Develop a Methodology for Cyber-Electronic Warfare Battle Damage Assessment (BDA) using Game Theory
Navy SBIR FY2014.1
||Navy SBIR FY2014.1|
||Develop a Methodology for Cyber-Electronic Warfare Battle Damage Assessment (BDA) using Game Theory|
250 Apollo Drive
Mount Airy, North Carolina 27030
||The Cyber Battle Damage Assessment Tool (CyBDAT) will allow warfighters to quantify the probabilities associated with CNA, CNE and EA and facilitate for the direct comparison of cyber attacks to kinetic attacks in mission planning.
CyBDAT creates a directed graph Markov attack model of the cyber attack mission from a user provided disablement chain for the mission and a model of the target network. This is accomplished using a multi-dimensional taxonomy of the CNA/CNE/EA attack space and a library of attack scenario templates.
CyBDAT include several significant innovations and enhancements to the current state-of-the-art. First, it is an application of gaming theory to offensive cyber. Second, CyBDAT used a very detailed attack model and simplified game theory models is rather than the opposite which is common in defense modeling. Third, a GUI based comparison tool allows users to view the results of cyber and kinetic attack modeling on a graphical view of the disablement chain with the probability of disablement and the comparison of non-weaponeering effects, and allows for the user to. The comparison tool will allow the warfighter to make a completely informed decision on the attack method to be chosen, and to plan for back-up attacks, if necessary, to increase the probability of overall mission success. In Phase I we will design the comparison tool and implement a bare bones version to enable the proof-of-concept prototype demonstration.|
||CyBDAT is designed to produce probability of disablement estimates for complex cyber-physical systems in a manner analogous to how these estimates are produced in kinetic weaponeering tools. In this hypothetical example, a power plant is targeted to be disabled. In Kinetic Weaponeering, the power plant is represented by a disablement chain and that model is used. The weapon damage mechanisms, accuracies, effectiveness, and target damage sustainment information, along with release conditions, number of weapons used, aim-points, and other information is input into the tool, and a kinetic probability of disablement for that scenario is produced. CyBDAT uses the same target disablement chain and links that with a model of the target network to produce a CNA/CNE/EA model (described below) that is used as inputs for a Gaming Engine to produce cyber probability of disablement. These probabilities of disablement for all cyber and kinetic scenarios are fed into a comparison tool which allows for inclusion of non weaponeering factors such as psychological affects, compromise of vulnerabilities, etc.|