EASY – Energy-efficient analysis and control processes in the dynamic edge-cloud continuum for industrial manufacturing
EASY is researching the application of federated learning to industrial manufacturing processes. This allows for the cost- and energy-efficient use of distributed computing capacities in the edge-cloud continuum. To this end, we are developing a scalable, open-source runtime environment that allows for the dynamic distribution and execution of services described according to industry standards. Salzburg Research is contributing to the analysis and prediction of connection quality between edge and cloud nodes, as well as to the detection of anomalies in industrial communication networks. These contributions will enable the dynamic placement of services in the ideal location. These contributions are being integrated into the runtime environment, and their application is being demonstrated in a laboratory prototype.
For manufacturing organizations based on the Industry 4.0 model, edge computing enables the sovereign, near-real-time processing of data at its point of origin. The significant reduction in latency associated with edge computing is driving the adoption of industrial analytics, control systems, and AI applications in production environments. This promises to increase productivity and resource efficiency throughout the manufacturing process. The EASY project aims to provide these benefits by creating an easy-to-use edge-cloud continuum that offers a runtime environment and services for executing AI-based analysis and control processes. Within this continuum, the execution of services is dynamically and distributedly optimized across the entire spectrum, from central cloud to decentralized edge instances, in terms of energy requirements, data usage, and data transfer. This makes data science and control processes easier to use for automation at the PLC and field levels, opening up new market opportunities for providers of these processes. Standardization and openness foster a comprehensive ecosystem that promotes rapid growth and international appeal while reducing market entry barriers, particularly for SMEs.
Important input parameters for optimization include available bandwidth and latency in the edge-cloud continuum. Therefore, Salzburg Research is researching efficient analysis processes to evaluate and predict the communication performance of nodes along the continuum. Additionally, federal machine learning methods will be employed to enhance anomaly detection in industrial communication networks. These services will be integrated into the runtime environment of the overall project and evaluated using a laboratory prototype. The target groups for the results include the manufacturing industry, SMEs, machine manufacturers, cloud service providers, and research institutions.
