D-Dos Attack Prediction Using Machine Learning Algorithms

Abstract

Abstract: The risk of cyber-attack keeps on growing irrespective of development of new technologies for protection. One of the most frequent cyber-attacks is the DOS attack. A Denial-of-Service (DoS) attack is an attack which tries to shut down a machine or network, by flooding the target with unwanted traffic or triggers a crash by sending it some information, which makes it challenging for the users to access their network. A higher version of DoS attacks is the DDoS attacks that have recently become quite severe in security companies. Many organizations have begun facing these issues. Such attacks are very well coordinated that disrupts the normal functioning of the networking system from large firms to small scale business. Hence, detecting such attacks has become a tedious task. However, such a classification problem can be resolved using machine learning. Also, the same problem can be addressed using the concepts of cloud computing in order to detect and identify the computational effort carried out by the attacks. A DoS is generally considered to be an organized attack by hackers that is implemented from a single source of origin and targeted towards the victim’s end. In order to attack these systems such attackers impersonate themselves as legit users and gain access from the users by asking them their personal credential and details. As compared to this, a DDoS attack is limited to a single source of origin and takes place on distributed computers all together. Hence the primary aim of this thesis is to identify such attacks caused by hackers and detect them using machine learning algorithms. Keywords: Denial-of-Service (DoS), Distributed Denial-of-Service (DDoS), Machine Learning, Bots, Botnets, flooding attacks