Domain-polymorphic language for privacy-preserving applications

Abstract

We present SecreC, a programming language for specifying privacy-preserving applications using a mix of techniques for secure multiparty computation. Building on the concept of protection domain as an abstraction of resources used to ensure the privacy of data, the SecreC language allows the specification of protection domains for different pieces of data, and the specification of the computation in domain-polymorphic manner. We have implemented the compiler for the language, integrated it with the existing SMC framework Sharemind, and are currently using it for new privacy-preserving applications.