The Westermo network traffic data set.

Area of Science:

• Cybersecurity
• Network Traffic Analysis
• Artificial Intelligence

Background:

• Existing network intrusion detection datasets are often outdated or unsuitable for contemporary industrial systems.
• Research in distributed anomaly detection requires specialized, modern network traffic data.
• The Westermo network traffic dataset addresses these limitations.

Purpose of the Study:

• Introduce a novel, comprehensive network traffic dataset for cybersecurity research.
• Facilitate research in intrusion detection, anomaly detection, and distributed artificial intelligence.
• Support studies on resource-constrained artificial intelligence in edge devices.

Main Methods:

• Collected 1.8 million network packets over 90 minutes from a twelve-device industrial network.
• Provided raw data in PCAP format and pre-processed network flows in CSV files.
• Included six distinct event types: SSH, IP address issues, port scans, and man-in-the-middle attacks.

Main Results:

• The Westermo dataset offers a contemporary and detailed resource for network security analysis.
• The dataset is suitable for studying various cyber threats and network anomalies.
• It enables research into distributed and federated artificial intelligence applications.

Conclusions:

• The Westermo network traffic dataset is a valuable contribution to the cybersecurity research community.
• It specifically supports research on distributed anomaly detection and AI in edge computing.
• The dataset's contemporary nature and detailed event logging enhance its utility for diverse security studies.