The Community for Technology Leaders
2016 IEEE 17th International Conference on Information Reuse and Integration (IRI) (2016)
Pittsburgh, Pennsylvania, United States
July 28, 2016 to July 30, 2016
ISBN: 978-1-5090-3207-5
pp: 591-600
ABSTRACT
In this paper, we consider the outsourcing model in which a third-party server provides data integration as a service. Identifying approximately duplicate records in databases is an essential step for the information integration processes. Most existing approaches rely on estimating the similarity of potential duplicates. The service provider returns all records from the outsourced dataset that are similar according to specific distance metrics. A major security concern of this outsourcing paradigm is whether the service provider returns sound and complete near-duplicates. In this paper, we design ARM, an authentication system for the outsourced record matching. The key idea of ARM is that besides the similar record pairs, the server returns the verification object (VO) of these similar pairs to prove their correctness. First, we design an authenticated data structure named MB-tree for VO construction. Second, we design a lightweight authentication method that can catch the service provider's various cheating behaviors by utilizing VOs. We perform an extensive set of experiment on real-world datasets to demonstrate that ARM can verify the record matching results with cheap cost.
INDEX TERMS
Servers, Outsourcing, Authentication, Data structures, Databases, Niobium, Nickel
CITATION

B. Dong and W. Wang, "ARM: Authenticated Approximate Record Matching for Outsourced Databases," 2016 IEEE 17th International Conference on Information Reuse and Integration (IRI), Pittsburgh, Pennsylvania, United States, 2016, pp. 591-600.
doi:10.1109/IRI.2016.86
94 ms
(Ver 3.3 (11022016))