Federated Learning for Distributed NGN Security.CASE STUDY: Camtel NGN Security

The rapid evolution of Next Generation Networks (NGNs) has brought unprecedented capabilities in connectivity, automation, and data-driven services. However, the distributed and heterogeneous nature of NGNs—spanning cloud, edge, and IoT domains—poses significant challenges to maintaining robust cybersecurity without compromising data privacy or system latency. Traditional centralized security models struggle to cope with massive data volumes, privacy constraints, and real-time threat detection requirements. To address these challenges, this paper proposes the integration of Federated Learning (FL) into NGN security architectures, enabling distributed intelligence across network nodes while preserving data locality. In the proposed framework, participating edge devices and network entities collaboratively train a shared intrusion detection or anomaly detection model without exchanging raw data. This approach enhances data privacy, scalability, and adaptability to emerging threats. Furthermore, it mitigates the risk of single-point failures and bandwidth bottlenecks inherent in centralized learning systems. Experimental results and simulations demonstrate that FL-based NGN security models can achieve comparable or superior detection accuracy to centralized methods while significantly reducing data transmission overhead. The proposed system architecture leverages secure aggregation, robust model updates, and blockchain-based trust mechanisms to ensure integrity and resilience against model poisoning and adversarial attacks. This study highlights Federated Learning as a transformative paradigm for distributed NGN security, paving the way for intelligent, privacy-preserving, and self-adaptive defense mechanisms in future 5G and 6G networks.