Home // International Journal On Advances in Security, volume 15, numbers 3 and 4, 2022 // View article

Design and Implementation of a Model-based Intrusion Detection System for IoT Networks using AI

Authors:
Peter Vogl
Sergei Weber
Julian Graf
Katrin Neubauer
Rudolf Hackenberg

Keywords: Intrusion Detection; Network Security; Internet of Things; Artificial Intelligence; Machine Learning; Deep Learning.

Abstract:
The rising digitalisation introduces many useful features, as a result of, which the vulnerability for cyber attacks increases. Internet of Things devices and networks can be used to monitor and process sensitive data but at the same time they often are not hardened against threats. This can subsequently put personal data at risk. In this paper the capabilities of an approach to combine artificial intelligence and static analysis in an intelligent Intrusion Detection System for Internet of Things networks are evaluated. The development of static and dynamic methods for attack detection in networks is additionally discussed. The architecture follows a layer-based concept. Methods of classic security analysis and artificial intelligence are therefore deployed in a modular manner. For the extraction of important features a block-based approach has been developed, in which the calculated entropy of the network traffic is used in the extraction process. Detailed insights into the methodologies to analyse port and address information as well as used tools like Snort and Snorkel are given respectively. The metadata of the network traffic and extracted features are then used in combination to further improve the performance of anomaly detection and attack classification. The various models and algorithms utilised in this process are also shown in detail. This approach demonstrates that the security of Internet of Things environments can be enhanced with the deployment of an intelligent Intrusion Detection System that uses combined methodologies of static analysis and artificial intelligence.

Pages: 86 to 95

Copyright: Copyright (c) to authors, 2022. Used with permission.

Publication date: December 31, 2022

Published in: journal

ISSN: 1942-2636