Semi-supervised approach for detecting distributed denial of service in SD-honeypot network environment

Distributed Denial of Service (DDoS) attacks is the most common type cyber-attack. Therefore, an appropriate mechanism needed to overcome those problems. This paper proposed integration method between honeypot sensor and software defined network (SDN) (SD-honeypot network). In terms attack detection process, server utilized Semi-supervised learning in classification process by combining Pseudo-labelling model (support vector machine (SVM) algorithm) subsequent with Adaptive Boosting method. The dataset used this monitoring data taken Suricata sensor. research experiment was conducted examining several variables, namely accuracy, precision, recall pointed at 99%, 66%, respectively. central processing unit (CPU) usage during relatively small, which around 14%. average time flow rule mitigation installation 40s. addition, packet/prediction loss occurred attack, caused packets not be classified 43%.