Scholarlink

In the complex and dynamic Internet environment, comprehensive monitoring and management of network assets are crucial for maintaining network security. However, many Internet users struggle to understand the full scope of their assets, leaving them vulnerable to numerous potential security risks. This paper proposes an innovative network asset discovery algorithm aimed at helping users identify all possible assets belonging to them across the entire internet. We start by obtaining an SEED asset list by leveraging passive DNS (PDNS) data, ensuring that these assets belong to the target user. Next, we analyze the network communication records of these seed assets to collect other network assets that potentially belongs to the target user, then vectorize their features. Finally, we apply a neural network-based clustering method to judge whether the assets actually belong to the target user. To validate our algorithm’s effectiveness, we constructed a dedicated dataset and conducted extensive experiments. The results demonstrate that our algorithm efficiently and accurately identifies most of the user’s assets on a global scale.

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 Novel Feature Clustering Approach for Comprehensive Network Asset Discovery: Methods, Implementation, and Validation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Xue Zhou,
  • Min Zhang,
  • Xuyang Yao,
  • Lingran Zhang,
  • Jiayi Zhou,
  • Jiaqi Wei

摘要

In the complex and dynamic Internet environment, comprehensive monitoring and management of network assets are crucial for maintaining network security. However, many Internet users struggle to understand the full scope of their assets, leaving them vulnerable to numerous potential security risks. This paper proposes an innovative network asset discovery algorithm aimed at helping users identify all possible assets belonging to them across the entire internet. We start by obtaining an SEED asset list by leveraging passive DNS (PDNS) data, ensuring that these assets belong to the target user. Next, we analyze the network communication records of these seed assets to collect other network assets that potentially belongs to the target user, then vectorize their features. Finally, we apply a neural network-based clustering method to judge whether the assets actually belong to the target user. To validate our algorithm’s effectiveness, we constructed a dedicated dataset and conducted extensive experiments. The results demonstrate that our algorithm efficiently and accurately identifies most of the user’s assets on a global scale.