A Lattice-Based Oblivious Transfer Protocol in Malicious Model

Bao Li, Hongda Li, Guangwu Xu +1 more

We first present a protocol which reduces 1-out-of-n oblivious transfer OT m to 1out-of-n oblivious transfer OT k for n > 2 in random oracle model, and show that the protocol is secure against malicious sender and semi-honest receiver. Then, by employing a cut-and-choose technique, we obtain a variant of the basic protocol which is secure against a malicious receiver.

Michael J. Fischer, Silvio M. Micali, Charles Rackoff

Xiaochao Wei, Han Jiang, Chuan Zhao +2 more

In secure two-party computation protocols based on garbled circuit, oblivious transfer (OT) plays an important role in transferring the garbled keys of the participants. In addition to the traditional OT primitive, many other variants of OT have also been presented, such as outsourced oblivious transfer (OOT), cut-and-choose oblivious transfer (CCOT), cut-and-choose bilateral oblivious transfer (CCBOT), etc. These new primitives significantly improve the efficiency and feasibility of secure two-party computation protocols, mainly in optimising computational complexity and interactive rounds. Among these primitives, CCBOT proposed by Zhao et al. in TrustCom 2015 is a novel one and helps to minimize the round complexity of the outer secure two-party computation protocols. In addition, they constructed a CCBOT protocol based on homomorphic encryption scheme in the malicious model. However, their protocol uses the cut-and-choose technique to guarantee security against malicious adversaries, as a result the protocol has an error probability. The commitment scheme is also inevitably involved in their protocol. In this paper, we present a CCBOT protocol with the security against malicious adversaries based on the Decisional Diffie-Hellman (DDH) assumption. Our proposed protocol avoids using the cut-and-choose technique and commitment scheme. In terms of efficiency, our protocol is much more efficient than the previous protocol.

Benoît Libert, San Ling, Fabrice Mouhartem +2 more

Benoît Libert, San Ling, Fabrice Mouhartem +2 more

Bao Li, Hongda Li, Guangwu Xu +1 more

We first present a protocol which reduces 1-of-n oblivious transfer OT m for n > 2 in random oracle model, and show that the protocol is secure against malicious sender and semi-honest receiver. Then, by employing a cut-and-choose technique, we obtain a variant of the basic protocol which is secure against a malicious receiver.