A practical secure outsourced computation scheme for resource constrained clients
Zihong Gong, Min Luo, Xiaolin Yang +2 more
Secure Outsourced Computation (SOC) is a computing paradigm where resource-constrained clients can offload their data to servers for compute-intensive tasks on a pay-per-work model, while ensuring the security and privacy of clients’ data. Rooted in Secure Multi-Party Computation (MPC), SOC supports an arbitrary number of clients and a wide range of computational tasks, thus attracting growing attention in recent years. Existing research has focused predominantly on specific features of SOC, failing to fully address multiple critical requirements for real-world deployment—including privacy, robustness, cost-effectiveness, and efficiency. This creates a disconnect between theoretical frameworks and practical application scenarios. In this paper, we propose a secure outsourced computation scheme based on Three-Party Computation (3PC) over a ring. Our scheme is distinguished by its support for multiple security settings, while guaranteeing Guaranteed Output Delivery (GOD) and achieving high efficiency. We further formally prove the security of our scheme under each security model in the setting of honest-majority. In addition, we present a detailed complexity analysis and perform extensive practical performance evaluations, wherein the results validate its efficiency.
Cyber Security and Applications•Jan 1, 2026•0 citations