Title: ReconsMap: a reliable controller-switch mapping in software defined networks

Authors: Wenbo Wang; Binqiang Wang; Yuxiang Hu; Bangzhou Liu

Addresses: China National Digital Switching System Engineering and Technological Research Center, No. 7, Jianxue Street, Jinshui District, Zhengzhou City, Henan, China ' China National Digital Switching System Engineering and Technological Research Center, No. 7, Jianxue Street, Jinshui District, Zhengzhou City, Henan, China ' China National Digital Switching System Engineering and Technological Research Center, No. 7, Jianxue Street, Jinshui District, Zhengzhou City, Henan, China ' China National Digital Switching System Engineering and Technological Research Center, No. 7, Jianxue Street, Jinshui District, Zhengzhou City, Henan, China

Abstract: The software defined networking (SDN) innovates the future network structure, decoupling the control plane and data plane. In large SDN networks, multiple controllers or controller domains are deployed, where each controller has a logically centralised vision while managing a set of switches. Recent studies focus more on controller placement, but simply assign switches to its closest controller. Such 'latency-first' controller-switch mapping may lead to controller overload and vulnerability of the spanning tree. In this paper, we illustrate this case and propose a reliable controller-switch mapping model (ReconsMap model). This model: 1) adjusts the number of mapping switches according to the network traffic; 2) ensures an accepted propagation delay; 3) builds a robust spanning tree for each controller to ensure the lowest data loss. Computational results on OS3E and topology zoo show that ReconsMap effectively reduces the data loss, improves the controller overload and the robustness of the spanning tree.

Keywords: software defined networking; SDN; controller-switch mapping; reliability; load balance.

DOI: 10.1504/IJHPCN.2017.084247

International Journal of High Performance Computing and Networking, 2017 Vol.10 No.3, pp.201 - 206

Received: 31 Jul 2015
Accepted: 07 Sep 2015

Published online: 22 May 2017 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article