Cloud datacenter workload estimation using error preventive time series forecasting models

  • Jitendra KumarEmail author
  • Ashutosh Kumar Singh


The workload estimation plays a vital role in efficient management of cloud resources. This paper introduces the error preventive score (EPS) in time series forecasting models to improve the prediction accuracy. The EPS analyzes the most recent estimations to capture the forecast error trend and uses it to achieve better forecasts. In addition, we have also proposed two metrics for accuracy evaluation namely predictions in error range (PER) and magnitude of predictions (MoP). These matrices favor a model that has maximum predictions close to actual values by evaluating the error and magnitude of individual forecast. The impact of EPS on the accuracy is evaluated using three workload estimation models. The experimental analysis is carried out over five data traces and performance is measured using correlation coefficient (CoC), sum of elasticity index (SEI), mean squared prediction error (MPE), PER and MoP metrics. The error preventive models achieved maximum improvement upto 183.9%, 95.4% and 100.0% over non error preventive models in CoC, SEI, and MPE respectively. The error preventive models significantly brought down the individual forecast error below 25% and under estimations are reduced by a maximum factor of 55.2%. The superiority of the proposed scheme is validated using a comprehensive statistical evaluation based on Wilcoxon signed rank test and Friedman test with Finner post-hoc analysis. We observed that error preventive weighted exponential smoothing model produced best forecasts.


Cloud computing Workload estimation Time series Auto regression Exponential smoothing 



The authors would like to thank the Ministry of Electronics & Information Technology (MeitY), Government of India for financial support to carry out this work.


