Semantic-driven Mining of Functional Dependencies in Large-Scale Systems-of-Systems
Sprache des Titels:
Proceeding of the 5th International Conference on Information Technology and Systems (ICITS), San Carlos, Costa Rica, February 2022
Large-scale Systems-of-Systems prevalent in the area of critical infrastructures such as Intelligent Transportation Systems are characterized by massive heterogeneities. Hence, cross-system oriented monitoring of the underlying operational technology (OT) is challenging, particular due to the lack of information about functional dependencies within and between systems being an indispensable prerequisite for efficient Operational Technology Monitoring. Since existing approaches to mine functional dependencies from log files, being often the only available source, rarely address the (i) different semantics hidden in the logs, as well as, (ii) challenges prevalent in large-scale SoS, sufficiently, we put forward a novel semantic-driven mining method being able to cope with those challenges aiming to identify functional dependencies between OT objects based on the co-occurrence of events from log files. The applicability of this approach with respect to accuracy, efficiency, and effectivity has been demonstrated on the basis of a systematic evaluation based on artificially generated real-world data.