Autonomous Botnet Detection and Defense in SDN Environments

The rapid growth of Internet of Things (IoT) devices has significantly changed the face of modern digital systems by facilitating the development of smart applications across multiple fields ranging from healthcare to smart homes and industry.Nevertheless, this rapid growth has resulted in serious security threats to modern digital systems.Due to limited computational power and poor security capabilities, IoT devices are highly vulnerable to cyber attacks.Attackers use this weakness to infiltrate IoT devices and develop botnets that can be used to carry out large-scale attacks such as Distributed Denial of Service (DDoS) by disrupting services by sending excessive traffic to the affected systems.Traditional security systems, ranging from signature-based systems to rule-based systems, are not effective in addressing the challenges associated with IoT devices due to their inability to efficiently deal with the dynamic and ever-changing nature of IoT systems.Additionally, traditional systems are not effective in dealing with the large amounts of data associated with IoT devices.As a solution to this challenge, this project proposes the use of machine learning for botnet attacks.The data collected by the system is network traffic data from IoT devices, which contains various features such as IP addresses, packet sizes, protocol types, and connection time.This data is then preprocessed.After the preprocessing of the data, various feature extraction and normalization techniques are employed.These techniques are used for the enhancement of the quality of the data.Also, data balancing techniques are employed.These techniques are necessary for the fair training of the model.After the data is prepared, multiple models are trained and tested for the classification of the network traffic as malicious or normal.The proposed system is useful for the enhancement of the accuracy of the results, the minimization of false positives, and the early identification of botnet attacks.