Service Level Agreement-Based Migration For Database-As-A-Service In The Cloud

Loading...
Thumbnail Image
Date
2015-10
Authors
Wong, Wong Jik Soon
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
With the emergence of cloud computing, comes a new kind of cloud service namely Database-as-a-Service. Database-as-a-Service removes users the hassle to own their own database servers. The responsibility of maintaining and monitoring the database workloads thus transferred to the Database-as-a-Service cloud operator. Service operator generates income through the database service they have provided by meeting the service level agreement (SLA). In order to maximize the income generated, a cloud service operator needs to utilize the least amount of machines to host the maximum amount of database workloads. In this regards, this thesis proposes a model which monitors the database workloads performance and server resources utilization. Database workloads migration will be triggered when the model detects any database workload that has failed to meet SLA agreement. The model is able to choose a server which serves the database workload better based on the workload profile and server profile generated by the monitoring module. Experimental results proof that the algorithm is able to detect workloads that violated SLA and migrate the affected workloads. The efficiency of decision making in workload placement is proofed to balance the resource utilization and reduce pending jobs during peak hour. Additionally, the proposed model can consolidate multiple database workloads into fewer machines without having the workloads violate SLA in the new environment.
Description
Keywords
Electronic Computer
Citation