A Cloud-Edge Microservices Architecture for Smart Healthcare: SDN-Based Medical Asset Management
Henok Berhanu Tsegaye, Petro Mushidi Tshakwanda, Michael Devetsikiotis, Petros Spachos · 2025
Real-time hospital asset tracking and predictive maintenance are critical for ensuring patient safety, operational efficiency, and cybersecurity in modern healthcare environ-ments. Traditional network architectures often struggle with scalability, security enforcement, and real-time telemetry processing, limiting the effectiveness of medical IoT systems. To address these challenges, this paper presents a novel microservices-based cloud-edge architecture that integrates Software-Defined Networking (SDN), edge-AI, and agents for real-time medical asset tracking, anomaly detection, and adaptive policy enforce-ment. The proposed system leverages SDN for centralized traffic control and dynamic flow reconfiguration, while KubeEdge extends analytics and enforcement capabilities to the edge, ensuring low-latency decision-making and improved telemetry accuracy. Telemetry collector and security enforcement agents, deployed as Kubernetes DaemonSets, continuously monitor medical devices such as ventilators, infusion pumps, and ECG monitors. Real-time inference is powered by Autoencoder-based models, which detect anomalies with 94.5% accuracy at an average latency of 35.2 ms. Open vSwitch (OVS), integrated with Calico networking, enables programmable flow control and security isolation, achieving 99.2% policy enforcement success and reducing unauthorized access attempts by 72%. The network was deployed and validated on a three-node Kuber-netes cluster, demonstrating quantifiable improvements in fault detection, packet loss reduction, and resource responsiveness. These results position the proposed architecture as a scalable, secure, and measurement-aware framework for advancing next-generation healthcare instrumentation and intelligent telemetry systems.