Publication Details
Abstract
The two greatest threats of insecurity in cyberspace are botnets and a Distributed Denial of Service (DDoS) attacks. When translated in understanding it is simply a composition of compromised computer systems that have been captured by malicious intruders and used for different heinous acts. This paper discusses how IoT has become an important tool as well as a threat vector within botnet attacks. Moreover, machine learning models are used for assessment of these threats in order to have an opportunity to prevent them. Here are different artificial intelligence techniques used in the identification of botnet activity on IoT devices: Bayesian, KNN, SVM, and Decision Tree. There has also been an interest on the graph-based machine learning models for botnet identification since they can mimic host communication and are hard to attack at zero dayado, The paper also looks at the application of deep learning techniques especially the Recurrent Neural Network (RNN) coupled with BiLSTM. Also, identifying botnets and employing permissioned Byzantine Fault-Tolerant (BFT) blockchain for dynamic modelling of IoT communities is suggested. Some of the research areas of cybersecurity are detection methods for DDoS attacks and botnets, as well as features of the graph and methods of creating general models for detecting botnets based on important characteristics. Thus, the fields of DDoS and botnet detection via machine learning and deep learning have laid a firm ground for the future work. These are concepts that were first demonstrated in this work as the field’s first application of advanced computational methods on complex network security problems showing the need for further innovation due to ever increasing threats.