Title: INFRDET: IoT network flow regulariser-based detection and classification of IoT botnet

Authors: Umang Garg; Santosh Kumar; Manoj Kumar

Addresses: Department of Computer Science and Engineering, Graphic Era Deemed to be University, Dehradun, Uttarakhand, India; Graphic Era Hill University, Dehradun, Uttarakhand, India ' Department of Computer Science and Engineering, Graphic Era Deemed to be University, Dehradun, Uttarakhand, India ' Department of Computer Science and Engineering, Graphic Era Deemed to be University, Dehradun, Uttarakhand, India

Abstract: Internet of Things (IoT) botnet is one of the attacks which affect the working of authentic IoT devices. In this paper, a novel light-weighted intelligent system has been devised by using traffic analysis and regulators to detect botnet-infected devices in the IoT network. The system operates on a low-powered Raspberry Pi device with network packet counts. Besides, an IoT Network Flow Regulariser (INFR) algorithm is proposed and embedded for transforming network flows to the uniform length traffic frame. The experimental results show the better performance of the proposed system with the INFR algorithm in comparison to the existing work. In addition, to classify the benign and malicious traffic, a novel method is used to visualise the network activities through graphical heatmaps. These heatmaps are further investigated using a hybrid Convolution Neural Network (CNN) model without and with the INFR algorithm and therefore receive remarkable differences in terms of better results.

Keywords: IoT botnet; deep learning; CNN; DDoS; VGG.

DOI: 10.1504/IJGUC.2023.135344

International Journal of Grid and Utility Computing, 2023 Vol.14 No.6, pp.606 - 616

Received: 28 Dec 2022
Accepted: 03 May 2023
Published online: 05 Dec 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article