Quantitative Analysis of Active Cyber Defenses Based on Temporal Platform Diversity∗

Active cyber defenses based on temporal platform diversity have been proposed as way to make systems more resistant to attacks. These defenses change the properties of the plat-forms in order to make attacks more complicated. Unfortu-nately, little work has been done on measuring the effective-ness of these defenses. In this work, we use four different approaches to quantitatively analyze these defenses; an ab-stract analysis studies the algebraic models of a temporal platform diversity system; a set of experiments on a test bed measures the metrics of interest for the system; a game theoretic analysis studies the impact of preferential selection of platforms and derives an optimal strategy; finally, a set of simulations evaluates the metrics of interest on the mod-els. Our results from these approaches all agree and yet are counter-intuitive. We show that although platform diversity can mitigate some attacks, it can be detrimental for others. We also illustrate that the benefit from these systems heav-ily depends on their threat model and that the preferential selection of platforms can achieve better protection