Implementation of the IIRA industrial internet of things architecture (information domain) in SysML version 2 notation

Abstract

The formalization of the information domain of the reference architecture for the Industrial Internet of Things (IIRA) by means of the SysML version 2 systems modeling language enables a precise definition of the rules for collecting, storing, transforming, analytically processing, and providing data received from peripheral devices into decision-making systems under real-time conditions with stringent reliability requirements. The central position of this domain within the IIRA structure necessitates its representation in a notation based on the native KerML metamodel and textual form, which opens prospects for automated model generation, version control using standard source code management systems, and integration with continuous integration and deployment tools. With respect to capital construction facilities and industrial infrastructure, such formalization acquires particular significance in the context of the digitalization of the life cycle of buildings and structures, where building information modeling technologies are supplemented by telemetry streams from sensors monitoring load-bearing structures, engineering systems, microclimate parameters, and energy consumption. The developed model encompasses key functional components, including subsystems for data collection, normalization and cleaning, storage, analytical processing, visualization and provision, metadata management, and data security, which are characterized by quantitative indicators of structural complexity: the number of part definitions, attributes, and flow connections, as well as a nesting depth reaching six levels in the analytical processing subsystem. The heterogeneity of telemetry streams generated by deformation, vibration, temperature, and humidity sensors, with data volumes ranging from 0.01 to 164.8 MB per day per sensor, determines architectural decisions regarding multi-level buffering, differentiated storage policies, and a three-stage organization of analytics (edge, platform, and enterprise) with corresponding constraints on processing latency, required computational resources, and a data reduction ratio of 2.88 between the extreme levels. The assessment of coverage of the functional requirements of the IIRA information domain demonstrates a value of 91.2 percent, with a total of 138 traceability links ensuring traceability from the moment of data generation by sensors to the formation of conclusions regarding the condition of structures. Analysis of the metric characteristics reveals a high correlation between the number of component definitions and communication links, variability in the cyclomatic complexity of subsystems, and statistically significant differences in nesting depth across functional groups. This approach facilitates the creation of verifiable architectural artifacts that meet regulatory requirements for information systems in construction and establishes a foundation for model-based systems engineering that promotes interoperability, scalability, and risk reduction in the implementation of continuous monitoring systems for the condition of construction facilities within the Industrial Internet of Things.