Time-Series Analysis for Price Prediction of Opportunistic Cloud Computing Resources

  • Sarah Alkharif
  • Kyungyong LeeEmail author
  • Hyeokman Kim
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 461)


Cloud computing resources are offered in various forms, and surplus of computing resources are provided at cheaper price. A leading cloud computing vendor, Amazon Web Services, provides such opportunistic resources as EC2 spot instance whose price changes dynamically based on the resource demand from users. We analyze the spot instance price logs and apply various predictive analysis algorithms to better predict future spot instance price. By applying various train dataset modeling heuristics, we uncover that the SARIMA algorithm achieves the best prediction accuracy in spot price prediction; it shows 17% more accuracy than other algorithms that are widely used for spot instance applications. By applying contributions in this paper, we expect that spot instance users can decrease monetary cost while improving system stability.



This work is supported by the National Research Foundation of Korea (NRF) Grant funded by the Korean Government (MSIP) (No. NRF-2015R1A5A7037615 and NRF-2016R1C1B2015135), the ICT R&D program of IITP (2017-0-00396), and the AWS Cloud Credits for Research program.


  1. 1.
    Agmon Ben-Yehuda, O., Ben-Yehuda, M., Schuster, A., Tsafrir, D.: Deconstructing amazon ec2 spot instance pricing. ACM Trans. Econ. Comput. 1(3), 16:1–16:20 (2013).
  2. 2.
    Gong, Y., He, B., Zhou, A.C.: Monetary cost optimizations for MPI-based HPC applications on amazon clouds: checkpoints and replicated execution. In: Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, SC 2015, pp. 32:1–32:12. ACM, New York (2015).
  3. 3.
    Hyndman, R.J., Athanasopoulos, G.: Forecasting: Principles and Practice (2012).
  4. 4.
    Javadi, B., Kondo, D., Vincent, J.M., Anderson, D.: Discovering statistical models of availability in large distributed systems: an empirical study of seti@home. IEEE Trans. Parallel Distrib. Syst. 22(11), 1896–1903 (2011). doi: 10.1109/TPDS.2011.50 CrossRefGoogle Scholar
  5. 5.
    Lee, K., Son, M.: Deepspotcloud: leveraging cross-region GPU spot instances for deep learning. In: 2017 IEEE 10th International Conference on Cloud Computing (CLOUD) (2017)Google Scholar
  6. 6.
    Sharma, P., Guo, T., He, X., Irwin, D., Shenoy, P.: Flint: Batch-interactive data-intensive processing on transient servers. In: Proceedings of the Eleventh European Conference on Computer Systems, EuroSys 2016, pp. 6:1–6:15, NY, USA. ACM, New York (2016). doi: 10.1145/2901318.2901319
  7. 7.
    Sharma, P., Irwin, D., Shenoy, P.: How not to bid the cloud. In: 8th USENIX Workshop on Hot Topics in Cloud Computing (HotCloud 2016). USENIX Association, Denver, CO (2016)Google Scholar
  8. 8.
    Taylor, S.J., Benjamin, L.: Forecasting at scale. In: Facebook Technical Report (2017)Google Scholar
  9. 9.
    Yan, Y., Gao, Y., Chen, Y., Guo, Z., Chen, B., Moscibroda, T.: Tr-spark: Transient computing for big data analytics. In: Proceedings of the Seventh ACM Symposium on Cloud Computing, SoCC 2016, pp. 484–496. ACM, New York (2016).
  10. 10.
    Zhao, H., Pan, M., Liu, X., Li, X., Fang, Y.: Optimal resource rental planning for elastic applications in cloud market. In: Proceedings of the 2012 IEEE 26th International Parallel and Distributed Processing Symposium, IPDPS 2012, pp. 808–819. IEEE Computer Society, Washington, DC (2012).
  11. 11.
    Zheng, L., Joe-Wong, C., Tan, C.W., Chiang, M., Wang, X.: How to bid the cloud. In: Proceedings of the 2015 ACM Conference on Special Interest Group on Data Communication, SIGCOMM 2015, pp. 71–84. ACM, New York (2015).

Copyright information

© Springer Nature Singapore Pte Ltd. 2018

Authors and Affiliations

  1. 1.Kookmin UniversitySeoulSouth Korea

Personalised recommendations