You are looking at a specific version 20201213:163840 of this paper. See the latest version.

Paper 2020/1541

A Framework of Private Set Intersection Protocols.

Ziyuan Liang and Weiran Liu and Fan Zhang and Bingsheng Zhang and Jian Liu and Lei Zhang and Kui Ren

Abstract

Private Set Intersection (PSI) is a specified protocol of secure Multi-Party Computation (MPC). PSI allows two parties to obtain the intersection of their private sets while nothing else is revealed. In contrast to the great demand for PSI in real-world applications, there is still no evaluation results of different general practical PSI framework. Most existing PSI implmentations are based on C/C++, which also makes them hard to compute in parallel. %We focus on OT-based PSI in this work. Oblivious transfer (OT) allows a party to obliviously choose messages from others. Lots of PSI protocols have been proposed in recent years, which achieve good performance and are regarded as one of the most potential PSI species. In this paper, we propose a generic Java-based PSI framework and implement all up-to-date OT-based PSI protocols within the framework until now. We evaluate these OT-based PSI protocols and the dependent cryptographic primitives and provide the best combination of primitives for constructing a best-performed OT-based PSI from the ground up. Additional optimizations are also applied to the protocols in our framework, including both generic and custom-tailored ones. We adopt filters to significantly reduce the communication of OT-based PSI protocols. The implementations in our framework support concurrence by using the natural feature of Java, which avoids to manurally allocate threads when using C/C++. We believe that our framework benefits a lot for future MPC and PSI researches and helps the promotion of PSI-based applications.

Metadata
Available format(s)
PDF
Category
Implementation
Publication info
Preprint. MINOR revision.
Keywords
Private Set IntersectionMulti-Party IntersectionCryptography and Securit
Contact author(s)
liangziyuan @ zju edu cn,weiran lwr @ alibaba-inc com,fanzhang @ zju edu cn
History
2024-02-28: revised
2020-12-13: received
See all versions
Short URL
https://ia.cr/2020/1541
License
Creative Commons Attribution
CC BY
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.