Date of Award


Degree Name

Master of Science


Computer Science


Luc Longpre


While the explosion of availability of information has become invaluable in modern society, it has also raised valid concerns about erosion of privacy. More and more, different entities are encouraged to share information in order to discover potential security risks, health pattern, or social behavior trends. However, some of this information is sensitive. Owners of information may be unwilling to share their whole databases with other entities, either to protect the privacy of the records, or because of the proprietary nature of the information.

We consider the following problem. How can we compute the intersection and the equijoin of databases owned by different parties, while revealing as little additional information as possible? One can address this challenge in different ways. Some protocols use trusted third parties, whereas some rely on the use of cryptographic techniques to achieve their goal without the use of a third party. An example of such an approach is a protocol proposed by Agrawal, Evfimievski and Srikant. In this thesis, we extend their protocol to apply to the intersection and the equijoin of three databases owned by three different entities. We also demonstrate the protocol with an implementation.




Received from ProQuest

File Size

42 pages

File Format


Rights Holder

Swapnil Suresh Samant