关键词: 云存储服务, 全同态密码技术, 文档相似度计算, simhash, 隐私保护

Abstract:

In order to preserve user privacy in cloud storage services, we propose a method for calculating the similarity of documents under the ciphertext environment. After the data owner uploads the document ID, the ciphertext of document and the ciphertext of document simhash to Cloud servers, the cloud server performs fully homomorphic addition operations on the simhash ciphertext of the document whose similarity is expected and the simhash ciphertext of the data owner's document. Then the ciphertext of the Hamming distance between documents  is obtained. The data owner can get the results of document similarity ranking by decrypting the ciphertext of the Hamming distance. The goal of privacy preservation can be achieved by this method because the cloud server can complete  similarity calculation without any plaintext information, neither the document text nor its simhash value. We explain the proposed method in detail and the related experimental data verify its feasibility and correctness.

Key words: cloud storage service, fully homomorphic encryption technology, document similarity calculation, simhash, privacy preservation