Scalable explicit path control in software-defined networks

Long Luo, Hongfang Yu, Shouxi Luo, Zilong Ye, Xiaojiang Du, Mohsen Guizani

Research output: Contribution to journalArticlepeer-review

3 Scopus citations

Abstract

In this paper, we present a scalable PAth COntrol (PACO) approach to meet the increasing demands of fine-grained and explicit paths in Software-Defined Networks. PACO builds upon Segment Routing and adopts the pathlet segment as the building blocks for constructing network paths. By proactively generating and pre-installing a small collection of segments in networks, PACO can explicitly and quickly set up a massive number of reactive paths by simply concatenating these segments. We theoretically investigate the proactive segment generation as well as the optimal concatenation for reactive paths and propose efficient algorithms based on the optimal solutions for each of them. Extensive simulations show that PACO outperforms the state-of-the-art SR-based approach by supporting more than 40% explicit paths and achieves a substantial reduction in switch memory (up to 94%) compared with the OpenFlow-based approach.

Original languageEnglish
Pages (from-to)86-103
Number of pages18
JournalJournal of Network and Computer Applications
Volume141
DOIs
StatePublished - 1 Sep 2019

Keywords

  • Explicit path control
  • Scalability
  • Segment routing
  • Software-defined networks

Fingerprint

Dive into the research topics of 'Scalable explicit path control in software-defined networks'. Together they form a unique fingerprint.

Cite this