SURVEY OF DEEP LEARNING TECHNIQUES FOR DDOS ATTACK DETECTION: METHODS, CHALLENGES, AND FUTURE DIRECTIONS

Abstract

Attacks known as distributed denial-of-service (DDoS) can seriously damage internet infrastructure by interfering with vital services and resulting in losses. Target systems are overloaded with traffic from these assaults, making them inaccessible to authorised users. This essay examines DDoS attacks, describing their several manifestations and the difficulties they pose. Next, we explore the vital field of DDoS detection, looking at both conventional and novel methods. Along with traffic pattern analysis and machine learning techniques, signature-based and anomaly-based techniques are investigated. In order to protect against the always changing threat of DDoS attacks, the paper's conclusion discusses the ongoing arms race between attackers and defenders, highlighting the necessity of constant modification and the creation of strong detection techniques.