PERFORMANCE MODELING OF NETWORK DEFENSE IN BREADTH SYSTEMS BY MATRIX REWRITING SRN WITH FUZZY PARAMETERS

We have defined in this paper, a new kind of stochastic reward network (SRN) by introducing matrix attributes and fuzzy parameters of timed transitions and of rewriting rules, called MFRSRN, allowing the dynamic marking-dependent reconfiguration of these models. Moreover, this formalism offers a descriptive language that allows managing the compact representations the model's size through the introduction of matrix structuring and rewriting mechanisms of the compositional model's behavior. As an example of application, we show how the proposed formalism can be applied to performance modeling of network “Defense in Breadth” system which includes the combination of firewall, IDS, honeypot and moving target defense (MTD) techniques, taking into account the probabilistic and epistemic uncertainty aspects.