NodeRank: An Algorithm to Assess State Enumeration Attack Graphs

An intruder needs to find and utilize a combination of vulnerabilities in order to launch a successful attack. Attack graphs have been very useful tools to model the vulnerabilities of network, and can demonstrate all possible routes that an attacker can utilize to achieve a specific goal. However, for real systems, Attack graph are not fit for human users to comprehend owing to its size and complexity, the user cann't distinguish which route is more susceptible to be attacked intuitively from the attack graph. So to solve this, we propose an algorithm named NodeRank in state enumeration attack graphs. The rank value of the nodes shows the likelihood of an intruder reaching this state. Given all the rank values of an attack graph, the users can focus on relevant critical paths to determine how to deploy security measures. In our algorithm, not only the state nodes' network topology relationship is put into consideration, but also we consider the effects of nodes' intrinsic attributes, and here we take the vulnerabilities' CVSS (Common Vulnerability Scoring System) marks as the algorithm's input parameters to describe nodes' intrinsic attributes. Then we did an experiment and put our algorithm into practice, the experiment results show that our algorithm is valuable for users of attack graph.