Article Details

Abstract

With the popularization of information technology and the rapid development of Internet applications, cloud computing technology is becoming more and more mature and widely used. While cloud computing provides flexible, efficient, secure and convenient IT services, the explosive growth of IT business also imposes new requirements on its performance. Virtualization technology is the core technology of the IaaS platform. Compared with CPU and memory virtualization, I/O virtualization has become the performance bottleneck for Virtual Machines. By analyzing the network principles of several I/O virtualization methods and combining the characteristics of high-concurrency services on the IaaS platform, this study proposes an application solution based on SR-IOV technology on the IaaS platform. Through experimental data analysis and verification in a production environment, it is proven that SR-IOV technology brings a qualitative improvement to the network performance of I/O virtualization, meeting the network performance requirements of network-intensive applications in a cloud computing environment.

Keywords

SR-IOV; cloud computing; KVM virtualization; Virtual Machine Network; High Performance Network

References

1. [1] Gong Shanshan Research and Implementation of Network Card Virtualization Based on SRIOV Technology [D]. Beijing: China Shipbuilding Research Institute, 2015
2. [2] Huang Zhiqiang Research and Optimization of High Performance I/O Virtualization Based on SRIOV [D]. Shanghai: Shanghai Jiao Tong University, 2013
3. [3] Yang Hongbo Research on High Performance Network Virtualization Technology [D]. Shanghai: Shanghai Jiao Tong University, 2012
4. [4] Liu Mingda, Cao Huiyuan, Shi Yijuan, etc Building a Trusted Virtual Environment for TCM Hardware Virtualization Based on SRIOV [J]. Journal of Wuhan University: Science Edition, 2017, 63 (2): 117-124
5. [5] Ni Pingbo, Zhang Jie, Yang Fan, etc Research on the Application of SRV-IOV Based on KVM and Infiniband [C]//2016 Smart Grid Development Seminar Beijing: China Electric Power Science Research, 2016:668-674
6. [6] Xun Zhongkai, Huang Hao, Jin Yincheng Design and Implementation of Virtual Machine Firewall Based on SRIOV [J]. Computer Engineering, 2014, 40 (5): 154-157
7. [7] Zhang Chi, Zhang Ao Application of SRIOV Technology in OpenStack [J]. Computer System Applications, 2017, 26 (9): 246-252
8. [8] Li Nan, Feng Weijun, Zhang Wu, etc A VF allocation model based on SRIOV extension [J]. Computer Applications and Software, 2013, 30 (11): 24-27
9. [9] OpenNebula. Docs Page [EB/OL].[2024-09-06].https://docs.opennebula.io.
10. [10] Kernel. Virtual Function I/O [EB/OL].[2024-09-06].https://www.kernel.org/doc/Documentation.

About the authors

Dong Kaikai (1991.02-), male, Han nationality, from Yuncheng, Shanxi Province, intermediate engineer, bachelor's degree, research direction: cloud computing, network security.