基于网卡虚拟化的高性能容器网络设计

doi:10.19678/j.issn.1000-3428.0061999

计算机工程 ›› 2022, Vol. 48 ›› Issue (7): 214-219. doi: 10.19678/j.issn.1000-3428.0061999

基于网卡虚拟化的高性能容器网络设计

臧迪¹, 杨志刚¹, 王晶¹, 姚治成², 张伟功³

1. 首都师范大学信息工程学院, 北京 100048;
2. 中国科学院计算技术研究所先进计算机系统研究中心, 北京 100080;
3. 首都师范大学高可靠嵌入式系统技术北京市工程中心, 北京 100048

收稿日期:2021-07-07 修回日期:2021-09-26 出版日期:2022-07-15 发布日期:2021-10-15
作者简介:臧迪(1995—),男,硕士,主研方向为计算机体系结构、智能芯片;杨志刚(通信作者),副教授、硕士;王晶,教授、博士;姚治成,工程师、硕士;张伟功,研究员、博士。
基金资助:
国家自然科学基金（62076168）。

Design of High-Performance Container Network Based on Network Interface Card Virtualization

ZANG Di¹, YANG Zhigang¹, WANG Jing¹, YAO Zhicheng², ZHANG Weigong³

1. Information Engineering College, Capital Normal University, Beijing 100048, China;
2. Center for Advanced Computer Systems, Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100080, China;
3. Beijing Engineering Research Center of High Reliable Embedded System, Capital Normal University, Beijing 100048, China

Received:2021-07-07 Revised:2021-09-26 Online:2022-07-15 Published:2021-10-15

摘要/Abstract

摘要： 单根I/O虚拟化技术为传统数据中心提供高效的服务器整合能力和灵活的应用部署能力，通过将多个网卡直通到虚拟机，减少额外包复制带来的性能损失，使得网络I/O具有接近主机的性能。然而，在网络功能虚拟化场景下单独使用单根I/O虚拟化技术会降低传统数据中心的网络I/O虚拟化性能。针对网络功能虚拟化长链场景，结合单根I/O虚拟化技术和软件虚拟化技术，设计基于网卡虚拟化的高性能容器网络。通过转发模块判断网络流量的目的地址，寻找最优的流量转发路径，实现流量的灵活转发。利用基于脚本程序的自动化部署模块，对每个节点业务进行支持动态增删服务的配置，便于用户对网络进行管理和修改。实验结果表明，在网络功能虚拟化长链场景下，相比单根I/O虚拟化技术，该网络延迟降低约20%，同时能够有效提高网络吞吐量，解决数据中心的网络I/O虚拟化问题。

关键词: 单根I/O虚拟化, 网络功能虚拟化, 网卡虚拟化, 容器网络, 数据中心

Abstract: Single Root I/O Virtualization(SR-IOV) technology provides traditional data centers with efficient server integration and flexible application deployment capabilities.To reduce the performance loss caused by additional packet replication, SR-IOV technology connects multiple network interface cards directly to the virtual machine, bringing the network I/O performance closer to that of the host.However, SR-IOV technology in the Network Function Virtualization (NFV) scenario degrades the performance of traditional data center virtualization.For the combined use of SR-IOV and software virtualization technology in such scenarios, a high-performance container network based on network interface card virtualization is designed.A forwarding module is used to determine the destination address, find the optimal forwarding path, and ensure the flexible forwarding of network traffic.A fully automatic deployment module based on a script program is used to support dynamic addition and deletion services configuration for the business of each node, facilitate users to manage and modify the network, and improve the module's adaptability.Experimental results show that the network's delay in the NFV long chain scenario is about 20% lower than that of SR-IOV technology, as the former improves throughput.This solves network I/O virtualization issues in data centers.

Key words: Single Root I/O Virtualization(SR-IOV), Network Function Virtualization(NFV), network interface card virtualization, container network, data center

中图分类号:

TP391

臧迪, 杨志刚, 王晶, 姚治成, 张伟功. 基于网卡虚拟化的高性能容器网络设计[J]. 计算机工程, 2022, 48(7): 214-219.

