A New Game Theoretic Scheme for Verifiable Cloud Computing

Outsourcing computation to cloud service providers (CSPs) has been convenient as the cloud computing paradigm becomes popular. One problem with computation outsourcing is how to efficiently verify if the returned computation result is correct. This paper studies the game theoretic approach to this problem. Specifically, we re-visit a state-of-the-art game theoretic scheme for this problem, and point out that one component of the scheme is too restrictive, which if replaced by another one, the scheme will become ineffective. To further address this flaw, we propose a new game theoretic scheme, which is probabilistic, and show that, a client can use this scheme to hire two CSPs who could collude based on any collusion contract, and the computation result can be verifiable as long as the CSPs are rational.