Secure Multiparty Sorting Protocols with Covert Privacy

Xiaobing Li, Yunyan Xiong, Cai Zhang

Yuval Ishai, Eyal Kushilevitz Z, Yehuda Lindell +1 more

Koteswararao Ch, Kunwar Singh, Anoop Kumar

Henry Carter, Patrick Traynor

Outsourced secure multiparty computation (SMC) protocols allow resource-constrained devices to execute input-private computation with great efficiency. Unfortunately, existing outsourced SMC protocols require that all parties know the function being evaluated, precluding applications where the function itself must remain private. We develop the first linear-complexity protocols for outsourcing private function evaluation (PFE), SMC protocols that provide input and function privacy. Assuming a semi-honest function holder, we build on existing two-party PFE constructions to develop outsourced protocols that are secure against a semi-honest, covert, or malicious outsourcing server and malicious mobile participants. To do this, we develop a garbling technique for combining public and private sub-circuits in a single computation. This allows us to apply auxiliary checks for malicious behaviour using only free-XOR gates. These protocols demonstrate the feasibility of outsourced PFE and provide a first step towards privacy-preserving applications for use in cloud computing.

Peeter Laud, Alisa Pankova

Abstract. This paper presents a generic method for turning passively secure protocols into protocols secure against covert attacks, adding an offline preprocessing and a cheap post-execution verification phase. The execution phase, after which the computed result is already available to the parties, has only negligible overhead. Our method uses shared verification based on precomputed multiplica-tion triples. Such triples are often used to make the protocol execution itself faster, but in this work we make use of these triples especially for verification. The verification preserves the privacy guarantees of the original protocol, and it can be straightforwardly applied to protocols over finite rings, even if the same protocol performs its computation over several distinct rings at once.

玄武 諸橋