数据中心光互连网络拓扑可重构研究

杨雯; 郭秉礼; 杨鸿珍; 王彦波; 范超; 孟泠宇; 黄善国

doi:10.13756/j.gtxyj.2022.01.003

您当前的位置：

首页 >

文章列表页 >

数据中心光互连网络拓扑可重构研究

光通信系统与网络技术 | 更新时间：2022-06-15

- 数据中心光互连网络拓扑可重构研究
- Study on Topology Reconfiguration of Optical Interconnection Network in Data Center
- 光通信研究 2022年第1期页码：12-17,57
- 作者机构：
  
  1.北京邮电大学　电子工程学院　信息光子学与光通信国家重点实验室，北京　100867
  2.国网浙江省电力有限公司信息通信分公司，杭州　310036
- 作者简介：
  
  杨雯(1995-)，女，黑龙江大庆人。硕士，主要研究方向为强化学习算法与数据中心光网络性能优化。
  郭秉礼，副教授。E-mail：guobingli@bupt.edu.cn
- 基金信息：
  
  国家重点研发计划资助项目(2018YFB1801702);国家自然科学基金资助项目(61771074)
- DOI：10.13756/j.gtxyj.2022.01.003
  中图分类号： TN915
- 纸质出版日期：2022-02-10，
  
  收稿日期：2021-01-18，
扫描看全文
杨雯, 郭秉礼, 杨鸿珍, 等. 数据中心光互连网络拓扑可重构研究[J]. 光通信研究, 2022,(1):12-17,57.

YANG Wen, GUO Bing-li, YANG Hong-zhen, et al. Study on Topology Reconfiguration of Optical Interconnection Network in Data Center[J]. Study on Optical Communications, 2022,(1):12-17,57.
杨雯, 郭秉礼, 杨鸿珍, 等. 数据中心光互连网络拓扑可重构研究[J]. 光通信研究, 2022,(1):12-17,57. DOI： 10.13756/j.gtxyj.2022.01.003.

YANG Wen, GUO Bing-li, YANG Hong-zhen, et al. Study on Topology Reconfiguration of Optical Interconnection Network in Data Center[J]. Study on Optical Communications, 2022,(1):12-17,57. DOI： 10.13756/j.gtxyj.2022.01.003.

摘要

针对传统数据中心电互连网络在应对业务动态流量时存在适应性差的问题，文章提出并验证了一种可以根据网络流量波动进行网络拓扑自优化重构的机制。文章所提机制通过网络仿真系统与深度强化学习模型的迭代交互，实现了对拓扑结构与业务流量分布关系的持续训练，进而在实际系统中，深度强化学习模型，根据软件定义网络控制器实时收集的业务流量分布信息，实现了网络拓扑的自动优化重构，进而提升了网络性能。实验结果表明，针对给定的流量强度，采用深度强化学习进行训练后的模型可以一步输出优化的网络拓扑结构，降低了平均网络延迟和丢包率。

Abstract

Aiming at preventing the network performance deterioration caused by fixed link topology connections in optically switched networks

this paper proposes and validates an interconnection and control mechanism that can automatically optimize and reconfigure network topology according to the network traffic volume. This mechanism combines deep reinforcement learning with software defined network technology. Based on the real-time network performance parameters collected by the software-defined network control plane and the performance feedback of the network simulation system

the continuous training of the topology structure can be realized in the model of deep reinforcement learning. The automatic network topology reconfiguration is achieved with improved network performance. The experimental results show that for a given traffic intensity

the model trained by deep reinforcement learning can learn the topology reconstruction strategy and realize automatic selection. The optimal network topology structure can be achieved in one step

with reduced average network delay and average packet loss.

关键词

光互联网络深度强化学习网络拓扑重构

Keywords

optical interconnect networkdeep reinforcement learningnetwork topology reconfiguration

references

Lee B G, Dupuis N, Pepeljugoski P, et al. Silicon Photonic Switch Fabrics in Computer Communications Systems[J]. Journal of Lightwave Technology, 2015, 33(4): 768-777.

余晓杉, 王琨, 顾华玺. 云计算数据中心光互连网络: 研究现状与趋势[J]. 计算机学报, 2015, 38(10): 1924-1945.

Deng N, Xue Q, Li M, et al. An Optical Multi-ring Burst Network for a Data Center[C]//2013 Optical Fiber Communication Conference and Exposition and the National Fiber Optic Engineers Conference. Anaheim, CA, USA: IEEE, 2013: OTh1A.5.

Xue X, Nakamura F, Prifti K, et al. SDN Enabled Elexible Optical Data Center Network with Dynamic Bandwidth Allocation based on Photonic Integrated Wavelength Selective Switch[J]. Optics Express, 2020, 28(6): 8949-8958.

Xue X, Wang F, Agraz F, et al. SDN-controlled and Orchestrated OPSquare DCN Enabling Automatic Network Slicing with Differentiated QoS Provisioning[J]. Journal of Lightwave Technology, 2020, 38(6): 1103-1112.

Raouf B, Salahuddin M A, Noura L, et al. A Comprehensive Survey on Machine Learning for Networking: Evolution, Applications and Research Opportunities[J]. Journal of Internet Services & Applications, 2018, 9(1): s13174-018-0087-2.

Yu C, Lan J, Guo Z, et al. DROM: Optimizing the Routing in Software-defined Networks with Deep Reinforcement Learning[J]. IEEE Access, 2018, 6: 64533-64539.

Wang J, Laat C, Zhao Z. QoS-aware Virtual SDN Network Planning[C]//2017 IFIP/IEEE Symposium on Integrated Network and Service Management. Lisbon, Portugal: IEEE, 2017: 7987350.

Parsaei M R, Mohammadi R, Javidan R. A New Adaptive Traffic Engineering Method for Telesurgery using ACO Algorithm over Software Defined Networks[J]. European Research in Telemedicine, 2017, 6(3-4): 173-180.

Li W, Li G, Yu X. A Fast Trafc Classication Method based on SDN Network[J]. Electronics, Communications and Networks, 2015, 5: 223229.

Stampa G, Arias M, Sánchez-Charles D, et al. A Deep-reinforcement Learning Approach for Software-defined Networking Routing Optimization[DB/OL]. (2017-09-20)[2021-01-18]. https://arxiv.org/abs/1709.07080v1https://arxiv.org/abs/1709.07080v1.

Dulac-Arnold G, Evans R, van Hasselt H, et al. Deep Reinforcement Learning in Large Discrete Action Spaces[DB/OL]. (2015-11-24)[2021-01-18]. https://arxiv.org/abs/1512.07679v2https://arxiv.org/abs/1512.07679v2.

Shang Yu, Guo B, Li X, et al. Traffic Pattern Adaptive Hybrid Electrical and Optical Switching Network for HPC System[J]. IEEE Communications Letters, 2019, 23(2): 270-273.

Clark D D, Partridge C, Ramming J C, et al. A Knowledge Plane for the Internet[C]//Proceedings of the 2003 Conference on Applications, Technologies, Architectures, and Protocols for Computer Communications. New York, NY, USA: IEEE, 2003: 3-10.

Sutton R S, Barto A G. Reinforcement Learning: An Introduction[J]. IEEE Transactions on Neural Networks, 1998, 9(5): 1054.

更多指标>

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于深度强化学习的V2X频谱资源管理方法

电力光通信网中风险均衡路由算法研究