Comparative Analysis of Cybersecurity Datasets in Industrial Control Systems

Operational Technology (OT) plays a crucial role in modern industrial systems, encompassing a wide array of devices and processes that facilitate critical infrastructures. The necessity for Intrusion Detection Systems (IDS) customized for these environments is essential. Machine Learning based solutions can analyze and flag attacks within OT environments. However, it’s important to note that the effectiveness of these solutions depends on the quality of the available datasets. This study conducts a comprehensive comparative analysis of OT-IDS datasets, considering factors such as dataset composition, coverage of attack scenarios, limitations, and the representation of these attacks within the dataset. This analysis reveals shared limitations within the existing datasets, such as attack range and imbalance, while identifying X-IIoTID as the most complete dataset despite some limitations, while also highlighting the lack of datasets that capture multistage attacks.