Playing hide-and-seek
Martin D. Chapman, Gareth Tyson, Peter McBurney, G. Flucke, Simon Parsons · 2014
In order to begin to solve many of the problems in the domain of cyber security, they must first be transformed into abstract representations, free of complexity and paralysing technical detail. We believe that for many classic security problems, a viable transformation is to consider them as an abstract game of hide-and-seek. The tools required in this game -- such as strategic search and an appreciation of an opponent's likely strategies -- are very similar to the tools required in a number of cyber security applications, and thus developments in strategies for this game can certainly benefit the domain. In this paper we consider hide-and-seek as a formal game, and consider in depth how it is allegorical to the cyber domain, particularly in the problems of attack attribution and attack pivoting. Using this as motivation, we consider the relative performance of several hide and seek strategies using an agent-based simulation model, and present our findings as an initial insight into how to proceed with the solution of real cyber issues.