FSS-DBSCAN: Outsourced Private Density-Based Clustering via Function Secret Sharing

被引:0
作者
Fu, Jiaxuan [1 ]
Cheng, Ke [1 ,2 ]
Song, Anxiao [1 ]
Xia, Yuheng [1 ]
Chang, Zhao [1 ]
Shen, Yulong [1 ]
机构
[1] Xidian Univ, Sch Comp Sci & Technol, Xian 710071, Shaanxi, Peoples R China
[2] Xian Univ Posts & Telecommun, Shaanxi Key Lab Informat Commun Network & Secur, Xian 710121, Peoples R China
基金
中国国家自然科学基金;
关键词
Clustering algorithms; Protocols; Cryptography; Graphics processing units; Additives; Outsourcing; Distributed databases; Secure clustering; DBSCAN; additive secret sharing; secure comparison;
D O I
10.1109/TIFS.2024.3446233
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
Density-based clustering algorithms such as DBSCAN, are highly effective in handling large datasets and identifying clusters of arbitrary shapes, playing a crucial role in data analysis fields like outlier detection and social networks. Outsourcing DBSCAN to the cloud brings substantial benefits but also raises major privacy concerns regarding the private input data of data owners. Existing private DBSCAN methods often face challenges of inefficiency or potential privacy leakage, hindering their practical deployment. To address these challenges, we introduce FSS-DBSCAN, a three-server MPC platform designed for outsourced private density-based clustering using function secret sharing (FSS). This solution guarantees clustering quality equivalent to plaintext algorithms, ensures comprehensive privacy protection, and achieves top-tier efficiency. The high performance of FSS-DBSCAN is driven by two pivotal strategies. First, we devise an MPC-friendly DBSCAN algorithm that is highly compatible with efficient secret-sharing-based cryptographic protocols and benefits from GPU acceleration. Second, we construct novel FSS-based protocols tailored for complex operations integral to our DBSCAN variant, such as Euclidean distance comparison and point assignment, and further optimize their computation through tensorization techniques. We implement our platform as an extensible system on top of PyTorch that leverages GPU hardware acceleration for cryptographic and tensorized operations. These innovations enable FSS-DBSCAN to significantly outperform ppDBSCAN (AsiaCCS 2021), reducing the clustering time for 5000 samples to approximately 2 hours, achieving an $83.4\times $ speed improvement.
引用
收藏
页码:7759 / 7773
页数:15
相关论文
共 42 条
[1]  
Anikin IV, 2017, IEEE INT SIBER CONF
[2]   RN-SMOTE: Reduced Noise SMOTE based on DBSCAN for enhancing imbalanced data classification [J].
Arafa, Ahmed ;
El-Fishawy, Nawal ;
Badawy, Mohammed ;
Radad, Marwa .
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES, 2022, 34 (08) :5059-5074
[3]   Function Secret Sharing for Mixed-Mode and Fixed-Point Secure Computation [J].
Boyle, Elette ;
Chandran, Nishanth ;
Gilboa, Niv ;
Gupta, Divya ;
Ishai, Yuval ;
Kumar, Nishant ;
Rathee, Mayank .
ADVANCES IN CRYPTOLOGY - EUROCRYPT 2021, PT II, 2021, 12697 :871-900
[4]   Function Secret Sharing: Improvements and Extensions [J].
Boyle, Elette ;
Gilboa, Niv ;
Ishai, Yuval .
CCS'16: PROCEEDINGS OF THE 2016 ACM SIGSAC CONFERENCE ON COMPUTER AND COMMUNICATIONS SECURITY, 2016, :1292-1303
[5]   Function Secret Sharing [J].
Boyle, Elette ;
Gilboa, Niv ;
Ishai, Yuval .
ADVANCES IN CRYPTOLOGY - EUROCRYPT 2015, PT II, 2015, 9057 :337-367
[6]   Privacy-preserving Density-based Clustering [J].
Bozdemir, Beyza ;
Canard, Sebastien ;
Ermis, Orhan ;
Moellering, Helen ;
Onen, Melek ;
Schneider, Thomas .
ASIA CCS'21: PROCEEDINGS OF THE 2021 ACM ASIA CONFERENCE ON COMPUTER AND COMMUNICATIONS SECURITY, 2021, :658-671
[7]  
Bunn P, 2007, CCS'07: PROCEEDINGS OF THE 14TH ACM CONFERENCE ON COMPUTER AND COMMUNICATIONS SECURITY, P486
[8]   Security and composition of multiparty cryptographic protocols [J].
Canetti, R .
JOURNAL OF CRYPTOLOGY, 2000, 13 (01) :143-202
[9]   ABY - A Framework for Efficient Mixed-Protocol Secure Two-Party Computation [J].
Demmler, Daniel ;
Schneider, Thomas ;
Zohner, Michael .
22ND ANNUAL NETWORK AND DISTRIBUTED SYSTEM SECURITY SYMPOSIUM (NDSS 2015), 2015,
[10]  
Derpanis K.G., 2005, Lecture Notes, V32, P1