A Model-driven approach to design and verify autonomic network behaviors

The domain of autonomic networking comes with its own set of architectural challenges and requirements that necessitates the knowledge of other domains like protocol design, system engineering and modern control theory for the design of autonomic behaviors. Thus, there is a need for an unified framework and methodology for designing the autonomic entities of such networks. A Model-driven framework or approach appears as a viable option, as it integrates and masks the differences between the domains through meta-models and model-to-model transformation services. In this paper, we focus on the instantiation of one such methodology and associated tool-chain for the design of the Auto-Configuration behavior in a network. We showcase how the application of the methodology helps in detecting and resolving conflicting and entangled control-loops during the design phase of autonomic behaviors.