考虑虚拟机间性能互扰的虚拟资源分配方法

doi:10.3969/j.issn.1000-436x.2014.09.008

Abstract

Abstract:

A method for allocating the virtual resource with the consideration of the performance interference among vir-tual machines was proposed.The model for evaluating the performance interference among virtual machines was formed and the method for predicting the performance interference was presented based on the workload pattern of the virtual machine.The predicted result of the performance interference among virtual machines was put into the evaluation of the resource allocation plan to be used as the basis for resource allocation.By doing this,the effectiveness of the resource al-location plan could be insured.The experiments show that the proposed virtual resource allocation method can insure the execution performance of the application deployed on the virtual machines.

Key words: performance interference among virtual machines, virtual resource allocation, load pattern, prediction of performance interference

Lei YANG,He-xu XING,Yu DAI,Sai XU,Hao WANG,Bin ZHANG. Virtual resource allocation method with the consideration of performance interference among virtual machines[J]. Journal on Communications, 2014, 35(9): 79-90.

Figures/Tables 14

References 12

[1]	罗军舟，金嘉晖，宋爱波．云计算—体系架构与关键技术[J]．通信学报， 2011,32(7):3-21. LUO J Z , JIN J H , SONG A B . Cloud computing: architecture and key technologies[J]. Journal on Communications, 2011,32(7):3-21.
[2]	CHEN M , ZHANG H , SU Y Y . Effective VM sizing in virtualized data centers[A]. 12th IFIP/IEEE International Symposium on Integrated Network Management (IM'2011)[C]. Dublin,Ireland, 2011.594-601.
[3]	LEITE D , PEIXOTO M , SANTANA M . Performance evaluation of virtual machine monitors for cloud computing[A]. Symposium on Computer Systems (WSCAD-SSC'2012)[C]. De outubro,Brasil, 2012.65-71.
[4]	HIROTOGU A . A new look at the statistical model identification[J]. IEEE Transactions on Automatic Control, 1974,19(6):716-723.
[5]	ULRIKE L . A tutorial on spectral clustering[J]. Statistics and Computing, 2007,17(4):395-476.
[6]	KOH Y , KNAUERHASE R , BRETT P . An analysis of performance interference effects in virtual environments[A]. IEEE International Symposium on Performance Analysis of Systems and Software (IS-PASS' 2007)[C]. San Jose,California,USA, 2007.200-209.
[7]	潘飞，蒋从锋，徐向华．负载相关的虚拟机放置策略[J]．小型微型计算机系统， 2013,34(3):520-524. PAN F , JIANG C F , XU X H . Placement strategy of virtual machines based on workload characteristics[J]. Journal of Chinese Computer Systems, 2013,34(3):520-524.
[8]	ANH V D , JUNLIANG C , CHEN W . Profiling applications for virtual machine placement in clouds[A]. 4th IEEE 2011 International Con-ference on Cloud Computing (CLOUD 2011)[C]. Honolulu,Hawaii,USA, 2011.660-667.
[9]	LIN J W , CHEN C H . Interference-aware virtual machine placement in cloud computing system[A]. International Conference on Computer ＆Information Science (ICCIS'2012)[C]. Kuala Lumpur,Malaysia, 2012.598-603.
[10]	CHIANG R C , HUANG H H . TRACON: interference-aware schedul-ing for data-intensive applications in virtualized environments[A]. In-ternational Conference for High Performance Computing,Networking,Storage and Analysis(SC'2011)[C]. New York,NY,USA, 2011.1-12.
[11]	VAN H N , TRAN F D , MENAUD J M . Performance and power man-agement for cloud infrastructures[A]. IEEE International Conference on Cloud Computing (CLOUD'2010)[C]. Miami,USA, 2010.329-336.
[12]	XU J , FORTES J A B . Multi-objective virtual machine placement in virtualized data center environments[A]. IEEE/ACM Int'l Conference on Green Computing and Communications (GreenCom'2010) ＆ Int'l Conference on Cyber,Physical and Social Computing (CPSCom'2010)[C]. Hangzhou,China, 2010.179-188.

Metrics

Recommended 0

No Suggested Reading articles found!

应用名称	应用类型
Super PI	CPU密集型
cat	Disk I/O密集型
Bzip2	混合型
Cp	Disk I/O密集型
Gzip	混合型

	Super PI	cat	Bzip2
Super PI	1.018	1.009	1.021
cat	1.465	2.108	1.114
Bzip2	1.066	1.760	1.113

App	cpuutil	memutil	rps	wps	await/s	svctm	Runtime/s
Bzip2	0.91	0.007	278.099	0.999	0.775	0.345	84.265
cat	0.001	0.044	335.158	433.111	137.643	1.58	27.801
Super PI	0.98	0.001	0.245	1.547	12.222	9.39	99.107
Iozone	0.264	0.036	370.25	392.95	151.99	9.93	108.724
Ccrypt	0.762	0.421	172.77	168.68	5.162	2.13	216.611
Gzip	0.912	0.053	295.72	11.83	8.44	1.18	218.95

	Bzip2	cat	Super PI	Iozone	Ccrypt	Gzip
Bzip2	93.887	148.28	89.787	146.988	142.82	144.168
cat	30.996	58.892	40.762	40.762	45.104	50.026
Super PI	101.141	99.981	100.892	121.99	105.21	103.54
Iozone	175.33	205.27	109.756	198.73	110.26	113.64
Ccrypt	245.03	296.55	233.50	311.73	257.91	254.03
Gzip	230.75	393.99	227.09	403.74	245.10	240.02

rps	wps	await	svctm	predicted PID	Observed PID	error
141.238	121.428	5.765	1.973	0.398 6	0.401 4	0.7%
175.498	31.716	3.634	2.033	0.189 1	0.198	4.5%
264.861	251.134	5.751	5.452	0.439 2	0.385 7	13.9%
0.762	0.468	9.172	5.874	0.077 4	0.080 2	3.5%
7.156	65.821	7.844	2.165	0.314 5	0.301 7	4.2%
264.894	335.418	38.551	2.968	0.731 9	1.042 9	29.8%
568.563	142.745	67.854	1.254	0.351 4	0.453 9	22.6%
10.495	3.154	24.264	3.746	0.138 8	0.137 2	1.2%

Virtual resource allocation method with the consideration of performance interference among virtual machines

RichHTML

PDF

Knowledge

Cited

Abstract

Cite this article

share this article

Figures/Tables 14

References 12

Related Articles 1

Metrics

Recommended 0

Bzip2	cat	Super PI	Iozone	Ccrypt	Gzip
0.549	0.229	0.515	0.216	0.337	0.526