Hybrid SDN Performance: Switching between Centralized and Distributed Modes under Unreliable Control Communication Channels

Author:

Osman MohammedORCID,Mangues-Bafalluy Josep

Abstract

Software-defined networking generally assumes ideal control channels between controller and network nodes. This may not be the case in challenged environments that are becoming more common due to dense and reduced-coverage 5G deployments and use cases requiring cost-effective wireless transport networks. In this paper, we evaluate the impact on network performance of unreliable controller-to-node communication channels, propose a hybrid SDN (hSDN) solution that switches between centralized and distributed operational modes depending on network conditions, and evaluate this solution under a variety of network scenarios (e.g., link impairments or packet loss ratios) designed to assess its operational limits. The results show that the proposed solution substantially improved the aggregated throughput, particularly when control channel packet loss ratios increased, while only showing a slight increase in average latency (e.g., 28% throughput improvement for 20% control packet losses). This enables network operation in hard conditions under which a canonical centralized SDN control would result in a nonoperational network.

Funder

Agència de Gestió d'Ajuts Universitaris i de Recerca

Publisher

MDPI AG

Subject

Control and Optimization,Computer Networks and Communications,Instrumentation

Reference28 articles.

1. Small cell backhaul requirements;Robson,2012

2. Small Cell Millimeter Wave Mesh Backhaul,2013

3. Small Cells Drive Microwave Backhaul Boom. Small Cells 2019—2026,2020

4. Report on SDN Usage in NFV Architectural Framework,2015

5. OpenFlow v1.3https://opennetworking.org/wp-content/uploads/2014/10/openflow-spec-v1.3.0.pdf

Cited by 2 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3