<![CDATA[Continuous water quality monitoring in remote and infrastructure-limited regions is constrained by the lack of reliable communication networks to support real-time IoT data transmission. Starlink offers a promising alternative due to its independence from terrestrial infrastructure, yet empirical evidence on its reliability as an end-to-end IoT communication backbone remains limited. This study therefore presents the design, implementation, and empirical network performance evaluation of a Starlink-based IoT system for real-time water quality monitoring. The proposed system integrates an ESP32 microcontroller with pH, total dissolved solids (TDS), and turbidity sensors, transmitting sensor data via a Starlink satellite link to a backend platform using the MQTT protocol with AES-128-GCM application-layer encryption. Received data are processed in Node-RED, stored in InfluxDB, and visualized through a Grafana real-time dashboard. Network performance was evaluated through five independent test iterations under both TCP and UDP transmission modes, measuring latency, jitter, packet loss, and throughput as key indicators of satellite link reliability for continuous IoT data transmission. The results demonstrate stable and reliable satellite connectivity, with latency consistently within 33–36 ms, jitter below 10 ms, zero packet loss across all configurations, and UDP throughput reaching up to 32.8 Mbps. TCP throughput was constrained to approximately 3.4–4.1 Mbps due to congestion control behavior over high-latency satellite links, a finding with direct implications for transport protocol selection in satellite-based IoT deployments. These results confirm that Starlink-based connectivity provides communication quality well in excess of the demands of periodic MQTT-based sensor transmission, demonstrating its feasibility as a reliable communication backbone for IoT-based water quality monitoring in environments where terrestrial infrastructure is unavailable or unreliable.]]>
