Paper 2024/074

PRIDA: PRIvacy-preserving Data Aggregation with multiple data customers

Abstract

We propose a solution for user privacy-oriented privacy-preserving data aggregation with multiple data customers. Most existing state-of-the-art approaches present too much importance on performance efficiency and seem to ignore privacy properties except for input privacy. Most solutions for data aggregation do not generally discuss the users’ birthright, namely their privacy for their own data control and anonymity when they search for something on the browser or volunteer to participate in a survey. Still, they are ambitious to secure data customers’ rights (which should come later). They focus on resulting in an efficiency-oriented data aggregation enabling input privacy only. We aim to give importance to user privacy, and we have designed a solution for data aggregation in which we keep efficiency in balance. We show that PRIDA provides a good level of computational and communication complexities and is even better in timing evaluation than existing studies published recently (i.e., Bonawitz et al. (CCS’17), Corrigan-Gibbs et al. (NSDI’17), Bell et al. (CCS’20), Addanki et al. (SCN’22)). We employ threshold homomorphic encryption and secure two-party computation to ensure privacy properties. We balance the trade-off between a proper design for users and the desired privacy and efficiency.