Formal Methods for Modeling, Refining and Verifying Autonomic Components of Computer Networks

The domain of autonomic and nature-inspired networking comes with its own set of design challenges and requirements for its architectures. This demands a tailored solution to model and design its components rather than a generic approach. In this paper, we provide a hybrid methodology consisting of formal methods to design, refine and verify the entities of autonomic networks. We focus our discussions on the methods for

meta-modeling

,

structural modeling

and

behavior modeling and design

of existing protocols and newly introduced autonomic components, that autonomically manage and adapt the behaviour of protocols to changing policy and network conditions. A case study, based on the recently introduced Hierarchical Autonomic Management and Control Architectural Framework called GANA, is used for highlighting the practical benefits and design choices available to modelers and autonomic components designers. The results of our case study are analyzed to explain the trade offs that future designers would be forced to make in order to achieve their design objectives for an autonomic network. A tool-chain to realize the methodology is also briefly discussed.