Metamodel Mashups: Synchronizing Across Modeling Languages with Model Hub
Models are everywhere. Our challenge is to figure out how to interpret them, even when they aren't expressed in the modeling language we expect.
Service design, service-oriented API management and master data management are among the areas most in need of a mapping and synchronization solution that can work with different types of models. We want to manage service interfaces and data types in a canonical form, but we need to recognize existing models that are already expressed in other forms: UML, XML Schema, ERD, and various DSLs. Moreoever, these models may be living documents, with ongoing evolution in their original forms, so a one-time import into a new API catalog may not be an option.
With a heterogeneous, constantly changing landscape of models, how can we hope to maintain a master catalog of service APIs and data type definitions that reflects the current state of these models?
Model Hub is an emerging commercial software platform for large-scale model integration. In the early stage of its evolution, the ModelSolv development team is focusing on model synchronization, using a flexible set of mapping structures and algorithms that work across different metamodels.
This talk will describe how Model Hub implements this solution using Eclipse modeling technologies including EMF, UML2, EMF Compare, ATL and XTend. We will describe the use cases driving this solution inside a leading European bank. We will discuss some possible applications of model mapping and synchronization in other contexts, and talk about the future evolutionary path for this technology.