Uncertainty in Interdependent Security Games
Benjamin Johnson, Jens Grossklags, Nicolas Christin and John Chuang
Even the most well-motivated models of information security have application limitations due to the inherent uncertainties involving risk. This paper exemplifies a formal mechanism for resolving this kind of uncertainty in interdependent security (IDS) scenarios. We focus on a single IDS model involving a computer network, and adapt the model to capture a notion that players have only a very rough idea of security threats and underlying structural ramifications. We formally resolve uncertainty by means of a probability distribution on risk parameters that is common knowledge to all players. To illustrate how this approach might yield fruitful applications, we postulate a well-motivated distribution, compute Bayesian Nash equilibria and tipping conditions for the derived model, and compare these with the analogous conditions for the original IDS model.