Detection of DDoS attacks on time based features using Stacking ensemble technique

DDoS attacks are now among the most common and costly risks in today’s constantly evolving cyber-security land- scape. A DDoS attack occurs when a large amount of internet traffic is directed at a single server or network in an effort to overwhelm it and disrupt regular operations. According to reports from Cloudfare HTTP DDoS attacks increased by a huge percentage of 111% from 2021 to 2022. In fact, current studies and statistics predict that in 2023, the number of DDOS attacks will rise by nearly 300%. This poses a substantial risk to both corporations and individuals because these assaults have the potential to seriously harm websites that are tied to both. A defense mechanism for DDoS has become crucial for key enterprises and government organizations due to their capacity to interrupt network services and cause damages. Previously related work established that shallow and deep learning classifiers were highly useful and detecting and classifying DDoS attacks. Despite this, there hasn’t been much research on the The time-based characteristics and classification of the various DDoS assault kinds. As a result, the proposed model examines the efficiency using time-based criteria in recognising and categorising different DDoS attack types with stacking ensemble technique combining various models for the best accuracy in order to detect the DDoS attack.