Distributed private matching and set operations

  • Authors:
  • Qingsong Ye;Huaxiong Wang;Josef Pieprzyk

  • Affiliations:
  • Centre for Advanced Computing, Algorithms and Cryptography, Department of Computing, Macquarie University, NSW, Australia;Centre for Advanced Computing, Algorithms and Cryptography, Department of Computing, Macquarie University, NSW, Australia and Division of Mathematical Sciences, Nanyang Technological University, S ...;Centre for Advanced Computing, Algorithms and Cryptography, Department of Computing, Macquarie University, NSW, Australia

  • Venue:
  • ISPEC'08 Proceedings of the 4th international conference on Information security practice and experience
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

Motivated by the need of private set operations in a distributed environment, we extend the two-party private matching problem proposed by Freedman, Nissim and Pinkas (FNP) at Eurocrypt'04 to the distributed setting. By using a secret sharing scheme, we provide a distributed solution of the FNP private matching called the distributed private matching. In our distributed private matching scheme, we use a polynomial to represent one party's dataset as in FNP and then distribute the polynomial to multiple servers. We extend our solution to the distributed set intersection and the cardinality of the intersection, and further we show how to apply the distributed private matching in order to compute distributed subset relation. Our work extends the primitives of private matching and set intersection by Freedman et al. Our distributed construction might be of great value when the dataset is outsourced and its privacy is the main concern. In such cases, our distributed solutions keep the utility of those set operations while the dataset privacy is not compromised. Comparing with previous works, we achieve a more efficient solution in terms of computation. All protocols constructed in this paper are provably secure against a semi-honest adversary under the Decisional Diffie-Hellman assumption.