移动云环境中数据流应用的Cloudlet选择策略研究

doi:10.11959/j.issn.1000-436x.2019020

通信学报 ›› 2019, Vol. 40 ›› Issue (1): 87-101.doi: 10.11959/j.issn.1000-436x.2019020

移动云环境中数据流应用的Cloudlet选择策略研究

刘伟¹,熊曙¹,杜薇¹,王伟¹

¹ 武汉理工大学计算机科学与技术学院，湖北武汉 430070
² 交通物联网技术湖北省重点实验室，湖北武汉 430070
³ 同济大学计算机科学与技术系，上海 200092

修回日期:2018-07-16 出版日期:2019-01-01 发布日期:2019-02-03
作者简介:刘伟（1978- ），男，湖北襄阳人，博士，武汉理工大学副教授，主要研究方向为云计算、边缘计算、绿色计算。|熊曙（1994- ），男，湖北黄冈人，武汉理工大学硕士生，主要研究方向为云计算。|杜薇（1978- ），女，湖北武汉人，博士，武汉理工大学副教授，主要研究方向为云计算、边缘计算。|王伟（1979- ），男，湖北武汉人，博士，同济大学副教授，主要研究方向为云计算、大数据。
基金资助:
国家自然科学面上基金资助项目(61672384);教育部人文社科基金资助项目(16YJCZH014);中央高校基本科研业务基金资助项目(2016Ⅲ028);中央高校基本科研业务基金资助项目(2017Ⅲ028-005)

Research on Cloudlet selection strategy for data streaming applications in mobile cloud environment

Wei LIU¹,Shu XIONG¹,Wei DU¹,Wei WANG¹

¹ School of Computer Science and Technology,Wuhan University of Technology,Wuhan 430070,China
² Hubei Key Laboratory of Transportation Internet of Things,Wuhan 430070,China
³ Department of Computer Science and Technology,Tongji University,Shanghai 200092,China

Revised:2018-07-16 Online:2019-01-01 Published:2019-02-03
Supported by:
The Chinese National Natural Science Foundation(61672384);The Fundamental Research Funds for the Central Universities(2016Ⅲ028);The Fundamental Research Funds for the Central Universities(2017Ⅲ028-005)

摘要/Abstract

摘要：

现有的Cloudlet选择策略大多只使用单个Cloudlet资源进行计算卸载，对于拥有较多可并行执行组件的移动数据流应用程序，性能提升有限。针对这一问题，提出一种基于化学反应优化算法的Cloudlet选择策略。该策略以减少应用的完成时间和移动设备能耗为目的，在满足应用程序组件间依赖关系的前提下，充分利用多Cloudlet 的计算资源使移动数据流应用的并行组件同时执行，提升了应用执行效率的同时降低了移动设备能耗。仿真实验表明，在多Cloudlet环境中应用程序的性能相较于POCSS策略平均提升了18.2%。

关键词: 微云选择, 能耗, 完成时间, 移动数据流应用, 化学反应算法

Abstract:

Most existing Cloudlet selection strategies only used the resources of one Cloudlet to compute offloading,which couldn’t obtain the superior performance improvement for mobile data streaming application with many parallel components.To address this issue,a Cloudlet selection strategy based on chemical reaction optimization algorithm was proposed.The strategy aims to reduce application’s completion time and energy consumption of mobile device.When the dependencies among application’s components was satisfied,the strategy can take full advantage of the computing capability of multi-cloudlet to execute the parallel components of mobile data stream application simultaneously.Therefore the strategy can improve the execution efficiency and reduce the energy consumption of mobile device.The simulation results reveal that the proposed strategy can achieves 18.3% on average performance improvement than POCSS strategy does in multi-Cloudlet environment.

Key words: Cloudlet selection, energy consumption, completion time, mobile data streaming application

中图分类号:

TP319

刘伟,熊曙,杜薇,王伟. 移动云环境中数据流应用的Cloudlet选择策略研究[J]. 通信学报, 2019, 40(1): 87-101.

Wei LIU,Shu XIONG,Wei DU,Wei WANG. Research on Cloudlet selection strategy for data streaming applications in mobile cloud environment[J]. Journal on Communications, 2019, 40(1): 87-101.

