Facilitating the Co-Evolution of Semantic Descriptions in Standards and Models

Context: Standardised specifications for sophisticated technologies are subdivided in multiple documents maintained by different working groups, typically accompanied by models and other formalised artefacts. As the specifications and the models evolve, ensuring their consistency at scale becomes challenging. Objective: While previous work developed a methodology for facilitating the co-evolution of models and standards, based on the Network Function Virtualisation (NFV) Information Model (IM) and models extracted from the related standardised specifications, the methodology focused on structural aspects only. This article refines the methodology, enabling the alignment of semantic descriptions of information elements and attributes, both across specifications and across information elements. Method: To enable the alignment of semantic descriptions, we extend the methodology by using statistical and visual analyses of terms used in the specifications. The underlying meta-model for the information extracted from the specifications is extended to accommodate the capturing of additional semantic information. Results: We report on our experiences with the application of a prototypical implementation of the methodology during the continued alignment and maintenance of the IM and the related standardised specifications. More than 400 potential inconsistencies were identified, leading to more than 100 contributions, some of which addressed multiple findings. Feedback from the working group provided insights on how to refine the methodology further. Conclusions: Models shall play a more central role and be better integrated throughout the specification development and implementation processes, helping to ensure and maintain consistency among specifications. Our experiences may provide useful insights into ongoing and future initiatives where similar challenges are faced.