An expanding reference library for Peer-to-Peer content

Many Peer-to-Peer (P2P) networks have emerged as scalable platforms for enabling file-sharing and content-distribution applications. Due to their reliance on decentralized models encompassing millions of users, P2P networks are subjected to being infested with contraband material, copyright infringing content, as well as malicious viruses and programs. Moreover, the dynamics involved as users join and leave these networks make investigating them quite complex and time-dependent. To that end, this paper presents PxP, a tool that creates an expanding reference library for content that appears on P2P networks. PxP continuously searches the P2P networks for new files, downloads and fingerprints them, deletes them, and add their meta data to a public reference library. We envision that this library will aid investigators in criminal investigations that involve contraband material, civil investigations with copyright infringement content. In addition, this library can be used to speed up forensics examination of media through skipping over some of the files shared on P2P networks and in validating carving tools. We assess the performance of our framework through analysis, simulations and real experiments for content collected from the Gnutella P2P network.