Diving into Decimals: Navigating Privacy Challenges in Floating-Point Data with Order-Preserving Encryption

Abstract

Order preservation is useful for encrypted databases because it provides a means for range queries on encrypted data while preserving the order of the plaintext in the ciphertext. This characteristic is very effective for applications that ask for data to be searched, and sorted and where other operations can be carried out within the encrypted data without the data being decrypted first. OPE is used in several encrypted database systems to enable efficient and secure search on the encrypted data keeping the original order of the data intact. This thesis proposes a novel hybrid OPE scheme based on the General Approximate Common Divisor (GACD) problem. Our approach enhances the earlier utilized OPE methods, which were restricted to solving problems of integer data nature in the past. Just like in previous work, our scheme also supports only integer data; in this work, we expand it to floating-point data, hence improving its applicability. Further, compared with other older OPE methods based on a less reliable security model, our scheme utilizes the computational hardness of the GACD problem. This approach not only improves the protection against some potential risks but also raises the standard of data protection. Our GACDP scheme shows high efficiency as it takes an average time for the encryption and decryption operation with minimal time complexity equals O (1). Results demonstrate that our scheme enhances the functioning of both integer and floating-point data types and proves to be superior to previous approaches.