Scholarlink

<p>Software-Defined Network (SDN) is a revolutionary approach in overall networking architecture, implementing separation between control and data planes. The idea of centralized control, however, exponentially enhances the attack surface area, thereby making SDN vulnerable to various cyber-attack threats. To address such threats, several innovative concepts, including Machine Learning techniques, are effectively employed to enhance SDN security by implementing automated attack detection and proactive mitigation measures. This paper provides a systematic review of machine learning techniques employed to enhance the security of SDN, based on cyber-attack studies published between 2020 and 2025. This review was conducted under dual themes : (i) Machine Learning strategies with respect to involved datasets, preprocessing, and performance measures, and (ii) attack characteristics concerning general aspects, such as involved SDN planes, attack type, and affected controllers. The outcome clearly indicates the robustness and limitations, as well as the research gaps, regarding more scalable and context-aware Machine Learning-based security systems in SDN environments.</p>

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

Enhancing SDN Security Through Machine Learning: A Comprehensive Review of Cyber-Attack Detection and Mitigation Approaches

  • 期刊论文

  • 发表日期：

    2026年6月30日

查看全文
  • Hind Sounni,
  • Youssef Houali,
  • Rkia Fajr,
  • Fatima Lakrami,
  • Najib EL Kamoun,
  • Chaimaa Belbergui,
  • Houda Lifi,
  • Abdelilah Jraifi

摘要

Software-Defined Network (SDN) is a revolutionary approach in overall networking architecture, implementing separation between control and data planes. The idea of centralized control, however, exponentially enhances the attack surface area, thereby making SDN vulnerable to various cyber-attack threats. To address such threats, several innovative concepts, including Machine Learning techniques, are effectively employed to enhance SDN security by implementing automated attack detection and proactive mitigation measures. This paper provides a systematic review of machine learning techniques employed to enhance the security of SDN, based on cyber-attack studies published between 2020 and 2025. This review was conducted under dual themes : (i) Machine Learning strategies with respect to involved datasets, preprocessing, and performance measures, and (ii) attack characteristics concerning general aspects, such as involved SDN planes, attack type, and affected controllers. The outcome clearly indicates the robustness and limitations, as well as the research gaps, regarding more scalable and context-aware Machine Learning-based security systems in SDN environments.