A Study On Deep Learning-based Multivariate Resource Estimation With Feature Selection In Cloud Computing

다국어 초록 (Multilingual Abstract)

Resource provisioning in cloud computing environment depends on different future resource utilization. Because the resource utilization trend may vary dynamically, we need to estimate the future resource utilization for effective resource provisioning decisions. The problem becomes more challenging since performance indicators for one resource may depend on other resources. This paper proposes a deep learning-based multivariate workload estimation with feature selection for cloud computing environment. First, we use the Pearson correlation method to select the best features as inputs for the multivariate model. Then, we propose to use the bidirectional long short-term memory (Bi-LSTM) to estimate future resource utilization. The results are conducted using a real workload dataset and show that the proposed multivariate Bi-LSTM model outperforms the multivariate LSTM model in prediction accuracy.

목차 (Table of Contents)

• Abstract
• Ⅰ. Introduction
• Ⅱ. Related work
• Ⅲ. Proposed models
• IV. Experiment and Evaluation
• V. Conclusion
• 참고문헌