Verifiable private equality test: enabling unbiased 2-party reconciliation on ordered sets in the malicious model

Abstract

In this paper we introduce the novel notion called Verifiable Private Equality Test (VPET) and propose an efficient 2-party protocol for its implementation. VPET enables two parties to securely perform an arbitrary number of comparisons on a fixed collection of (key, value) pairs and thus it is more generic than existing techniques such as Private Equality Test and Private Set Intersection. In addition, we demonstrate how higher-level protocols such as Privacy-Preserving Reconciliation on Ordered Sets (PROS) can be implemented using VPET. Using simulation-based techniques, our new protocols are proven secure in the malicious model. Furthermore, we present a theoretical complexity analysis as well as a thorough experimental performance evaluation of the C++ implementation of our new VPET and PROS protocols.