Security Analysis on One-to-Many Order Preserving Encryption-Based Cloud Data Search

Abstract

For ranked search in encrypted cloud data, order preserving encryption (OPE) is an efficient tool to encrypt relevance scores of the inverted index. When using deterministic OPE, the ciphertexts will reveal the distribution of relevance scores. Therefore, Wang et al. proposed a probabilistic OPE, called one-to-many OPE, for applications of searchable encryption, which can flatten the distribution of the plaintexts. In this paper, we proposed a differential attack on one-to-many OPE by exploiting the differences of the ordered ciphertexts. The experimental results show that the cloud server can get a good estimate of the distribution of relevance scores by a differential attack. Furthermore, when having some background information on the outsourced documents, the cloud server can accurately infer the encrypted keywords using the estimated distributions.