In cloud storage, to ensure data confidentiality, the data owner (DO) typically encrypts files before uploading them to the cloud server (CS) for storage. By leveraging data integrity auditing, the DO can delegate a third-party auditor (TPA) to perform integrity verification on the cloud-stored data. DO stores a substantial volume of documents in the cloud, and distinct documents hold varying levels of significance for DO. Occasionally, DO may need only to audit a specific category of documents. Keyword-based data integrity auditing schemes precisely fulfill this requirement. However, existing schemes still exhibit specific flaws. In this paper, we concretely demonstrate that the recently proposed scheme (AAA-KSC) is vulnerable to tag forgery and proof spoofing attacks, and we analyze the specific limitations present in other schemes. To address these vulnerabilities, we propose a keyword-based data integrity auditing scheme (KB-DIAS), with security proofs demonstrating achievement of tag unforgeability and proof non-deceptiveness, while further incorporating dual anonymity for data owner identities and audit tasks. Finally, we compare KB-DIAS with existing keyword-based data integrity auditing schemes. The results show that KB-DIAS has better performance in terms of computational efficiency and storage overhead.
