Abstract Amid the rapid evolution of cloud computing, safeguarding data privacy has become a core priority for both academic and industrial communities. To protect user data from unauthorized access, vast volumes of sensitive information are encrypted before being transmitted to and stored on cloud platforms. Nevertheless, encryption inherently limits the platform’s ability to manipulate encrypted data, like direct search or matching on ciphertexts. Traditional Public Key Encryption with Equality Test (PKEET) schemes address this issue by enabling ciphertext equivalence verification without decryption, but they lack support for multi-ciphertext scenarios and fine-grained security. We propose a novel quantum-resistant Public Key Encryption with Multi-ciphertexts Equality Test (PKE-MET) scheme, the first to be constructed based on the Learning with Rounding (LWR) problem in lattice-based cryptography. We prove that the scheme achieves Chosen Ciphertext Attack (CCA2) security under the standard model, addressing the limitation of the existing Learning With Error (LWE)-based PKE-MET scheme which only reaches Chosen-Plaintext Attack (CPA) security. Compared with LWE-based solution, our scheme eliminates the need for complex discrete Gaussian sampling and adopts a smaller modulus. Theoretical analysis demonstrates that our proposed scheme exhibits good ciphertext scalability. Compared with the LWE-based solution, it requires only approximately 1/3 of the ciphertext storage. This reduces storage and communication resource consumption, thus lightening the operational load on cloud servers.

Abstract With the rise of cloud storage and the looming threat of quantum computing, traditional encryption methods are encountering significant challenges that hinder data manipulation without decryption. To counter quantum attacks while maintaining data manipulation capabilities, new architectures such as quantum-resistant public key encryption with equality test (PKEET) must be developed. Our study presents the initial PKEET that leverages the Learning with Rounding (LWR) problem, which provides security within standard model. We also introduce its variants, public key encryption with delegated equality test (PKE-DET) and PKEET supporting flexible authorization (PKEET-FA). Our proposals could achieve fine-grained delegation at the ciphertext-specified level compared to previous PKE-DET schemes. For example, our PKE-DET supports a delegated tester function while ensuring security against quantum computing threats. Our PKEET-FA could accord users even more controls over what ciphertexts they want to compare. Our schemes’ security is founded on the LWR problem which avoids the need for discrete Gaussian sampling, unlike the Learning with Errors (LWE) problem. This distinction renders our methods both simpler and more efficient compared to those based on LWE. Moreover, our schemes enjoy smaller-sized ciphertexts.