A Composite Pipeline for Forwarding Low-Latency Traffic in SDN Programmable Data Planes

Author:

Ling ZhiyuanORCID,Chen Xiao,Song Lei

Abstract

With the rapid evolution of network technologies over recent years, emerging network services, especially industrial control networks, video conferencing, intelligent driving, and other scenarios, have put forward higher demand for the low-latency forwarding of network traffic. The existing flow caching and hardware acceleration methods only improve the overall forwarding performance of data-plane devices but cannot separate the forwarding process of low-latency traffic from others to reflect the priority of these flows. In this paper, we extend the POF southbound interface protocol and propose a marking method for low-latency flows, based on which we design a composite pipeline to achieve fast processing for low-latency traffic by introducing a fast-forwarding path. The experiments show that the fast path has a higher forwarding capability than the MAT pipeline in the POF Switch and can reduce the forwarding delay of low-latency flows by 62–68%. In a real network environment with a mixed traffic simulation, the reduction reaches 17–20% with no delay increment for the non-low-latency part.

Funder

Strategic Leadership Project of Chinese Academy of Sciences

Publisher

MDPI AG

Subject

Electrical and Electronic Engineering,Computer Networks and Communications,Hardware and Architecture,Signal Processing,Control and Systems Engineering

Reference25 articles.

1. Ultra-low latency (ull) networks: The ieee tsn and ietf detnet standards and related 5g ull research;Nasrallah;IEEE Commun. Surv. Tutor.,2018

2. Seanet: Architecture and technologies of an on-site, elastic, autonomous network;Wang;J. Netw. New Media,2020

3. Future data center networking: From low latency to deterministic latency;Han;IEEE Netw.,2022

4. A survey on architectures and energy efficiency in data center networks;Hammadi;Comput. Commun.,2014

5. (2020). Cisco Annual Internet Report (2018–2023) White Paper, Cisco.

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