Edge Computing Architecture for Real-Time Water Quality Monitoring in Aquaculture Ponds

Abstract

Shrimp and fish farmers need early warning when dissolved oxygen or pH drifts outside safe ranges, but rural ponds often have unreliable internet connectivity. We propose an edge computing architecture in which low-cost sensor nodes report to a pond-side gateway that detects anomalies locally and synchronises summaries to the cloud when a connection is available. The prototype was deployed in eight ponds for 90 days. Local processing reduced alert latency from a median of 41 seconds with a cloud-only design to 3 seconds, and cut cellular data use by 86%. Farmers received 97% of critical dissolved-oxygen alerts within one minute, including during network outages. The architecture shows that edge processing makes continuous water quality monitoring practical in areas with limited connectivity.