期刊
COMPUTER STANDARDS & INTERFACES
卷 87, 期 -, 页码 -出版社
ELSEVIER
DOI: 10.1016/j.csi.2023.103764
关键词
Private set intersection; Multi-party; Semi-honest secure; Small set setting; Collusion resistant; Malicious secure
Private set intersection (PSI) is widely studied for preserving privacy in collaborative data analytics. However, existing multi-party PSI (MP-PSI) protocols often face performance and scalability issues when the number of participants increases. To address this, we propose two efficient MP-PSI protocols for scenarios with a large number of participants and small set size.
Private set intersection (PSI) has attracted the researchers and the developers both from academia and industry in cooperating the private data to achieve privacy preserving. Traditional PSI protocols allow the two participants with each private sets interactively calculating the intersection without revealing any additional information. Multi-party private set intersection (MP-PSI) is quite different from the conventional two-party PSI, which could not directly obtain from the two-party PSI, and become an emerging topic in collaborative data analytics from different private data owners. However, the most of the existing MP-PSI protocols always consider the small number of the participants with large set size and the performance drops with the increasing number of participants. In addition, the participants with small set size also bring the heavy communication and computation overhead. To overcome the above drawbacks, we propose two efficient MP-PSI protocols for large number of participants and small set size, and formally prove secure against collision attack in the semi-honest model and malicious model. The experiments show that when the participant number increases from 5 to 50 and the set size scales from 27 to 210, our protocols are faster than the most efficient MP-PSI protocol. Thus, our MP-PSI protocols are considering to be practical solutions, which fit the scenarios in small set size with increasing participants.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据