Alexander Viand (ETH Zurich)
September 9, 2022 at 11:00 AM on Zoom / Soda Hall
Useable FHE: Challenges & Opportunities
Abstract: Fully Homomorphic Encryption (FHE) allows a third party to perform arbitrary computations on encrypted data, learning neither the inputs nor the computation results. Hence, it provides protection in situations where computations are carried out by an untrusted party. This powerful concept was first conceived by Rivest et al. in the 1970s but remained unrealized until Craig Gentry presented the first feasible FHE scheme in 2009. Thanks to algorithmic improvements and the development of dedicated hardware accelerators, FHE has since then gone from theoretical achievement to practical real-world deployments. However, major obstacles to democratizing its use remain, since developing secure and efficient FHE applications is complex and currently requires significant understanding of the underlying crypto. In this talk, I walk through the inherent cryptographic and engineering challenges in developing FHE applications and cover our efforts towards designing an end-to-end software stack for FHE with a focus on the developer ease of use. Importantly, this ease of use must not come at the expense of application performance, which remains critical even as FHE itself becomes more efficient. Therefore, the challenge is to enable non-experts to enjoy the same level of performance as state-of-the-art solutions by experts. We propose a series of abstractions and associated optimizations that allow us to build an End-to-End development toolchain that can compile high-level programs written by non-experts into efficient FHE-based applications.
Bio: Alexander Viand is a doctoral student & research assistant in the Applied Cryptography Group at ETH Zürich and a member of the Privacy Preserving Systems Lab. His interests include usable security and privacy, privacy enhancing technologies, and the interactions between these technologies and society. In his research, he works with secure computation technologies including Fully Homomorphic Encryption, trying to make these techniques more accessible to non-experts by developing new systems, tools and abstractions.