Advertisement

Sibyl: Host Load Prediction with an Efficient Deep Learning Model in Cloud Computing

  • Zhiyuan Zhang
  • Xuehai Tang
  • Jizhong Han
  • Peng Wang
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11335)

Abstract

Prediction of host load is essential in Cloud computing for improving resource utilization and achieving service-level agreements. However, accurate prediction of host load remains a challenge in Clouds because the type of load varies differently. Furthermore, selecting metrics for host load prediction is also a difficult task. With so many metrics in the Cloud systems, it is hard to determine which metrics are going to be useful. To address these challenges, this paper proposes an efficient deep learning model named Sibyl to improve the accuracy and efficiency of prediction. Sibyl includes two parts: a metrics selection module and a neural network training module. Sibyl first selects metrics by filtering out irrelevant metrics. Afterwards, Sibyl applies a powerful neural network model built with bidirectional long short-term memory to predict actual load one-step-ahead. We use Sibyl to analyze a 40-day load trace from a data center with 176 machines. Experiments show that Sibyl can reduce training metrics while maintaining prediction accuracy. Besides, Sibyl significantly improves prediction accuracy compared to other state-of-the-art methods based on autoregressive integrated moving-average and long short-term memory.

Keywords

Cloud computing Host load prediction Time series analysis Bidirectional long short-term memory 

Notes

Acknowledgements

This work was supported by Grant 2017YFB 1010000 from the National Key R&D Program of China.

References

  1. 1.
    Nazarko, J., Jurczuk, A., Zalewski, W.: ARIMA models in load modelling with clustering approach. In: Proceedings of IEEE Russia Power Tech (2005)Google Scholar
  2. 2.
    Tran, N., Reed, D.A.: Automatic ARIMA time series modeling for adaptive I/O prefetching. In: Proceedings of International Conference on Supercomputing (2002)Google Scholar
  3. 3.
    Zhu, B., Sastry, S.: Revisit dynamic ARIMA based anomaly detection. In: Proceedings of Published in International Conference on Social Computing (2002)Google Scholar
  4. 4.
    Wang, J., Chen, J.W., Wang, Y., Zheng, D.: Intelligent load balancing strategies for complex distributed simulation applications. In: Proceedings of International Conference on Computational Intelligence and Security (2009)Google Scholar
  5. 5.
    Di, S., Kondo, D., Cirne, W.: Host load prediction in a google compute cloud with a Bayesian model. In: Proceedings of the International Conference on High Performance Computing, Networking, Storage and Analysis (2012)Google Scholar
  6. 6.
    Dabrowski, C., Hunt, F.: Using Markov chain analysis to study dynamic behaviour in large-scale grid systems. In: Proceedings of the Seventh Australasian Symposium on Grid Computing (2009)Google Scholar
  7. 7.
    Akioka, S., Muraoka, Y.: Extended forecast of CPU and network load on computational grid. In: Proceedings of International Symposium on CLUSTER Computing and the Grid (2004)Google Scholar
  8. 8.
    Byun, E.J., Choi, S.J., Baik, M.S.: MJSA: markov job scheduler based on availability in desktop grid computing environment. Future Gener. Comput. Syst. 23(4), 616–622 (2007)CrossRefGoogle Scholar
  9. 9.
    Yong, W.F, Goh, C., Hong, C.L., Zhan, Z.H., Li, Y.: Evolutionary neural network based energy consumption forecast for cloud computing. In: Proceedings of International Conference on Cloud Computing Research and Innovation (2015)Google Scholar
  10. 10.
    Xue, J., Yan, F., Birke, R., Chen, L.Y., Scherer, T.: PRACTISE: robust prediction of data center time series. In: Proceedings of International Conference on Network and Service Management (2015)Google Scholar
  11. 11.
    Lang, K., Zhang, M., Yuan, Y., Yue, X.: Short-term load forecasting based on multivariate time series prediction and weighted neural network with random weights and kernels. In: Proceedings of Cluster Computing (2018)Google Scholar
  12. 12.
    Huang, Z., Peng, J., Lian, H., Guo, J., Wei, Q.: Deep recurrent model for server load and performance prediction in data center. Complexity 2017(99), 1–10 (2007)Google Scholar
  13. 13.
    Saripalli, P., Oldenburg, C., Walters, B., Nanduri, R.: Implementation and usability evaluation of a cloud platform for scientific computing as a service (SCaaS). In: Proceedings of International Conference on Utility and Cloud Computing (2011)Google Scholar
  14. 14.
    Hussain, W., Hussain, F.K., Hussain, O., Chang, E.: Profile-based viable service level agreement (SLA) violation prediction model in the cloud. In: Proceedings of International Conference on P2P, Parallel, Grid, Cloud and Internet Computing (2016)Google Scholar
  15. 15.
    Thalheim, J., Rodrigues, A., Akkus, I.E.: Sieve: actionable insights from monitored metrics in distributed systems. In: Proceedings of the 18th ACM/IFIP/USENIX Middleware Conference (2017)Google Scholar
  16. 16.
    Amazon CloudWatch. https://aws.amazon.com/de/cloudwatch. Accessed 30 May 2018
  17. 17.
    Paparrizos, J., Gravano, L.: k-Shape: efficient and accurate clustering of time series. In: ACM SIGMOD International Conference on Management of Data, vol. 45, no. 1, pp. 1855–1870 (2015)Google Scholar
  18. 18.
    Ballesteros, M., Dyer, C., Smith, N.A.: Improved transition-based parsing by modeling characters instead of words with LSTMs. In: Proceedings of Empirical Methods in Natural Language Processing (2015)Google Scholar
  19. 19.
    Zhang, Y., Chen, G., Yu, D., Yaco, K.: Highway long short-term memory RNNS for distant speech recognition. In: Proceedings of International Conference on Acoustics, Speech and Signal Processing (2016)Google Scholar
  20. 20.
    Microsoft Cloud Monitoring. https://www.microsoft.com/en-us/cloud-platform/operations-management-suite. Accessed 28 May 2018
  21. 21.
    Google Stackdriver.https://cloud.google.com/stackdriver. Accessed 1 June 2018
  22. 22.
    Cao, J., Fu, J., Li, M.: CPU load prediction for cloud environment based on a dynamic ensemble model. Softw. Pract. Exp. 44(7), 793–804 (2014)CrossRefGoogle Scholar
  23. 23.
    Song, B., Yu, Y., Zhou, Y.: Host load prediction with long short-term memory in cloud computing. J. Supercomput. 23(2), 1–15 (2017)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  • Zhiyuan Zhang
    • 1
    • 2
  • Xuehai Tang
    • 1
  • Jizhong Han
    • 1
  • Peng Wang
    • 1
  1. 1.Institute of Information EngineeringChinese Academy of SciencesBeijingChina
  2. 2.School of Cyber SecurityUniversity of Chinese Academy of SciencesBeijingChina

Personalised recommendations