EUVIMEDEUVIMED entdecken
Crossref

NS-ZTFedIDS: A Neuro-Symbolic Federated Zero Trust Framework for Explainable Intrusion Detection in SDN and Optical Communication Networks

Applied IT & Engineering · 01.12.2025 · 3 (1) · 1-8

DOI: 10.25163/engineering.3110775 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Applied IT & Engineering, online erschienen am 01.12.2025.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Modern software-defined and fiber-optic communication networks face an accelerating threat landscape that traditional, centralized intrusion detection systems were simply not designed to handle. The architectural concentration that makes SDN operationally powerful simultaneously expands its attack surface, while privacy constraints, regulatory boundaries, and the sheer heterogeneity of distributed infrastructure make centralized data pooling increasingly impractical as a security strategy.

Methods

This study proposes NS-ZTFedIDS — a Neuro-Symbolic Zero Trust Federated Intrusion Detection System integrating Federated Learning with FedAvg aggregation and differential privacy (ε = 0.5), Graph Attention Networks for topology-aware anomaly detection, Zero Trust contextual features (Policy Compliance Score, Identity Confidence Score, Micro-Segment Boundary Crossing, Session Risk Tier), neuro-symbolic guardrails for physics-aware autonomous response validation, and SHAP/LLM-based explainability. The framework was evaluated across three benchmark datasets — NSL-KDD, CICIDS2017, CIC-IDS2018 — and simulated SDN/optical traffic, comprising 1,200,000 flow records distributed across ten non-IID federated nodes.

Results

NS-ZTFedIDS achieved 97.3% classification accuracy, F1-score of 0.969, ROC-AUC of 0.988, and a false positive rate of 2.8%, outperforming all five baselines — including a centralized equivalent. Lateral movement detection reached F1 = 0.962. Ablation confirmed that Zero Trust features and graph encoding were the primary performance drivers.

Conclusion

NS-ZTFedIDS demonstrates that federated privacy preservation, graph-based reasoning, policy-aware symbolic constraints, and explainability can be co-integrated without architectural compromise, offering a credible path toward autonomous, trustworthy cyber defense for critical communication infrastructure.

Zitieren

APA

(2025). NS-ZTFedIDS: A Neuro-Symbolic Federated Zero Trust Framework for Explainable Intrusion Detection in SDN and Optical Communication Networks. Applied IT & Engineering, 3(1), 1-8. https://doi.org/10.25163/engineering.3110775

Vancouver

NS-ZTFedIDS: A Neuro-Symbolic Federated Zero Trust Framework for Explainable Intrusion Detection in SDN and Optical Communication Networks. Applied IT & Engineering. 2025;3(1):1-8. doi:10.25163/engineering.3110775

BibTeX
@article{euvimed2025,
  title = {NS-ZTFedIDS: A Neuro-Symbolic Federated Zero Trust Framework for Explainable Intrusion Detection in SDN and Optical Communication Networks},
  journal = {Applied IT & Engineering},
  year = {2025},
  volume = {3},
  number = {1},
  pages = {1--8},
  doi = {10.25163/engineering.3110775}
}
Kostenlos starten →