Encrypted Traffic Dataset (BCCC-DarkNet-2025)

BCCC-DarkNet-2025 is an augmented, research-driven dataset that supports encrypted traffic analysis and threat detection across anonymized communication networks. It integrates and extends two benchmark datasets, CIC-Darknet2020 and Darknet-Dataset-2020, selected for their robust coverage of encryption protocols and darknet-specific traffic behaviors. The dataset includes diverse encrypted traffic types like VPN, Tor, I2P, Freenet, and ZeroNet, with multi-class labeling and protocol-specific annotations. These sources were chosen based on well-defined criteria, including support for time-dependent patterns, entropy measures, and behavioral features critical to identifying obfuscated malicious activities. The dataset has been standardized to ensure consistency and scalability using NTLFlowLyzer-V3, a custom flow-based feature extractor. This preprocessing step harmonizes temporal, statistical, and protocol-level features across both datasets, enabling seamless integration into machine learning pipelines. The result is BCCC-DarkNet-2025, a unified and enriched dataset that significantly improves classifier performance in encrypted traffic scenarios by capturing structural and behavioral anomalies. It is a valuable resource for developing AI-powered cybersecurity solutions in dynamic and evasive threat environments.