Privacy Preserving Database Application Testing

University Of North Carolina At Charlotte, Charlotte NC

Investigators

Abstract

Government agencies and private organizations are accumulating a vast amount of sensitive information on citizens. These organizations often develop new applications to utilize the information stored in the databases. An essential phase of developing such a new application is to test it thoroughly prior to deployment. Traditionally, application developers carry out tests on their local development databases. A major shortcoming of this approach is that these local databases usually have only a small number of data samples and hence cannot simulate satisfactorily a live environment. On the other hand, the practice of testing applications against live production databases is becoming increasingly undesirable due to the fact that such use of live production databases may disclose sensitive data to an unauthorized tester and incorrectly update the underlying databases. This project investigates techniques to generate mock databases for application software testing without revealing any confidential information on the live production databases. Cryptographic techniques such as indistinguishability will be used to model privacy preservation as well as the degree of similarity between a live and a mock database. Success of the project will open up a new approach to secure application software testing and produce a prototype system for future deployment.