Scholarlink

<p>Cloud environments enable scalable multi-tenant computing but introduce security risks like covert channels, making their detection and classification essential for maintaining cloud security. Initially, covert channels exploit shared resource dynamics to mimic normal workload behavior, allowing malicious data transmission to go unnoticed by standard security measures. In addition, covert channels embed signals in encrypted or obfuscated traffic through multi-layered encryption and protocol tunneling, creating inherent noise that attackers exploit to sustain covert communication channels quietly. Hence, to tackle theses drawbacks, a Fourier-Warp Entropic Reinforcement Graph Detector is introduced, combining the Fourier-Warp Convolutional IsoForest Graph Detector and Adaptive Entropic Reinforcement Graph Transformer. This integrated system analyzes temporal and spatial workload patterns, detects abnormal timing behaviors, and identifies hidden communications. It then adaptively refines decisions, ensuring reliable distinction between normal workload fluctuations and covert activity, even when communications are encrypted or obfuscated. Thus, the model learns and classifies diverse covert threats with high precision by mapping inter-tenant relationships, delivering robust and adaptive protection for multi-tenant cloud environments. This method improves cloud security by constantly learning and neutralizing covert threats, achieving high accuracy, recall, and low detection errors, while significantly reducing RMSE.</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

Detecting hidden communication threats in cloud systems using advanced pattern and threat propagation analysis

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月7日

查看全文
  • Gavini Sreelatha,
  • Tan Kuan Tak,
  • Rajkumar Kalimuthu,
  • Pravin R. Kshirsagar,
  • Balajee Maram,
  • T. Venkatakrishnamoorthy

摘要

Cloud environments enable scalable multi-tenant computing but introduce security risks like covert channels, making their detection and classification essential for maintaining cloud security. Initially, covert channels exploit shared resource dynamics to mimic normal workload behavior, allowing malicious data transmission to go unnoticed by standard security measures. In addition, covert channels embed signals in encrypted or obfuscated traffic through multi-layered encryption and protocol tunneling, creating inherent noise that attackers exploit to sustain covert communication channels quietly. Hence, to tackle theses drawbacks, a Fourier-Warp Entropic Reinforcement Graph Detector is introduced, combining the Fourier-Warp Convolutional IsoForest Graph Detector and Adaptive Entropic Reinforcement Graph Transformer. This integrated system analyzes temporal and spatial workload patterns, detects abnormal timing behaviors, and identifies hidden communications. It then adaptively refines decisions, ensuring reliable distinction between normal workload fluctuations and covert activity, even when communications are encrypted or obfuscated. Thus, the model learns and classifies diverse covert threats with high precision by mapping inter-tenant relationships, delivering robust and adaptive protection for multi-tenant cloud environments. This method improves cloud security by constantly learning and neutralizing covert threats, achieving high accuracy, recall, and low detection errors, while significantly reducing RMSE.