Representing Technical Standards as Knowledge Graph to Guide the Design of Industrial Systems

Abstract

Technical standards help software architects to identify relevant requirements and to facilitate system certification, i.e., to systematically assess whether a system meets critical requirements in fields like security, safety, or interoperability. Despite their usefulness, standards typically remain vague on how requirements should be addressed via solutions like patterns or reference architectures. Thus, software architecture design remains a time-consuming human-centered process. In this work, we propose an approach on how to use knowledge graphs for supporting software architects in the design of complex industrial systems. We discuss how project-generic knowledge (e.g., technical standards) and project-specific knowledge like the description of a concrete system can be modeled as knowledge graph. Making the architectural knowledge, which is currently present in technical standards and other resources, machine-readable, enables the support of the software architect through expert systems and therefore, improve the quality of the overall system design. However, since architectural knowledge is currently presented in many different formats, the transformation to a uniform, machine-readable form is required. We demonstrate the applicability of our approach with a representative example of an industrial client-server architecture and outline research challenges for future work.