Security concerns are progressively increasing with the unstoppable penetration of technology in everyday life, as the world becomes ever more interconnected. Hackers, attracted by lucrative targets, perform more and more sophisticated attacks. Serious doubts are surfacing regarding the effectiveness of many (if not most) security strategies and techniques.

Semantic approaches to security want to overcome many evident limitations of common security mechanisms (for example, lack of semantics) by applying contextual reasoning and inference. In this paper, the authors propose an ontology-based technique to detect and classify different categories of attacks. This semantic detection system is capable of making intelligent decisions by analyzing the context of a web application and its underlying protocols. This dynamic semantic analysis, performed on incoming and outgoing messages, shows good results supporting complex and zero-day attack detection.

The paper is very interesting and inspiring. The proposed approach intrinsically enables analysis at a semantic level inside a promising technological environment. The representation over rich data models looks convincing, and the dynamic contextual processing proposes great potentialities in perspective.