Anonymity Issues in Peer-to-Peer Systems
Sponsors: Ling Liu, Aameek Singh
CCB 221/ CCB 260
Last few years has seen a dramatic rise of the Peer-to-Peer (P2P) technology. Shedding away the legacy client-server model, P2P systems have made enormous amounts of information and data accessible to casual users. They have the ability to function on their own without any commanding/external authority. One of their most important features is the anonymity it provides to a user searching for content. It is desired because it protects user privacy and escapes censorship. The P2P protocols like Gnutella follow message forwarding mechanisms which make the search completely anonymous. Users can search for any kind of data without exposing their identity.
However, the anonymity is not so easy to maintain on the publisher/service provider side. Any exchange of resources will require the peers to expose their IP addresses. There are a number of schemes which are trying to achieve this. Freenet is one of the most popular ones. They "route" the data through other nodes which are the part of the system. There are other schemes similar to WWW anonymizers which involve use of proxies. Another issue for anonymity is the rise of new kinds of P2P system topologies like Distributed Hash Tables based P2P systems (Chord, Pastry). These systems, in their current form, provide no anonymity. Also the use of anonymity in specific P2P applications like trust management (TrustMe) is worth exploring.
In this project, the student will do a literature survey on all issues related to anonymity in P2P systems and come up with new ideas for domain specific or general P2P anonymity.
Background: No prior knowledge of P2P systems is required.
Deliverables: At the end of the mini project, students will deliver a report of 10-15 pages providing a brief survey of the anonymity issues in decentralized distributed computing such as P2P including any idea and technical development of such ideas you come up with. Students will be judged on the quality of the report and novelty of the ideas brought up and the novelty of your organization of the issues discussed in your report.
1. Gnutella <http://www.gnutella.com>
2. Freenet <http://freenet.sourceforge.net>
3. Chord <http://www.pdos.lcs.mit.edu/chord/> [The signature paper appeared in IEEE/ACM Transactions on Networking]
4. Pastry <http://research.microsoft.com/~antr/Pastry/>
5. TrustMe <http://www.cc.gatech.edu/~aameek/publications/trustme-p2p03.pdf>
6. "Mutual Anonymity Protocols for Hybrid P2P Systems" by L. Xiao et al, in proceedings of ICDCS 2003.