Deep belief network and support vector machine fusion for distributed denial of service and economical denial of service attack detection in cloud

Abstract

Cloud computing is a progressive technology that offers computing resources as Internet-based services, revolutionized information, and communication technologies.From an economic standpoint, this transformation is beneficial because it allows them to streamline technology infrastructure and capital costs. However, economical denial of service (EDoS) potential is a crucial impediment to cloud computing success. Several improved ways to detect EDoS and distributed denial of service (DDoS) attacks in the cloud have been presented; nevertheless, these approaches still result in a considerable reduction in detection accuracy when employed in a cloud setting. Because selecting relevant features and precise classifiers for attack detection is a challenge.We recommend using an EDoS and DDoS attack identification framework in the cloud based on optimized deep learning techniques for higher detection accuracy. The experimental results reveal True Positive Rate (TPR) varies from 98.9% to 99.8% when using deep belief network with support vector machine as a learning mechanism, while True Negative Rate (TNR) from 99.6% to 99.9%. TPR and TNR were found to have average values of 99.32% and 99.67%, respectively. At 1600 requests/s, the maximum accuracy achieved and the overall accuracy of the proposed strategy was 99.78%. K E Y W O R D Sdeep belief network (DBN), distributed denial of service (DDoS), economical denial of service (EDoS), support vector machine (SVM) INTRODUCTIONCloud computing is currently widely regarded as the most cost-effective platform for delivering big data and artificial intelligence capabilities to businesses and cloud customers through the Internet. As a result, there is no need to invest in an extensive computer system because cloud computing delivers inexpensive and scalable on-demand system requirements. 1 In contrast to the commencement of the cloud computing era, which occurred shortly after the high-performance era; today's cloud computing has evolved into a significant associate for Internet of Things (IoT), big data analytics, and intelligence business solutions, in addition to its traditional fundamental facilities. 2However, in this new technology, security is a significant worry. 3 A denial of service (DoS) attack is a common challenge in cloud security. DoS assaults raise the load on the server and render the system unusable. Because of these security concerns, many clients are hesitant to use it for their business and other uses. 4 As a result, cloud computing's popularity has been hampered by security concerns. The network that interconnects the system in a cloud, for example, must be secure. Without a doubt, sharing your data and executing your software on someone else's hard drive can look to be hazardous, posing significant hazards to an organization's data. 5 Cloud computing allows an attacker attempting to extort money from

Cloud Network Anomaly Detection Using Machine and Deep Learning Techniques— Recent Research Advancements

“…A variety of evaluation metrics such as detection rate, accuracy, recall, computational time, and precision are considered when simulating the TEHO-DBN classifier. However, the proposed system demonstrated improved performance metrics compared to existing techniques.The study[61] focused on detecting DDoS and economic denial of service (EDoS) attacks within cloud computing by combining Deep Belief Network (DBN) and Support Vector Machine (SVM) technologies. It tackled challenges such as accurately recognizing different forms of EDoS and DDoS attacks, preventing attacks from moving between Virtual Machines (VMs) and the hypervisor, and estimating attack percentages while determining sensitivity thresholds based on system requirements.…”