Scholarlink

<p>Distributed denial of service (DDoS) attacks are numerous and potentially devastating for Internet-connected machines, which cause possible threats and exhaust the network resources to make them unavailable for legitimate users. Therefore, implementing algorithms for detecting and classifying DDoS attacks is crucial for security. Although some techniques have been introduced for detecting attacks, they are still limited in dealing with advanced attacks, whose methodologies constantly change and evolve over time. According to the work in this paper, an algorithm has been introduced to detect and classify DDoS attacks. The proposed model is considered a hybrid of the Gated Recurrent Units (GRU) and Bidirectional Long Short-Term Memory (BiLSTM) deep learning model, called GRU-BiLSTM. It aims to extract deeper spatial and temporal features of the input data to achieve accurate detection and obtain a low False Positive Rate using the CIC-DDoS2019 dataset. This dataset provides comprehensive coverage of diverse DDoS attacks and addresses vulnerabilities in existing datasets. The proposed GRU-BiLSTM model achieves a significant improvement in detecting attacks when compared to other standard algorithms by achieving 99.99% accuracy and 0.01% False Positive Rate. Thus, the proposed model strongly guarantees to secure networks against DDoS threats. </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

A Deep Learning GRU-BiLSTM for DDoS Attack Detection

  • 期刊论文

  • 发表日期：

    2025年6月30日

查看全文
  • Amal M. Al-Eryani,
  • Fatma A. Omara,
  • Eman Hossny

摘要

Distributed denial of service (DDoS) attacks are numerous and potentially devastating for Internet-connected machines, which cause possible threats and exhaust the network resources to make them unavailable for legitimate users. Therefore, implementing algorithms for detecting and classifying DDoS attacks is crucial for security. Although some techniques have been introduced for detecting attacks, they are still limited in dealing with advanced attacks, whose methodologies constantly change and evolve over time. According to the work in this paper, an algorithm has been introduced to detect and classify DDoS attacks. The proposed model is considered a hybrid of the Gated Recurrent Units (GRU) and Bidirectional Long Short-Term Memory (BiLSTM) deep learning model, called GRU-BiLSTM. It aims to extract deeper spatial and temporal features of the input data to achieve accurate detection and obtain a low False Positive Rate using the CIC-DDoS2019 dataset. This dataset provides comprehensive coverage of diverse DDoS attacks and addresses vulnerabilities in existing datasets. The proposed GRU-BiLSTM model achieves a significant improvement in detecting attacks when compared to other standard algorithms by achieving 99.99% accuracy and 0.01% False Positive Rate. Thus, the proposed model strongly guarantees to secure networks against DDoS threats.