NORMA eResearch @NCI Library

Time Series Forecasting of Database Workloads in Hybrid Cloud

McDonald, Richard (2021) Time Series Forecasting of Database Workloads in Hybrid Cloud. Masters thesis, Dublin, National College of Ireland.

[img]
Preview
PDF (Master of Science)
Download (1MB) | Preview
[img]
Preview
PDF (Configuration manual)
Download (1MB) | Preview

Abstract

The two biggest factors for the Enterprise in moving to a Public Cloud offering over existing offerings are control of cost and ensuring availability. The challenge is to ensure that both factors are managed wisely, even when both compete for attention at the expense of the other. In this work well-trusted statistical techniques were applied to aid that decision-making process. System logs are collected and maintained in a non-invasive manner and ARIMA and ES models are then applied to these logs to build a profile of a workload that identifies the sweet spot for configuration and satisfies cost control. In this work large Database workloads were created in a Private Cloud lab and then a simple method was used to extract and maintain the information before transforming it into a Time-Series Data-set. This work extends research in this field by keying in on Relational Database workloads, and by using non-invasive methods to collect the necessary information to complete the project. The results of the work shows a 96.9 percent level of accuracy in forecasting CPU activity as well as profiling workload that ensures optimum Quality of Service.

Item Type: Thesis (Masters)
Subjects: Q Science > QA Mathematics > Electronic computers. Computer science
T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science

T Technology > T Technology (General) > Information Technology > Cloud computing
Divisions: School of Computing > Master of Science in Cloud Computing
Depositing User: Clara Chan
Date Deposited: 14 Oct 2021 09:11
Last Modified: 14 Oct 2021 09:11
URI: http://norma.ncirl.ie/id/eprint/5090

Actions (login required)

View Item View Item