A Practical Parallel Computation in a Scalable Multiparty Private Set Intersection

Abstract

Medical data is often managed individually by multiple medical institutions, and it is used to analyze patients’ diseases. However, for more detailed analysis, it is necessary to integrate such scattered data without violating the privacy of patients and companies. Multiparty Private Set Intersection (MPSI) is a kind of privacy protection protocol that can integrate the same person’s data managed by several independent institutions without information leakage. MPSI is an indispensable technology for the use of big data containing sensitive information. However, usability and performance improvements of MPSI have been a problem for its practical use. In this paper, we implemented a parallelization of the existing MPSI protocol and discussed the characteristics of the parallelization through experiments.