图/表 18

图1

图2

表1

图3

表2

表3

表4

图4

表5

图5

表6

图6

表7

图7

表8

图8

表9

图9

参考文献 33

[1]	ITU-T. ICT facts and figures 2017[R]. International Telecommunication Union, 2017.
[2]	SHAHZAD M , LIU A X , SAMUEL A . Secure unlocking of mobile touch screen devices by simple gestures:you can see it but you can not do it[C]// International Conference on Mobile Computing ＆ NETWORKING. 2013: 39-50.
[3]	AKARIMAN Q , JATI A N , NOVIANTY A . Face recognition based on the Android device using LBP algorithm[C]// International Conference on Control,Electronics,Renewable Energy and Communications. 2015: 166-170.
[4]	曹洋, 江涛, 杨世永 ,等. 移动云计算网络中的最优资源分配研究[J]. 通信学报, 2011,32(9A): 42-48.
	CAO Y , JIANG T , YANG S Y ,et al. Optimal resource allocation in mobile cloud computing network[J]. Journal on Communications, 2011,32(9A): 42-48.
[5]	崔勇, 宋健, 缪葱葱 ,等. 移动云计算研究进展与趋势[J]. 计算机学报, 2017,40(2): 273-295.
	CUI Y , SONG J , MIAO C C ,et al. Mobile cloud computing research progress and trends[J]. Chinese Journal of Computers, 2017,40(2): 273-295.
[6]	CUERVO E , BALASUBRAMANLAN A , Cho D ,et al. MAUI:making smartphones last longer with code offload[C]// Proceedings of the 8th international conference on Mobile systems,applications,and services. 2010: 49-62.
[7]	CHUN B G , IHM S , MANIATIS P ,et al. Clonecloud:elastic execution between mobile device and cloud[C]// Proceedings of the sixth conference on Computer systems. 2011: 301-314.
[8]	RA M R , SHETH A , Mummert L ,et al. Odessa:enabling interactive perception appplications on mobbile devices[C]// Proceedings of the 99th international conference on Moobile systems,appliications,and servicces. 2011: 43-56.
[9]	张文丽, 郭兵, 沈艳 ,等. 智能移动终端计算迁移研究[J]. 计算机学报, 2016,339(5): 1021-1038.
	ZHANG W L , GUO B , SHEN Y ,et al. Computation offloading on iintelligent mobbile terminal[J]. Chinese Journal oof Computers, 2016,39(5): 1021-1038.
[10]	SATYANARAYANAN M , BAHL P , CACERES R ,et al. The case ffor VM-based Cloudlets in mobile computing[J]. IEEE Pervasive Compputing, 2009,8((4): 14-23.
[11]	华夏进, 董瑞志, 彭鑫 ,等. 基于统计预测的 Cloudlet 调度机制研究[J]. 小型微型计算机系统, 2016,37(3): 406-411.
	HUA X J , DONG R Z , PENG X ,et al. Cloudleet scheduling mechhanism research based on the staatistical forecastingg[J]. Journal of Chhinese Mini-Micro Computer Sysstems, 2016,37(3): 406-411.
[12]	YANG L , CAO J , TANG S ,et al. Run time appliication repartitioninng in dynamic mobile cloud envvironments[J]. IEEE Transactions oon Cloud Compuuting, 2016,4(3): 336-348.
[13]	PILLAI P S , MUMMERT L B , SCHLOSSER S W ,et al. SLIPstream:scalable low-latency interactivee perception on strreaming data[C]// IInternational Workshop on Netwoork and Operating Systems Support ffor Digital Audioo and Video. 2009: 43-48.
[14]	MUKHERJEE A , DE D , ROY D . A power and lattency aware Cloudllet selection straategy for multi-Clloudlet environmennt[J]. IEEE Transaactions on Clouud Computing, 2016: 1.
[15]	PARMAR D , KUMAR A S , NIVANGUNE A ,et al. Discovery annd selection meechanism of Clouddlets in a decentrallized MCC enviroonment[C]// IEEE/ACM Internatioonal Conference oon Mobile Softwaare Engineering and Systems. 2016: 15-16.
[16]	TAWALBEH L , JARARWEH Y , ABABNEH F ,et al. Large scaale Cloudlets depployment for efficient mobile cloud computing[J]. Journnal of Networks,, 2015,10(1): 70-76.
