Scholarlink

<p>Distributed Denial of Service attack detection is critical due to the growing number of compromised Internet of Things devices used to flood target networks with overwhelming traffic, disrupting services, and preventing legitimate access. Researchers have developed various detection strategies, yet the approaches faced drawbacks of generalizability issues, computational complexities, lack of scalability, and flexibility. Hence, the research proposes a Stochastic Foraging Movement Optimized-Bottleneck Cross Attention-based Convolutional Network and Bidirectional Long Short-Term Memory (SBC2TM) model for DDoS attack detection. However, the proposed SBC2TM model determines the subtle and complex patterns characteristic of Distributed Denial of Service attacks in dynamic Internet of Things network environments effectively. Moreover, the integration of the Bottleneck Cross Attention (BnCA) mechanism in the model enhances the accuracy while detecting the attacks, and it distinguishes between normal and malicious traffic. Additionally, a Stochastic Foraging Movement Optimization (SFgMO) algorithm is introduced for tuning the hyperparameters of the model that resulted in optimal detection performance. Notably, the SBC2TM model exhibited significant enhancement over the existing approaches by achieving an accuracy of 98.42%, F1-score of 98.13%, precision of 98.31%, recall of 98.15%, sensitivity of 98.15%, and specificity of 98.88%, by utilizing the Bot-IoT dataset, respectively.</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

SBC2TM: Stochastic Foraging Movement Optimized-Bottleneck Cross Attention-based Deep Learning Model for DDoS Attack Detection in IoT Networks

  • 期刊论文

  • 发表日期：

    2026年5月4日

查看全文
  • Manjusha V. Khond,
  • Mahesh R. Sanghavi

摘要

Distributed Denial of Service attack detection is critical due to the growing number of compromised Internet of Things devices used to flood target networks with overwhelming traffic, disrupting services, and preventing legitimate access. Researchers have developed various detection strategies, yet the approaches faced drawbacks of generalizability issues, computational complexities, lack of scalability, and flexibility. Hence, the research proposes a Stochastic Foraging Movement Optimized-Bottleneck Cross Attention-based Convolutional Network and Bidirectional Long Short-Term Memory (SBC2TM) model for DDoS attack detection. However, the proposed SBC2TM model determines the subtle and complex patterns characteristic of Distributed Denial of Service attacks in dynamic Internet of Things network environments effectively. Moreover, the integration of the Bottleneck Cross Attention (BnCA) mechanism in the model enhances the accuracy while detecting the attacks, and it distinguishes between normal and malicious traffic. Additionally, a Stochastic Foraging Movement Optimization (SFgMO) algorithm is introduced for tuning the hyperparameters of the model that resulted in optimal detection performance. Notably, the SBC2TM model exhibited significant enhancement over the existing approaches by achieving an accuracy of 98.42%, F1-score of 98.13%, precision of 98.31%, recall of 98.15%, sensitivity of 98.15%, and specificity of 98.88%, by utilizing the Bot-IoT dataset, respectively.