Predictive Autoscaling for Kubernetes Using Multivariate Time-Series Forecasting
DOI:
https://doi.org/10.63282/3050-922X.IJERET-V2I2P110Keywords:
Kubernetes, Predictive Autoscaling, Time-Series Forecasting, Cloud Computing, Resource Management, Lstm, Machine Learning, Container Orchestration, Sla OptimizationAbstract
Applications built on Kubernetes clusters tend to be cloud-native and have to deal with highly dynamic loads whilst being constrained under a rigorous performance and cost-efficiency benchmark. Historical reactive autoscaling tools, including the Horizontal Pod Autoscaler (HPA) deliver a response that largely depends on the real-time resource consumption indicators and are not always able to react to abrupt workload changes. This is constraining, performance is affected, service-level agreement (SLA) is not met and resources are not used efficiently. A new trend of predictive autoscaling has come into picture as a format to rise above these issues by predicting future requirement of resources and preempting pooling of resources. The current paper has provided a detailed model of predictive autoscaling in Kubernetes systems based on multivariate time-series forecasting algorithms. The approach proposed will combine historical workload data, system level metrics, and application-specific data in order to come up with predictions of future resource requirements that are accurate. The system predicts the patterns of workload in advance with the help of machine-learning-based forecasting models, such as Long Short-Term Memory (LSTM) networks, Autoregressive Integrated Moving Average (ARIMA), and Multivariate Regression to adjust cluster resources accordingly. The framework is deployed as a planetary controller connected to the Kubernetes control plane, which would not create an issue with integration with the existing infrastructure. Large-scale experiments were done on benchmark workloads and real-world traces. The findings indicate that the suggested strategy can spend much less time to reply, make compliance with SLA significant, and increase the efficiency of resources as a whole as compared to traditional methods of reactive autoscaling. The results affirm that multivariate time-series forecasting offers a strong basis of intelligent and adaptable resource management in existence in cloud-native platforms. The research takes a step towards the development of autonomous cloud systems and offers some useful recommendations to implement the predictive autoscaling systems in the production setting.
References
[1] Lorido-Botrán, T., Miguel-Alonso, J., & Lozano, J. A. (2012). Auto-scaling techniques for elastic applications in cloud environments. Department of Computer Architecture and Technology, University of Basque Country, Tech. Rep. EHU-KAT-IK-09, 12, 2012.
[2] Ghani, N., Ahmad, I., Mohsin, M., & Khan, M. A. (2018). Predictive autoscaling for cloud resources using machine learning techniques. Journal of Cloud Computing: Advances, Systems and Applications, 7(1), 1–18.
[3] Cortes, C., & Vapnik, V. (1995). Support-vector networks. Machine learning, 20(3), 273-297.
[4] Mao, M., & Humphrey, M. (2012, June). A performance study on the vm startup time in the cloud. In 2012 IEEE Fifth International Conference on Cloud Computing (pp. 423-430). IEEE.
[5] Box, G. E., Jenkins, G. M., Reinsel, G. C., & Ljung, G. M. (2015). Time series analysis: forecasting and control. John Wiley & Sons.
[6] Hyndman, R. J., & Athanasopoulos, G. (2018). Forecasting: principles and practice. OTexts.
[7] Abdi, H., & Williams, L. J. (2010). Principal component analysis. Wiley interdisciplinary reviews: computational statistics, 2(4), 433-459.
[8] Masdari, M., & Khoshnevis, A. (2020). A survey and classification of the workload forecasting methods in cloud computing. Cluster Computing, 23(4), 2399-2424.
[9] LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. nature, 521(7553), 436-444.
[10] Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural computation, 9(8), 1735-1780.
[11] Nikravesh, A. Y., Ajila, S. A., & Lung, C. H. (2015, May). Towards an autonomic auto-scaling prediction system for cloud resource provisioning. In 2015 IEEE/ACM 10th International Symposium on Software Engineering for Adaptive and Self-Managing Systems (pp. 35-45). IEEE.
[12] Cho, K., Van Merriënboer, B., Gulcehre, C., Bahdanau, D., Bougares, F., Schwenk, H., & Bengio, Y. (2014). Learning phrase representations using RNN encoder-decoder for statistical machine translation. arXiv preprint arXiv:1406.1078.
[13] Alshamrani, A., & Malek, S. (2019). A survey on performance prediction and auto-scaling techniques for cloud-based services. Journal of Systems and Software, 156, 110–124.
[14] Nikravesh, A. Y., Ajila, S. A., & Lung, C. H. (2015, May). Towards an autonomic auto-scaling prediction system for cloud resource provisioning. In 2015 IEEE/ACM 10th International Symposium on Software Engineering for Adaptive and Self-Managing Systems (pp. 35-45). IEEE.
[15] Breiman, L. (2001). Random forests. Machine learning, 45(1), 5-32.
[16] Karim, F., Majumdar, S., Darabi, H., & Chen, S. (2017). LSTM fully convolutional networks for time series classification. IEEE access, 6, 1662-1669.
[17] Baldan, F. J., Ramirez-Gallego, S., Bergmeir, C., Herrera, F., & Benitez, J. M. (2016). A forecasting methodology for workload forecasting in cloud systems. IEEE Transactions on Cloud Computing, 6(4), 929-941.
[18] Imdoukh, M., Ahmad, I., & Alfailakawi, M. G. (2020). Machine learning-based auto-scaling for containerized applications. Neural Computing and Applications, 32(13), 9745-9760.
[19] Yin, S., Liu, L., & Hou, J. (2016). A multivariate statistical combination forecasting method for product quality evaluation. Information sciences, 355, 229-236.
[20] Lu, S., Lu, H., & Kolarik, W. J. (2001). Multivariate performance reliability prediction in real-time. Reliability Engineering & System Safety, 72(1), 39-45.
[21] Chen, L., Bahsoon, R., & Yao, X. (2018). Self-adaptive and self-aware autoscaling for cloud resource management: A survey. ACM Computing Surveys, 51(3), 1–41.
[22] Imdoukh, M., Ahmad, I., & Alfailakawi, M. G. (2020). Machine learning-based auto-scaling for containerized applications. Neural Computing and Applications, 32(13), 9745-9760
[23] Obannagari, C. K. R. N., & Nangi, P. R. (2020). Deep Learning-Driven Compliance Automation for Continuous Monitoring of Security Controls in Regulated Cloud Systems. International Journal of Artificial Intelligence, Data Science, and Machine Learning, 1(3), 21-32. https://doi.org/10.63282/3050-9262.IJAIDSML-V1I3P104
[24] Obannagari, C. K. R. N., & Nangi, P. R. (2020). Advanced Data Science Frameworks for Predictive Cyber-Risk Assessment and Adaptive Security Policy Optimization in Zero Trust Networks. International Journal of Emerging Trends in Computer Science and Information Technology, 1(4), 67-78. https://doi.org/10.63282/3050-9246.IJETCSIT-V1I4P108
