›› 2017,Vol. 32 ›› Issue (5): 936-956.doi: 10.1007/s11390-017-1774-3

• Special Section on Selected Paper from NPC 2011 • 上一篇    下一篇

面向仅有稀疏数据的移动云应用的一种自动化性能预测方法

Wei-Qing, Liu Jing Li*, Member, CCF, IEEE   

  1. School of Computer Science and Technology, University of Science and Technology of China, Hefei 230026, China
  • 收稿日期:2016-06-17 修回日期:2016-11-02 出版日期:2017-09-05 发布日期:2017-09-05
  • 通讯作者: Jing Li,lj@ustc.edu.cn E-mail:lj@ustc.edu.cn
  • 作者简介:Wei-Qing Liu received his B.E. degree in computer science and technology from the University of Science and Technology of China (USTC), Hefei, in 2011. He is currently a Ph.D. candidate at the School of Computer Science and Technology in USTC, Hefei. His research interests include cloud computing, mobile computing and big data processing.
  • 基金资助:

    This paper was supported by the National High Technology Research and Development 863 Program of China under Grant No. 2014AA01A302.

An Approach to Automatic Performance Prediction for Cloud-enhanced Mobile Applications with Sparse Data

Wei-Qing, Liu Jing Li*, Member, CCF, IEEE   

  1. School of Computer Science and Technology, University of Science and Technology of China, Hefei 230026, China
  • Received:2016-06-17 Revised:2016-11-02 Online:2017-09-05 Published:2017-09-05
  • Contact: Jing Li,lj@ustc.edu.cn E-mail:lj@ustc.edu.cn
  • About author:Wei-Qing Liu received his B.E. degree in computer science and technology from the University of Science and Technology of China (USTC), Hefei, in 2011. He is currently a Ph.D. candidate at the School of Computer Science and Technology in USTC, Hefei. His research interests include cloud computing, mobile computing and big data processing.
  • Supported by:

    This paper was supported by the National High Technology Research and Development 863 Program of China under Grant No. 2014AA01A302.

在移动云计算中,将移动应用中计算密集的部分迁移到云上执行从而提升应用的性能有着很大的吸引力。而为了做出好的迁移决策,已有的研究工作利用机器学习技术基于历史数据预测应用在不同迁移方案下将达到的不同性能。然而,在真实的移动云计算场景中,应用的历史数据往往都是稀疏的,这一特性所带来的问题在已有的研究中却很少被关注。在本文中,我们针对数据稀疏这一特性设计了两阶段的混合框架来预测移动云应用的性能。第一阶段利用已有的历史数据训练若干个多层神经网络以预测一系列的运行中间参数。此训练得来的一部分神经网络能够在多个应用之间复用,从而减轻数据稀疏带来的影响。基于第一阶段预测的中间参数,第二阶段利用确定性的算法估计应用总体的性能。即便对于新发布的应用,在历史执行数据极少的情况下,确定性的算法也能够在一定程度上保证预测的准确度。我们将这一框架应用于一个带有云加速的物体识别应用上进行验证,结果显示我们的方案能够准确地预测应用的性能并且对数据稀疏问题有较好的鲁棒性。

Abstract: In mobile cloud Computing (MCC), offloading compute-intensive parts of a mobile application onto the cloud is an attractive method to enhance application performance. To make good offloading decisions, history-based machine-learning techniques are proposed to predict application performance under various offloading schemes. However, the data sparsity problem is common in a realistic MCC scenario but is rarely the concern of existing works. In this paper, we employ a two-phase hybrid framework to predict performance for cloud-enhanced mobile applications, which is designed to be robust to the data sparsity. By training several multi-layer neural networks with historical execution records, the first phase automatically predicts some intermediate parameters for each execution of an application. The models learned by these neural networks can be shared among different applications thus alleviating the data sparsity. Based on these predicted intermediate parameters and the application topology, the second phase deterministically calculates the estimated values of the performance metrics. The deterministic algorithm can partially guarantee the prediction accuracy of newly published applications even with no execution records. We evaluate our approach with a cloud-enhanced object recognition application and show that our approach can precisely predict the application performance and is robust to data sparsity.

