Browse Theses

Database querying has become an inevitable part of modern life. From DNS querying to Internet searches, almost all forms of information retrieval involve database querying. However, not until recently have privacy issues caught people’s attention. These issues apply to database owners as well as users. They also hinder database deployment in economical cloud environments. This motivates the development of new techniques to defend against privacy leaks.

In this thesis, we deal with several privacy challenges in traditional and cloud database environments. They include: protecting user query content privacy against both the database owner and database server, protecting database content privacy, allowing the database owner to exert access control over user queries and proving user query legitimacy to the database owner.

We propose several schemes to address these privacy challenges. First, we present a novel range query scheme that mitigates privacy leaks of user query content. Next, we propose a privacy-preserving scheme for sharing sensitive information that prevents the database server from learning any database content and hides user query content from the database owner and database server. The proposed scheme allows the database owner to verify that a user query is authorized by a certification authority without revealing query content. It also allows the database owner to impose access control over user queries. Then, we extend this scheme to support more flexible types of access control, e.g., arbitrary combination of conjunctive and disjunctive queries. Finally, we propose a range predicate encryption scheme and apply it to a system that supports logarithmic-time search without compromising user query privacy, database privacy and database owner’s access control ability. Additional features, such as query result integrity checks and provable data updates, are also supported.