A Thorough Cybersecurity Dataset for Intrusion Detection in Smart Water Networks
Riccardo Taormina, Andrés Murillo, Nils Ole Tippenhauer, Stefano Galelli · 2022
The increase in the number and complexity of cyber-physical attacks on water distribution systems requires better intrusion detection systems. So far, the design and validation of such systems has relied on datasets, such as the EWRI 2017 BATtle of the Attack Detection ALgorithms (BATADAL), that provide a detailed representation of hydraulic processes in response to cyber-physical attacks. However, the BATADAL, generated with epanetCPA, does not include an equivalent and detailed representation of the processes occurring within the industrial communication system. Here, we fill in this gap by presenting the BATADAL 2.0 dataset, generated with the DHALSIM simulator, a novel co-simulation environment that can represent the hydraulic processes, digital control, and network communication of smart water networks. The dataset includes a broad variety of attacks and network anomalies. Most importantly, the availability of both process and network data is expected to pave the way to more advanced and accurate detection algorithms.