Scholarlink

<p>Smart assistants, the Internet of Things (IoT), and Edge devices are widely adopted, enhancing daily life, economic growth, and environmental sustainability. Edge computing improves communication by making it faster and more secure, reducing the communication burden, and enabling smoother system operations and more effective work. However, these technologies also introduce security and privacy risks in Edge-IoT environments. Unauthorized access to network-stored data can lead to the misuse of sensitive information and raise privacy breaches, necessitating robust detection mechanisms. This study presents a framework, a machine learning-based hybrid PETDA2C-EC, a privacy-enhancing technique for detecting attacks against confidentiality in edge computing. The framework achieves a maximum accuracy of 99.93% when applied to four benchmark datasets. Additionally, the framework detects intrusion, distributed denial of service (DoS/DDoS), man-in-the-middle (MiTM), injection, information gathering, heartbleed, and infiltration, which impact confidentiality in the edge-IoT environment, with a maximum accuracy of 99.98%. The proposed detection framework is trained on benchmark datasets such as UNSW-NB15, CIC-IDS2017, TON_IoT, and Edge-IIoTset. The results demonstrate that our framework outperforms existing state-of-the-art methods, reinforcing its effectiveness in securing Edge-IoT 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

PETDA2C-EC: a privacy-enhancing technique to detect attacks against confidentiality in edge computing

  • 期刊论文

  • 发表日期：

    2025年4月9日

查看全文
  • Vipin Kumar,
  • Vivek Kumar

摘要

Smart assistants, the Internet of Things (IoT), and Edge devices are widely adopted, enhancing daily life, economic growth, and environmental sustainability. Edge computing improves communication by making it faster and more secure, reducing the communication burden, and enabling smoother system operations and more effective work. However, these technologies also introduce security and privacy risks in Edge-IoT environments. Unauthorized access to network-stored data can lead to the misuse of sensitive information and raise privacy breaches, necessitating robust detection mechanisms. This study presents a framework, a machine learning-based hybrid PETDA2C-EC, a privacy-enhancing technique for detecting attacks against confidentiality in edge computing. The framework achieves a maximum accuracy of 99.93% when applied to four benchmark datasets. Additionally, the framework detects intrusion, distributed denial of service (DoS/DDoS), man-in-the-middle (MiTM), injection, information gathering, heartbleed, and infiltration, which impact confidentiality in the edge-IoT environment, with a maximum accuracy of 99.98%. The proposed detection framework is trained on benchmark datasets such as UNSW-NB15, CIC-IDS2017, TON_IoT, and Edge-IIoTset. The results demonstrate that our framework outperforms existing state-of-the-art methods, reinforcing its effectiveness in securing Edge-IoT environments.