Modelling Correct Operation of Webcams for Security Purposes
Blaine Billings, Xenia Mountrouidou · 2018
In October of 2016, we saw a Denial of Service (DoS) attack, the Mirai botnet, which made use of machines on a global scale, primarily targeting often-unprotected devices like webcams and routers. Due to the widespread use of the Internet of Things (IoT), and, specifically, webcams, the attack surface available to malicious actors has increased dramatically. Whereas some researchers tackle this problem by measuring and increasing the efficiency of existing Intrusion Detection Systems (IDSs) or by creating models for the purpose of characterizing cyber-attacks, such solutions do not investigate the problem of identifying when a system itself is behaving under incorrect operation. Through our research, we established a set of stochastic models that are able to accurately and efficiently model the correct operation and behavior of webcams. In order to verify the efficacy and validity of such models, we ran a multitude of normal-operation scenarios and cyber-attacks against webcams using an isolated network. Using the data from these emulated experiments, we correlated network traffic data and audit logs to verify the correctness and accuracy of our models.