Related Experiment Videos
BlockFedZTA: a trust-aware federated learning framework for secure multi-organizational intrusion detection
Reem Alshenaifi1, Shailendra Mishra2, Shams Tahzib1
1Department of Information Technology, College of Computer and Information Sciences, Majmaah University, Al Majmaah, 11952, Saudi Arabia.
Scientific Reports
|July 2, 2026
Summary
BlockFedZTA enhances supply chain security with a privacy-preserving intrusion detection system. This federated learning framework ensures robust threat detection despite data drift and malicious attacks.
Area of Science:
- Cybersecurity
- Machine Learning
- Supply Chain Management
Background:
- Designing privacy-preserving intrusion detection systems (IDS) for supply chains is complex due to data privacy needs, varied data distributions, and unreliable nodes.
- Existing methods struggle with data heterogeneity and potential malicious participants in federated learning environments.
Purpose of the Study:
- To propose BlockFedZTA, a novel framework integrating federated learning, XGBoost, trust-aware aggregation, and integrity verification for privacy-preserving IDS in supply chains.
- To evaluate the robustness and performance of BlockFedZTA against data drift and label poisoning attacks.
Main Methods:
- BlockFedZTA utilizes federated learning where participants train local models and share only commitment hashes (SHA-256) to protect privacy.
- A trust-aware aggregation mechanism weights updates based on validation performance, mitigating low-quality or malicious contributions.
- Integrity is verified using a lightweight commitment-based mechanism.
Main Results:
- BlockFedZTA achieved high average accuracies (0.966, 0.964, 0.963) across no-drift, moderate-drift, and severe-drift scenarios.
- The trust-aware aggregation significantly outperformed FedAvg under drift conditions (p < 0.01) and demonstrated superior performance (0.9647 accuracy) compared to other methods.
- The framework showed high resistance to label poisoning (accuracy > 0.962 at 60% poisoning) and maintained performance with scalability up to 50 clients.
Conclusions:
- BlockFedZTA provides a robust and privacy-preserving federated intrusion detection system suitable for supply chain environments.
- The trust-aware aggregation and integrity verification mechanisms enhance security and reliability against various threats and data variations.
- The framework offers a practical solution for enhancing supply chain network security without compromising data privacy.