臺灣博碩士論文加值系統

隨著硬體及網路的進步，1966年提出的運算資源即是公用資源的概念（Computing as a Utility），終於得以實現，並在近年內取得了商業上的矚目，它即是現今的雲端運算。雲端運算提供了一個嶄新的IT服務模型──應用程式及服務供應商不需要自行購置數據中心，轉而向雲端供應商依需求租賃運算資源。然而，現今雲端供應商眾多，各家的硬體資源及計價方式都不盡相同，在缺乏標準化及資訊不公開的情況下，雲端使用者難以比較不同的雲端服務的效能及價格，並確保雲端的移植確實帶來成本的節約。在此等前提下，對一個欲將自身Web應用程式移植至雲端平台的使用者來說，如何依程式的特性，來選擇適當的雲端平台，是一件困難的事。本研究致力於提供一套成本及效能導向的工具組，依程式的特性及使用者的效能品質需求，來選擇適當的雲端平台及預測所需要的部署成本。

With cloud business gowning, many companies are joining the market to be a cloud service provider. Most providers offer similar services with slightly different charging models, and very little known performance comparison between them is published. This leaves cloud service users with the puzzle of guessing what cost they will need to pay to run their legacy application in a cloud environment.

CloudGuide is a tool suite that provides user an estimated cost of running a legacy application with QoS guarantee on different cloud providers. CloudGuide predicts the cloud computing resources required by a targeted application based on queuing model, meanwhile estimates monetary deployment cost for the application. CloudGuide also allows the user to explore cloud configuration can guarantee different levels of QoS. In our evaluation, we conducted experiments of a multi-tiered network application, RUBiS, and showed that CloudGuide can choose cloud configuration that guarantee QoS and provide cost estimation.

口試委員會審定書 i
致謝 ii
Abstract iii
中文摘要 iv
1 Introduction 1
2 Problem Statement 5
2.1 Cost calculation 6
2.2 Performance requirements 7
2.3 Satisfying requirements based on policy 8
2.3.1 Policy I: Minimize deployment cost while satisfying peak through-put 8
2.3.2 Policy II: Maximize throughput given monetary budget 9
2.3.3 Policy Example: Improve throughput by half and keep minimum cost 9
2.4 Design goals 10
3 Meeting Performance Requirement 12
3.1 Establishing base capacity 13
3.2 Determining capacity for different instance type 14
3.3 Finding speedup factor 14
4 Implementation 18
4.1 Overview 18
4.2 Cloud Providers Pricing Schemes and Performance 20
4.2.1 Chosen cloud providers 20
4.2.2 Performance Comparison Baseline 21
4.3 Estimating local and cloud server capacity 22
4.4 Finding configuration and cost that satisfy policy 23
5 Evaluation 24
5.1 Testbed application 24
5.2 Determining initial cloud configuration 25
5.3 Dynamic provisioning when workload changes 27
6 Related Work 28
7 Conclusion 30
Bibliography 31

