Strategic-level Framework with Combined Feature Engineering, Bio-inspired Optimization and ML for Detecting DDos Flooding Attacks in Cloud

Business and public enterprises have transitioned to cloud-based services. Most of the service-based companies do their core business activities using the cloud services allowing more accessibility, allowing multiple users to share the data regardless of their location. These new technologies also attract a lot of intruders and they use DDos as one of the main methods of attack. With increased amount of data and services being on cloud services Early detection of such attacks are dominant to any business. New innovation also attracts a rampant increase in the volume and frequency of DDoS attacks. New generation mechanism and technique are required for deal with new types of DDos attacks. With Increasing number of businesses, educational institutions, cultural organizations relying on cloud services, the detection system should not only be effective but also affordable to be masses. Thus, in this work, a proposed strategic-level framework that combines feature engineering mechanisms with Bio-Inspired optimization technique and machine learning. The models also are verified with opposite confirmation and tested for area-under-curve analysis. It provides a wide range of reliable solutions to avoid data overload problems while detecting direct DDoS flood attacks on Cloud. The framework is tested on the new DDoS test database (CICDDoS2019) with the DT classifier. To validate the proposed system, The other machine learning techniques are selected from literature and compared the obtained results. Test results showed that the use of feature engineering techniques with the Dragonfly optimization algorithm and Decision Tree separator achieved a high accuracy rate (98.69%) in the detection of DDoS flood attacks in a cloud environment. The results suggest that the proposed strategic level framework could achieve better results in DDoS flood attacks detection on Cloud by promising reductions in cargo loads and times.