Inbound traffic engineering for multihomed ASs using AS path prepending

Kow Chuen Chang, Michael Lo

Research output: Journal article publicationJournal articleAcademic researchpeer-review

19 Citations (Scopus)

Abstract

This article considers the AS path prepending approach to engineer inbound traffic for multihomed ASs. The AS path prepending approach artificially inflates the length of the AS path attribute on one of the links in hopes of diverting some of the traffic to other links. Unlike the current practice that determines the prepending length in a trial-and-error way, we propose AutoPrepend, a systematic and automated procedure for conducting prepending. The entire process consists of four components: passive measurement, active measurement, traffic change prediction, and AS path update. The main idea is to predict the traffic change due to a new prepending value before effecting the change. We have also employed a number of mechanisms to minimize intrusion into the normal operation of the Internet. We have deployed AutoPrepend on a noncommercial site and evaluated its effectiveness based on the measurements collected over six months.
Original languageEnglish
Pages (from-to)18-25
Number of pages8
JournalIEEE Network
Volume19
Issue number2
DOIs
Publication statusPublished - 1 Mar 2005

ASJC Scopus subject areas

  • Software
  • Information Systems
  • Hardware and Architecture
  • Computer Networks and Communications

Cite this