Optimizing Distributed Denial of Service (DDoS) Attack Detection Techniques on Software Defined Network (SDN) Using Feature Selection

Distributed denial of service (DDoS) attacks refer to a category of cyber attacks in which an attacker attempts to overwhelm a resource, such as a server, destination, or bandwidth through various means. This is usually done by either flooding the traffic or sending a heavy crunching job to exhaust the CPU or heavy amounts of data to pileup the storage. By the rise of virtualization and cloud technologies, networks have also evolved from hardware to a piece of software which virtualizes the operation of a network equipment. This virtualization is termed as a Software Defined Network (SDN). Due to the sheer impact of cyberattacks on the performance of networks, the study of DDoS attacks on SDN is vital in the present era. Contemporary approaches such as machine learning and deep learning have proven to be effective in attack detection of DDoS in SDNs. In this paper, a preliminary analysis is carried out to study the impact of feature selection to increase the throughput and response time of Machine learning techniques in detection of DDoS attacks on SDN. The results indicate that feature selection, when applied to machine learning algorithms, is an effective approach for DDoS attack detection in SDNs.