PDR-SFP:一种快速高效的私有文档检索方法

董方明; 梅迪菲; 安宝宇; 周川; 周强; 畅东升; 王平辉

引用本文：

董方明,梅迪菲,安宝宇,周川,周强,畅东升,王平辉.PDR-SFP:一种快速高效的私有文档检索方法[J].信息安全学报,2026,11(3):45-60 [点击复制]
DONG Fangming,MEI Difei,AN Baoyu,ZHOU Chuan,ZHOU Qiang,CHANG Dongsheng,WANG Pinghui.PDR-SFP: A Fast and Efficient Protocol for Private Document Retrieval[J].Journal of Cyber Security,2026,11(3):45-60 [点击复制]

本文已被：浏览 322次下载 208次	码上扫一扫！
PDR-SFP:一种快速高效的私有文档检索方法
董方明^1,2, 梅迪菲^3,4,5, 安宝宇^3,4,5, 周川^2,6, 周强^3,4,5, 畅东升^3,4,5, 王平辉^2,7
字体:加大+\|默认\|缩小-
(1.西安交通大学自动化科学与工程学院西安中国 710049;2.西安交通大学智能网络与网络安全教育部重点实验室西安中国 710049;3.中移雄安信息通信科技有限公司西安中国 710000;4.中移系统集成有限公司西安中国 710000;5.中移信息系统集成有限公司西安中国 710000;6.西安交通大学软件学院西安中国 710049;7.西安交通大学网络空间安全学院西安中国 710049)

摘要:

文档检索是信息系统中的一项基本功能,旨在从数据库中检索出与用户查询最相关的一个或多个文档。在传统的查询模式中,待查询的原始数据与查询索引全部存放在本地,当数据积少成多,在本地进行检索会带来高昂的本地存储以及计算成本。随着云计算和类似技术的出现,越来越多的计算任务被转移到云服务器上。这一转变引发了重大的隐私问题,因为用户查询文档中的敏感信息可能会暴露给服务器,导致严重后果。为了解决这一问题,保护用户查询隐私变得至关重要。此外,在以往的工作中,研究者使用如全同态加密等密码学技术,协议完成一次查询的成本极高。为了解决在文档搜索过程中的隐私问题以及效率问题,本研究提出了一种旨在保护用户隐私的文档检索方法。利用Regev同态加密来保护查询文档的隐私,确保查询内容即使在服务器上处理也能保持机密。此外,本方法采用Sign-Full随机投影来压缩数据库存储并减少计算开销,从而使检索过程更加高效。与现有方法相比,本方法在速度和存储效率上都有显著改进。具体而言,它的查询时间大约快了101倍,存储空间需求减少了约30倍,同时保持了相当的查询准确性。这些改进使该方法在隐私和效率至关重要的实际应用中非常适用。采用这种方法可以显著提高云环境中文档检索系统的安全性。

关键词: 私有文档检索隐私查询同态加密随机投影私有信息检索

DOI：10.19363/J.cnki.cn10-1380/tn.2026.05.04

投稿时间：2024-11-13修订日期：2025-02-08

基金项目:本课题得到西安交通大学-中国移动数字政府联合研究院2023年“揭榜挂帅”数字政府领域课题(No.XJTU-CMCC-YF202301010)资助。

PDR-SFP: A Fast and Efficient Protocol for Private Document Retrieval

DONG Fangming^1,2, MEI Difei^3,4,5, AN Baoyu^3,4,5, ZHOU Chuan^2,6, ZHOU Qiang^3,4,5, CHANG Dongsheng^3,4,5, WANG Pinghui^2,7

(1.School of Automation Science and Engineering, Xi'an Jiaotong University, Xi'an 710049, China;2.Key Laboratory for Intelligent Networks and Network Security Ministry of Education, Xi'an Jiaotong University, Xi'an 710049, China;3.China Mobile Xiong'an Information and Communication Technology Co., Ltd., Xi'an 710000, China;4.China Mobile System Integration Co., Ltd., Xi'an 710000, China;5.China Mobile Information System Integration Co., Ltd., Xi'an 710000, China;6.School of Software Engineering, Xi'an Jiaotong University, Xi'an 710049, China;7.School of Cyber Security, Xi'an Jiaotong University, Xi'an 710049, China)

Abstract:

Document retrieval is a fundamental function in the field of information systems, designed to retrieve one or more documents most relevant to a user's query from a separate database. In traditional query models, both the raw data to be queried and the query index are stored locally. As the data grows over time, performing retrieval locally incurs high costs in terms of storage and computational resources. With the advent of cloud computing and similar technologies, an increasing number of computational tasks have been offloaded to cloud servers. This shift has raised significant privacy concerns, as sensitive information in the documents queried by users could potentially be exposed to servers, leading to severe consequences. Consequently, protecting the privacy of user queries has become a critical issue.
Moreover, in previous studies, researchers have employed cryptographic techniques, such as fully homomorphic encryption, to address privacy concerns; however, the cost of executing a single query with such protocols is prohibitively high. To resolve the privacy and efficiency issues in the document retrieval process, this study proposes a method designed to protect user privacy during document retrieval. The method utilizes Regev homomorphic encryption to safeguard the privacy of queried documents, ensuring that the query content remains confidential even when processed on the server. Additionally, the method employs Sign-Full random projection to compress database storage and reduce computational overhead, thereby making the retrieval process more efficient. Compared to existing methods, this approach demonstrates significant improvements in both speed and storage efficiency. Specifically, the query time is approximately 101 times faster, and the storage space requirement is reduced by about 30 times, while maintaining comparable query accuracy. These advancements make the method highly applicable to practical scenarios where privacy and efficiency are crucial. Adopting this approach can substantially enhance the security of document retrieval systems in cloud environments.

Key words: private document retrieval private query homomorphic encryption random projection private information retrieval