Anticipation of DDoS Assault pattern using Deep Learning

Abstract

From the past few decades, the interconnected digital technology has offered a complete virtual atmosphere by complecting many information processing systems and smart devices. This virtual space has also been prone to numerous threats and security breaches. Despite gigantic efforts made by protectors, zero-day and other complex assaults are being dispatched consistently. A Distributed denial of service attack or DDoS attack is a network breach with abnormal and nonintrusive effort, which makes impractical for a service to be delivered to the client. This attack is the most common threat for smart meters and smart home devices due to Internet of Things (IoT). These attacks are viewed as the most well-known and frequently the most decimating ones. Attacks like these are difficult to identify and alleviate. our proposed DDoS Attack Detection problem, the code will run the code for a small number of epochs and it will complete various number of iterations where the batch size is 32 by default and we have used test set and training set to derive the computational values.