An Effective DOS Attack Detection Model in Cloud Using Artificial Bee Colony Optimization

Abstract

Denial of service (DOS) attack is a serious threat in the cloud which causes unavailability of cloud services to genuine users. Firewalls alone are not able to detect DOS attack because of the dynamic nature of the attack and masked identity. This paper proposes a cloud-based DOS attack detection model (CDOSD) which is setup through key feature selection using a new binary version of Artificial bee colony optimization (BABCO) and decision tree (DT) classifier. The DT classifier is utilized because it has superior learning speed than other classification algorithms and BABCO is used for feature selection from the dataset. The real-time DOS attack tools are used to perform the attacks on cloud host. It has been observed that the CDOSD detects DOS attack on cloud host with high accuracy and a very low false positive rate. The features of the dataset are significantly reduced by BABCO which provides a low dimension of computation space for training and classification. The proposed scheme is also compared with the other existing models and found superior in performance. The proposed methodology may help the cloud service providers to design more secure cloud environment.