the pollution levels of investigated titles. Analyzing the false positives and false negatives we conclude that the methodology is efficient and accurate.