Machine Learning-Based Approaches for Detecting and Mitigating Distributed Denial of Service (DDoS) Attacks to Improved Cloud Security
Navya Vattikonda, Anuj Kumar Gupta, Genesis Alkali 2 Oracle ERP Senior Business Analyst, Achuthananda Reddy Polu, Cloudhub IT Solutions Senior SDE, Bhumeka Narra, Statefarm Sr Software Developer, Dheeraj Varun Kumar Reddy Buddula, Hari Hara Sudheer Patchipulusu, Walmart Senior Software Engineer · International Journal of Innovative Research in Multidisciplinary Education · 2024
Cloud environments encounter massive service disruptions together with security breaches and substantial financial losses through Distributed Denial of Service (DDoS) attacks. Detecting and mitigating DDoS assaults is the focus of this research, which examines the efficacy of ML models, particularly the CNN-LSTM model and the ID3 decision tree method. The CIC DDoS2019dataset was used for both training and evaluation, employing a train-test data split of 80:20. The hybrid CNN-LSTM model achieved superior performance than the ID3 decision tree method when subjected to comparison because it integrates CNN spatial extraction with LSTM sequence learning. A CNN-LSTM model using 0.97 recall together with 0.98 precision and 0.98F1 score achieved 98.5% accuracy in detecting DDoS attacks. Analyses indicate that the ID3 model delivered below-average results yet remained a usable solution for detection of DDoS attacks in cloud environments. These findings provide light on the utilization of decision tree algorithms such as ID3 in cloud security applications and highlight the potential of the CNN-LSTM hybrid model as a strong solution for DDoS attack detection.