Extending the analysis of the FlipIt game

Abstract

In this thesis, we extend the game theoretical analysis of the FlipIt game¹. The game was first articulated and analyzed by Marten van Dijk, Ari Juels, Alina Oprea and Ronald L. Rivest. FlipIt (or otherwise The Game of "Stealthy Takeover") is a game-theoretic framework for modeling computer security scenarios. Such scenarios include targeted attacks, cryptographic key rotation, password changing policies and cloud auditing. What we are particularly interested in are situations in which an attacker repeatedly takes control over a system or critical resource completely, learns all its secret information and this is not immediately detected by the system owner. As mentioned in the original paper', FlipIt is a two-player game between an attacker and a defender, in which both players try to control a shared resource. The players do not know who is currently in charge of the resource until they move; this is why it is called "Stealthy Takeover." Moreover, each player pays a specific cost every time he moves. The objective of each player is to maximize his benefit; this involves controlling the resource for as large fraction of time as possible and, at the same time, minimizing the total move cost. Here, we are only considering scenarios in which the defender plays with a renewal strategy; in particular, the intervals between his consecutive moves are given by a fixed probability distribution. However, the attacker can be more sophisticated than that and deploy adaptive strategies; i.e. he moves based on feedback received during the game. The main strategy we are considering for the defender is the biperiodic as stated in Section 4. For the attacker, we are considering various non-adaptive and adaptive strategies. Finally, we compare it to the relevant strategies already analyzed in the original FlipIt paper¹.