ZANG Di, YANG Zhigang, WANG Jing, YAO Zhicheng, ZHANG Weigong. Design of High-Performance Container Network Based on Network Interface Card Virtualization[J]. Computer Engineering, 2022, 48(7): 214-219.

https://www.ecice06.com/CN/Y2022/V48/I7/214

图/表 9

20221029175040

20221029175044

20221029175048

20221029175052

20221029175055

20221029175059

20221029175103

20221029175107

20221029175111

参考文献

[1] PHAM C, TRAN N H, REN S L, et al.Traffic-aware and energy-efficient VNF placement for service chaining:joint sampling and matching approach[J].IEEE Transactions on Services Computing, 2020, 13(1):172-185.
[2] MECHTRI M, GHRIBI C, SOUALAH O, et al.NFV orchestration framework addressing SFC challenges[J].IEEE Communications Magazine, 2017, 55(6):16-23.
[3] Mark Beierl.OPNFV[EB/OL].[2021-06-05].http://www.creader.com/news/20011219/200112190019.html.
[4] 徐江, 孟相如, 韩晓阳, 等.博弈论组合赋权的虚拟网络映射算法[J].小型微型计算机系统, 2020, 41(7):1464-1469. XU J, MENG X R, HAN X Y, et al.Virtual network embedding algorithm based on game theory combination weighting[J].Journal of Chinese Computer Systems, 2020, 41(7):1464-1469.(in Chinese)
[5] 黄强, 李宁.5G边缘计算演进[J].邮电设计技术, 2018(11):68-73. HUANG Q, LI N.MEC in 5G evolution[J].Designing Techniques of Posts and Telecommunications, 2018(11):68-73.(in Chinese)
[6] VIZARRETA P, CONDOLUCI M, MACHUCA C M, et al.QoS-driven function placement reducing expenditures in NFV deployments[C]//Proceedings of IEEE International Conference on Communications.Washington D.C., USA:IEEE Press, 2017:1-7.
[7] BEN JEMAA F, PUJOLLE G, PARIENTE M.QoS-aware VNF placement optimization in edge-central carrier cloud architecture[C]//Proceedings of IEEE Global Communications Conference.Washington D.C., USA:IEEE Press, 2016:1-7.
[8] BREMLER-BARR A, HARCHOL Y, HAY D.OpenBox:a software-defined framework for developing, deploying, and managing network functions[C]//Proceedings of ACM SIGCOMM Conference.New York, USA:ACM Press, 2016:511-524.
[9] SUN C, BI J, ZHENG Z L, et al.NFP:enabling network function parallelism in NFV[C]//Proceedings of the Conference of the ACM Special Interest Group on Data Communication.New York, USA:ACM Press, 2017:1-10.
[10] GARG P, WANG Y S. NVGRE:network virtualization using generic routing encapsulation[EB/OL].[2021-06-05].http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.697.7171&rep=rep1&type=pdf.
[11] Intel Company.Single root I/O virtualization and sharing 1.0 S-pecification[EB/OL].[2021-06-05].https://compo-ster.com.ua/documents/sr-iov1_1_20Jan10_cb.pdf.
[12] LEIVADEAS A, FALKNER M, PITAEV N.Analyzing service chaining of virtualized network functions with SR-IOV[C]//Proceedings of the 21st International Conference on High Performance Switching and Routing.Washington D.C., USA:IEEE Press, 2020:1-6.
[13] Linux Foundation.OpenVSwitch, an open virtual switch[EB/OL].[2021-06-15].https://www.openvswit-ch.org/.
[14] ZHANG T Z, LINGUAGLOSSA L, GIACCONE P, et al.Performance benchmarking of state-of-the-art software switches for NFV[J].Computer Networks, 2021, 188:1-10.
[15] ABRAMSON D, JACKSON J, MUTHRASANALLUR S, et al.Intel virtualization technology for directed I/O[J]. Intel Technology Journal, 2006, 10(3):1-10.
[16] RICHTER A, HERBER C, RAUCHFUSS H, et al. Performance isolation exposure in virtualized platforms with PCI passthrough I/O sharing[C]//Proceedings of International Conference on Architecture of Computing Systems.Berlin, Germany:Springer, 2014:171-182.
[17] DONG Y Z, YANG X W, LI J H, et al.High performance network virtualization with SR-IOV[J].Journal of Parallel and Distributed Computing, 2012, 72(11):1471-1480.
[18] 张懿, 禹忠, 王军选.基于OpenFlow的控制器部署可靠性方案设计[J].计算机工程, 2018, 44(8):74-78. ZHANG Y, YU Z, WANG J X.Design of controller deployment reliability scheme based on OpenFlow[J].Computer Engineering, 2018, 44(8):74-78.(in Chinese)
[19] 刘海客, 李集林, 尤启迪, 等.一种OpenFlow网络的动态负载均衡方法[J].计算机工程, 2016, 42(8):85-90. LIU H K, LI J L, YOU Q D, et al.A dynamic load balancing method for OpenFlow network[J].Computer Engineering, 2016, 42(8):85-90.(in Chinese)
[20] Hewlett-Packard Company.Netperf Manual[EB/OL].[2021-06-15].http://www.cs.kent.edu/~farrell/dist/ref/Netper-f.html.
[21] Willy Tarreau.Haproxy[EB/OL].[2021-06-15].https://www.haproxy.com/.
[22] 杨鹏, 马志程, 彭博, 等.集成Docker容器的OpenStack云平台性能研究[J].计算机工程, 2017, 43(8):26-31. YANG P, MA Z C, PENG B, et al.Performace research of OpenStack cloud platform integrated with Docker container[J].Computer Engineering, 2017, 43(8):26-31.(in Chinese)
[23] JAMS T.The Docker book[M].Cambridge USA:MIT Press, 2016.

