Sustainable Smart Facility: A Real‐Time Internet of Things‐Based Environmental Monitoring and Predictive Control System Using Amazon Web Services Cloud Services

ABSTRACT

Industrial facilities face critical environmental hazards—including temperature surges, humidity imbalances and combustible gas leaks—that existing monitoring solutions fail to anticipate, responding only after threshold violations occur and leaving a window of risk between hazard onset and system response. This paper presents Sustainable Smart Facility (SSF), a novel Internet of Things (IoT)–cloud architecture combining real‐time environmental sensing with a predictive anomaly‐detection layer to enable proactive hazard mitigation. An ESP32 microcontroller continuously samples DHT11 temperature/humidity and MQ2 gas sensors, transmitting readings via secure MQTT over transport layer security (TLS) to AWS IoT Core. A serverless AWS Lambda function applies both rule‐based threshold logic and a lightweight Z ‐score anomaly detection algorithm to classify environmental states and issue actuator commands—controlling a ventilation fan and a servo‐operated window—before critical thresholds are breached. Sensor data are persisted in Amazon DynamoDB for real‐time querying and archived in Amazon S3 in JSON format for longitudinal analysis via AWS Glue and Amazon Athena. Amazon simple notification service (SNS) delivers immediate email alerts to facility managers upon threshold violation or anomaly detection. Comprehensive experiments over a 30‐day deployment demonstrate end‐to‐end latency of  ms, MQTT delivery reliability of 99.3%, Lambda execution time of  ms, anomaly‐detection F 1‐score of 93.0% and a false‐positive rate of 2.1%. Scalability tests confirm sub‐linear latency growth to 50 concurrent sensor nodes below 800 ms. Comparative evaluation against five state‐of‐the‐art IoT monitoring systems shows SSF achieves the best balance of responsiveness, reliability, predictive capability and cost efficiency ($0.0023 per 1000 readings).