Bibliography
[1] Amazon web service. http://aws.amazon.com.
[2] Hicloud. http://hicloud.hinet.net.
[3] Ntu cloud. https://www.cloud.ntu.edu.tw/.
[4] Rackspace. http://www.rackspace.com/.
[5] Armbrust, M., Fox, A., Griffith, R., Joseph, A. D., Katz, R. H., Konwinski, A., Lee,
G., Patterson, D. A., Rabkin, A., Stoica, I., and Zaharia, M. Above the clouds: A
berkeley view of cloud computing. Tech. Rep. UCB/EECS-2009-28, EECS Depart-
ment, University of California, Berkeley, Feb 2009.
[6] Cecchet, E., Chanda, A., Elnikety, S., Marguertie, J., and Zwaenepoel, W. Perfor-
mance comparison of middleware architectures for generating dynamie web content.
In Proceedings of the 4th ACM/IFIP/USENIX International Middleware Confernce
(Rio de Janeiro, Brazil, June 2003).
[7] Cloudharmony. http://www.cloudharmony.com/benchmarks.
[8] Csirik, J., Frenk, J. B. G., Labbe, M., and Zhang, S. Heuristics for the 0-1 min-
knapsack problem. Acta Cybernetica 10, 1--2 (September 1991), 15--20.
[9] Dean, J., and Ghemawat, S. Mapreduce: Simplified data processing on large clusters.
In Proceedings of the 6th Symposium on Operating Systems Design and Implemen-
tation (San Francisco, CA, December 2004).
31[10] Doyle, R. P., Chase, J. S., Asad, O. M., Jin, W., and Vahdat, A. M. Model-based
resource provisioning in a web service utility. In Proceedings of the 4th conference
on USENIX Symposium on Internet Technologies and Systems - Volume 4 (Berkeley,
CA, USA, 2003), USITS''03, USENIX Association, pp. 5--5.
[11] Garfinkel, S. L. An evaluation of amazon''s grid computing services: Ec2, s3, and
sqs. Tech. rep., Harvard University, 2007.
[12] Hajjat, M., Sun, X., and)David Maltz, Y.-W. E. S., Rao, S., Sripanidkulchai, K., and
TawarmalaFni, M. Cloudward bound: Planning for beneficial migration of enterprise
applications to the cloud. In Proceedings of SIGCOMM 2010 (New Helhi, INDIA,
2010).
[13] Hoste†, K., Phansalkar, A., Eeckhout, L., Georges†, A., John, L. K., and Bosschere,
K. D. Performance prediction based on inherent program similarity. In Proceedings
of PACT''06 (Seattle, Washington, USA, September 2006).
[14] Kleinrock, L. Queueing Systems Volume II: Computer Applications. Wiley-
Interscience, 1976, pp. 29--50.
[15] Kossmann, D., Kraska, T., and Loesing, S. An evaluation of alternative architec-
tures for transaction processing in the cloud. In The Proceedings of SIGMOD ''10
(Indianapolis, Indiana, USA, June 2010), pp. 579 -- 590.
[16] Kundu, S., Rangaswami, R., Gulati, A., Zhao, M., and Dutta, K. Modeling vir-
tualized applications using machine learning techniques. In Proceedings of Eighth
Annual International Conference on Virtual Execution Environments (VEE 2012)
(London, UK, March 2012), pp. 3--14.
32[17] Li, A., Yang, X., Sandula, S., and Zhang, M. CloudCmp: Comparing Public Cloud
Providers. In Internet Measurement Confernce (November 2010).
[18] Li, A., Zong, X., Zhang, M., Kandula, S., and Yang, X. Cloudprophet: Predicting
web application performance in the cloud. Tech. rep., Duke University, 2011.
[19] Li, X., Li, Y., Liu, T., Qiu, J., and Wang, F. The method and tool of cost analysis for
cloud computing. In Proceedings of 2009 IEEE International Conference on Cloud
Computing (Bangalore, India, September 2009).
[20] Oprofile. http://oprofile.sourceforge.net.
[21] Schad, J., Dittrich, J., and Quiane-Ruiz, J.-A. Runtime measurements in the cloud:
Observing, analyzing, and reducing variance. Proceedings of the VLDB Endowment
3, 1 (2010).
[22] Sharma, U., Shenoy, P., Sahu, S., and Shaikh, A. A cost-aware elasticity provisioning
system for the cloud. In The Proceedings of the 31st International Conference on
Distributed Computing Systems (ICDCS) (June 2011).
[23] Shimizu, S., Rangaswami, R., Duran-Limon, H. A., and Corona-Perez, M. Platform-
independent modeling and prediction of application resource usage characteristics.
J. Syst. Softw. 82, 12 (Dec. 2009), 2117--2127.
[24] Stewart, C., Kelly, T., Zhang, A., and Shen, K. A dollar from 15 cents: Cross-
platform management for internet services. In In USENIX (2008).
[25] Stewart, C., and Shen, K. Performance modeling and system management for multi-
component online services. In Proceedings of the 2nd Symposium on Networked
Systems Design and Implementation (NSDI ''05) (Boston, MA, May 2005).
33[26] sysstat. http://sebastien.godard.pagesperso-orange.fr/t.
[27] Tak, B. C., Urgaonkar, B., and Sivasubramaniam, A. To move or not to move: The
economics of cloud computing. In Proceedings of 3rd USENIX Workshop on Hot
Topics in Cloud Computing (Portland, OR, June 2011).
[28] Tergowda, P. B., Urgaonkar, B., and Giles, C. L. Citeseerx: A cloud perspective. In
Proceedings of HotCloud 2010 (Boston, MA, June 2010).
[29] Urgaonkar, B., Shenoy, P., Chandra, A., and Goyal, P. Agile dynamic provisioning
of multi-tier internet applications. ACM Transactions on Autonomous and Adaptive
Systems (TAAS), 1 (March 2008).
[30] Walker, E. Benchmarking amazon ec2 for high-performance scientific computing.
USENIX Login (2008).
[31] Wieder, A., Bhatotia, P., Post, A., and Rodrigues, R. Conductor: Orchestrating the
clouds. In Proceedings of the 4th Workshop on Large Scale Distributed Systems and
Middleware (LADIS) (Zurich, Switzerland, July 2010).