System Catalogue design for a Privacy Preserving Relational Database Management System

Abstract

The thesis presents a privacy system catalogue design for a relational database management system (RDBMS). The design includes the predicates of a privacy taxonomy [BAB+09] such as purpose (p), visibility (v), granularity (g) and retention (r). The aim here is to propose a normalized and implementable design for an RDBMS. The main contributions include a study of privacy predicates representation, the basic design of system tables, query processing, an analysis of the design and a partial implementation. Since additional privacy features affect query processing, an algorithm for the SELECT process is described, implemented and analyzed. An alternate design has been suggested as well to fully consider other design possibilities. Lastly, an example application is provided to help understand the proposed design. Design, analysis and implementation of the proposed design, and the query process along with the example system have allowed us to conclude that the design is normalized and implementable.