Year: 2026
Type: conference-paper
Abstract: The paper presents the development and implementation of a hybrid cloud-fog architecture for real-time IoT analytics in pharmaceutical applications. Cloud-centric approaches face latency limitations in industrial scenarios, while pure edge solutions lack the resources for complex analytics. The proposed architecture combines the computing power of the cloud with the low latency of fog computing, enabling near-real-time processing while keeping cloud-based analytics. The implementation uses ESP32 microcontrollers with TAL220B strain gauge sensors for data acquisition, EdgeX Foundry as fog computing middleware for local preprocessing, and InfluxDB with Grafana for cloud-based storage and visualization. Experiments with a dual-platform pharmaceutical weighing system showed a mean latency of 155.55 ms from message submission to storage in the cloud database and about 440 processed messages per second at an offered load of 500 messages per second. Local buffering kept the system operating during network disruptions. The prototype hardware cost approximately 98% less than commercial weighing systems, which suggests that low-cost hybrid architectures are viable for industrial IoT deployments requiring near-real-time analytics.
Language: English
Topic: IoT and Edge/Fog Computing
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W7223268618