A Comprehensive Review of Public Datasets for Machine Learning-Based Intrusion Detection in IoT and OT Networks

The rapid growth of networks for Operational Technology (OT) and the Internet of Things (IoT) has increased their susceptibility to cyberattacks. Since many years, traditional intrusion detection systems (IDS) are in place. However, they are falling short, as they are insufficient against dynamic threats. However, ML-based models are only as good as their training datasets, so these datasets are of paramount importance. Consequently, this paper reviews publicly accessible datasets for anomalies in OT- and IoTnetworks and evaluates their applicability to ML-based IDS. It discusses the strengths and weaknesses of the datasets, including data imbalance, oversimplified attack scenarios, and restrictions on protocol diversity. In order to increase the effectiveness of ML-based IDS and strengthen cybersecurity in OT and IoT environments, the paper suggests increasing dataset complexity and realism.