Method and system for managing resources in a data center

• US 7,308,687 B2
• Filed: 02/04/2003
• Issued: 12/11/2007
• Est. Priority Date: 02/07/2002
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Scarth, Michael, Iszlai, Gabriel, Mihaescu, Mircea, Li, Michael, Trossman, Andrew, Vytas, Paul, Hill, Duncan
• Primary Examiner(s): BULLOCK JR, LEWIS ALEXANDER
• Application Number:
  US10/357,352
• Publication Number:
  US 20030149685A1
• Time in Patent Office: 1,771 Days
• Field of Search: 718100-108, 706/14
• US Class Current: 718/104
• CPC Class Codes: G06F 11/3006 where the computing system is distributed, e.g. networked systems, clusters, multiprocessor systems multiprogramming arrangements G06F9/46; allocation of resources G06F9/50
• CPC Class Codes: G06F 11/3055 Monitoring arrangements for monitoring the status of the computing system or of the computing system component, e.g. monitoring if the computing system is on, off, available, not available error or fault processing without redundancy G06F11/0703; error detection or correction by redundancy in data representation G06F11/08; error detection or correction by redundancy in operation G06F11/14; error detection or correction by redundancy in hardware G06F11/16
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 11/3447 Performance evaluation by modeling
• CPC Class Codes: G06F 11/3466 Performance evaluation by tracing or monitoring
• CPC Class Codes: G06F 2209/5019 Workload prediction
• CPC Class Codes: G06F 9/5011 the resources being hardware resources other than CPUs, Servers and Terminals
• CPC Class Codes: G06F 9/5061 Partitioning or combining of resources

Abstract

The present invention provides dynamic configuration and allocation of resources in a data center. These resources are monitored for availability and performance information according to their assigned execution (or application) environments. The measured performance information for each environment is used to predict levels of demand for an application in the environment. From the predicted levels of demand resource requirements can be determined to provide a predetermined level of performance under the demand changes. The resources can then be reconfigured or reallocated to different environment so that the requirements of each environment can be met.

Claims

1. A method of managing an application environment having an operating state according to an operating objective, the application environment having a computing resource with a characteristic representative of an operating state of the computing resource, said method comprising:

(a) determining a future operating state of the application environment based on a current status of the characteristic of the computing resource, wherein such step includes;

estimating a future time-varying status of the characteristic based on a time-varying component of the current status;

estimating a future time stationary status of the characteristic based on a time stationary component of the current status;

combining the future time-varying status and the future time stationary status to form a future status of the characteristic; and

determining a response of the application environment to the future status of the characteristic, wherein the future operating state of the application environment is based on the response;

(b) determining a difference between the future operating state of the application environment and the operating objective;

(c) generating a selected set of changes to the application environment for reducing the difference; and

(d) repeating steps (a) to (c) to monitor the future operating state of the application environment.

2. The method according to claim 1 wherein the step of estimating a future time-varying status includes:

obtaining the time-varying component from the current status of the characteristic;

determining a time-varying periodic model from the time-varying component; and

extrapolating the future time-varying status from the time-varying periodic model based on periodicity of the time-varying periodic model.

3. The method according to claim 1 wherein the step of estimating a future time stationary status includes:

extracting the time-varying component from the current status of the characteristic to obtain the time stationary component of the current status;

determining a time stationary periodic model from the time stationary component; and

extrapolating the future time stationary status from the time stationary periodic model based on periodicity of the time stationary periodic model.

4. The method according to claim 1 wherein the operating state of the computing resource includes the characteristic representing a current demand for the computing resource, and a current performance state for the computing resource given the current demand and wherein the future operating state of the application environment includes a future demand for the application environment based on a future demand for the computing resource and a future performance state for the application environment given the future demand for the application environment.

5. The method according to claim 4 wherein the future performance state from the future operating state of the application environment is used to determine the difference.

6. The method according to claim 1 wherein step (c) includes:

creating a plurality of sets of changes to the application environment, each of the plurality of sets of changes resulting in a reduction of the difference;

assessing each of the plurality of sets of changes to determine a quantitative preference for the effect on the future operating state of the application environment of each of the plurality of sets of changes; and

effecting the selected set of changes on the application environment.

