TON_IoT: Heterogeneous IIoT Data for Cybersecurity ML

TON_IoT datasets collect heterogeneous data from a dedicated Industrial Internet of Things testbed. The testbed uses multiple virtual machines running Windows, Linux, and Kali Linux to manage interconnections between IIoT, Cloud, and Edge/Fog layers. Author Nour Moustafa created these datasets for evaluating cybersecurity applications like intrusion detection and adversarial machine learning.

Use Cases

• Training intrusion detection models based on collected network, telemetry, and operating system data.
• Evaluating threat intelligence systems based on heterogeneous data sources from the IIoT testbed.
• Testing adversarial machine learning robustness based on the multi-source dataset.
• Developing privacy-preserving models based on the described telemetry and operational data.

Strengths

• Data is collected from a dedicated IIoT testbed designed for cybersecurity research.
• Datasets integrate multiple data sources: Telemetry, Operating Systems, and Network data.
• Initial statistical evaluation suggests capability for evaluating multiple cybersecurity applications.

Limitations

• Row count, file formats, and column-level documentation are unknown.
• Last update date is unknown; freshness unverified.
• The description lacks specifics on data volume, feature dimensions, or collection time range.

Provenance

Source

Nour Moustafa

Collection Method

Collected from a proposed IIoT network testbed deployed using multiple virtual machines.