| Abstrakt: |
Industrial communication networks are fundamental to the development of Industry 4.0. In particular, those based on real-time Ethernet can meet the demands of time, data transmission, availability, and resources required by modern automation. However, these networks have vulnerabilities. Thus, this article aims to propose a method for detecting attacks on PROFINET networks, by extracting and selecting features for the further classification of attacks. For this, complementing the gap of related works, this methodology does not need anomalous data for the training of the intelligent classifier, proposes the extraction of original feature, and employs cloud computing to guarantee the integration between the attack detection systems and the information management. The results show approximately 100% accuracy in the proposed test scenarios. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink