DIT: A Dynamic Bandwidth Isolated Transmission System for Large-Scale Inter-DC Wireless Communication Network

Author:

Zhang Shiyan1ORCID,Zhang Yuchao1,Wang Ran1,Gong Xiangyang1,Xiong Yongping1

Affiliation:

1. Beijing University of Posts and Telecommunications, Beijing 100089, China

Abstract

Large companies are increasingly providing their services over interdomain data centers. Such large-scale wireless communication networks carry various applications, including both online time-sensitive services and offline bandwidth-sensitive services. These applications share the same physical links but have different bandwidth requirements at different time, making the coscheduling problem a key challenge in improving the overall network utility. Most existing commercial systems reserve peak value bandwidth for online time-sensitive services to ensure the quality of service and use the leftover bandwidth to transmit offline data in a best-effort manner. Such systems cannot obtain high link utilizations because of the nonstationarity of online traffic. To solve this problem, we present a dynamic isolated transmission (DIT) system in this paper, where a sliding- k algorithm is designed to predict online traffic, and a bottleneck bypass routing scheme is proposed to schedule the mixed traffic in a shared inter-DC wireless communication network. We conduct a series of experiments in an experimental inter-DC wireless communication network to analyze for efficient network operation of the DIT and compare its performance with typical existing fix-bandwidth-based solutions including Microsoft’s SWAN. The results show that DIT outperforms existing solutions, improving the interdomain link utilization by 18% and the intradomain link utilization by 65%.

Funder

Fundamental Research Funds for the Central Universities

Publisher

Hindawi Limited

Subject

Electrical and Electronic Engineering,Computer Networks and Communications,Information Systems

Reference30 articles.

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