One of the main concerns of users in using current Internet based search techniques such as Google Search is lack of privacy. Such services can learn user preferences and they log search history of users. In current services, users need to trust the service that it would not leak user's identity and user's interest in particular keywords to entities that can misuse such data.
One possible approach for solving this problem is to exploit the fact that users have real-world trust relations with other users and they may be friends with them: A user trusts the friend not to reveal the user's interactions or identity to any non-authorized entity. Hence, a search service that only employs interactions between friends can possibly provide a privacy preserving search techniques.
Social networks model the friend-to-friend relationship between all users in the system. Studies have shown that social networks exhibit small world properties (the network is highly clustered with small diameters) and users have small paths to other users in the network. Studies have also shown that if users can estimate distance between users, then they can find short paths to distant users using decentralized greedy routing where each user is only aware of his immediate friends.
We consider the project for an application scenario that focuses on providing a location service for searching data or users. The aim of this project is to design algorithms and techniques for data lookup in a distributed and privacy preserving manner. The project also involves evaluating various lookup techniques, both existing and new, under different churn conditions.
References:
1. I. Clarke, O. Sandberg, M. Toseland, V Verendel. Private Communication
Through a Network of Trusted Connections: The Dark Freenet
2. O. Sandberg. Distributed Routing in Small World. http://freenetproject.org/papers.html
3. A. M. Kermarrec, V. Leroy & G. Tredan. Distributed Social Graph Embedding