A Theoretical Framework for Distributed Secure Outsourced Computing Using Secret Sharing

Outsourced computing gives computationally-limited data owners opportunities to have third parties process the data. However, data privacy is a major concern for outsourced computing. To the end of secure outsourced computing (SOC), the information-theoretic secure multiparty computation (IT-SMC) derived from secret sharing (SS) schemes is among the most computationally efficient techniques that are candidates for secure computation. Carried out in a distributed way, a drawback of SS is its proneness to collusion attacks among computing agents. In this paper, we propose a theoretical framework using Shamir’s secret sharing (SSS) with collusion deterrence for distributed SOC. The main idea is to design communication routes between computing agents so that the collusion chain is broken, plus a mechanism that selects good agents and assigns agents in different positions in the route. The framework models real world scenarios like the cloud computing. We implement the secure millionaire protocol in our framework on a real network to demonstrate its practicality.