Imagine a factory or research laboratory filled with machines, each equipped with its own sensors and controllers. Traditionally, data from these devices is sent to central servers for processing. While effective, this approach often introduces delays and network bottlenecks. Industrial Edge Computing addresses the problem by bringing computation closer to where data is generated. Instead of relying on distant servers, applications run directly on edge devices installed near the equipment, allowing for real-time decision-making. These industrial PCs can host applications that collect diagnostics, perform analytics, or even take direct control of equipment. For example, a ventilation system can adjust airflow more efficiently by running advanced control algorithms locally on an edge device using nearby sensor readings.

At CERN’s CMS experiment, data quality has traditionally been certified by human operators reviewing run after run. However, with the increasing volume of data, this manual process is becoming a bottleneck, paving the way for machine learning to take a more prominent role.

To ensure datasets are reliable for physics analyses, the CMS collaboration uses Data Quality Monitoring (DQM) software. This software analyzes raw detector output and generates concise summaries, known as monitor elements. These include histograms of sensor signals, basic statistics about detector performance, and plots that highlight unusual or unexpected behavior.