Track: Cyber Security

Abstract

Uncertainties in the systems are increasing at an unprecedented rate, with the advent of fast-paced changes in technology. A report which mentions the proliferation of IoT and industrial internet projects estimates that almost 50% of the manufacturing organizations will adopt Industry 4.0 standards in the upcoming years. Encapsulating the uncertain aspects of IoT and cybersecurity has become the prime focus of organizations working with digital networks. Organizations transmit sensitive information across networks and cyber security is dedicated to protecting the information systems. Dynamic modus operandi of the attackers pose a great challenge in detecting the anomalies. There are several layers of security having different levels of vulnerability and independent attributes.

In this paper, we have modeled the layers of the system as states of a Markov model. We have attempted to predict the mean time spent in the transient states. Each state of the model corresponds to vulnerable layers on the network which possess individual factors of vulnerability. The model also incorporates the effect of interactions among various layers of security. The dependency of layers is modeled as a Markov Graph. The transition probabilities of security layers are obtained using Bayesian techniques which provides computational tractability and flexibility for prior elicitation. This work will help organizations to design preventive strategies based on the attributes of the attack.