of 3 Index of Terms - - Information Assurance , Kolmogorov Complexity , and Active Networks

The vulnerability analysis technique presented in this paper takes into account the innovation of an attacker. A metric for innovation is not new; 700 years ago William of Occam suggested a technique [6]. The salient point of Occam’s Razor and complexity-based vulnerability analysis is that the better one understands a phenomenon, the more concisely the phenomenon can be described. This is the essence of the goal of science: to develop theories that require minimum size to be fully described. Ideally, all the knowledge required to describe a phenomenon can be algorithmically contained in formulae; formulae that are larger than necessary lack a full understanding of the phenomenon. Consider an attacker as a scientist trying to learn more about his environment, that is, the target system. Parasitic computing [1] is a literal example of a scientist studying the operation of a communication network and utilizing it to his advantage in an unintended manner. The attacker as scientist generates hypotheses and theorems. Theorems are attempts to increase understanding of a system by assigning a cause to an event, rather than assuming all events are randomly generated. If theorem x, described in bits, is of length l(x), then a theorem of length l(m), where l(m) is much less than l(x), is not only much more compact, but also ( ) ( ) 2 x l m − times more likely to be the actual cause than pure chance [6]. Thus, the more compactly a theorem can be stated, the more likely one understands the underlying cause. A very literal example of an attacker as a scientist who is trying to understand a phenomenon, i.e. the system being attacked, can be seen in [1].