Logistic Regression Based Multizone Classification Framework to Detect Mitigate and Prevent HTTP DDoS Attacks in the Cloud
Loading...
Date
item.page.authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Cloud computing is the modern technological trend today. The cloud computing offers many advantages such as easy to use, pay for the usages, anywhere access, availability, low operational cost investment, elasticity, etc. Though the cloud gives many advantages still has its issues and concerns. The important issues and concerns related to the cloud are Data Security,Distributed Denial of Services (DDoS) Attacks, Data location, Vendor lock-in, etc. The application layer DDoS attacks are complicated to distinguish, lessen and block as they look like normal behaviour and cannot be detected in any of the previous Open System interconnection (OSI) layers. The HTTP DDoS attacks detection, mitigation and prevention are one of the serious and most challenging issues in the cloud environment due to impersonating the normal behaviour of the user. The characteristics of cloud such as Multitenancy, Virtualization, Computing Model, etc, also make it complex to detect those attacks. This research work focused on designing a framework to cover the complexities involved along with all possible HTTP DDoS attack paths in the cloud as well as detecting, mitigating and preventing such attacks in the cloud environment. The novel logistic regression based multizone classification model is proposed to detect HTTP DDoS attacks. The same has been integrated into the designed cloud framework to safeguard online web-based applications in the cloud ecosystem. The proposed solution is implemented using OpenStack cloud software and FIFA WorldCup 98 real-time data set is used in the experiment for generating HTTP DDoS flooding attacks