Search Condition-Hiding Query Evaluation on Encrypted Databasesopen access
- Authors
- Kim, M.; Lee, H.T.; Ling, S.; Ren, S.Q.; Tan, B.H.M.; Wang, H.
- Issue Date
- Nov-2019
- Publisher
- Institute of Electrical and Electronics Engineers Inc.
- Keywords
- Encrypted database; homomorphic encryption; private queries
- Citation
- IEEE Access, v.7, pp 161283 - 161295
- Pages
- 13
- Journal Title
- IEEE Access
- Volume
- 7
- Start Page
- 161283
- End Page
- 161295
- URI
- https://scholarworks.bwise.kr/cau/handle/2019.sw.cau/63708
- DOI
- 10.1109/ACCESS.2019.2951695
- ISSN
- 2169-3536
- Abstract
- Private database query (PDQ) is a protocol between a client and a database server, designed for processing queries to encrypted databases. Specifically, PDQ enables a client to submit a search query and to learn a resulting set satisfying its search condition, without revealing sensitive information about a query statement. The whole query can be protected from the server, but for efficiency reasons known PDQ solutions generally consider to hide the constants only in a query statement. In this paper, we provide two fully homomorphic encryption (FHE)-based PDQ protocols that hide type of queries as well as the constants of a query statement. Particularly, our constructions focus on conjunctive, disjunctive, and threshold conjunctive queries. To this end, we first build a single compact logical expression to cover both conjunctive and disjunctive queries. On top of the logical expression, we design a PDQ protocol that enables to evaluate conjunctive and disjunctive queries without revealing any information on a given query. The second PDQ protocol comes from our observation that if a threshold conjunctive query has a particular threshold value, it results in either a conjunctive query or a disjunctive query. Because the PDQ protocol writes the three types of queries into a single polynomial expression, the resulting protocol can evaluate the three types of query statements without revealing any information on queries. To demonstrate their efficiency, we provide proof-of-concept implementation results of our proposed PDQ protocols. According to our rudimentary experiments, it takes 37.57 seconds to perform a query on 316 elements consisting of 16 attributes of 64 bits using Brakerski-Gentry-Vaikuntanathan's leveled FHE with SIMD techniques for 149-bit security, yielding an amortized rate of just 0.119 seconds per element.
- Files in This Item
-
- Appears in
Collections - College of Software > School of Computer Science and Engineering > 1. Journal Articles
![qrcode](https://api.qrserver.com/v1/create-qr-code/?size=55x55&data=https://scholarworks.bwise.kr/cau/handle/2019.sw.cau/63708)
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.