Design and Performance Evaluation of a Low-Cost Edge-To-Cloud Telemetry System for Industrial Monitoring
Abstract
Industrial facilities increasingly rely on continuous telemetry acquisition to support condition monitoring, operational awareness, and predictive maintenance. However, many existing edge-to-cloud monitoring systems employ complex communication architectures involving message brokers and middleware, which can increase deployment complexity and computational overhead for low-cost embedded platforms. This study presents the design, implementation, and experimental evaluation of a lightweight edge telemetry system that directly transmits industrial sensor measurements from an ESP32 DevKit V1 to the InfluxDB Cloud time-series database using HTTP. The developed platform integrates an MLX90614 infrared temperature sensor for real-time temperature monitoring, while vibration and electrical measurements were incorporated to validate the complete multi-parameter telemetry pipeline. Telemetry data were transmitted over Wi-Fi, stored in InfluxDB Cloud, and visualized through Grafana for continuous monitoring. Experimental validation was conducted for approximately eleven hours using a one-second sampling interval under representative industrial operating conditions. Performance evaluation demonstrated a mean sensor acquisition time of 1.62 ms, stable free heap memory averaging 230,129 bytes, a mean HTTP communication latency of 2.63 s, and a mean Wi-Fi received signal strength of −22.20 dBm. Successful HTTP response code 204 and continuous telemetry storage confirmed successful edge-to-cloud communication throughout the monitoring period. Analysis of the exported telemetry dataset further demonstrated stable sensor acquisition, firmware execution, wireless connectivity, and cloud data storage. Compared with conventional broker-based architectures, the developed system provides a simpler direct edge-to-cloud telemetry architecture while supporting continuous communication and scalable time-series data storage. The findings demonstrate the capability of low-cost embedded hardware to support continuous industrial telemetry and provide a practical foundation for scalable edge-to-cloud industrial monitoring applications.