Abstract
With the rapid development and popularization of Internet of Things (IoT) devices, an increasing number of cyber-attacks are targeting such devices. It was said that most of the attacks in IoT environments are botnet-based attacks. Many security weaknesses still exist on the IoT devices because most of them have not enough memory and computational resource for robust security mechanisms. Moreover, many existing rule-based detection systems can be circumvented by attackers. In this study, we proposed a machine learning (ML)-based botnet attack detection framework with sequential detection architecture. An efficient feature selection approach is adopted to implement a lightweight detection system with a high performance. The overall detection performance achieves around 99% for the botnet attack detection using three different ML algorithms, including artificial neural network (ANN), J48 decision tree, and Naïve Bayes. The experiment result indicates that the proposed architecture can effectively detect botnet-based attacks, and also can be extended with corresponding sub-engines for new kinds of attacks.
Highlights
The traditional attack detection systems cannot be competently relocated in the Internet of Things (IoT) environments because of the different nature of such devices, and the diverse architecture of the underlying network methodologies with the conventional network
A botnet attacks detection framework with sequential architecture based on machine learning (ML) algorithms is proposed for dealing with attacks in IoT environments
We found that the detection performance is the best if the rectified linear unit (ReLU) function [23] was applied on hidden nodes, and the sigmoid function [23] was assigned on the output node
Summary
The Internet of Things (IoT) devices have been rapidly developing in recent years, and it makes our daily activities more convenient. The average of IoT devices was attacked once every two minutes, according to the Symantec report [2]. According to the Kaspersky report [3], they could collect 121,588 malware samples for IoT devices in 2018, around four times more than in 2017. Many of such malware is strong and dangerous to IoT devices. The attacker can circumvent the signature-based approaches, and these mechanisms cannot guarantee to detect the unknown attacks and the variances of known attacks. We proposed a machine learning-based botnet attack detection architecture. The experiment results indicate that the detection accuracy of our proposed system is high enough to detect the botnet attacks. It can support the extension for detecting the new distinct kinds of attacks
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
