An SLA and Operation Cost Aware Performance Re-tuning Algorithm for Cloud Databases

Cloud database, also called Database as a Service (DbaaS), is defined as a pay-per-use model for enabling on-demand access to reliable and configurable services. Tenants pay to the cloud service provider only for the services they use through a service level agreement (SLA). Due to changes in queries and/or databases, tenants' performance SLA may be violated at some point in time. This paper proposes a performance tuning algorithm, called AutoClustC, which estimates the costs of resource provisioning and database re-partitioning and chooses the lower cost approach to tune the system in order to re-guarantee the performance SLA when a performance violation occurs. When database partitioning is chosen, the algorithm takes both performance SLA and operation cost into consideration, and uses data mining techniques to automatically and dynamically re-partition the databases for a tenant to generate better partitioning solutions.