A Deep Learning Approach Against Botnet Attacks to Reduce the Interference Problem of IoT

Abstract

Purpose: Today we are witnessing a world where hacking into a user’s computer using tiny bots or intercepting a group of interconnected devices is no more impossible. These tiny bots are called botnets which are a group of malicious codes that can hamper the whole security system without the knowledge of the user. As Internet of Things (IoT) is emerging rapidly, the interconnected devices are susceptible to breach, as one affected device can hamper the whole network. The security threat increases as botnet attacks increase their presence to the interconnected devices. In this paper, we are implementing Restricted Boltzmann Machine (RBM) algorithm of deep learning approach on the CTU-13 dataset (The CTU-13 is a dataset of botnet traffic that was captured in the CTU University, Czech Republic, in 2011. The goal of the dataset was to have a large capture of real botnet traffic mixed with normal traffic and background traffic.) to train the algorithm about the botnet attack patterns in IoT and to prevent the botnet attacks on IoT devices, thus reducing the interference problem in the network.

Methodology: In this paper, we worked on a deep learning-based botnet detection algorithm, which trains the IoT devices few ways for future research have been distinguished. To show the capacity of our created model to identify new varieties of botnets, a changed adaptation of the Torii sample will be utilized in the next phase of the work, to produce a second combined dataset and will be looked at against existing mark and stream-based oddity location strategies.

Conclusion: This paper proposes a solution to the detection of botnet activity within consumer IoT devices and networks. A novel detection algorithm was developed based on Deep learning mechanism. Earlier detection was performed at the packet level with Wireshark by creating a fake network using ApateDNS.

Paper Type: View Point