[1] Cuervo E, Balasubramanian A, Cho D K, Wolman A, Saroiu S, Chandra R, Bahl P. MAUI:Making smartphones last longer with code offload. In Proc. the 8th International Conference on Mobile Systems, Applications, and Services, Jun. 2010, pp.49-62.

[2] Chun B G, Ihm S, Maniatis P, Naik M, Patti A. CloneCloud:Elastic execution between mobile device and cloud. In Proc. the 6th Conference on Computer Systems, Apr. 2011, pp.301-314.

[3] Gordon M S, Jamshidi D A, Mahlke S, Mao Z M, Chen X. COMET:Code offload by migrating execution transparently. In Proc. the 10th USENIX Symposium on Operating Systems Design and Implementation, Oct. 2012, pp.93-106.

[4] Ra M R, Sheth A, Mummert L, Pillai P, Wetherall D, Govindan R. Odessa:Enabling interactive perception applications on mobile devices. In Proc. the 9th International Conference on Mobile Systems, Applications, and Services, Jun. 28-Jul. 1, 2011, pp.43-56.

[5] Kosta S, Aucinas A, Hui P, Mortier R, Zhang X W. ThinkAir:Dynamic resource allocation and parallel execution in the cloud for mobile code offloading. In Proc. IEEE INFOCOM, Mar. 2012, pp.945-953.

[6] Zhang X W, Kunjithapatham A, Jeong S, Gibbs S. Towards an elastic application model for augmenting the computing capabilities of mobile devices with cloud computing. Mobile Networks and Applications, 2011, 16(3):270-284.

[7] Kumar K, Lu Y H. Cloud computing for mobile users:Can offloading computation save energy? Computer, 2010, 43(4):51-56.

[8] Wang C, Li Z Y. Parametric analysis for adaptive computation offloading. ACM SIGPLAN Notices, 2004, 39(6):119-130.

[9] Ipek E, de Supinski B R, Schulz M, McKee S A. An approach to performance prediction for parallel applications. In Lecture Notes in Computer Science 3648, Cunha J C, Medeiros P D (eds.), Springer-Verlag, 2005, pp.196-205.

[10] Narayanan D, Flinn J, Satyanarayanan M. Using history to improve mobile application adaptation. In Proc. the 3rd IEEE Workshop on Mobile Computing Systems and Applications, Dec. 2000, pp.31-40.

[11] Lee B C, Brooks D M, de Supinski B R, Schulz M, Singh K, McKee S A. Methods of inference and learning for performance modeling of parallel applications. In Proc. the 12th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, Mar. 2007, pp.249-258.

[12] Flinn J, Park S, Satyanarayanan M. Balancing performance, energy, and quality in pervasive computing. In Proc. the 22nd International Conference on Distributed Computing Systems, Jul. 2002, pp.217-226.

[13] Hecht-Nielsen R. Theory of the back-propagation neural network. Neural Networks, 1988, 1(Supplement1):445-448.

[14] Satyanarayanan M, Bahl P, Caceres R, Davies N. The case for VM-based cloudlets in mobile computing. IEEE Pervasive Computing, 2009, 8(4):14-23.

[15] Giurgiu I, Riva O, Alonso G. Dynamic software deployment from clouds to mobile devices. In Lecture Notes in Computer Science 7662, Narasimhan P, Triantafillou P (eds.), Springer, 2012, pp.394-414.

[16] Newton R, Toledo S, Girod L, Balakrishnan H, Madden S. Wishbone:Profile-based partitioning for sensornet applications. In Proc. the 6th USENIX Symposium on Networked Systems Design and Implementation, Apr. 2009, pp.395-408.

[17] Kim K, De La Garza J M. Evaluation of the resourceconstrained critical path method algorithms. Journal of Construction Engineering and Management, 2005, 131(5):522-532.

[18] Bengio Y, Ducharme R, Vincent P, Janvin C. A neural probabilistic language model. Journal of Machine Learning Research, 2003, 3:1137-1155.

