Supporting fine-grained generative model-driven evolution

Abstract

In the standard generative Model-driven Architecture (MDA), adapting the models of an existing system requires re-generation and restarting of that system. This is due to a strong separation between the modeling environment and the runtime environment. Certain current approaches remove this separation, allowing a system to be changed smoothly when the model changes. These approaches are, however, based on interpretation of modeling information rather than on generation, as in MDA. This paper describes an architecture that supports fine-grained evolution combined with generative model-driven development. Fine-grained changes are applied in a generative model-driven way to a system that has itself been developed in this way. To achieve this, model changes must be propagated correctly toward impacted elements. The impact of a model change flows along three dimensions: implementation, data (instances), and modeled dependencies. These three dimensions are explicitly represented in an integrated modeling-runtime environment to enable traceability. This implies a fundamental rethinking of MDA.