Real time performance analysis on DDoS attack detection using machine learning

Abstract

In recent years, Distributed Denial of service (DDoS) attacks have led to a tremendous financial loss in some industries and governments. Such as banks, universities, news and media publications, financial services, political or governmental servers. DDoS attack is one of the biggest threats for cyber security nowadays. It is a malicious act that slows down the server, makes loss of confidential data and makes reputation damage to a brand. With the advancement of developing technologies for example cloud computing, Internet of things (IoT), Artificial intelligence attackers can launch attacks very easily with lower cost. However, it is challenging to detect DDoS trafic as it is similar to normal trafic. In this era, we rely on the internet services. Attackers send a huge volume of trafic at the same time to a speci c network and make the network null and void. So that the server cannot respond to the actual users. As a result, clients cannot get the services from that server. It is very essential to detect DDoS attacks and secure servers from losing important information and data. However, many detection techniques are available for preventing the attack. But it is very challenging to choose one method among those as some are time efficient and some are result oriented. In our paper, we mainly focused on the top machine learning classification algorithms and evaluated the best model according to the dataset. The experimental result shows that the Decision Tree algorithm achieved the excellent accuracy of 98.50 percent with very less time consumption. Therefore, we are using a better approach to detect DDoS attacks in real time.