An improved scheme for outsourced computation with attribute-based encryption

摘要

With the wide deployment of cloud computing, outsourcing complicated computational tasks to cloud service providers has attractedmuch attention. An increasing number of clients with computationally constrained devices choose to outsource their heavy tasks to cloud servers to reduce the computational overhead in local. However, how to preserve the integrity of computational results becomes a challenge since commercial cloud servers are not trusted. Public verifiability is an effective mechanism to allow clients to verify the integrity of the results returned by the servers. Because the results are sensitive in many applications, it raises the problem of privacy leakage in the public verification process. In this paper, we propose an efficient verifiable computation schemewhile keeping outputprivacy.Theproposedschemeachieves blind verifiability such that the verifiers who have not the additional information (retrieve key) can verify the integrity of the result without learning the result. Furthermore, by combining with (k, n)-threshold sharing, our scheme allows the clients jointly learn the results.