Scholarlink

<p>Open datasets of Internet backbone topologies are essential for the reproducible evaluation of various types of routing algorithms. Thus, numerous research papers have been using available network repositories such as Topology Zoo and SNDlib. However, a lesson learned in the past years is that, in many cases, it is not enough to treat backbone networks as purely logical topologies, but they should be treated as geometric graphs instead (e.g., natural or man-made disasters fail network elements sharing the same geographic region). While TopoHub, a recent repository, makes a significant step in solving this issue, its graph representations still neglect the (almost) planar nature of backbone networks, a feature many recent routing algorithms rely on fundamentally. This paper introduces T-Garden, a curated and openly accessible collection of (nearly) planar Internet backbone topologies. The dataset consists of (i) fixed and standardized versions of selected physical-layer backbone networks sourced from available repositories, and (ii) synthetic planar topologies that scale to several thousand nodes while preserving key structural and geometric properties observed in real networks. By providing full geometric information on over 120 (nearly) planar backbones, T-Garden provides a comprehensive benchmark dataset for research on geography-aware telecommunication networks.</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

A curated dataset of real and synthetic quasi-planar Internet backbone network topologies

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年8月1日

查看全文
  • Tamás-Zsolt Képes,
  • Noémi Gaskó,
  • János Tapolcai,
  • Balázs Vass

摘要

Open datasets of Internet backbone topologies are essential for the reproducible evaluation of various types of routing algorithms. Thus, numerous research papers have been using available network repositories such as Topology Zoo and SNDlib. However, a lesson learned in the past years is that, in many cases, it is not enough to treat backbone networks as purely logical topologies, but they should be treated as geometric graphs instead (e.g., natural or man-made disasters fail network elements sharing the same geographic region). While TopoHub, a recent repository, makes a significant step in solving this issue, its graph representations still neglect the (almost) planar nature of backbone networks, a feature many recent routing algorithms rely on fundamentally. This paper introduces T-Garden, a curated and openly accessible collection of (nearly) planar Internet backbone topologies. The dataset consists of (i) fixed and standardized versions of selected physical-layer backbone networks sourced from available repositories, and (ii) synthetic planar topologies that scale to several thousand nodes while preserving key structural and geometric properties observed in real networks. By providing full geometric information on over 120 (nearly) planar backbones, T-Garden provides a comprehensive benchmark dataset for research on geography-aware telecommunication networks.