GOSVM: Gannet optimization based support vector machine for malicious attack detection in cloud environment

Cloud computing is the most useful computing technology for the new service progression. Due to the distributed nature of cloud computing, security threats and cyber attacks are major problems that penetrate the network and cause sudden harm to the financial and business accounts by affecting the servers. Commonly, the vigorous and stable growth of cloud computing is mainly affected due to the security issues in a cloud computing environment. Various malware variants generate cyberattacks. Thus, intrusion detection technology is employed for securing cloud computing from malicious attacks. Cloud Intrusion Detection System (CIDS) identifies malicious attack behavior and guarantees the security and reliability of cloud computing. But, virtual network flow is uncontrollable and unnoticeable among the virtual machines. This paper proposes a novel Gannet Optimization Algorithm-based hybrid Support vector Machine-Extreme Learning Machine (GOA-optimized hybrid SVM-ELM) technique to identify and prevent malicious attacks in a cloud computing atmosphere. Gannet Optimization Algorithm (GOA) optimizer is adopted for selecting optimal features and for minimizing the loss of information. The parameters of the hybrid SVM-ELM model are optimized by the GOA algorithm. The proposed architecture improves the overall security and performance of a cloud-based Intrusion Detection System (IDS). This proposed technique is used for classifying the different attacks like Normal, web attack, Brute Force, Infiltration, Portscan, DoS/DDoS, and Botnet ARES and is executed using Matlab through employing the CICIDS2017 dataset, and the evidence detection in cloud forensics dataset. The results revealed that high precision, F-Measure, and recall rate are obtained in this technique. Then the training time decreased to 389 ms on the evidence detection in the cloud forensics dataset and 399 ms on the CICIDS2017 dataset.