Multi-Class OpenFlow Attack Detection Data for LAN-SDN

Featuring synthetic, tabular data for benign traffic and five attack types in a LAN-SDN environment. It has 36 features, including OpenFlow-dependent values, and is structured across six CSV files from different network topology experiments.

Use Cases

• Train a multi-class classifier to distinguish between benign traffic and the five attack types (link fabrication, host hijacking, host injection, ddos, portscan) using the 36 OpenFlow-dependent features.
• Analyze the impact of different network topologies on attack detection by comparing the five experiment-specific CSV files.
• Build an anomaly detection model for SDN security using the recommended combined file (LAN_SDN_NIDS.csv) which aggregates data from four topologies.

Strengths

• Includes data for five distinct attack types, enabling multi-class classification.
• Contains 36 features per record, providing a detailed feature set for analysis.
• Data is derived from five different experimental network topologies, offering scenario variety.

Limitations

• Data is synthetic and emulated, which may not fully represent real-world network traffic patterns.
• The dataset is split across six files, requiring integration for a unified view.
• The sample size (row count) is unknown, making it difficult to assess statistical power.

Provenance

Source

e-cienciaDatos Harvested Dataverse

Collection Method

Synthetic data generated by emulating a LAN-SDN with Mininet and the OpenDaylight controller.

Time Range

null

Freshness

null

Geography

null