E.V. Zharikov
National Technical University of Ukraine «Igor Sikorsky Kyiv Polytechnic Institute»

Èlektron. model. 2018, 40(5):49-66


Providing a given quality of cloud services under non-stationary workload is one of the main tasks in managing a cloud data center. To ensure a given service quality, it is necessary to apply a proactive approach to managing computing resources. It is possible to prevent problems of inadequate allocation of resources or excessive allocation of resources by forecasting the consumption of resources by virtual machines or containers. In this paper, the author proposes an adaptive method for forecasting the consumption of computational resources, which provides a smaller forecasting error than using a single forecasting method with a model obtained with fixed-size training data. The evaluation of the proposed method show that the accuracy of the forecast increases on average from 2.4% to 23.6%, depending on the statistical characteristics of the time series presented by monitoring data. Increasing the accuracy of the computing resources consumption forecast allows to reduce power consumption and to reduce the number of service-level agreement violation by more accurately allocating the necessary resources to the virtualized ap- plications of the cloud data center.


cloud computing, data center, forecasting, time series, virtualization, energy efficiency.


  1. Barham, Dragovic, B., Fraser, K., Hand, S., Harris, T., Ho, A., and Warfield, A. (2003), «Xen and the art of virtualization», In ACM SIGOPS operating systems review, vol. 37, no. 5, pp. 164—177.
  2. Chen, , He, W., Liu, J. et al (2008), «Energy-Aware Server Provisioning and Load Dispatching for Connection-Intensive Internet Services», In NSDI, Vol. 8, pp. 337-350.
  3. Padala, Hou, K. Y., Shin, K. G., et al (2009), «Automated control of multiple virtualized resources», Proc. of the ACM European conference on Computer systems (EuroSys'09).
  4. Gross, , and Galiana, F. D. (1987), «Short-term load forecasting», Proceedings of the IEEE, 75(12), pp. 1558—1573.
  5. Xiao, , Song, W., and Chen, Q. (2013), «Dynamic resource allocation using virtual machines for cloud computing environment», Parallel and Distributed Systems, IEEE Transac-tions on, vol. 24, no. 6, pp. 1107—1117.
  6. Xue,, Yan, F., Birke, R. et al (2015), «PRACTISE: Robust prediction of data center time series», 11th International Conference on Network and Service Management (CNSM), pp. 126—134.
  7. Farahnakian, , Liljeberg, P., and Plosila, J. (2013), «LiRCUP: Linear regression based CPU usage prediction algorithm for live migration of virtual machines in data centers», 39th EUROMICRO Conference on Software Engineering and Advanced Applications (SEAA), pp. 357—364.
  8. Islam, Keung, J., Lee, K., and Liu, A. (2012), «Empirical prediction models for adaptive resource provisioning in the cloud», Future Generation Computer Systems, vol. 28, no. 1, pp. 155—162.
  9. Dabbagh, M., Hamdaoui, B., Guizani, M., and Rayes, A. (2015), «Energy-efficient resource allocation and provisioning framework for cloud data centers», IEEE Transactions on Network and Service Management, vol. 12, no. 3, pp. 377—391.
  10. Naseera, Rajini, G. K., Prabha, N. A., and Abhishek, G. (2015), «A comparative study on CPU load predictions in a computational grid using artificial neural network algorithms», Indian Journal of Science and Technology, vol. 8, no. 35.
  11. Naseera, , Rajini, G. K., and Reddy, P. S. K. (2016), «Host CPU Load Prediction Using Statistical Algorithms a comparative study», International Journal of Computer Technology and Applications, 9 (12). pp. 5577—5582.
  12. Dinda, A. (2006), «Design, implementation, and performance of an extensible toolkit for resource prediction in distributed systems», IEEE Transactions on Parallel and Distributed Systems, vol. 17, no. 2, pp. 160—173.
  13. Box, E., Jenkins, G. M., Reinsel, G. C., and Ljung, G. M. (2015), «Time series analysis: fore-casting and control», 5th ed. Hoboken, NJ, USA: John Wiley & Sons.
  14. Montgomery, Douglas , Elizabeth A. Peck, and G. Geoffrey Vining, (2015). «Introduction to linear regression analysis», John Wiley & Sons.
  15. Park, , and Pai, V. S. (2006), «CoMon: a mostly-scalable monitoring system for PlanetLab», ACM SIGOPS Operating Systems Review, pp. 65-47.
  16. Telenyk, Zharikov, E., Rolik, O. (2017), «Architecture and Conceptual Bases of Cloud IT Infrastructure Management», Advances in Intelligent Systems and Computing, Springer, vol. 512, pp. 41—62.
  17. R Core Team (2018), «R: A language and environment for statistical computing», R Foundation for Statistical Computing, Vienna, Austria. URL
  18. Jorgensen, (1995), «Experience with the accuracy of software maintenance task effort prediction models», IEEE Transactions on Software Engineering, vol. 21, pp. 674—681.
  19. Hyndman, , Koehler, A. B., Ord, J. K., and Snyder, R. D. (2008). «Forecasting with exponential smoothing: the state space approach», Springer Science & Business Media.
  20. Holt C. (2004), «Forecasting seasonals and trends by exponentially weighted moving averages», International journal of forecasting, vol. 20, no. 1, pp. 5-10.
  21. Gardner Jr S., McKenzie E. D. (1985), «Forecasting trends in time series», Management Science, vol. 31, no. 10, pp. 1237-1246.
  22. GWA-T-12 Bitbrains [Online] Available from: bitbrains [Accessed September 12, 2018].
  23. Hyndman, J., and Khandakar, Y. (2008), «Automatic time series forecasting: The forecast package for R», Journal of Statistical Software, 27(1), pp. 1—22. Retrieved from
  24. Shen, , van Beek, V., and Iosup, A. (2015), «Statistical characterization of business-critical workloads hosted in cloud datacenters», 15th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid), pp. 465—474.
  25. Hyndman R, Bergmeir C, Caceres G, et al (2018), «_forecast: Forecasting functions for time series and linear models_», R package version>.

Full text: PDF