Dynamically Adaptive Systems are Product Lines too: Using Model-Driven Techniques to Capture Dynamic Variability of Adaptive Systems
Nelly Bencomo, Pete Sawyer, Gordon S. Blair, Paul Grace · Lancaster EPrints (Lancaster University) · 2008
In this paper we propose an approach to support the design and operation of dynamically adaptive systems. We apply the concept of variability modeling from software product lines to define how systems adapt at runtime to changes in their environment. Our approach models two dynamic variability dimensions; environment variability, which defines the conditions under which a system must adapt, and structural variability which defines the resulting architectural configurations. The variability dimensions identified are modeled using domain-specific languages (DSMLs) tailored to adaptive middleware technologies that provide support for runtime variability according to reconfiguration policies. We describe our experience with applying this approach through a case study; the design of a flood warning system.