System Architecture for Microservice-Based Data Exchange in the Manufacturing Plant Design Process

Abstract

The data exchange between numerous sub-processes and involved suppliers in the manufacturing plant design process is complex due to the use of proprietary data formats from a variety of vendor-specific tools. Consequently, the current data transfer in the plant design process is characterized by data inconsistencies and increased error rates, requiring additional revisions when incorporating multidisciplinary proprietary data formats. Therefore, this publication proposes a microservice-based system architecture approach that aims to simplify data exchange between cross-manufacturer sub-processes in the manufacturing plant design process. For this purpose, the required components of the system architecture are specified, namely the microservices, an ontology database, and a backbone. The open standard AutomationML is used as the data exchange format, as it converts the information into an object-oriented data structure that summarizes the structure, topology, attributes, and roles of the objects being described. An exemplary processing with an initial set of six microservices is presented to demonstrate the functional overview for a simplified data exchange, covering basic data management functionalities and the conversion from AML to JSON. In order to validate ideas at an early stage of the development process, Python is selected for initial development. For the selection of a suitable software framework, a list of criteria is created to evaluate different solutions to build microservice architectures. A comparison of Flask, Django, and FastAPI, three well-established Python libraries, indicates that FastAPI meets the criteria to cover database integration, security and scalability with its built-in features. The resulting system architecture shows the potential to speed up the manufacturing plant design processes and indicates flexibility and scalability through the use of microservices.