MP Wellman and A Prakash
This report is superseded by a follow-on paper presented at the 2015 MTD workshop.
We investigate an adaptive cyber-defense scenario, where an attacker’s ability to compromise a targeted server increases progressively with probing, and the defender can erase attacker progress through a moving-target technique. The environment includes multiple resources, interdependent preferences, and asymmetric stealth. By combining systematic simulation over a strategy space with game-theoretic analysis, we identify equilibria for six versions of this environment. The results show how strategic outcomes vary qualitatively with environment conditions, and demonstrate the value of reliable probe detection in setting up an effective deterrent to attack.