S2MPC @ EuroSys’26

Secure MPC from a Systems Perspective:
A Hands-on Tutorial with ORQ

Abstract

Cryptographically secure multiparty computation (MPC) enables mutually distrusting entities to perform computations on their collective private data while keeping the data siloed from each other with provable security guarantees. MPC has been deployed in healthcare, education, finance, advertising, and more. The field of MPC has reached a turning point, where advancements in cryptography necessitate the development of efficient systems to achieve practical performance. At the same time, performing systems research in MPC requires a deep understanding of cryptographic protocols. This interplay between the two fields is a critical juncture, at which innovations in one area are closely intertwined with the development of meaningful solutions in the other. 

The proposed tutorial aims to educate systems researchers on recent advances in MPC and open challenges in developing, operating, and deploying MPC applications. The tutorial will consist of two parts: an introduction to fundamentals of MPC and state-of-the-art systems, and a hands-on programming session, where attendees will have the opportunity to use and deploy MPC applications in ORQ (SOSP’25). The intended audience are, but not limited to, academic and industrial computer scientists interested in cryptography, secure computation, and distributed systems. No prior knowledge of MPC concepts is required for attendance.

Schedule

9:00 – 10:30: Fundamentals of Multi-Party Computation and ORQ [Slides – Part 1]

  • Introduction to outsourced multi-party computation
  • Real-world use cases for secure analytics in untrusted clouds
  • Oblivious computations and source of information leakage
  • Vectorization and efficient communication in MPC
  • Oblivious relational analytics and oblivious multi-way joins with ORQ

10:30 – 11:00: Break
11:00 – 12:30: Hands-on Programming Session with ORQ [Slides – Part 2]

  • Deploying ORQ on CloudLab
  • ORQ example code walk-through
  • Implement and run your own secure analytics pipeline with ORQ’s API