ENSURING DATA QUALITY AT THE EDGE

Automation, artificial intelligence, and data exchange are of primary importance in the fourth industrial revolution era. Almost every process is built starting from data; according to the "garbage in, garbage out" principle, data quality is essential to derive valuable insights for both the company and clients. This thesis presents the design and development of a Rust microservice that performs rule-based anomaly detection to ensure data integrity in an Industrial Internet of Things (IIoT) scenario. The microservice performs analysis at the edge to minimize latency and keep data local. Detected errors will be communicated to the system to make a technician aware of them and sent to the machine, too, so that it can take action whenever possible. The developed service represents only one aspect of the work; this thesis also proposes a state-of-the-art technique for deploying the service within a Kubernetes cluster. This ensures interesting capabilities such as self-healing and the possibility of a remote update of the services with automatic rollouts and rollbacks. The result will be a performant, robust service that checks data integrity at the edge, allowing flexible, customizable configuration for specific machine needs.