An efficient quasi-identifier index based approach for privacy preservation over incremental data sets on cloud

Abstract

Cloud computing provides massive computation power and storage capacity which enable users to deploy applications without infrastructure investment. Many privacy-sensitive applications like health services are built on cloud for economic benefits and operational convenience. Usually, data sets in these applications are anonymized to ensure data ownersʼ privacy, but the privacy requirements can be potentially violated when new data join over time. Most existing approaches address this problem via re-anonymizing all data sets from scratch after update or via anonymizing the new data incrementally according to the already anonymized data sets. However, privacy preservation over incremental data sets is still challenging in the context of cloud because most data sets are of huge volume and distributed across multiple storage nodes. Existing approaches suffer from poor scalability and inefficiency because they are centralized and access all data frequently when update occurs. In this paper, we propose an efficient quasi-identifier index based approach to ensure privacy preservation and achieve high data utility over incremental and distributed data sets on cloud. Quasi-identifiers, which represent the groups of anonymized data, are indexed for efficiency. An algorithm is designed to fulfil our approach accordingly. Evaluation results demonstrate that with our approach, the efficiency of privacy preservation on large-volume incremental data sets can be improved significantly over existing approaches.