选择文件类型/文献管理软件名称

选择包含的内容

基于网卡虚拟化的高性能容器网络设计

Design of High-Performance Container Network Based on Network Interface Card Virtualization

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	张冠莹, 伊鹏, 李丹, 朱棣, 毛明. 面向大规模网络的服务功能链部署方法[J]. 计算机工程, 2023, 49(8): 122-129.
[2]	张青青, 汤红波, 游伟, 普黎明. 基于演化博弈的NFV拟态防御架构动态调度策略[J]. 计算机工程, 2022, 48(4): 30-38,49.
[3]	刘雅丽, 史久根. 基于成本的服务链组成与部署联合优化策略[J]. 计算机工程, 2022, 48(4): 106-112.
[4]	都繁杰, 李静, 郭志勇, 任颖文, 尹晓宇, 董小菱. 基于瓶颈感知的多级反馈队列Coflow调度机制[J]. 计算机工程, 2022, 48(10): 193-201,211.
[5]	赵季红, 张文娟, 乔琳琳, 张梦雪. 服务功能链中基于机器学习的QoE评估与预测[J]. 计算机工程, 2022, 48(1): 163-169.
[6]	钱声攀, 于洋, 翟天一, 张徐东, 常彦博. 基于深度学习的主机负载在线预测模型研究[J]. 计算机工程, 2021, 47(9): 84-89.
[7]	左攀, 束永安. DCN中基于前馈神经网络的动态多路径负载均衡方法[J]. 计算机工程, 2021, 47(9): 113-119.
[8]	潘成胜, 梁芷铭, 石怀峰, 孔志翔. 面向并发业务的卫星网络服务功能链优化算法[J]. 计算机工程, 2021, 47(3): 196-201.
[9]	朱国晖, 刘璐, 雷兰洁. 基于VNF组合的服务功能链设计及映射算法[J]. 计算机工程, 2020, 46(4): 183-188,197.
[10]	樊玉琦, 张蓓, 王伦飞. 面向电耗与网络同步代价优化的数据副本放置研究[J]. 计算机工程, 2020, 46(2): 110-117.
[11]	徐超, 邓君华, 姜丽丽, 吉明涛, 李鑫. 服务质量感知的云系统数据副本恢复算法[J]. 计算机工程, 2019, 45(9): 17-22.
[12]	臧韦菲,兰巨龙,胡宇翔,陈文涛. 基于多目标离散差分进化的网络感知虚拟机放置算法[J]. 计算机工程, 2019, 45(6): 96-102.
[13]	张钊,李海龙,董思歧,胡磊,马景仁. 基于SDN的数据中心网络流概率路径选择方法[J]. 计算机工程, 2019, 45(4): 36-40.
[14]	海梅生,伊鹏,江逸茗. 基于服务质量与资源约束的服务链部署策略[J]. 计算机工程, 2019, 45(3): 54-59.
[15]	邱航,游伟,汤红波,王琛,牛犇. 基于多路径的可靠性服务功能链部署方案[J]. 计算机工程, 2019, 45(3): 101-106,112.

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

基于网卡虚拟化的高性能容器网络设计

Design of High-Performance Container Network Based on Network Interface Card Virtualization

RichHTML

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献

相关文章 15

编辑推荐

Metrics

本文评价