面向网络性能优化的虚拟计算资源调度机制研究

doi:10.11959/j.issn.1000-436x.2016161

摘要/Abstract

摘要：

针对基于Xen的vCPU调度机制对虚拟机网络性能的影响进行了深入研究和分析。提出一种高效、准确、轻量级的网络排队敏感类型虚拟机（NSVM）识别方法，可根据当前虚拟机I/O传输特征将容易受到影响的虚拟机进行准确识别和区分。进而设计一种新型虚拟计算资源调度和分配机制 Diff-Scheduler，将不同类型虚拟机的vCPU实施分池隔离调度，同时提高NSVM类型虚拟机vCPU的调度频率。原型系统实验结果表明，相比Xen默认的调度机制，Diff-Scheduler能够大幅提高虚拟机网络性能，同时保证计算资源分配的公平性。

关键词: 云计算, 虚拟化, vCPU调度, 网络排队敏感型, 虚拟计算资源分配

Abstract:

A deep insight into the relationship between vCPU scheduling and I/O transmit in Xen was provided.Then an effective and lightweight recognition method,through which could identify the so-called NSVM (network queuing sensi-tive virtual machine) that was more vulnerable to the congestion in I/O transmit was put forward.Furthermore,a novel mechanism for resource assignment and scheduling called Diff-Scheduler was proposed.It could schedule the vCPU of the NSVM more frequently than other VM in different pools independently.Evaluations based on a prototype of Xen platform featured Diff-Scheduler show that the proposed mechanism significantly improves the network performance of VM.Specifically,comparing with the default mechanism of Xen,Diff-Scheduler proposed jointly enhances throughput,latency remarkably and ensures the fairness of resource allocation at the same time.

Key words: cloud computing, virtualization, vCPU scheduling, network queuing sensitive, virtual computing resource assignment

王煜炜,刘敏,房秉毅,秦晨翀,闫小龙. 面向网络性能优化的虚拟计算资源调度机制研究[J]. 通信学报, 2016, 37(8): 105-118.

Yu-wei WANG,Min LIU,Bing-yi FANG,Chen-chong QIN,Xiao-long YAN. Study on virtual computing resource scheduling for network performance optimization[J]. Journal on Communications, 2016, 37(8): 105-118.

图/表 15

图1

图2

图3

图4

图5

图6

图7

图8

图9

图10

图11

图12

图13

图14

图15

参考文献 21

[1]	Amazon elastic compute cloud[EB/OL]. .
[2]	Google cloud platform[EB/OL]. .
[3]	Microsoft server and cloud platform[EB/OL]. .
[4]	Aliyun.com[EB/OL]. .
[5]	Wo cloud[EB/OL]. .
[6]	Ecloud[EB/OL]. .
[7]	E Cloud[EB/OL]. .
[8]	The XEN project[EB/OL]. .
[9]	OSTERMANN S , IOSUP A , YIGITBASI N , et al. A performance analysis of EC2 cloud computing services for scientific comput-ing[C]// International Conference on Cloud Computing. c2010:115-131.
[10]	WHITEAKER J , SCHNEIDER F , TEIXEIRA R . Explaining packet delays under virtualization[J]. ACM Sigcomm Computer Communica-tion Review, 2011,41(1):38-44.
[11]	WANG G , NG T . The impact of virtualization on network performance of amazon EC2 data center[C]// The 29th IEEE International Confer-ence on Computer Communications. c2010:1-9.
[12]	SHEA R , WANG F , WANG H , et al. A deep investigation into network performance in virtual machine based cloud environments[C]// The 33th IEEE International Conference on Computer Communications. c2014:1285-1293.
[13]	GAMAGE S , KANGARLOU A , KOMPELLA R R , et al. Opportunis-tic flooding to improve TCP transmit performance in virtualized clouds[C]// The 2nd ACM Symposium on Cloud Computing. c2011:24-.
[14]	KANGARLOU A , GAMAGE S , KOMPELLA R R , et al. vSnoop:improving TCP throughput in virtualized environments via acknowl-edgement offload[C]// The 2010 ACM/IEEE International Conference for High Performance Computing,Networking,Storage and Analysis. c2010:1-11.
[15]	XU C , GAMAGE S , RAO P N , et al. vSlicer:latency-aware virtual machine scheduling via differentiated-frequency CPU slicing[C]// The 21st International Symposium on High-Performance Parallel and Dis-tributed Computing. c2012:3-14.
[16]	XU Y , MUSGRAVE Z , NOBLE B , et al. Bobtail:avoiding long tails in the cloud[C]// The USENIX Conference on Network Systems De-sign and Implementation. c2013:329-341.
[17]	XU Y . Characterizing and mitigating virtual machine interference in public clouds[D]. Michigan:University of Michigan, 2014.
[18]	XU Y , BAILEY M , NOBLE B , et al. Small is better:avoiding latency traps in virtualized data centers[C]// The 4th Annual Symposium on Cloud Computing. c2013.
[19]	Credit2 scheduler[EB/OL]. .
[20]	Iperf[EB/OL]. .
[21]	Sysbench[EB/OL]. .