Gotham Dataset 2025: Reproducible IoT Security Research

The Gotham Dataset (2025) was created to solve the 'reproducibility crisis' in IoT security research. It provides a standardized environment for testing Intrusion Detection Systems (IDS) against modern threats. The dataset includes a wide array of network traffic captures from smart home environments, simulating various device interactions and attack scenarios. It focuses on high-fidelity labeling and provides detailed metadata for every packet, including protocol flags and inter-arrival times. Gotham is particularly useful for evaluating the generalizability of machine learning models across different IoT network configurations, offering a robust platform for both academic and industrial security benchmarking.