Oblivious Keyword Search Protocols in the Public Database Model

Abstract

Databases associated with keywords, can be public, private or hybrid, as a result the solutions to keyword search protocols for each type are different. In this paper, we study the problem of privacy-preserving keyword search in the public database model where the data-item is public but a client wishes to retrieve some data-item or search data-item, without revealing to the server which item it is. The contribution of this paper is of three fold. In the first fold, an efficient implementation of oblivious linear function evaluation protocols for secure two-party computation of a linear function over a finite field that does not emulate the circuit for computing oblivious linear function while at the same time it is provably secure against malicious adversary is presented. We show our implementation is provably secure assuming that the underlying Fujisaki-Okamoto's commitment scheme is informational hiding and computational binding as well Paillier's encryption scheme is semantically secure in the common reference string model. In the second fold, we further extend the techniques to non-linear polynomials and thus provide a secure reduction of OPE protocols of a polynomial of degree m, to m OPEs of linear polynomials. In the third fold, we use the OPEs for solving the oblivious keyword search problem along the Ogata and Kurosawa's ad hoc methodology, and show that our implementation is provably secure assuming that the underlying Fujisaki-Okamoto's commitment scheme is informational hiding and computational binding, Paillier's encryption scheme is semantically secure, G is a pseudo-random function in the common reference string model.