Model Integrated Computing in the Large

As model integrated computing gains wider and wider acceptance, the scalability of the supporting tools becomes a significant issue. Large, complex projects involve many developers who create large and complex models. Supporting large-scale model integrated computing requires two key features currently lacking in available tools: (1) distributed, simultaneous multi-user access to the models; and (2) model versioning. Our proposed solution is based on storing the models at a relatively fine granularity in XML files on a server under the control of a traditional source code control system. The model builder tool then only needs to implement the client functionality of the version control tool. However, there is one significant technical challenge that needs to be solved. Models are captured as complex interdependent data structures in a model editor. Containment, binary and n-ary relations are kept as a consistent set of data by the model builder. The model builder tool needs to keep track of these interdependences and ensure consistency at all times even when multiple users edit the models