An optimized feature selection guided light-weight machine learning models for DDoS attacks detection in cloud computing
Rahul Rajendra Papalkar, A.S. Alvi, Shabir Ali, Mohan Awasthy, Reshma Ramakant Kanse · 2023
The investigation highlights the need of lightweight machine learning models for efficient and effective cloud based VNF intrusion detection. The study suggests utilising lightweight machine learning models guided by feature selection to spot distributed denial of service (DDoS) attacks in the cloud. The proposed strategy uses the Extreme Gradient Boosting (XGBoost) approach to create a lightweight machine learning model, and a genetic algorithm to choose relevant characteristics for it. We put the proposed technique through its paces by simulating DDoS attacks on the cloud with a data set. The results showed that out of all the machine learning models evaluated, XGBoost&s;s optimised feature selection guided performance was the most effective and efficient. This study contributes to the field of cybersecurity by examining the necessity of intrusion detection in conjunction with the increasing significance of cloud-based VNF systems for high-speed network operations. The proposed lightweight machine learning approach may direct the development of more efficient and effective intrusion detection systems to protect cloud based VNFs from security threats. To evaluate the method&s;s usefulness in real-world cloud-based VNF systems and to explore its potential application to other kinds of cyber threat, more research is necessary.