Issue |
E3S Web Conf.
Volume 391, 2023
4th International Conference on Design and Manufacturing Aspects for Sustainable Energy (ICMED-ICMPC 2023)
|
|
---|---|---|
Article Number | 01072 | |
Number of page(s) | 15 | |
DOI | https://doi.org/10.1051/e3sconf/202339101072 | |
Published online | 05 June 2023 |
An Ensemble Learning Approach For Task Failure Prediction In Cloud Data Centers
1 Department of Computer Science and Engineering, Chaitanya Bharathi Institute of Technology, Hyderabad - 500075
2 Department of Computer Science and Engineering, Chaitanya Bharathi Institute of Technology, Hyderabad - 500075
3 Department of Computer Science and Engineering, Chaitanya Bharathi Institute of Technology, Hyderabad - 500075
4 Department of Information and Technology, Gokaraju Rangaraju Institute of Technology, Hyderabad - 500090
* Corresponding author: raman.vsd@gmail.com
Due to cloud computing’s extensive use and diverse nature, they experience failures in terms of software, service, and platform, which lead to the failure of task execution, resource waste and performance deterioration. Most studies focused on failure prediction resulted in lower prediction accuracies due to limited attributes and a single prediction model. Hence, in this paper, an efficient ensemble model for task failure prediction is put forth. Initially, the input dataset is collected and pre-processed. In pre-processing, the dataset is cleaned up of all null values. Then, the dimensionality of the pre-processed dataset is reduced by using the PCA algorithm. Thus, the reconstructed dataset is split into training and testing sets to train failure prediction models. The proposed model employs an ensemble learning approach based on different ML and DL algorithms. Then, a comparative study is performed, and the results show that task failure in the cloud system can be effectively predicted using the proposed ensemble method.
© The Authors, published by EDP Sciences, 2023
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.