Evaluation commercial cloud services performance
In this project, the objective was to design and implement an experiment to collect execution time of Machine Learning training on Amazon Elastic Cloud Computing machines. The collected data was then analysed to discover underlying patterns and trends. From the analysis done, it was observed that e...
Saved in:
Main Author: | |
---|---|
Other Authors: | |
Format: | Final Year Project |
Language: | English |
Published: |
2017
|
Subjects: | |
Online Access: | http://hdl.handle.net/10356/70151 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
Summary: | In this project, the objective was to design and implement an experiment to collect execution time of Machine Learning training on Amazon Elastic Cloud Computing machines. The collected data was then analysed to discover underlying patterns and trends.
From the analysis done, it was observed that execution time of Machine Learning training on Amazon Elastic Cloud Computing machines decreases as the number of virtual Central Processing Unit used increases. The execution time of Machine Learning training on Amazon Elastic Cloud Computing machines stop decreasing after the network capability limit was reached. Hence, it was concluded that the execution time of Machine Learning training was affected by the total number of virtual Central Processing Unit used and the capability of the network used.
A prediction algorithm was developed and implemented using the information found by analysing the collected data. The prediction algorithm developed can predict the execution time of Machine Learning training on Amazon Elastic Cloud Computing machines with an accuracy score of 71.37% and provide suggestions on machine configuration based on the cheapest cost, fastest execution time and most optimised configuration.
As a recommendation for future continuation of the project, it is recommended that the project can be carried out by collecting more on Amazon Elastic Cloud Computing machines using a bigger set of different Amazon Elastic Cloud Computing machine configurations. Also, the project can be carried out on other cloud computing providers such as Google Cloud Enterprise and Microsoft Azure. The data collected can also be further analysed using other methods to gain a better understanding of the characteristics of the execution time of Machine Learning training on Amazon Elastic Cloud Computing machines. |
---|