Labeled synthetic distributed banking- traffic dataset/code

Labeled synthetic distributed banking- traffic dataset whose per-flow feature schema (Table II) is designed to mirror the statistical structure used in established public IDS benchmarks (CICIDS2017 [12], NSLKDD [13], UNSW-NB15 [14]). The dataset spans N=30 simulated banking nodes (branches, ATM clusters, POS gateways, regional data centers), each contributing 2,000 flow records (60,000 total), with 12% of records labeled as one of five attack classes: DDoS, APT lateral movement, ransomware propagation, ATM/POS skimming, and credential brute-force. The dataset generator (generate_dataset.py) is included in the supplementary code and can be substituted with real datasets using the same feature schema.