A procedural and flexible approach for specification, modeling, definition, and analysis for self‐adaptive systems

Abstract

AbstractAn adaptive system can modify its settings at runtime as a response to changes in its operational environment. To analyse this kind of systems at design time is a difficult task since it requires considering the system together with the adaptation operations, and taking into account how such adaptations act on the system. In order to use simulation‐based techniques for the analysis of such systems, we not only need precise executable models of the systems to be analyzed, but also to capture the semantics of their adaptation mechanisms. Given the wide range and flexibility of adaptation operations, we need ways to allow the definition of new operations. We present a flexible approach for the definition and simulation‐based analysis in design‐time of adaptive component‐based systems. Our approach combines an extension of the Palladio component model in e‐Motions, a model of the adaptation mechanisms, and elastic requirements specification using the SYBL language. From the model of the system, its adaptation mechanisms, and its requirements, an executable Maude specification is generated for simulation. The application of the approach is illustrated on a use case that comprises some components and adaptations rules. The example is then analyzed using simulations. It is also shown that it is indeed possible to define additional metrics, specify adaptation requirements and rules which conduct simulations of the models in a more flexible way, and that the results of the simulation performed from these definitions can be used to carry on a valuable predictive performance analysis.