Benchmark and Design Support for Demand-Oriented Cloud-Communication Architectures of Cyber-Physical Production Systems

The usage of cloud applications in industrial automation domain is boosted by the desire to apply computational-intensive machine-learning (ML) models and access production data from all over the world. Accomplished by the demand for powerful hardware platforms for ML and centralized data storage, the connectivity performance also increases and allows high data transmission rates. Nevertheless, requirements for data transmission, like safety, security, and real-time must be considered in line with the production use case. Finding a suitable infrastructure with communication protocols for data transmission to transform legacy production plants into cyber-physical production systems (CPPS) is overwhelming due to the many available communication methods. Design support for CPPS would increase the acceptance of cloud-based solutions by engineers of automated production systems that are not specifically familiar with high-level information technology systems. This paper introduces an investigation and design recommendations for application-layer communication protocols usually used in industrial applications. The overall goal is to support the engineering for the different automation levels on field-, edge-, and cloud-level. In this paper, design measures are firstly derived from related work. Secondly, an analysis of the timing behavior and the CPU resources are carried out. Finally, the findings are collected in a summarizing rating table that briefly suggests adjusting the performance and the impact on the overall system design by selecting suitable communication protocols.