Privacy violation in limited disclosure of database having functional dependency

Upholding the privacy of data has gathered increased interest in the current industrial scenario. The databases that hold personal information and the individuals´ choices of privacy are often subjected to attacks from mischievous sources which try to access undisclosed data. The means used to achieve such ends may be in the form of complex queries that combine sensitive data with the non-sensitive ones. Access control mechanisms are already in place to prevent such malicious accesses. But the problem is highly compounded when there is a functional dependency among some of the attributes of the data. In such cases often the disclosed data of one person allows the user to infer the undisclosed data of another person. In this paper we propose a model to prevent the breach of privacy through such inference channels of functional dependency. Then we describe the architecture of our model and the corresponding implementation. Finally we carry out a performance analysis to evaluate the cost overheads of time and space and look out topics for future research.