Link-level interpretation of eigenanalysis for network traffic flows

Irfan Lateef, Ali N. Akansu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Scopus citations

Abstract

This paper presents a novel approach to interpret eigenanalysis of network statistics at the link level in order to identify traffic flows efficiently. It jointly uses and interprets eigencoefficients (frequency) and components of eigenvectors (time) to quantify their importance on each sample (each component of link traffic vector) in eigensubspace representation. We apply the proposed method to analyze the traffic data obtained from Internet2 network. Its merit and superiority over eigenflow based traditional analysis methods are displayed for a few network scenarios with anomalies. It is highlighted that the link-level resolution provided by the method offers advantages also for multi-layer traffic engineering, and it is currently being studied by the authors.

Original languageEnglish (US)
Title of host publication2017 51st Annual Conference on Information Sciences and Systems, CISS 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781509047802
DOIs
StatePublished - May 10 2017
Event51st Annual Conference on Information Sciences and Systems, CISS 2017 - Baltimore, United States
Duration: Mar 22 2017Mar 24 2017

Publication series

Name2017 51st Annual Conference on Information Sciences and Systems, CISS 2017

Other

Other51st Annual Conference on Information Sciences and Systems, CISS 2017
Country/TerritoryUnited States
CityBaltimore
Period3/22/173/24/17

All Science Journal Classification (ASJC) codes

  • Signal Processing
  • Information Systems and Management
  • Computer Networks and Communications
  • Information Systems

Keywords

  • Eigen Analysis
  • Eigenvalue
  • Eigenvector
  • Network Traffic Engineering
  • Principal Component Analysis (PCA)

Fingerprint

Dive into the research topics of 'Link-level interpretation of eigenanalysis for network traffic flows'. Together they form a unique fingerprint.

Cite this