NSF-BSF: SaTC: CORE: Small: Secure Massively Parallel Computations: Foundations and Constructions

Over the past decade, there has been increasing interest in processing enormous data-sets using large-scale clusters such as MapReduce and Spark. The most widely-accepted theoretical abstraction of such computations is called the Massively Parallel Computation (MPC) model. Since MPC algorithms are often executed on sensitive data, it is imperative to develop privacy-preserving techniques for efficient MPC-style computations.

The project explores a new theoretical foundation that enables efficient secure computation on MPC frameworks, as well as efficient implementations. The solutions developed provide privacy-preserving, large-scale data analytics without leaking private information.

The project combines cryptographic and algorithms techniques, and is expected to lead to new theoretical understanding of secure MPC and new concretely efficient secure MPC algorithms. The theoretical thrust explores questions such as the setup and cryptographic assumptions needed for realizing efficient secure computation on an MPC architecture, and understanding the round complexity, communication efficiency, and other suitable performance metrics.

The practical thrust explores concretely efficient instantiations of the theoretical paradigms developed, and validates their practical scalability. The project organizes joint workshops on cybersecurity for faculty and students in the US and Israel. Software tools developed in the research project are made publicly available through open source efforts. The project involves graduate and undergraduates students in this research.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.