Cybersecurity Dataset: IoT Telemetry and Simulated Attack Data for AI Research
SHIELD Project Dataset Overview
The SHIELD project generates a curated set of cybersecurity research datasets designed to support the evaluation of advanced AI-driven penetration-testing and anomaly-detection tools. All data is produced within i46’s dedicated and isolated IoT testing environment, ensuring a fully controlled, privacy-safe dataset suitable for research, model development, and benchmarking.
What the Data Contains
The data is composed of simulated attack data in PCAP format.
To replicate realistic threat conditions, the project performs controlled cybersecurity simulations. These include:
- Port scanning
- Distributed denial-of-service (DDoS) attacks (both inbound and outbound)
- Ping-based probing
- Email-based data exfiltration
- System operations such as OS updates
Each scenario produces corresponding network traffic and system-state records that capture the behavior of devices under stress.
Compliance and Data Safety
All SHIELD datasets are synthetic and produced exclusively within a closed experimental environment.
No personal data is processed at any stage, and the dataset is fully compliant with EU cybersecurity and data-protection regulations.
Access the Dataset
The dataset is publicly accessible for research, testing, and non-commercial development.
You can download it here:
English