Scholarlink

The growing popularity of 5G networks has resulted in distinctive features such as network slicing, which allows for the establishment of several virtualized networks on a shared physical infrastructure. While network slicing improves flexibility, scalability, and service customization, it also creates severe security flaws, particularly for Distributed Denial-of-Service (DDoS) attacks. This research analyzes the effect of DDoS attacks on network slice performance in SDN/NFV-enabled 5G settings and proposes a predictive protection mechanism known as the Network Slice Service Protection System (NSSPS). The system uses machine learning-based classifiers, notably Decision Tree (DT) and Support Vector Machine (SVM), to detect and mitigate DDoS attacks on the NSL-KDD dataset. A real-time simulated setting replicating metropolitan infrastructure was utilized to assess the system’s resilience to several attack types, such as TCP SYN Flood, UDP flooding, HTTP flooding, and RADIUS Attack. Performance measures, including accuracy, throughput, latency, and slice isolation, were evaluated. Experimental results demonstrate that the proposed NSSPS significantly improves detection accuracy (up to 98.4%) and maintains stable service quality under attack conditions, outperforming traditional and deep learning-based models. The study contributes to 5G network security by offering a lightweight, predictive, and scalable solution to protect critical slice resources in highly dynamic environments.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Securing 5G Network Slicing: A Hybrid Machine Learning Approach for DDoS Threat Detection and Performance Evaluation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Sahar Ebadinezhad,
  • Ibrahim Khalaf Hussein

摘要

The growing popularity of 5G networks has resulted in distinctive features such as network slicing, which allows for the establishment of several virtualized networks on a shared physical infrastructure. While network slicing improves flexibility, scalability, and service customization, it also creates severe security flaws, particularly for Distributed Denial-of-Service (DDoS) attacks. This research analyzes the effect of DDoS attacks on network slice performance in SDN/NFV-enabled 5G settings and proposes a predictive protection mechanism known as the Network Slice Service Protection System (NSSPS). The system uses machine learning-based classifiers, notably Decision Tree (DT) and Support Vector Machine (SVM), to detect and mitigate DDoS attacks on the NSL-KDD dataset. A real-time simulated setting replicating metropolitan infrastructure was utilized to assess the system’s resilience to several attack types, such as TCP SYN Flood, UDP flooding, HTTP flooding, and RADIUS Attack. Performance measures, including accuracy, throughput, latency, and slice isolation, were evaluated. Experimental results demonstrate that the proposed NSSPS significantly improves detection accuracy (up to 98.4%) and maintains stable service quality under attack conditions, outperforming traditional and deep learning-based models. The study contributes to 5G network security by offering a lightweight, predictive, and scalable solution to protect critical slice resources in highly dynamic environments.