The paper presents a set of open research issues about coupling model transformations with Cloud-based architectures, with the main objective to handle very large models.
The paper focused on two key aspects for executing model transformations on the Cloud:
– Model storage in the Cloud: how the models are stored, i.e., how to store partial models distributedly and how to hide that to the user of a model management API.
– Model transformation in the Cloud: how the transformation rules are executed distributedly, and how the distributed partial results are combine to produce a coherent output model.
We propose a set of solutions for both problems, and we think we can use existing APIs for data processing in the Cloud, such as MapReduce, to implement a Cloud-based MDE solution.
See the presentation below for a summary of the paper or read the full text