[17]	SAAD H B , KASSAR M , SETHOM K . Utility-bbased Cloudlet seleection in mobille cloud computingg[C]// 2016 Global Summit on Computter＆ Informatioon Technology, 2016: 91-96.
[18]	CHILUKURI S , BOLLAPRAGADA S , KOMMINENI S ,et al. RaiinCloud-Clouddlet selection for effective cyber foraging[C]// Wireleessf Communicattions and NETWORKING Conferencee. 2017.
[19]	GAI K,QIU M , ZHAO H , ett al . Dynamic eneergy-aware Cloudleetbased mobilee cloud computing model for green computing[J]. Journnal of Network＆ Computer Appliccations, 2016,59(C): 46-54.
[20]	ROY D G , DE D , MUKHERJEE A ,et al. Applicaation-aware Cloudllet selection for computation offloaading in multi-Clouudlet environment[J]. Journal of Suupercomputing, 2017,73(4): 1-19.
[21]	SHU G , ZHENG X , XU H . Cloudlet-assistedd heuristic offloadinng for mobile innteractive applications[C]// IEEE Interrnational Conferennce on Mobile Clloud Computing,Services,and Engineering. 2017: 66-733.
[22]	LIU W , CAO J , QIU X ,et al. Improving performaance of mobile inteeractive data-sstreaming applicatiions with multiple Cloudlets[C]// IEEE International Conference on Cloud Computing Technology and Sccience. IEEE Computer Society, 2014: 46-53.
[23]	RAVI A , PEDDOJU S K . Mobbility managed energy efficient androoid mobile devicces using Cloudlet[[C]// Students' Techhnology Symposium. 2014: 402-407.
[24]	PERRUCCI G P , FITZEK F H P , WIDMER J . Survey on eenergy consumption entities on t the smartphone plaatform[C]// Vehiculaar Technologyy Conference. 2011: 1-6.
[25]	蒋廷耀, 李庆华 . DAG任务图的一种调度算法[J]. 小型微型计算机系统, 2003,24(10): 1796-1799.
	JIANG T Y , LI Q H . A scheduling algorithm for dag task graphhs[J]. Journal of Chinese Mini-Micro Computer Systems, 2003,24(10): 1796-1799.
[26]	GAREY M R , JOHNSON D S . Computters and intractabiliity:a guide to thhe theory of NP-completeness[M]. W.H. Freeman, 1986.
[27]	陈乃金, 江建慧 . 融合面积估算和多目标优化的硬件任务划分算法[J]. 通信学报, 2013(2): 40-55.
	CHEN N J , JIANG J H . Hardware-task partitioning algoritthm merged area estimation with multi-objective optimization[J]. Journal on Communications, 2013(2): 40-55.
[28]	LAM A Y S , LI V O K . Chemical-reaction-inspired metahheuristic for optiimization[J]. IEEE Transactions onn Evolutionary Coomputation, 2010,14(3): 381-399.
[29]	XU J , LAM A Y S , LII V O K . Chemical reaction optimizatiion for task scheeduling in grid computing[J]. IEEE Trransactions on Paraallel ＆ Distribuuted Systems, 2011,22(10): 1624-1631.
[30]	JAMES J Q , LAM A Y S , LI V O K . Evolutionary artiffiicial neural network based on chemical reaction optiimization[C]// IEEE Congress on Evolutionary Compputation. 2011: 2083-2090.
[31]	刘全, 王晓燕, 傅启明 ,等. 双精英协同进化遗传算法[J]. 软件学报, 2012,23(4): 765-775.
	LIU Q , WANG X Y , FU Q M ,et al. Doouble elite coevoluutionary genetic algorithm[J]. Jouurnal of Software, 2012,23(4): 765-755.
[32]	LAM A Y S , LI V O K , XU J . On the cconvergence of chemical reactionn optimization for combinatorial optimization[J]. IEEE Transactions on Evolutionary Compputation, 2013,17(55): 605-62
[33]	PAN Q K , TASGETIIREN M F , LIANG Y C . A discreete particle swaarm optimization algorithm for thee permutation floowshop sequencing problem wiith makespan criteerion[M]. Londonn: Springer, 2007: 19-31

