A Threat Computation Model using a Markov Chain and Common Vulnerability Scoring System and its Application to Cloud Security

Abstract

Securing cyber infrastructures has become critical because they are increasingly exposed to attackers while accommodating a huge number of IoT devices and supporting numerous sophisticated emerging applications. Security metrics are essential for assessing the security risks and making effective decisions concerning system security. Many security metrics rely on mathematical models, but are mainly based on empirical data, qualitative methods, or compliance checking, and this renders the outcome far from satisfactory. Computing the probability of an attack, or more precisely a threat that materialises into an attack, forms an essential basis for a quantitative security metric. This paper proposes a novel approach to compute the probability distribution of cloud security threats based on a Markov chain and Common Vulnerability Scoring System. Moreover, the paper introduces the method to estimate the probability of security attacks. The use of the new security threat model and its computation is demonstrated through their application to estimating the probabilities of cloud threats and types of attacks.