A Multilayer Evolutionary Homomorphic Encryption Approach for Privacy Preserving over Big Data

Abstract

One of the biggest impediments that prevent the evolution of big data is the privacy of users. Many advanced researches are done within this topic and a lot of concepts had seen the light. One is a cryptographic concept known as homomorphic encryption which allows the application of operations on ciphered data without need to decipher it. However, from the cryptographic aspect, the homomorphic encryption has its defects which make it a potentially solution, in fact some researches proved the inefficiency of those cryptosystems against some kind of attacks such as attacks with chosen plaintext (IND-CPA) and attacks with chosen ciphered text (IND-CCA) and even for the majority of homomorphic cryptosystems which use user's identity attacks of chosen identity. On the other, a new type of cryptosystems was recently introduced where he aim is to improve the classic cryptography techniques, such as substitution and transposi-tion using evolutionary methods of data mining, e.g., genetic algorithms. The efficiency of this kind of schemes was proved IND-CPA and IND-CCA. In this paper, we improve the efficiency of a homomorphic cryptosystem known as TSZ (To, Safavi-Naini, and Zhang) by proposing a new approach that combines between it and evolutionary cryptography in order to use the advantages of these two categories.