Efficient searchable public key encryption against keyword guessing attacks for cloud-based EMR systems

Due to the tremendous benefits of cloud computing, an increasing number of health care providers tend to deploy the electronic medical record (EMR) storages and application services into cloud. To protect patients' privacy, sensitive EMRs have to be encrypted before being uploaded to cloud. This makes effective utilization of EMRs, such as plaintext keyword search, a very challenging problem. Public key encryption with keyword search (PEKS) provides a promising cryptographic solution to encrypted EMR data retrieval, because it allows one to delegate to an untrusted storage server the capability of searching on publicly encrypted EMR data without compromising the security of original EMR data. Recently, two secure channel free PEKS schemes were proposed for cloud-based EMR systems. However, our cryptanalysis demonstrates that both of these two schemes suffer from the security vulnerabilities caused by the keyword guessing attack. To deal with this problem, a novel secure channel free PEKS scheme is developed in this paper. The proposed scheme not only provides the resistance to the existing known three types of keyword guessing attacks, but also has the merit of no designated server. In the standard model, it is formally proven to achieve both the keyword ciphertext indistinguishability and the trapdoor indistinguishability under the adaptive chosen-keyword attacks. The comparisons indicate that the scheme is secure and practicable.