An Empirical Analysis of Kernel-Level Security Techniques: An IoT Perspective

Abstract

Securing devices connected in an Internet of Things (IoT) based network majorly refers to the application of cryptographic algorithms on the application layer. The cryptographic layer can provide a high level of node-to-node communication security, but if an IoT node is attacked internally then there are limited chances for the node to survive any further communication attacks. There are many causes of internal attacks, which include but are not limited to physical attack, bugs in the IoT operating system which can lead to remote code execution, inherent open-ports kept by the programmers for future Trojan horse attacks, etc. Due to these internal attacks, application-level security techniques might be insufficient. Thus, there is a need for kernel-level security for IoT devices. In this review, we have compared various kernel-level security algorithms and evaluated their performance when applied to real-time scenarios. Moreover, some of these algorithms are applied to IoT systems while some are not. So, in this text, we have proposed ways to import the non-IoT based security algorithms and apply them to a generalized IoT system. We have also made some fine observations about these techniques and recommended ways in order to improve them using state-of-the-art algorithms.

 Keywords: IoT, security, kernel, internal attacks, Trojan horse