Scholarlink

With the rise of Edge Computing (EC), more applications are being deployed on edge servers using containers to serve users. These containers rely on image files composed of multiple layers, which often harbor vulnerabilities that can typically be resolved through upgrades. However, due to resource limitations, geographical distribution, and latency sensitivity in EC, traditional cloud upgrade strategies are not directly applicable. Therefore, efficiently updating containers to address vulnerabilities is a pressing challenge. In this paper, we present a joint cluster upgrade and task scheduling algorithm using layer sharing to quickly and efficiently address container vulnerabilities. By modeling the edge cluster upgrade problem, we establish the upgrade sequence based on node load, thereby reducing the impact on ongoing tasks. Our layer-sharing task scheduling mechanism takes into account node resource status, layer distribution, and task characteristics. Based on this, we present the Layer-aware Cluster upgrade and Task scheduling (LCT) algorithm that optimizes layer transmission, download delays, and task computation delays to minimize the disruption caused by patching. Experiments conducted with both simulated and real datasets demonstrate that our algorithm significantly outperforms baseline methods.

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

Fast and Efficient Layer-Aware Container Vulnerability Patching in Edge Computing

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Xuan Xie,
  • Yuting Hu,
  • Xinpu Chen,
  • Yumin Lin,
  • Zhiqing Tang,
  • Jianxiong Guo,
  • Tianle Zhang,
  • Tian Wang

摘要

With the rise of Edge Computing (EC), more applications are being deployed on edge servers using containers to serve users. These containers rely on image files composed of multiple layers, which often harbor vulnerabilities that can typically be resolved through upgrades. However, due to resource limitations, geographical distribution, and latency sensitivity in EC, traditional cloud upgrade strategies are not directly applicable. Therefore, efficiently updating containers to address vulnerabilities is a pressing challenge. In this paper, we present a joint cluster upgrade and task scheduling algorithm using layer sharing to quickly and efficiently address container vulnerabilities. By modeling the edge cluster upgrade problem, we establish the upgrade sequence based on node load, thereby reducing the impact on ongoing tasks. Our layer-sharing task scheduling mechanism takes into account node resource status, layer distribution, and task characteristics. Based on this, we present the Layer-aware Cluster upgrade and Task scheduling (LCT) algorithm that optimizes layer transmission, download delays, and task computation delays to minimize the disruption caused by patching. Experiments conducted with both simulated and real datasets demonstrate that our algorithm significantly outperforms baseline methods.