Effective and Efficient DDoS Attack Detection Using Deep Learning Algorithm, Multi-Layer Perceptron

Distributed denial of service (DDoS) attacks pose an increasing threat to businesses and government agencies. They harm internet businesses, limit access information services, damage corporate brands. Attackers use application layer DDoS that are not easily detectable because impersonating authentic users. In this study, we address novel by analyzing the characteristics incoming packets, including size HTTP frame number Internet Protocol (IP) addresses sent, constant mappings ports, IP using proxy IP. We analyzed client behavior in public standard datasets, CTU-13 dataset, real weblogs (dataset) from our organization, experimentally created datasets attack tools: Slow Lairs, Hulk, Golden Eyes, Xerex. A multilayer perceptron (MLP), a deep learning algorithm, is used evaluate effectiveness metrics-based detection. Simulation results show proposed MLP classification algorithm has efficiency 98.99% detecting attacks. The performance technique provided lowest value false positives 2.11% compared conventional classifiers, i.e., Naïve Bayes, Decision Stump, Logistic Model Tree, Bayes Updateable, Multinomial Text, AdaBoostM1, Attribute Selected Classifier, Iterative OneR.