Deep Reinforcement Learning Based Coflow Scheduling in Data Center Networks

MA Teng; HU Yu-xiang; ZHANG Xiao-hui

doi:10.3969/j.issn.0372-2112.2018.07.011

您当前的位置：

首页 >

文章列表页 >

Deep Reinforcement Learning Based Coflow Scheduling in Data Center Networks

更新时间：2025-07-16

- Deep Reinforcement Learning Based Coflow Scheduling in Data Center Networks
- Acta Electronica Sinica Vol. 46, Issue 7, Pages: 1617-1624(2018)
- 作者机构：
  
  国家数字交换系统工程技术研究中心,河南,郑州,450002
- 作者简介：
- 基金信息：
- DOI：10.3969/j.issn.0372-2112.2018.07.011
  CLC： TP393
- Published Online：25 July 2018，
  
  Published：2018
- 稿件说明：
移动端阅览
MA Teng, HU Yu-xiang, ZHANG Xiao-hui. Deep Reinforcement Learning Based Coflow Scheduling in Data Center Networks[J]. Acta Electronica Sinica, 2018, 46(7): 1617-1624.
DOI：

MA Teng, HU Yu-xiang, ZHANG Xiao-hui. Deep Reinforcement Learning Based Coflow Scheduling in Data Center Networks[J]. Acta Electronica Sinica, 2018, 46(7): 1617-1624. DOI： 10.3969/j.issn.0372-2112.2018.07.011.

摘要

最小化语义相关流的平均完成时间是数据中心网络流量管理面临的难题之一.受人工智能领域深度增强学习方向的最新研究进展启发，本文提出一种的新的语义相关流调度机制.将带宽约束的语义相关流调度问题转化为连续的学习过程，通过学习以往策略实现最佳调度.引入反向填充和有限复用机制，保证系统的工作保持性和无饥饿性.仿真结果表明，在不同的网络负载下，本文提出的调度机制均使得语义相关流的平均完成时间小于其他调度机制，尤其是网络负载较大时，相比最先进的调度机制，性能提升约50%.

Abstract

Coflow completion time minimization is one of the challenges of traffic management in data center networks.Inspired by the newest research progress in deep reinforcement learning

which is one direction of artificial intelligence

this paper proposes a novel coflow scheduling mechanism.It translates the coflow scheduling problem with bandwidth constraint into a continuous learning process.By learning the previous decisions

the best scheduling is obtained.By introducing back filling and limited multiplexing mechanisms

the system is work-conserving and starvation-free.Simulation results show that

under different network load

compared with other scheduling mechanisms

the average coflow completion time is reduced.Especially when the network load is heavy

the proposed mechanism achieves about 50% performance improvement than the state-of-the-art scheduling mechanism.

关键词

Keywords

references

Views

672

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Data Center Network Flow Scheduling Based on DPSO Algorithm

A Hierarchical Fault Diagnosis Algorithm for Data Center Networks

Research on Energy-Saving Routing Devices and Protocols in the Internet

Slack Time and Accumulation-Based Mix-flow Scheduling in Data Center Networks

Related Author

LIN Zhi-hua

GAO Wen

WU Chun-ming

LI Yong-yan

XUAN Heng-nong

ZHANG Run-chi

ZUO Miao

LIU Tian-tian

Related Institution

College of Electronics Information Engineering, Fujian Jiangxia University

College of Computer Science,Zhejiang University

University of Shaoxing at Shangyu, Shangyu

College of Electronics Information Engineering Fujian Jiangxia University Fuzhou Fujian China

College of Computer Science Zhejiang University Hangzhou Zhejiang China

⁰