Data Security Paradigms: Paillier Homomorphic Encryption and AES in the Context of Privacy-Preserving Computation

Cloud computing offers many services to users like storing, processing, and sharing data in their day-to-day applications. The sensitive information is not protected if the data is sent directly to the cloud. Therefore, AES encryption is the de facto standard for achieving security. Though AES provides data security during transit and storage, it may not ensure privacy during data processing over the cloud as it does not possess encrypted domain processing capabilities. Alternatively, homomorphic encryption (HE) schemes achieve data privacy as they allow data processing without decrypting them. However, most of the HE schemes are still computationally very expensive compared to AES. Moreover, every application demands disparate performance, efficiency, privacy, and security requirements, and there is always a trade-off among them. Therefore, our work proposes a simple cloud architecture framework that is flexible enough to accommodate both AES and the Paillier HE encryption scheme to support varying needs of application security and privacy requirements. This work also evaluates the applicability of the proposed system with a feedback application using AES and the Paillier HE schemes with comparative performance analysis in terms of the computation and communication efficiency of the two schemes.