Protecting XML Databases against Ontology-Based Inference Attack

In this paper we study the ontology-based inference attack problem in large distribute XML databases. Supported by the semantically rich ontologies, the ontology-based inference attack can lead to undesired information disclosure via replicated XML data under different syntactic formats and inconsistent secure classifications. To address this new security problem, we propose a framework called Semantic-based XML Security Inference Engine (Semsie) to detect replicated XML data exposed to ontology-based inference attack. Compared to the existing works, our main technical contribution is the development of a semantic matching module based on the successful matching methods. Supported by this module, Semsie can detect the inference channels resulting from the inconsistent security classifications of replicated data automatically without basing the detecting process on any knowledge bases pre-defined manually.