A rational and reliable model for outsourcing polynomial two-party computation

Abstract

Outsourcing computation is a key technology for optimizing resource utilization and handling complex data tasks, especially when local resources are insufficient. However, service providers may trigger dishonest computation and data leakage due to selfishness, while clients may be reluctant to outsource computation due to high processing costs and malicious behaviour of servers. Addressing these issues, we propose a polynomial computation scheme based on game theory that achieves privacy computation and verifiability. Specifically, we formally construct a traditional two-party computation game model, analyse the benefits and motivations of the participants, and conclude that servers will generate selfish behaviours to break the protocol in order to maximize their benefits, resulting in damage to clients’ interests. Next, we propose a rational two-party polynomial computation protocol for efficient privacy computation between servers, and ensure the correctness of the computation based on a sampling verification technique and a deposit mechanism. Finally, game analysis proves that our scheme effectively constrains the selfish behaviour of service providers and conserves clients’ verification costs. Simulation experiments show that our scheme reduces the computation cost by more than 30% compared to other schemes.