A2Cloud-H: A Multi-tiered Machine Learning Framework for Cost-Effective Cloud Resource Selection
Document Type
Conference Presentation
Department
Electrical and Computer Engineering
Conference Title
Lecture Notes in Networks and Systems
Date of Presentation
1-1-2022
Abstract
We present the A2Cloud-Hierarchy (A2Cloud-H) framework that recommends Cloud instances to users for high-performance scientific computing. The framework comprises four components: training database, supervised learning (SL) module, unsupervised learning (USL) module, and a decision module. The training database contains the testing traces of previous applications and Cloud instances; these are contributed by the scientific community. The SL module comprises three popular supervised learning models: logistic regression, support vector machine, and random forest, which train using the database to qualitatively assess the instance performance for the target application. The USL module contains three collaborative filtering methods: application-based, instance-based, and rank-based, which use the database to estimate the instances’ performance ratings for the target application. The decision module includes multiple tiers of analytic hierarchy processing, which consolidate the instance recommendations from the SL and USL modules into a final instance recommendation. Our comprehensive testing with 20 Cloud instances and over 8 scientific applications yield more than 90% modeling accuracy. The framework’s ultimate goal is to encourage the scientific community to adopt the Cloud for green computing.
ISSN
23673370
Volume
360 LNNS
First Page
272
Last Page
291
DOI
10.1007/978-3-030-89912-7_21
Recommended Citation
Ai, X.,
Jena, T.,
Khan, S.,
Hughes, R.,
&
Pallipuram, V. K.
(2022).
A2Cloud-H: A Multi-tiered Machine Learning Framework for Cost-Effective Cloud Resource Selection.
Paper presented at Lecture Notes in Networks and Systems.
https://scholarlycommons.pacific.edu/soecs-facpres/466
