Two Stochastic Models for Security Evaluation Based on Attack Graph

Multiple-prerequisite graph (MP graph) is a type of attack graph that has been developed to help defending large scale enterprise network. As a middle-level attack graph, it has its unique advantages. However, quantitative security evaluations based on MP graph has not been proposed yet. In this paper, we present two stochastic models for quantitative security evaluation using MP graphs. These models are constructed based on the use of Markov Decision Process to model the attackerpsilas behaviors. The network administrators can use these two models respectively to evaluate security metrics at network designing stage and network defending stage.