Verona (VERONA)

Verona is a software platform that brings secure computation (SC) to data scientists and other programmers who are not experts in cryptography. Verona allows users to securely compute on sensitive data while it remains opaque, by providing a programming language and tool suite for developing, resource-estimating, deploying, executing, and verifying execution of applications that use SC. Verona interacts with users to recommend SC configuration trade-offs that minimize and optimize secure (and thus often expensive) portions of computation, and balances security and performance in those portions. 

Verona answers HECTOR’s call for new capabilities so users can (1) program distributed, multi-user SC applications without having deep cryptographic expertise; (2) explore the trade-space of multiple SC candidate solutions to data science challenges; (3) understand feasibility and performance of those solutions; and (4) make SC easy to deploy. 

We do this by (1) augmenting an existing mainstream programming language, Julia, with SC features to indicate which inputs/outputs are private/revealed to whom; (2) providing multiple optimized secure computation backends that address different configuration, security, and performance needs: homomorphic encryption (HE), garbled circuits (GC-MPC), linear secret sharing (LSS-MPC), functional encryption (FE) with bounded keys, oblivious RAM (ORAM), verifiable computation (VC), and hybrids of these; (3) providing resource estimation tools for each backend; and (4) automatically managing infrastructure to stand up and tear down SC execution environments, and seamlessly integrating SC results into application execution.