# Los Alamos National Laboratory Cyber-Security Event Datasets (2015 and 2017)

## Facts

- Provider: Los Alamos National Laboratory (Alexander D. Kent and others); mirrored by Taqui
- License: CC0-1.0 (https://csr.lanl.gov/data/cyber1/)
- Size: 102 files
- Stored at: Hugging Face, https://huggingface.co/datasets/Taqui/lanl-cyber-datasets
- Access: public, readable without an account
- Category: Machine Learning & AI
- Source: https://doi.org/10.17021/1179829
- Page: https://scigantic.com/archive/68e4d90c-614d-53b2-8fe7-caa19b8e7e12
- Keywords: Machine Learning & AI, Hugging Face

## About

Byte-identical copies of two public cybersecurity datasets from Los Alamos National Laboratory, acquired from the official csr.lanl.gov source. The 2015 Comprehensive, Multi-Source Cyber-Security Events dataset covers 58 consecutive days of de-identified events from LANL's internal enterprise network: Windows authentication events (auth), process start and stop events (proc), network flows (flows), DNS lookups (dns) and a labelled set of known red-team compromise events (redteam) that serve as ground truth. The 2017 Unified Host and Network Data Set adds Windows host event logs as JSON lines (wls_day-01 to wls_day-30) and network flow CSVs (netflow_day-02 to netflow_day-30). Each file is stored both as the original .gz or .bz2 archive and decompressed; in lanl-2015/ (measured 2026-11-14) auth.txt alone is about 32 GB uncompressed, and the repository uses about 744 GB. Uses include intrusion and lateral-movement detection, anomaly detection on authentication graphs, user and host behaviour modelling and temporal graph learning. Read with pandas, polars, DuckDB or Spark. LANL waived all copyright to both datasets (CC0); cite Kent (2015) and the 2017 Unified Host and Network paper.

## Load the data

```python
from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="Taqui/lanl-cyber-datasets", filename="lanl-2015/auth.txt.gz", repo_type="dataset")
print(path)
```

## Data structure

Example file: `lanl-2015/auth.txt.gz`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/68e4d90c-614d-53b2-8fe7-caa19b8e7e12
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/68e4d90c-614d-53b2-8fe7-caa19b8e7e12
- Over MCP: `get_schema_card` with id `68e4d90c-614d-53b2-8fe7-caa19b8e7e12` at https://api.scigantic.com/mcp

## Related topics

- [Machine Learning & AI datasets](https://scigantic.com/topics/machine-learning-and-ai)
