Privacy Preserving Query Answering in Peer Data Management Systems
Azade Nazi, Donggang Liu, Sajal Kumar Das · 2013
Peer Data Management Systems (PDMS) provide data sharing between heterogeneous databases using peer-to- peer schema mapping, where intermediate peers are used to translate the queries as well as the query results. However, such translation leaks not only the exchanged data but also the private mapping information used by intermediate peers for translation. The privacy of such information is critical in most applications, specially in healthcare. Researchers have proposed to inject dummy values into the query result to confuse the adversary. However, existing solutions overlooked the privacy of the exchanged query. Furthermore, they can only provide partial privacy for the private mapping information and are computationally expensive. This paper presents a novel privacy preserving query answering protocol for PDMS using query decoupling and random value mapping encoding. The former decouples the query and the query result to prevent the leakage of sensitive data, while the latter randomly encodes the private value mappings. The analytical and experimental results confirm that the proposed protocol provides better privacy and also removes the need for expensive commutative encryption, thus greatly reducing the computation overhead.