Attack Detection and Analysis with Deep Learning in Cloud Computing

Despite the rapidly growing field of cloud technologies, security remains an ongoing issue prioritizing the detection of cyber attacks. In this paper, we seek to leverage deep learning to address the detection and analysis of the attacks within cloud computing layers. Specifically, we have developed a comprehensive set of deep learning models to identify and classify the va- riety of cyber-threats targeting different vectors of the systems. We followed our methodology by pre-processing the data, thus ensuring balanced and normalized datasets, and developed sophisticated model architectures that could discern the understated featuers and patterns in the huge wave of network traffic. For experimentation, we selected two significant datasets, the WSNBFSF and KDD dataset. These were selected due to their accurate representations of real-life network behaviors and attack vectors occurring. Using multiple essential metrics such as accuracy rate, error rate, precision, and recall, we were able to evaluate the perfor- mance of all four models. In summary, our study produced exciting results, demonstrating some exceptional performance from CNNs, which achieved an accuracy of 95% and 98% for the WSNBFSF and KDD datasets, respectively. Importantly, CNNs have shown great per- formance in the aspect of static pattern recognition between diverse network environment. Thus, all of our experiment and research proved valuable in contributing to the cybersecurity field. Our study provided detailed insight into what deep learning strategy excels and should be implemented for the efficient improvement of the current and potential attack detection in cloud security systems. In this way, the cloud services can be provided safely and efficiently, making cloud systems more available with sustainable performance. The exceptional perfor- mance of CNN’s highlights it is, and other deep learning strategies could be integrated into current cloud computing security systems to face a substantial amount of attacks and threats ravaging the cyber ecosystems.