Paper 2024/194

Helium: Scalable MPC among Lightweight Participants and under Churn

Abstract

We introduce Helium, a novel framework that supports scalable secure multiparty computation for lightweight participants and tolerates churn. Helium relies on multiparty homomorphic encryption (MHE) as its core building block. While MHE schemes have been well studied in theory, prior works fall short of addressing critical considerations paramount for adoption such as supporting resource-constrained participants and ensuring liveness and security under network churn. In this work, we systematize the requirements of MHE-based MPC protocols from a practical lens, and we propose a novel execution mechanism, that addresses those considerations. We implement this execution in Helium, which makes it the first implemented solution that effectively supports sub-linear-cost MPC among lightweight participants and under churn. This represents a significant leap in applied MPC, as most previously proposed frameworks require the participants to have high bandwidth and to be consistently online. We show that a Helium network of $30$ parties connected with a $100$Mbits/s link and experiencing a system-wide churn rate of $40$ failures per minute can compute the product of a fixed secret $512\times512$ matrix (e.g., a collectively trained model) with an input secret vector (e.g., a feature vector) $8.3$ times per second. This is $\sim1500$ times faster than a state-of-the art MPC implementation without churn.