Privacy-Preserving Search in Data Clouds Using Normalized Homomorphic Encryption

By the rapid growth of computer systems, many IT applications that rely on cloud computing have appeared; one of these systems is the data retrieval systems, which need to satisfy various requirements such as the privacy of the data in the cloud. There are many proposed Privacy-Preserving search (PPS) techniques that uses homomorphic encryption to process the data after encryption, but these techniques did not take into account the possibility of repetition of some values of the features table (especially zero), even after the encryption, which makes them vulnerable to frequency attacks. On the other hand, the non-inclusion of these values may lead to the ability to infer some statistical information about the data. In this paper, we took the advantages of homomorphic encryption to encrypt the data as well as preventing any ability to infer any kind of information about the data by normalizing the histogram of the features table while maintaining the quality of the retrieval. The results showed that the proposed technique gave better retrieval efficiency than the previously proposed techniques while preventing frequency attacks.