Subversion-Resistant and Consistent Attribute-Based Keyword Search for Secure Cloud Storage

Kai Zhang,Zhe Jiang,Jianting Ning,Xinyi Huang

DOI: https://doi.org/10.1109/tifs.2022.3172627

IF: 7.231

2022-01-01

IEEE Transactions on Information Forensics and Security

Abstract:Secure cloud search service allows resource-constrained clients to effectively search over encrypted cloud storage. Towards enabling owner-enforced search authorization, the notion of attribute-based keyword search (ABKS) has been introduced and widely deployed in practice. To enhance traditional security of ABKS, two state-of-the-art solutions are presented to address keyword guessing attacks or setup inconsistency for secret key. Nevertheless, they have not simultaneously considered the following threats to a data user: (i) inconsistent secret key/cipher-index caused by outside dishonest authority and/or data owner; (ii) algorithm substitution attacks (ASA) launched by inside adversarial eavesdropping. These attacks may unfortunately lead to cloud data breach and user information exposure. To tackle such outside and inside threats, we introduce subversion-resistance and consistency for secure and fine-grained cloud document search services. In particular, we propose a consistent ABKS system with cryptographic reverse firewalls (CRF). Technically, we refer to verifiable functional encryption and employ non-interactive zero-knowledge proofs of discrete logarithm equality to ensure strong input consistency for ABKS. In addition, we build a trusted CRF zone for sanitizing algorithm outputs against ASA attacks. Moreover, we formalize the security model and formally prove security of our system. To clarify practical performance, we implement state-of-the-art solutions and our system in real cloud environment based on Enron dataset. The results show that our system achieves more enhanced security properties without obviously sacrificing performance. In particular, our system achieves comparable time and storage cost for document-index encryption and document search, as compared to state-of-the-art solutions.