BigFlow-NIDS

BigFlow-NIDS, a large-scale, NetFlow-based dataset and accompanying analysis pipeline designed for intrusion-detection research in big-data environments. BigFlow-NIDS was created by merging four major benchmark NetFlow datasets (NF-UNSW-NB15-v3, NF-ToN-IoT-v3, NF-BoT-IoT-v3, NF-CSE-CIC-IDS2018-v3) using an Apache Spark preprocessing pipeline that performs deduplication, missing-value handling, label encoding, and feature harmonization. The final release contains 66,935,021 flows, 55 flow attributes, and 34 fine-grained attack categories, available in both CSV and Parquet formats to support scalable ML and streaming analyses.