Persistent store-based dual replication system for distributed SDN controller

Sangyun Han, Sungwon Lee

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Citations (Scopus)

Abstract

In Software Defined Networking, the role of the centralized controller is paramount. Some SDN controllers adopt a distributed clustering mechanism for keeping normal operation of the whole network. Unfortunately, the data loss and the lack of consistency are unavoidable in this approach, because many controllers depend on volatile memory to store most of the important data. In case of restarting or machine failure, all data in memory can't be recovered. Therefore, we propose persistent store-based dual replication system for maintaining and backing a data of distributed SDN controller. Storing data in the persistent store can reduce Mean Time To Repair (MTTR) and consequently enhance the reliability and availability of the system.

Original languageEnglish
Title of host publication2016 International Conference on Selected Topics in Mobile and Wireless Networking, MoWNeT 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781509017430
DOIs
Publication statusPublished - 20 Jun 2016
Event2016 International Conference on Selected Topics in Mobile and Wireless Networking, MoWNeT 2016 - Cairo, Egypt
Duration: 11 Apr 201613 Apr 2016

Publication series

Name2016 International Conference on Selected Topics in Mobile and Wireless Networking, MoWNeT 2016

Conference

Conference2016 International Conference on Selected Topics in Mobile and Wireless Networking, MoWNeT 2016
Country/TerritoryEgypt
CityCairo
Period11/04/1613/04/16

Bibliographical note

Publisher Copyright:
© 2016 IEEE.

Keywords

  • 5G networks
  • Distributed SDN controller
  • SDN ONOS
  • distributed key-value store
  • google etcd
  • persistent store

Fingerprint

Dive into the research topics of 'Persistent store-based dual replication system for distributed SDN controller'. Together they form a unique fingerprint.

Cite this