A novel cloud architecture approach to detect network intrusions using an enhanced artificial neural network

The recent technologies like artificial intelligence and cloud services are rapidly expanding, and virtualized data centers are gaining favor as a practical infrastructure for the telecommunications sector. End customers, including numerous private and public organizations, have widely adopted and used infrastructure as a service (IaaS), platform as a service (PaaS), and software as a service (SaaS). Security remains the primary issue in cloud computing systems despite its widespread acceptance. Customers of cloud services live in constant worry of availability problems, information theft, security breaches, and data loss. With the development of machine learning (ML) tools, security applications are currently becoming more and more prominent in the literature. In this study, we investigate the applicability of enhanced artificial neural networks (ANN), a well-known ML method, to identify intrusions or unusual behavior in the cloud environment. We have designed machine learning (ML) models using an enhanced ANN approach and compared the results. The UNSW-NB-15 dataset was used to train and test the models. To reduce the complexity and training time of the ML model, we have also conducted feature engineering and parameter tuning to identify the best set of features with the highest level of accuracy. We see that the accuracy of anomaly detection achieved by an enhanced ANN approach with the right features set is 99.91%. This accuracy is greater than that seen in the literature and requires fewer features to train the model.