[19] Collobert R, Weston J. A unified architecture for natural language processing:Deep neural networks with multitask learning. In Proc. the 25th International Conference on Machine Learning, Jul. 2008, pp.160-167.

[20] Carbone M, Rizzo L. Dummynet revisited. ACM SIGCOMM Computer Communication Review, 2010, 40(2):12-20.

[21] Jain M, Dovrolis C. End-to-end available bandwidth:Measurement methodology, dynamics, and relation with TCP throughput. IEEE/ACM Transactions on Networking, 2003, 11(4):537-549.

[22] Bergstra J S, Bardenet R, Benjio Y, Kégl B. Algorithms for hyper-parameter optimization. In Proc. the 24th International Conference on Neural Information Processing Systems, Dec. 2011, pp.2546-2554.

[23] Bergstra J, Bengio Y. Random search for hyper-parameter optimization. The Journal of Machine Learning Research, 2012, 13:281-305.

[24] Balan R K, Satyanarayanan M, Park S Y, Okoshi T. Tactics-based remote execution for mobile computing. In Proc. the 1st International Conference on Mobile Systems, Applications and Services, May 2003, pp.273-286.

[25] Cormen T H, Leiserson C E, Rivest R L, Stein C. Introduction to Algorithms (3rd edition). The MIT Press, 2009.

[26] Huang L, Jia J Z, Yu B, Chun B G, Maniatis P, Naik M. Predicting execution time of computer programs using sparse polynomial regression. In Proc. the 23rd International Conference on Neural Information Processing Systems, Dec. 2010, pp.883-891.

[27] Kwon Y, Lee S, Yi H, Kwon D, Yang S, Chun B G, Huang L, Maniatis P, Naik M, Paek Y. Mantis:Automatic performance prediction for smartphone applications. In Proc. the USENIX Conference on Annual Technical Conference, Jun. 2013, pp.297-308.

[28] Shi C, Habak K, Pandurangan P, Ammar M, Naik M, Zegura E. COSMOS:Computation offloading as a service for mobile devices. In Proc. the 15th ACM International Symposium on Mobile Ad Hoc Networking and Computing, Aug. 2014, pp.287-296.
No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
[1] 韩建超; 史忠植;. Formalizing Default Reasoning[J]. , 1990, 5(4): 374 -378 .
[2] Harald E. Otto;. UNDO, An Aid for Explorative Learning?[J]. , 1992, 7(3): 226 -236 .
[3] I.V.Vel bitsky; A.L.Kovalev; I.V.Kasatkina; 王镭;. R-Technology of Programming: Basic Notions and Implementation[J]. , 1992, 7(4): 345 -355 .
[4] 魏华; 罗予频; 杨士元;. Fault Tolerance of Reconfigurable Bi-Directional Double-Loop LANs[J]. , 1999, 14(4): 379 -385 .
[5] . 半封闭数据立方体:一种有效平衡数据立方体体积和查询响应时间的实现方法[J]. , 2005, 20(3): 367 -372 .
[6] . 水彩的实时动画[J]. , 2006, 21(2): 159 -165 .
[7] . 暂缺[J]. , 2008, 23(5 ): 851 -861 .
[8] Leonardo Liao, Yong-Qiang Wu. 用于监督学习的具有层次结构的多边体自适应共振理论映射网络[J]. , 2010, 25(5): 1071 -1082 .
[9] Ke-Qing He, Jian Wang, Peng Liang . 服务需求精化中的语义互操作性聚合方法[J]. , 2010, 25(6): 1103 -1117 .
[10] Jun-Cheng Huang (黄俊成), Member, ACM, IEEE, Xiu-Qi Li (李秀琦), Member, ACM, IEEE and Jie Wu (吴杰), Member, ACM, Fellow, IEEE. 一个用于点对点网络的基于语义的多样化内容搜索模型[J]. , 2011, 26(6): 925 -941 .
版权所有 © 《计算机科学技术学报》编辑部
本系统由北京玛格泰克科技发展有限公司设计开发 技术支持:support@magtech.com.cn
总访问量: