Robust Intrusion Detection System to Protect Cloud Services

Cloud data centers are normally designed to handle dynamic work load and any type of large scale operation. But some of the characteristics of the datacenters make them vulnerable to Distributed Denial of Service attacks. An attack which involves data centers as target not only affect that center but also all the services hosted within that data center. Service availability will be severely affected by this attack. For a robust network defense mechanism, it is critical to detect and classify these attacks. The era that exists today is of Generative Artificial Intelligence and other technologies, which is very helpful to complete various tasks. Which can also be used for the effective detection of Distributed Denial of Service attack. To detect these attacks this project, propose a novel approach using BERT model. Transformer based architectures like BERT can be used to analyze various attack scenarios and predict potential attacks even it can specify the type of the attack. BERT is fine-tuned using a dataset containing both normal and Denial of Service attack from that the model can learn the subtle patterns which is helpful in distinguishing the types of these attack. In this project we compared the results with traditional models and BERT. From that we can see how BERT performs very well in the detection. Real time detection will provide a solution to overcome these attacks and provide a more robust defense against Distributed Denial of Service attacks.