A Comprehensive Survey on Deep Learning-Based DoS/DDoS Detection and Resource Management in 5G Network Slicing

Abstract

With the rapid advancement of 5G communication technology, network slicing has emerged as a key enabler for supporting diverse applications with varying Quality of Service (QoS) requirements. It allows multiple virtual networks to operate on a shared physical infrastructure, ensuring efficient resource utilization and service flexibility. However, this flexibility also introduces significant security challenges, particularly from Denial-of-Service (DoS) and Distributed Denial-of-Service (DDoS) attacks that can disrupt services and degrade network performance. This paper presents a literature survey on network slicing, intrusion detection systems, deep learning- based attack detection, and intelligent resource allocation in 5G environments. Various techniques such as Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), and Reinforcement Learning (RL) are analyzed for traffic classification and attack mitigation. The study highlights their advantages, limitations, and identifies future directions for improving security and efficiency in 5G networks.

Introduction

Key insights from the literature:

• Deep learning for traffic classification and attack detection: DNNs, CNNs, and LSTMs are widely applied for application-specific RAN slicing and real-time DDoS detection. CNNs capture spatial traffic patterns, while LSTMs model temporal dependencies, achieving high detection accuracy (up to 99.97%). Techniques like DeepSecure and DeepDefense integrate these models with SDN for traffic isolation and mitigation.
• Intrusion detection and mitigation: ML-based IDS monitor network traffic anomalies in real time, improving detection accuracy and reducing false positives. Slice isolation mechanisms prevent attacks from propagating across slices, maintaining service continuity.
• Dynamic resource management: Reinforcement learning (CMDP-based) and hierarchical allocation models improve adaptability, throughput, and fairness in multi-tenant slicing environments.
• Architectural approaches: End-to-end slicing frameworks integrating SDN and NFV enable flexible resource allocation, scalability, and slice isolation, complementing ML/DL-based methods.

Conclusion

This literature survey highlights the significant role of deep learning and intelligent techniques in enhancing the security and efficiency of 5G network slicing. Approaches such as CNN, LSTM, reinforcement learning, and slice isolation mechanisms effectively improve DDoS detection, resource allocation, and overall network performance. Among these, deep learning models demonstrate high accuracy in identifying complex attack patterns, while reinforcement learning offers adaptability in dynamic network conditions. The integration of these techniques contributes to improved reliability, scalability, and service quality in modern 5G environments.

References

[1] P. Du and A. Nakao, \"Deep Learning-based Application Specific RAN Slicing for Mobile Networks,\" The University of Tokyo, Tokyo, Japan, IEEE, 2018. [2] G. Iashvili, M. Iavich, R. Bocu, R. Odarchenko, and S. Gnatyuk, \"Intrusion Detection System for 5G with a Focus on DoS/DDoS Attacks,\" International Conference on Information and Telecommunication Technologies and Radio Electronics (UkrMiCo), 2021, pp. 1–6. [3] M. P. Novaes, L. F. Carvalho, E. R. Hruschka, and J. P. Papa, “Long short-term memory and fuzzy logic for anomaly detection and mitigation in software-defined network environment,” IEEE Access, vol. 8, pp. 83765–83781, 2020. [4] N. A. E. Kuadey, S. Khan, R. Ullah, and S. H. Ahmed, “DeepSecure: Detection of distributed denial of service attacks on 5G network slicing—Deep learning approach,” IEEE Wireless Communications Letters, vol. 11, no. 3, pp. 488–492, Mar. 2022 [5] D. Sattar and A. Matrawy, \"Towards Secure Slicing: Using Slice Isolation to Mitigate DDoS Attacks on 5G Core Network Slices\", IEEE Conference on Communications and Network Security (CNS), 2020, pp. 1–9. [6] Y. Liu, J. Ding, and X. Liu, \" Resource Allocation Method for Network Slicing Using Constrained Reinforcement Learning,\" IFIP Networking Conference (IFIP Networking), 2021, pp. 1–3. [7] S. O. Oladejo and O. E. Falowo, \"5G Network Slicing: A Multi-Tenancy Scenario,\" Global Wireless Summit (GWS), 2017, [8] pp. 88–92. [9] A. Ahalawat, S. S. Dash, A. Panda, and K. S. Babu, \"Entropy Based DDoS Detection and Mitigation in OpenFlow Enabled SDN,\" International Conference on Vision Towards Emerging Trends in Communication and Networking (ViTECoN), 2019, [10] pp. 1–5. [11] X. Yuan, C. Li, and X. Li, “DeepDefense: Identifying DDoS attack via deep learning,” in 2017 IEEE International Conference on Smart Computing (SMARTCOMP), pp. 1–8, 2017. [12] P. Du and A. Nakao, \"End-to-End Network Slicing for 5G Mobile Networks,\" IEEE Conference, University of Tokyo, Tokyo, Japan, 2018. [13] X. Li et al., \"Network Slicing for 5G: Challenges and Opportunities,\" IEEE Internet Computing, vol. 21, no. 5, 2017. [14] T. Yoo, \"Network Slicing Architecture for 5G Network,\" International Conference on Information and Communication Technology Convergence (ICTC), 2016, pp. 1010–1014. [15] S. Wang, X. Zhang, Y. Zhang, et al., \"A Survey on Mobile Edge Networks: Convergence of Computing, Caching and Communications,\" IEEE Access, 2017. [16] D. Sattar and A. Matrawy, \"Security in Software-Defined Networking: DDoS Detection and Mitigation,\" IEEE Confernece on Communications and Network Security (CNS), 2019.

Copyright

Copyright © 2026 Ambidi Naveena, Meghana Madipalli, Narayandas Shreya Vandana, Srivarsha Pochampally, Sneha Battula. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.