International Association for Cryptologic Research

International Association
for Cryptologic Research


Rahul Rachuri


Le Mans: Dynamic and Fluid MPC for Dishonest Majority
Rahul Rachuri Peter Scholl
Most MPC protocols require the set of parties to be active for the entire duration of the computation. Deploying MPC for use cases such as complex and resource-intensive scientific computations increases the barrier of entry for potential participants. The model of Fluid MPC (Crypto 2021) tackles this issue by giving parties the flexibility to participate in the protocol only when their resources are free. As such, the set of parties is dynamically changing over time. In this work, we extend Fluid MPC, which only considered an honest majority, to the setting where the majority of participants at any point in the computation may be corrupt. We do this by presenting variants of the SPDZ protocol, which support dynamic participants. Firstly, we describe a \emph{universal preprocessing} for SPDZ, which allows a set of $n$ parties to compute some correlated randomness, such that later on, any subset of the parties can use this to take part in an online secure computation. We complement this with a \emph{Dynamic SPDZ} online phase, designed to work with our universal preprocessing, as well as a protocol for securely realising the preprocessing. Our preprocessing protocol is designed to efficiently use pseudorandom correlation generators, thus, the parties' storage and communication costs can be almost independent of the function being evaluated. We then extend this to support a \emph{fluid online phase}, where the set of parties can dynamically evolve during the online phase. Our protocol achieves \emph{maximal fluidity} and security with abort, similarly to the previous, honest majority construction. Achieving this requires a careful design and techniques to guarantee a small state complexity, allowing us to switch between committees efficiently.
Improved Primitives for MPC over Mixed Arithmetic-Binary Circuits 📺
This work introduces novel techniques to improve the translation between arithmetic and binary data types in multi-party computation. To this end, we introduce a new approach to performing these conversions, using what we call \emph{extended doubly-authenticated bits} (edaBits), which correspond to shared integers in the arithmetic domain whose bit decomposition is shared in the binary domain. These can be used to considerably increase the efficiency of non-linear operations such as truncation, secure comparison and bit-decomposition. Our eDaBits are similar to the \emph{daBits} technique introduced by Rotaru et al.~(Indocrypt 2019). However, our main observations are that (1) applications that benefit from daBits can also benefit from edaBits in the same way, and (2) we can generate edaBits directly in a much more efficeint way than computing them directly from a set of DaBits. Technically, the second contribution is much more challenging, and involves a novel cut and choose technique that may be of independent interest, and requires taking advantage of natural tamper-resilient properties of binary circuits that occur in our construction to obtain the best level of efficiency. Finally, we show how our eDaBits can be applied to efficiently implement various non-linear protocols of interest, and we thoroughly analyze their correctness for both signed and unsigned integers. The results of this work can be applied to any corruption threshold, although they seem best suited to dishonest majority protocols such as SPDZ. We implement and benchmark our constructions, and experimentally verify that our technique yield a substantial increase in effiency. Our eDaBits save in communication by a factor that lies between $2$ and $170$ for secure comparisons with respect to a purely arithmetic approach, and between $2$ and $60$ with respect to using daBits. Improvements in throughput per second are more subdued but still as high as a factor of $47$. We also apply our novel machinery to the tasks of biometric matching and convolutional neural networks, obtaining a noticeable improvement as well.