Abstract

Privacy requirements have an increasing impact on the real-world applications. Technical considerations and many significant commercial and legal regulations demand that privacy guarantees be provided whenever sensitive information is stored, processed, or communicated to external parties. It is therefore crucial to design solutions able to respond to this demand with a clear integration strategy for existing applications and a consideration of the performance impact of the protection measures. In this paper, we propose a solution to enforce data privacy and user privacy over outsourced database services. The approach starts from a flexible definition of privacy constraints on a relational schema, applies encryption on information in a parsimonious way and mostly relies on attribute partition to protect sensitive information. Based on the approximation algorithm for the minimal encryption attribute partition with quasi-identifier detection, the approach allow storing the outsourced data on a single database server and minimizing the amount of data represented in encrypted format. Meanwhile，by applying cryptographic technology on the auxiliary random server protocol, the approach can solve the problem of private information retrieval to protect user privacy. The theoretical analysis and experimental results show that our new model can provide efficient data privacy protection and query processing, efficient in computational complexity and dose not increase the cost of communication complexity of user privacy protection.