Precise Estimation of Local Probabilities for Bayesian Attack Graph Analysis

Abstract

A Bayesian Attack Graph (BAG) is an essential model for red teams in cyber security to detect the most vulnerable components of a system. It is a probabilistic graphical model in which each node is initially assigned a probability value called local probability. For realistic and better analysis of BAGs, it is essential to evaluate local probabilities precisely. For that purpose, in this paper, we use the Common Vulnerability Scoring System (CVSS) to estimate temporal and environmental scores. We further consider various factors reflecting attackers' characteristics in BAG analysis. In this respect, we inaugurated a new environmental variable named “host type” that influences an attacker's motivation and abolishes the need for earlier network architecture knowledge to determine the factor values.