Abstract:

With the rapid expansion of cloud computing and the exponential growth of network traffic, cybersecurity threats such as ransomware, malware, and advanced persistent attacks have become increasingly sophisticated. Traditional signature-based intrusion detection systems (IDS) are often unable to detect zero-day attacks or adapt to evolving network patterns in real time. This study proposes areal-time intrusion detection framework that leverages deep learning (DL) techniques and big data analytics to identify malicious activities in large-scale cloud networks.

The framework integrates network traffic preprocessing, feature extraction, and DL-based classification models, including Convolutional Neural Networks (CNNs) and Long Short-Term Memory (LSTM) networks, to capture both spatial and temporal characteristics of network flows. To address the challenges of imbalanced datasets and high-dimensional traffic data, the system employs data augmentation techniques and feature selection, optimizing model performance without compromising computational efficiency.

Extensive experiments are conducted using publicly available datasets and simulated cloud network environments. Preliminary results demonstrate that the proposed model achieves high detection accuracy, low false positive rates, and near real-time responsiveness, outperforming traditional machine learning-based IDS and baseline deep learning models. Feature importance analysis also provides insights into the network attributes most indicative of malicious behavior, enhancing model interpretability and explainability for cybersecurity operators.
Furthermore, the framework is designed to scale across distributed cloud architectures, supporting real-time monitoring of multi-tenant environments. By combining deep learning with big data processing pipelines, the system enables proactive threat detection, early mitigation, and improved resilience of cloud infrastructures.

In conclusion, this study illustrates how AI-powered IDS can transform cybersecurity in cloud networks, offering a scalable, interpretable, and real-time solution for detecting and mitigating evolving threats. The proposed framework provides both a theoretical and practical foundation for future research and deployment in enterprise-scale cloud systems.

Keywords: Intrusion Detection, Deep Learning, Big Data, Cloud Computing, Network Security, Real-Time Analytics

Biography

Yousef Farhaoui is a professor of computer science at the Faculty of Sciences and Techniques (FSTE), part of Moulay Ismail University, located in Errachidia, Morocco. He holds a Ph.D. in computer security and is actively involved in teaching, research, and academic leadership. He serves as the head of the IDMS research team and director of the STI laboratory, contributing to the development of advanced research in areas such as computer security, e-learning, big data analytics, and business intelligence. In addition to his academic roles, he is also a research and publishing coordinator with international academic organizations and has authored several books and book chapters with well-known publishers like Springer and IGI. Beyond his research contributions, he is highly active in the global academic community, frequently serving as a program chair, organizer, and scientific committee member for numerous international conferences related to artificial intelligence, data science, and smart technologies. He has supervised doctoral research in emerging fields such as artificial intelligence, cybersecurity, and smart agriculture, reflecting his engagement in cutting-edge technological advancements. Overall, he is recognized as an experienced academic and researcher with a strong international presence in computer science and information technology.