SOFI dataset: Symptom-fault relationship for IP-network Articles uri icon

authors

  • Vargas Arcila, Angela
  • Corrales, Juan Carlos
  • SANCHIS ARELLANO, ALICIA
  • Rendon, Alvaro

publication date

  • October 2022

issue

  • 109233

volume

  • 216

International Standard Serial Number (ISSN)

  • 1389-1286

Electronic International Standard Serial Number (EISSN)

  • 1872-7069

abstract

  • This article describes a dataset of network performance and well-known faults. It is a dataset of symptom-fault causal relationships for an IP-based network. This information is periodically gathered from edge nodes of an emulated enterprise network. The dataset contains about 649 h of network monitoring data, 10 of which refer to artificially induced faults. Those faults are impact failures as faulty cables, fiber cuts, high link utilization, line card failure, link down, misconfigurations, protocol issues, and distribution layer going down. Data were methodically structured by an authoring tool. It is an original dataset created on purpose because of no availability of this dataset type. It is of interest to researchers in data mining, fault diagnosis, and overall network management.

subjects

  • Computer Science

keywords

  • fault diagnosis; link failure; network management; network monitoring