Agent for managing power among electronic systems

• US 20070067656A1
• Filed: 09/22/2005
• Published: 03/22/2007
• Est. Priority Date: 09/22/2005
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Ranganathan, Parthasarathy, Flynn, Thomas, Shaver, Charles, Leech, Phillip, Alzien, Khaldoun
• Granted Patent: US 7,581,125 B2
• Time in Patent Office: Days
• US Class Current: 713/320
• CPC Class Codes: G06F 1/3203 Power management, i.e. event-based initiation of a power-saving mode
• CPC Class Codes: G06F 1/324 by lowering clock frequency
• CPC Class Codes: G06F 1/3296 by lowering the supply or operating voltage
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

A power management agent for managing power among electronic systems includes a module for predicting a future power consumption level of the electronic systems, a module for determining a power budget threshold for the electronic systems, and a module for determining whether a predicted future power consumption level will exceed the power budget threshold. The power management agent also includes a module for selecting one or more of the electronic systems to throttle in response to a determination that the predicted future power consumption level will exceed the power budget threshold and a module for selecting a throttle level to be applied to the selected one or more of the electronic systems to substantially prevent the future power consumption level from exceeding the power budget threshold.

Claims

1. A power management agent for managing power among electronic systems, said power management agent comprising:

a module for predicting a future power consumption level of the electronic systems;

a module for determining a power budget threshold for the electronic systems;

a module for determining whether a predicted future power consumption level will exceed the power budget threshold;

a module for selecting one or more of the electronic systems to throttle in response to a determination that the predicted future power consumption level will exceed the power budget threshold; and

a module for selecting a throttle level to be applied to the selected one or more of the electronic systems to substantially prevent the future power consumption level from exceeding the power budget threshold.

2. The power management agent according to claim 1, further comprising:

a module for enabling communications between the power management agent and the electronic systems; and

a module for monitoring power consumption levels of the electronic systems.

3. The power management agent according to claim 2, wherein the module for enabling communications is further configured to enable communications between the power management agent and at least one other power management agent, to thereby enable management of power among the electronic systems through the power management agent and the at least one other power management agent.

4. The power management agent according to claim 1, wherein the electronic systems comprise servers.

5. The power management agent according to claim 4, wherein the servers are housed in at least one enclosure.

6. The power management agent according to claim 4, wherein at least two of the servers are positioned in different geographic locations with respect to each other.

7. The power management agent according to claim 1, wherein the electronic systems comprise blade PCs.

8. The power management agent according to claim 1, wherein the module for determining a power budget threshold for the electronic systems is configured to determine the power budget threshold at one or both of design time and run-time.

9. The power management agent according to claim 1, wherein the module for selecting one or more of the electronic systems to throttle is configured to select the one or more of the electronic systems according to a round-robin selection scheme.

10. The power management agent according to claim 1, wherein the module for selecting one or more of the electronic systems to throttle is configured to select the one or more of the electronic systems according to the relative priority of tasks running on the electronic systems and a prediction of resource usage among the electronic systems.

11. The power management agent according to claim 1, wherein the module for selecting a throttle level is configured to select a throttling level according to a rigidity associated with enforcement of the power budget.

12. A system for operating a plurality of servers at a power consumption level below a power budget threshold, the system comprising:

a power management agent configured to communicate with the plurality of servers, said power management agent being further configured to determine a power budget threshold, to monitor current power consumption levels of the plurality of servers, to predict future violations of the power budget threshold by the plurality of servers, and to make policy-driven recommendations for power throttling of the plurality of servers to thereby maintain power consumption levels of the plurality of servers substantially within the power budget threshold.

13. The system according to claim 12, wherein the power management agent is distributed at a plurality of locations, wherein the plurality of locations comprises at least one of hardware and software.

14. A method for managing power among electronic systems, said method comprising:

determining a power budget threshold;

monitoring current power consumption levels of the electronic systems;

predicting a future violation of the power budget threshold;

selecting one or more of the electronic systems to throttle;

determining throttling levels for the selected one or more of the electronic systems; and

recommending the determined throttling levels to the selected one or more of the electronic systems to thereby substantially prevent operating the electronic systems at a power consumption level that substantially exceeds the power budget threshold.

15. The method according to claim 14, wherein determining a power budget threshold further comprises determining the power budget threshold at least one of at design time and at run-time.

16. The method according to claim 15, wherein determining a power budget threshold further comprises determining a power budget threshold at design time when the targeted benefits of the power budget enforcement are geared towards at least one of reducing provisioning of cooling and power delivery and increasing flexibility in a selection of components for the electronic systems.

17. The method according to claim 15, wherein determining a power budget threshold further comprises determining the power budget threshold based upon a trigger external to the electronic systems.

18. The method according to claim 14, wherein predicting a future violation of the power budget threshold further comprises implementing a prediction algorithm to predict a future violation of the power budget threshold, wherein the prediction algorithm is selected based upon error limits allowable with mispredictions of possible future violations.

19. The method according to claim 14, wherein selecting one or more of the electronic systems to throttle further comprises selecting one or more of the electronic systems to throttle according to a round-robin allocation scheme.

20. The method according to claim 14, wherein selecting one or more of the electronic systems to throttle further comprises selecting one or more of the electronic systems to throttle based upon a relative priority of tasks running on the electronic systems.

21. The method according to claim 14, wherein determining throttling levels for the selected one or more of the electronic systems further comprises determining throttling levels for the selected one or more of the electronic systems based upon the rigidity associated with enforcement of the power budget.

22. The method according to claim 21, wherein determining throttling levels further comprises selecting conservative throttling levels to account for worst-case scenarios when enforcement of the power budget is relatively strict.

23. The method according to claim 21, wherein determining throttling levels further comprises selecting relatively aggressive throttling levels when the enforcement of the power budget comprises an allowable error range.

24. A power management agent for controlling power consumption levels of electronic systems to be within a power budget threshold, the power management agent comprising:

means for enabling communications between the agent and the electronic systems;

means for monitoring power consumption levels of the electronic systems;

means for predicting a future power consumption level of the electronic systems;

means for determining a power budget threshold for the electronic system;

means for determining whether a predicted future power consumption level will exceed the power budget threshold;

means for selecting one or more of the electronic systems to throttle in response to a determination that the predicted future power consumption level will exceed the power budget threshold; and

means for selecting a throttle level to be applied to the selected one or more of the electronic systems to substantially prevent the future power consumption level from exceeding the power budget threshold.

25. A computer program product embodied on a computer-readable medium and comprising code that, when executed, causes a computer to perform the following:

determine a power budget threshold;

monitor current power consumption levels of electronic systems;

predict a future violation of the power budget threshold;

select one or more of the electronic systems to throttle;

determine throttling levels for the selected one or more of the electronic systems; and

recommend the determined throttling levels to the selected one or more of the electronic systems to thereby substantially prevent operating the electronic systems at a power consumption level that substantially exceeds the power budget threshold.