  1. 1.
  2. 2.
    Adegboyega, A.: Time-series models for cloud workload prediction: a comparison. In: 2017 IFIP/IEEE Symposium on Integrated Network and Service Management (IM), pp. 298–307 (2017).
  3. 3. announces first quarter sales up 23% to \$35.7 billion. (2017)
  4. 4.
    Amiri, M., Mohammad-Khanli, L.: Survey on prediction models of applications for resources provisioning in cloud. J. Netw. Comput. Appl. 82(C), 93–113 (2017)CrossRefGoogle Scholar
  5. 5.
    Ardagna, D., Casolari, S., Colajanni, M., Panicucci, B.: Dual time-scale distributed capacity allocation and load redirect algorithms for cloud systems. J. Parallel Distrib. Comput. 72(6), 796–808 (2012)CrossRefGoogle Scholar
  6. 6.
    Baldán, F.J., Ramirez-Gallego, S., Bergmeir, C., Benitez-Sanchez, J.M., Herrera, F.: A forecasting methodology for workload forecasting in cloud systems. IEEE Trans. Cloud Comput. PP(99), 1–1 (2016). CrossRefGoogle Scholar
  7. 7.
    Box, G.E.P., Jenkins, G.: Time Series Analysis. Forecasting and Control. Holden-Day Inc, San Francisco, CA (1990)zbMATHGoogle Scholar
  8. 8.
    Box, G.E.P., Pierce, D.A.: Distribution of residual autocorrelations in autoregressive-integrated moving average time series models. J. Am. Stat. Assoc. 65(332), 1509–1526 (1970)MathSciNetCrossRefGoogle Scholar
  9. 9.
    Buyya, R., Broberg, J., Goscinski, A.M.: Cloud Computing Principles and Paradigms. Wiley, Hoboken (2011)CrossRefGoogle Scholar
  10. 10.
    Calheiros, R.N., Masoumi, E., Ranjan, R., Buyya, R.: Workload prediction using arima model and its impact on cloud applications’ qos. IEEE Trans. Cloud Comput. 3(4), 449–458 (2015)CrossRefGoogle Scholar
  11. 11.
    Cao, J., Zhu, X., Dong, F., Liu, B., Ma, Z., Min, H.: Time series based bandwidth allocation strategy in cloud datacenter. In: 2016 International Conference on Advanced Cloud and Big Data (CBD), pp. 228–233 (2016).
  12. 12.
  13. 13.
    Conejo, A.J., Plazas, M.A., Espinola, R., Molina, A.B.: Day-ahead electricity price forecasting using the wavelet transform and arima models. IEEE Trans. Power Syst. 20(2), 1035–1042 (2005)CrossRefGoogle Scholar
  14. 14.
    Ediger, V.S., Akar, S.: Arima forecasting of primary energy demand by fuel in turkey. Energy Policy 35(3), 1701–1708 (2007)CrossRefGoogle Scholar
  15. 15.
    Finner, H.: On a monotonicity problem in step-down multiple test procedures. J. Am. Stat. Assoc. 88(423), 920–923 (1993)MathSciNetCrossRefGoogle Scholar
  16. 16.
    Friedman, M.: The use of ranks to avoid the assumption of normality implicit in the analysis of variance. J. Am. Stat. Assoc. 32(200), 675–701 (1937)CrossRefGoogle Scholar
  17. 17.
    Khan, A., Yan, X., Tao, S., Anerousis, N.: Workload characterization and prediction in the cloud: a multiple time series approach. In: 2012 IEEE Network Operations and Management Symposium, pp. 1287–1294 (2012).
  18. 18.
    Khan, M.N.A.H., Liu, Y., Alipour, H., Singh, S.: Modeling the autoscaling operations in cloud with time series data. In: Proceedings of the 2015 IEEE 34th Symposium on Reliable Distributed Systems Workshop (SRDSW), SRDSW’15, pp. 7–12. IEEE Computer Society, Washington, DC, USA (2015).
  19. 19.
    Kim, I.K., Wang, W., Qi, Y., Humphrey, M.: Empirical evaluation of workload forecasting techniques for predictive cloud resource scaling. In: 2016 IEEE 9th International Conference on Cloud Computing (CLOUD), pp. 1–10 (2016)Google Scholar
  20. 20.
    Kuang, S.R., Wu, K.Y., Ke, B.C., Yeh, J.H., Jheng, H.Y.: Efficient architecture and hardware implementation of hybrid fuzzy-kalman filter for workload prediction. Integr. VLSI J. 47(4), 408–416 (2014)CrossRefGoogle Scholar
  21. 21.
    Kumar, A.S., Mazumdar, S.: Forecasting hpc workload using arma models and ssa. In: 2016 International Conference on Information Technology (ICIT), pp. 294–297 (2016).
  22. 22.
    Kumar, U., Jain, V.K.: Arima forecasting of ambient air pollutants (o3, no, no2 and co). Stoch. Env. Res. Risk Assess. 24(5), 751–760 (2010)CrossRefGoogle Scholar
  23. 23.
    Leesatapornwongsa, T., Stuardo, C.A., Suminto, R.O., Ke, H., Lukman, J.F., Gunawi, H.S.: Scalability bugs: When 100-node testing is not enough. In: Proceedings of the 16th Workshop on Hot Topics in Operating Systems, HotOS ’17, pp. 24–29. ACM, New York, NY, USA (2017)Google Scholar
  24. 24.
    Liu, C., Liu, C., Shang, Y., Chen, S., Cheng, B., Chen, J.: An adaptive prediction approach based on workload pattern discrimination in the cloud. J. Netw. Comput. Appl. 80(C), 35–44 (2017)CrossRefGoogle Scholar
  25. 25.
    Mao, M., Li, J., Humphrey, M.: Cloud auto-scaling with deadline and budget constraints. In: 2010 11th IEEE/ACM International Conference on Grid Computing, pp. 41–48 (2010)Google Scholar
  26. 26.
    Messias, V.R., Estrella, J.C., Ehlers, R., Santana, M.J., Santana, R.C., Reiff-Marganiec, S.: Combining time series prediction models using genetic algorithm to autoscaling web applications hosted in the cloud infrastructure. Neural Comput. Appl. 27(8), 2383–2406 (2016)CrossRefGoogle Scholar
  27. 27.
    Nordrum, A.: Popular internet of things forecast of 50 billion devices by 2020 is outdated. (2016)
  28. 28.
    Pai, P.F., Lin, C.S.: A hybrid arima and support vector machines model in stock price forecasting. Omega 33(6), 497–505 (2005)CrossRefGoogle Scholar
  29. 29.
    Rahman, Z.U., Hussain, O.K., Hussain, F.K.: Time series qos forecasting for management of cloud services. In: 2014 Ninth International Conference on Broadband and Wireless Computing, Communication and Applications, pp. 183–190 (2014).
  30. 30.
    Reiss, C., Wilkes, J.: Google cluster-usage traces: format + schema (2011)Google Scholar
  31. 31.
    Roy, N., Dubey, A., Gokhale, A.: Efficient autoscaling in the cloud using predictive models for workload forecasting. In: Proceedings of the 2011 IEEE 4th International Conference on Cloud Computing, CLOUD ’11, pp. 500–507. IEEE Computer Society, Washington, DC, USA (2011).
  32. 32.
    Shi, R., Gan, Y., Wang, Y.: Evaluating scalability bottlenecks by workload extrapolation. In: 2018 IEEE 26th International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS), pp. 333–347 (2018)Google Scholar
  33. 33.
    Shukur, O.B., Lee, M.H.: Daily wind speed forecasting through hybrid kf-ann model based on arima. Renew. Energy 76, 637–647 (2015)CrossRefGoogle Scholar
  34. 34.
    Traces available in the internet traffic archive.
  35. 35.
    Tran, V.G., Debusschere, V., Bacha, S.: Hourly server workload forecasting up to 168 hours ahead using seasonal arima model. In: 2012 IEEE International Conference on Industrial Technology, pp. 1127–1131 (2012)Google Scholar
  36. 36.
    Wilcoxon, F.: Individual comparisons by ranking methods. Biom. Bull. 1(6), 80–83 (1945)CrossRefGoogle Scholar
  37. 37.
    Yang, L., Foster, I., Schopf, J.M.: Homeostatic and tendency-based cpu load predictions. In: Proceedings International Parallel and Distributed Processing Symposium, p. 9 (2003)Google Scholar
  38. 38.
    Ye, Z., Mistry, S., Bouguettaya, A., Dong, H.: Long-term qos-aware cloud service composition using multivariate time series analysis. IEEE Trans. Serv. Comput. 9(3), 382–393 (2016). CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC, part of Springer Nature 2019

Authors and Affiliations

  1. 1.Department of Computer ApplicationsNational Institute of Technology, KurukshetraKurukshetraIndia

Personalised recommendations