Publication
Detection and Classification of Network Traffic Attacks in Cloud Computing Using ML
Ihab Alzalam; Anasuya Chattopadhyay; Christoph Lipps; Hans Dieter Schotten
In: IEEE International Black Sea Conference on Communications and Networking 2026. IEEE International Black Sea Conference on Communications and Networking (BlackSeaCom-2026), June 8-11, Bucharest, Romania, IEEE, 6/2026.
Abstract
Cloud Computing is considered a decisive enabler in the architecture of fifth generation (5G) systems, contributing to the fulfillment of the stringent quality of service (QoS) requirements of current and future use cases. Despite the significant benefits, it also poses a number of security challenges and vulnerabilities that can compromise system integrity and may lead to significant losses. Integration of machine learning (ML) techniques into traditional detection systems has proven to be an effective approach to learning traffic patterns and detecting anomalies. The use of up-to-date and environment-specific datasets to train ML models significantly improves system reliability. In this work, we compare and evaluate several ML, deep learning (DL), and hybrid models that are trained on a dataset specially developed to evaluate anomaly-based network intrusion detection system (IDS) in cloud environments. The performance of the models is analyzed with metrics derived from the confusion matrix. The results demonstrate that random forest is a suitable choice for resource-constrained use cases with good overall performance, while the long short-term memory (LSTM) model and hybrid approaches that integrate LSTM outperform conventional ML models in identifying potential attacks within the used dataset.