At the Dagstuhl Seminar on Models@run.time Professor Giese gave a talk on »Models at Runtime for Adaptive and Self-managing Software« together with Thomas Vogel.
Software systems have to be continuously adapted to their changing requirements or environments. This is typically done by maintenance, while self-aware, context-aware, ultra-large-scale, or mission-critical software systems often have to be adapted during operation (self-adaptive software). Model-driven engineering and models at runtime play crucial roles in supporting maintenance, enabling self-adaptation, and integrating both, maintenance and self-adaptation.
In our talk, we outline our perspective on "models at runtime" as any models used on-line, i.e., internal to the running system, to represent running software, to represent the software's environment, or to manipulate or analyze any of the former two. Moreover, we sketch the benefits of models at runtime for maintenance, self-adaptation, and for an integrated setting of maintenance and self-adaptation. Considering different kinds of runtime models required for self-adaptation and self-adaptation activities as model operations, we discuss the role of a megamodel for specifying adaptation loops for self-adaptive software and operationalizing the treatment of runtime models as means for monitoring, analysis, planning, and executing planned changes.