arrow
Return

Pattern-Hiding Encrypted Multi-Maps With Support for Join Queries

delete2025-01-01
delete0
PRE
AI
李晋国 cover
李晋国 (Jinguo Li)
D
Delong Cui
黄俊钦 (Junqin Huang)
L
Linghe Kong
DOI:10.1109/TIFS.2025.3627000delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
The recently proposed Join Cross-Tags Protocol (JXT) addresses the long-standing issue of excessive query overhead in table joins within Searchable Symmetric Encryption (SSE). As a purely symmetric-key solution, JXT supports efficient conjunctive queries over equi-joins of encrypted tables without requiring any pre-computation during the setup phase. However, JXT has a potential limitation: it may inadvertently reveal the actual volumes of identifiers corresponding to attribute-value pairs, as well as the result values of the join queries. In this paper, we propose JXTMM (JXT multi-map), the first join query scheme designed to hide both volume patterns and result patterns. JXTMM is capable of concealing identifier volumes, preventing the server from learning the actual volumes of attribute-value pairs, and shifting the checkability of join results to the client side, thereby eliminating result pattern leakage. We provide a formal security proof for JXTMM, along with a comprehensive efficiency analysis. Experimental results demonstrate that JXTMM not only performs efficiently on table join queries but also effectively achieves volume-hiding in such queries.
Keywords:
Searchable encryption
volume-hiding
join queries
encrypted multi-map

Journal

IEEE Transactions on Information Forensics and Security cover
IEEE Transactions on Information Forensics and Security
IF:
8
Papers:
5.2K
Citations:
2.3W

Organization

S
shanghai jiao tong university
Scholars:
15.5W
Papers: 11.6W
Citations: 159
S
Shanghai University of Electric Power
Scholars:
5.2K
Papers: 3.4K
Citations: 4.9K