7. The method according to claim 6 wherein step (c) further includes:

determining the selected set of changes from the plurality of sets of changes based on the quantitative preference prior to effecting the selected set of changes.

8. The method according to claim 6 wherein the step of assessing includes:

determining a quantitative assessment of the effect on the future operating state of the application environment of a set of changes from the plurality of sets of changes for each of a plurality of properties wherein the plurality of properties includes the reduction of the difference and use of computing resources;

combining the quantitative assessment from each of the plurality of properties for the set of changes to form the quantitative preference for the set of changes; and

repeating the steps of determining a quantitative assessment and combining the quantitative assessment for each of the plurality of sets of changes.

9. The method according to claim 6 wherein the step of effecting includes:

forming a workflow for each change in the selected set of changes; and

performing the workflow for each change in the selected set of changes to invoke the selected set of changes.

10. The method according to claim 9 wherein the step of forming a workflow includes:

decomposing each change in the selected set of changes into a set of steps that can be performed to invoke each change;

assembling the workflow from the set of steps for each change in the selected set of changes.

11. The method according to claim 1 wherein each change in the selected set of changes to the application environment is one of adding a new computing resource to the application environment, removing the computing resource from the application environment or reconfiguring the computing resource in the application environment.

12. The method according to claim 1 wherein step (d) includes:

repeating steps (a) to (c) when a change in the current status of the characteristic of the computing resource is detected.

13. The method according to claim 1 wherein step (d) includes:

repeating steps (a) to (c) at a repeating time interval.

14. A method of managing a plurality of application environments according to an operating objective for each of the plurality of application environments, each of the plurality of application environments having an operating state and being assigned a computing resource from a plurality of computing resources, each of the plurality of computing resources having a characteristic representative of an operating state of the computing resource, said method comprising:

(a) estimating a future time-varying status of the characteristic of the assigned computing resource for the specific application environment based on a time-varying component of the current status;

(b) estimating a future time stationary status of the characteristic of the assigned computing resource for the specific application environment based on a time stationary component of the current status;

(c) combining the future time-varying status and the future time stationary status to form a future status of the characteristic of the assigned computing resource for specific application environment;

(d) determining a response of the specific application environment to the future status of the characteristic of the assigned computing resource for the specific application environment, wherein the future operating state of the specific application environment is based on the response;

(e) determining a difference between the future operating state of the specific application environment and the operating objective for the specific application environment;

(f) creating a plurality of sets of changes to the specific application environment, each of the plurality of sets of changes resulting in a reduction of the difference;

(g) assessing each of the plurality of sets of changes to determine a quantitative preference for the effect on the future operating state of the specific application environment of each of the plurality of sets of changes based a property of the effect;

(h) determining the selected set of changes from the plurality of sets of changes based on the quantitative preference;

(i) effecting the selected set of changes on the selected application environment; and

(j) repeating steps (a) to (i) for each of the plurality of application environments to monitor the future operating state of each of the plurality of application environments.

15. The method according to claim 14 wherein the operating state of each of the plurality of computing resources includes the characteristic representing a current demand for the computing resource, and a current performance state for the computing resource given the current demand and wherein the future operating state of each of the plurality of application environment includes a future demand for the application environment based on a future demand for the computing resource assigned to the application environment and a future performance state for the application environment given the future demand for the application environment.

16. The method according to claim 14 wherein the step of effecting includes:

forming a workflow for each change in the selected set of changes including;

decomposing each change in the selected set of changes into a set of steps that can be performed to invoke each change;

assembling the workflow from the set of steps for each change in the selected set of changes; and

performing the workflow for each change in the selected set of changes to invoke the selected set of changes.

17. The method according to claim 14 wherein each change in the selected set of changes to the specific application environment is one of assigning one of the plurality of computing resources to the specific application environment, removing the assigned computing resource from the specific application environment or reconfiguring the assigned computing resource in the specific application environment.