Performance analysis method and management computer

1. A method for analyzing the performance of a computer system using a management computer, wherein the computer system includes a plurality of resources, the method comprising:

providing a management computer including a processor and a memory device in which a program to be executed by the processor is stored,

utilizing the processor to divide the plurality of resources of the computer system into a plurality of resource groups based on deviations in performance data between the resources,

utilizing the processor to analyze the performance data for each of the divided resource groups,

wherein the plurality of resources are constituted by a plurality of hierarchies with a resource belonging to one of the plurality of hierarchies as a node,

utilizing the processor to determine a number of the resource groups based on a number of the nodes, and utilizing the processor to divide the plurality of resources into the plurality of resource groups based on the deviations in performance data between the resources and the number of the nodes,

utilizing the processor to acquire values of a plurality of metrics for each of the resources as the performance data of the resource from the computer system,

utilizing the processor to accumulate the acquired value for each of the metrics in a time-series manner, and

utilizing the processor to calculate a correlation between a change of the accumulated value of the metric along with a lapse of time and a change of the accumulated value of the other metric along with the lapse of time to obtain the deviations in performance data between the resources.

2. The performance analysis method according to claim 1 ,

wherein the plurality of resources constitute a plurality of resource hierarchies in the computer system, and

wherein the method additionally comprises utilizing the processor to set the number of resource groups as a number of nodes which is a minimum number of resources for each of the resource hierarchies.

3. The performance analysis method according to claim 1 ,

wherein the computer system comprises a storage device including a plurality of storage pools, and

wherein the method additionally comprises utilizing the processor to set a number of the plurality of storage pools as the number of the nodes.

4. The performance analysis method according to claim 1 ,

wherein the method additionally comprises utilizing the processor to set a preset number as the number of the nodes.

5. The performance analysis method according to claim 1 ,

wherein the method additionally comprises utilizing the processor to select one representative metric from among the plurality of metrics of one of the accumulated resources to select the representative metric for each of the plurality of accumulated resources, and to calculate a correlation between a change of a value of the representative metric along with a lapse of time and a change of a value of the other representative metric along with the lapse of time to obtain the deviations in performance data between the resources.

6. The performance analysis method according to claim 5 ,

wherein the method additionally comprises utilizing the processor to calculate the correlation between the change of the value of the representative metric along with the lapse of time and the change of the value of the other representative metric along with the lapse of time to obtain the deviations in performance data between the resources, to divide the plurality of representative metrics into the plurality of resource groups based on the obtained deviations in performance data between the resources, and to divide the other metrics of the resource to which each of the representative metrics belongs into the same resource group as each of the representative metrics to divide the plurality of resources into the plurality of resource groups.

7. The performance analysis method according to claim 1 ,

wherein the method additionally comprises utilizing the processor to record the plurality of resource groups, to compare the plurality of recorded resource groups, and to change the resource group for analysis of the performance data when there is a difference exceeding a preset threshold between the plurality of compared resource groups.

8. The performance analysis method according to claim 7 ,

wherein the method additionally comprises utilizing the processor to perform learning for each of the divided resource groups to create a mathematical model, and to perform evaluation for each of the divided resource groups using the created mathematical model to analyze the performance data.

9. The performance analysis method according to claim 8 ,

wherein the method additionally comprises utilizing the processor to change the resource group for analysis of the performance data when there is the difference exceeding the preset threshold between the plurality of compared resource groups, and to stop the evaluation when the resource group is changed in a course of the evaluation.

10. A management computer that analyzes performance of a computer system constituted by a plurality of resources, the management computer comprising:

a processor; and

a memory device in which a program to be executed by the processor is stored,

wherein the processor divides the plurality of resources of the computer system into a plurality of resource groups based on deviations in performance data between the resources, and analyzes the performance data for each of the divided resource groups

wherein

the plurality of resources are constituted by a plurality of hierarchies with a resource belonging to one of the plurality of hierarchies as a node,

the processor determines a number of the resource groups based on a number of the nodes, and divides the plurality of resources into the plurality of resource groups based on the deviations in performance data between the resources and the number of the nodes, and,

wherein the processor acquires values of a plurality of metrics for each of the resources as the performance data of the resource from the computer system, accumulates the acquired value for each of the metrics in a time-series manner, and calculates a correlation between a change of the accumulated value of the metric along with a lapse of time and a change of the accumulated value of the other metric along with the lapse of time to obtain the deviations in performance data between the resources.