Scholarlink

<p>Domain Name System (DNS) amplification attacks are the most rampant and harmful Distributed Denial-of-Service (DDoS) threats, exploiting open DNS resolvers to devastate targets with massive amount of traffic volumes to disrupt online services and triggering significant financial damage. Classical detection and mitigation methods often grapple with keeping pace with the evolving complexity of these attacks; prompting the investigation of Deep Learning, (DL) approaches for enhanced accuracy and efficiency. This Systematic Literature Review (SLR) employed standard Preferred Reporting Items and Meta-Analysis (PRISMA) guideline for scrutinizing the recent developments in Deep Learning-based techniques for identifying and mitigating DNS amplification attacks, analyzing state-of-the-art methods, model architectures and dataset used in their effectiveness, challenges, and future directions. Scopus, Association of Computing Machinery (ACM). Google Scholar, ieeeXplore, Science Direct, Multidisciplinary Digital Publishing Institute (MDPI), and Springer are the identified database repositories. Through a structured survey of recent peer-reviewed articles published between 2019 and 2025, this study ascertains the search strings including (“Domain Name System (DNS) amplification” OR “Domain Name System (DNS) reflection”) AND (“Deep Learning (DL)” OR “neural network” OR “Convolutional Neural Network (CNN)” OR “Long Short-Term Memory (LSTM)” OR “Recurrent Neural Network (RNN)”) AND (“Detect” OR “Mitigate” OR “Prevent”) in order to identify key Deep Learning models which have better performance that have been engaged for attack identification. The systematic review emphasizes performance metrics used in the literatures to evaluate model efficiency. Moreover, it examines real-world deployment challenges, such as dataset constraints, computational expense, scalability, and adversarial bypass techniques. Finally, it recommends directions in leveraging Deep Learning (DL) as contribution in filling the gap of detection and mitigation Domain Name System (DNS) amplification attack.</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

Deep learning-based detection and mitigation of DNS amplification attacks: a systematic literature review

  • 期刊论文

  • 发表日期：

    2026年5月23日

查看全文
  • Getachew Kefelegn,
  • Ketema Adere

摘要

Domain Name System (DNS) amplification attacks are the most rampant and harmful Distributed Denial-of-Service (DDoS) threats, exploiting open DNS resolvers to devastate targets with massive amount of traffic volumes to disrupt online services and triggering significant financial damage. Classical detection and mitigation methods often grapple with keeping pace with the evolving complexity of these attacks; prompting the investigation of Deep Learning, (DL) approaches for enhanced accuracy and efficiency. This Systematic Literature Review (SLR) employed standard Preferred Reporting Items and Meta-Analysis (PRISMA) guideline for scrutinizing the recent developments in Deep Learning-based techniques for identifying and mitigating DNS amplification attacks, analyzing state-of-the-art methods, model architectures and dataset used in their effectiveness, challenges, and future directions. Scopus, Association of Computing Machinery (ACM). Google Scholar, ieeeXplore, Science Direct, Multidisciplinary Digital Publishing Institute (MDPI), and Springer are the identified database repositories. Through a structured survey of recent peer-reviewed articles published between 2019 and 2025, this study ascertains the search strings including (“Domain Name System (DNS) amplification” OR “Domain Name System (DNS) reflection”) AND (“Deep Learning (DL)” OR “neural network” OR “Convolutional Neural Network (CNN)” OR “Long Short-Term Memory (LSTM)” OR “Recurrent Neural Network (RNN)”) AND (“Detect” OR “Mitigate” OR “Prevent”) in order to identify key Deep Learning models which have better performance that have been engaged for attack identification. The systematic review emphasizes performance metrics used in the literatures to evaluate model efficiency. Moreover, it examines real-world deployment challenges, such as dataset constraints, computational expense, scalability, and adversarial bypass techniques. Finally, it recommends directions in leveraging Deep Learning (DL) as contribution in filling the gap of detection and mitigation Domain Name System (DNS) amplification attack.