A Scalable MQTT-Based Edge IoT Architecture for Real-Time Distributed Solar PV Panel Monitoring

Distributed Photovoltaic (PV) installations require panel-level high-frequency monitoring to detect transient electrical and thermal anomalies that are not observable through aggregated inverter-level measurements. Conventional Supervisory Control and Data Acquisition (SCADA) systems rely on centralized architectures, polling-based industrial protocols, and aggressive data aggregation, making them unsuitable for distributed and rooftop PV deployments operating under intermittent network connectivity. Although Internet of Things (IoT)-based monitoring solutions offer low-cost and flexible alternatives, most existing implementations remain cloud-centric and rely solely on Message Queuing Telemetry Transport (MQTT) broker-level Quality of Service (QoS), which does not guarantee durable persistence at the application database layer. This study presents a scalable MQTT-based edge IoT architecture for real-time distributed solar PV monitoring that guarantees true end-to-end data reliability. The proposed system integrates panel-level electrical and environmental sensing with disk-backed edge persistence and an application-level acknowledgment mechanism (MQTT-ACK) that issues delivery confirmation only after successful fog-side database insertion. By coupling acknowledgment with durable storage rather than broker reception, the proposed architecture explicitly decouples sensing reliability from network availability. Experimental evaluation using real-world PV data demonstrates complete data recovery during network disruptions, bounded synchronization delay upon reconnection, and low computational and communication overhead on both edge and fog nodes. The results confirm that combining lightweight MQTT communication with edge-level persistence and application-level acknowledgment provides a practical, low-cost, and resilient alternative to centralized SCADA platforms and cloud-centric IoT PV-monitoring systems.