Cloud based private data analytic using secure computation over encrypted data

Secure and efficient analytic over sensitive data becomes nowadays a big challenge for some enterprises (medical health care, insurance, etc). While these companies are collecting a large amount of sensitive data, cloud system is offering important opportunities for outsourcing of storage and computation. Thus, preserving the data privacy while maintaining the efficiency in computation at the cloud is considered an attractive research topic for computer scientists in academia as well as in industry. Homomorphic Encryption (HE) came as a new cryptographic topic that allows computation over encrypted data at the cloud side. Paillier is a famous HE scheme, that allows secure addition and average calculation over encrypted data. It presents the efficiency in implementation while preserving the security. In this paper, a new HE scheme competent for Paillier crypto-system is proposed. The presented scheme is called Secure Analytic using Vector based Homomorphic Operation (SAVHO). A deep security analysis has shown that the new scheme is resistant against several types of attacks. In addition, performance analysis has shown that SAVHO is a practical scheme by comparison with Pailler crypto-system. Different implementations are achieved under Python using SageMath Library.