Better Preprocessing for Secure Multiparty Computation

We present techniques and protocols for the preprocessing of secure multiparty computation (MPC), focusing on the so-called SPDZ MPC scheme [19] and its derivatives [16,18,1]. These MPC schemes consist of a so-called preprocessing or offline phase where correlated randomness is generated that is independent of the inputs and the evaluated function, and an online phase where such correlated randomness is consumed to securely and efficiently evaluate circuits. In the recent years, it has been shown that such protocols (such as [4,31,27] ) turn out to be very efficient in practice. While much research has been conducted towards optimizing the online phase of the MPC protocols, there seems to have been less focus on the offline phase of such protocols (except for [16]). With this work, we want to close this gap and give a toolbox of techniques that aim at optimizing the preprocessing. We support both instantiations over small fields and large rings using somewhat homomorphic encryption an...

• Type: Journal
• Year: 2016
• Where: IACR
• Authors: Carsten Baum, Ivan Damgård, Tomas Toft, Rasmus Winther Zakarias
• Keywords: Biometrics, IACR 2016