Modular Consistency Checking Between Heterogeneous Models Without Direct Data Exchange Between Collaborators

abstract

When using model-based systems engineering to develop a system with different models, inconsistencies can occur between these models based on their information overlaps. Multiple approaches exist how to handle such multi-model inconsistencies, but all of them rely in some form on having complete access to all models or exchanging model data directly between all models. This is critical regarding confidentiality, especially when different parties, e.g., different companies, are involved in the model-based development and do not want to share all their model data with each other. Therefore, we present a concept on how to check consistency between models from different parties without directly exchanging model data between them. The concept comprises a central neutral instance that handles the consistency checks, adapters for the involved modeling tools that can trigger checks, and minimal data exchange between the central instance and the adapters based on viewtypes. The feasibility of this concept is demonstrated with a prototype implementation and evaluated on a real model environment with heterogeneous models. The concept and prototype are modular and extensible, and thus can be used to test further confidentiality preservation techniques.