Dynamic reconfiguration of service-oriented resources in cyber–physical production systems by a process-independent approach with multiple criteria and multiple resource management operations

Abstract

During the last years, the convergence of classical automation and modern information technology within cyber–physical production systems is an ongoing process. With this development, service-oriented architectures and technologies of Fog/edge and Cloud computing are now moving into the next generations of production systems. However, these innovations also create new challenges in terms of complexity and fragility. In order to ensure reliable production operation, it is necessary to adapt to any disruptions and disturbances caused by system-internal or external events. This leads to a dynamic reconfiguration of both, the operational technology and the services including the dependent information technology. The goal of this paper is to provide a system- and process-independent method for the dynamic evaluation, embedment and placement of services on the processing IT resources of a cyber–physical production system during operation. For this purpose, both production- and IT-specific requirements and properties of all resource elements involved are mapped in a unified formal system description. This includes the individual process steps as part of process control, the necessary services and their dependencies as well as the processing IT resources such as computing nodes and transmission links in the field, Fog/edge and Cloud. For the evaluation and comparison of different combinations, a suitable multi-criteria evaluation metric is defined and used for resource configuration. The unified modeling and evaluation of process steps and services enables a dynamic embedment of the most suitable available services for each individual process step. In addition, modeling and evaluating of services and processing IT resources enables services to be dynamically placed on the most appropriate IT resources. The result is a system of service embedments and placements in a heterogeneous system landscape that adapts dynamically to production- and IT-specific events. The validation of the approach is based on a case study in which dynamic changes of a process engineering production process are examined and discussed. This shows how the reliability of production in dynamically changing system landscapes can be increased.