Model integration by virtualization

By Robert Bill .
Advised by Gerti Kappel

Model integration has been a recurring problem for several years now and gets even more impact due to the large amount of big, heterogenous model used in today’s industries. Virtualization has been successfully applied to data integration for a long time now and already received some attention in the modeling community, but is typically limited to providing read–only integrated information. Thus, it is proposed to go one step further and use virtualization for more aspects of model integration like synchronization and transformation. Models to be synchronized are de- fined as virtual models over their common and custom parts which might reduce the need of explicit synchronization. This new approach will need a theoretical foundation of virtual models, a new virtualization language, and implementation of this language in a framework and various strategies concerning the handling of virtual models