方案编号	移动设备	C₁	C₂	C₃	完成时间/s
1	0,6	1,4,5	2	3	0.286
2	0,1,6	2	3	4,5	0.324
3	0,1,2,3,6	4,5	—	—	0.498
4	0,6	2,3,4,5	1	—	0.329
5	0,6	1,2,3,4,5	—	—	0.370
6	0,1,2,3,4,5,6	null	—	—	0.660

Cloudlet	内存/GB	CPU
Cloudlet₁	16	Intel(R) Xeon(R) E5-2407 0@ 2.20 GHz
Cloudlet₂	32	Intel(R) Xeon(R)E5-2620v3@ 2.40 GHz
Cloudlet₃	32	Intel(R) Xeon(R)E5-2620v2@ 2.10 GHz

移动设备	内存/GB	CPU	CPU功率		Wi-Fi功率
移动设备	内存/GB	CPU	运行/mw	空闲/mw	发送/mw	接收/mw
MI 5	3	MSM8996@1.8 GHz	610	8	120	120

实验环境	参数值
Cloudlet个数/个	3
移动设备与Cloudlet之间带宽/（Mbit.s^-1）	8
并行组件指令数占比	40%
参数α	0.5
Cloudlet之间的带宽/（Mbit.s^-1）	10,30,50

实验环境	参数值
Cloudlet之间的带宽/(Mbit.s^-1)	210
移动设备与Cloudlet之间的带宽/(Mbit.s^-1)	8
并行组件指令数占比	40%
参数α	0.5
Cloudlet个数/个	1,2,3,4

移动云环境中数据流应用的Cloudlet选择策略研究

Research on Cloudlet selection strategy for data streaming applications in mobile cloud environment

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 18

参考文献 33

相关文章 15

Metrics

推荐阅读 0

算法	寻找到全局最优解的概率	平均适应度	最低适应度	最大适应度	标准差
GA	13.33%	0.466 9	0.455 666	0.527 732	0.163 5
DPSO	43.33%	0.458 7	0.455 666	0.465 812	0.003 2
CROCS	60%	0.452 7	0.455 666	0.462 26	0.002 4

[1]	鲁蔚锋, 李宁, 徐佳, 徐力杰, 徐建. 多接入边缘计算中相关性任务的联合调度算法[J]. 通信学报, 2023, 44(4): 87-98.
[2]	余雪勇, 邱礼翔, 宋家宁, 朱洪波. 无人机辅助边缘计算中安全通信与能效优化策略[J]. 通信学报, 2023, 44(3): 45-54.
[3]	李翠然, 王雪洁, 谢健骊, 吕安琪. 基于改进PSO的铁路监测线性无线传感器网络路由算法[J]. 通信学报, 2022, 43(5): 155-165.
[4]	任品毅,许茜. 基于移动边缘计算的时延能耗最小化安全传输[J]. 通信学报, 2020, 41(11): 52-63.
[5]	蒲勇霖,于炯,鲁亮,李梓杨,卞琛,廖彬. 基于Storm平台的数据迁移合并节能策略[J]. 通信学报, 2019, 40(12): 68-85.
[6]	毛艳艳,程大鹏,冯烟利,窦全胜,李大社. C3S：基于相长干涉的智能传感系统并发传输策略研究[J]. 通信学报, 2019, 40(1): 180-194.
[7]	梁俊斌,周翔,李陶深. 移动低占空比无线传感网中低能耗的主动邻居发现算法[J]. 通信学报, 2018, 39(4): 45-55.
[8]	蒲勇霖,于炯,鲁亮,卞琛,廖彬,李梓杨. storm平台下工作节点的内存电压调控节能策略[J]. 通信学报, 2018, 39(10): 97-117.
[9]	何灏,陈永锐,易卫东,李鸣. 无线可充电传感器网络中固定充电器的部署策略[J]. 通信学报, 2017, 38(Z1): 156-164.
[10]	黄焱,王鹏,谢高辉. 基于PE方法的数据中心需量费用优化算法[J]. 通信学报, 2016, 37(3): 90-97.
[11]	虞贵财,龙承志,向满天. 认知无线传感器网络中能耗有效的协作频谱感测算法[J]. 通信学报, 2015, 36(3): 52-62.
[12]	肖鹏,胡志刚,屈喜龙. 面向数据密集型工作流的能耗感知调度策略[J]. 通信学报, 2015, 36(1): 149-158.
[13]	谷勇浩，郭达，林九川. 低能耗的隐私数据安全融合方法[J]. 通信学报, 2014, 35(Z2): 15-116.
[14]	谷勇浩,郭达,林九川. 低能耗的隐私数据安全融合方法[J]. 通信学报, 2014, 35(Z2): 112-116.
[15]	丁麟轩,黄昆,张大方. 基于TCAM的低能耗正则表达式匹配算法[J]. 通信学报, 2014, 35(8): 162-168.