Privacy Preserving Aggregate Query of OLAP for Accurate Answers

Abstract

In recent years, privacy protection has become an important topic when cooperative computation is performed in distributed environments. This paper puts forward efficient protocols for computing the multi-dimensional aggregates in distributed environments while keeping privacy preserving. We propose a novel model, which contains two crucial stages: local computation and cooperative computation based on secure multiparty computation protocols for privacy-preserving on-line analytical processing. According to the new model, we develop approaches to privacy-preserving count aggregate query over both horizontally partitioned data and vertically partitioned data. We, meanwhile, propose an efficient sub-protocol Two-Round Secure Sum Protocol. Theoretical analysis indicates that our solutions are secure and the answers are exactly accurate, that is, they can securely obtain the exact answer to aggregate query without revealing anything about their confidential data to each other. We also analyze detailedly the communication cost and computation complexity of our schemes in the paper and it shows that the new solutions have good linear complexity. No privacy loss and exact accuracy are two main significant advantages of our new schemes.