Evaluation of Homomorphic Primitives for Computations on Encrypted Data for CPS systems

In the increasingly connected world, cyber-physical systems (CPS) have been quickly adapted in many application domains, such as smart grids or healthcare. There will be more and more highly sensitive data important to the users being collected and processed in the cloud computing environments. Homomorphic Encryption (HE) offers a potential solution to safeguard privacy through cryptographic means while allowing the service providers to perform computations on the encrypted data. Throughout the process, only authorized users have access to the unencrypted data. In this paper, we provide an overview of three recent HE schemes, analyze the new optimization techniques, conduct performance evaluation, and share lessons learnt from the process of implementing these schemes. Our experiments indicate that the YASHE scheme outperforms the other two schemes we studied. The findings of this study can help others to identify a suitable HE scheme for developing solutions to safeguard private data generated or consumed by CPS.