Power management of multi-processor servers

• US 20060149985A1
• Filed: 12/16/2004
• Published: 07/06/2006
• Est. Priority Date: 12/16/2004
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Dubinsky, Dean V.
• Granted Patent: US 7,308,591 B2
• Time in Patent Office: Days
• US Class Current: 713/324
• CPC Class Codes: G06F 1/3203 Power management, i.e. event-based initiation of a power-saving mode
• CPC Class Codes: G06F 1/3287 by switching off individual functional units in the computer system
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management
• CPC Class Codes: Y02D 30/50 in wire-line communication networks, e.g. low power modes or reduced link rate

Abstract

A method and system are provided for dynamically managing delivery of power to partitionable elements in a computer system while supporting terms of a Service Level Agreement (SLA). Parameters of the SLA are gathered in conjunction with the topology of the computer system. Transactions associated with the SLA are monitored and high and low usage periods are predicted based upon a history of transactions. Power to partitionable elements of the computer system may be adjusted during high and low usage periods. In addition, dynamic management of the partitionable elements is provided in response to current demands. Management of the partitionable elements are all made in compliance with the SLA.

Claims

1. A method for managing power delivery to a partitionable element within a multiple element computer system comprising:

collecting data pertaining to a service level agreement associated with usage of said computer system;

monitoring a transaction on said computer system; and

adjusting delivery of power to at least one of said elements to ensure compliance with said service level agreement.

2. The method of claim 1, further comprising monitoring response time of said multiple element computer system once power has been adjusted to at least one of said elements.

3. The method of claim 1, wherein said element is selected from a group consisting of:

multiple CECs, multiple blades, a processor housed within a CEC, a group of processors housed within a CEC, a processor housed within a blade, and a group of processors housed within a blade.

4. The method of claim 1, wherein the step of adjusting delivery of power to at least one of said elements is responsive to work load selected from a group consisting of:

current work load, prediction of work load, and terms of said service level agreement.

5. The method of claim 1, further comprising analyzing past performance associated with said service level agreement to predict anticipated usage periods.

6. The method of claim 1, wherein the step of adjusting delivery of power to at least one of said elements includes an in-band call to an operating system in communication with said element.

7. The method of claim 1, wherein the step of adjusting delivery of power to at least one of said elements includes an out-of-band call to a service processor in communication with said element.

8. The method of claim 1, wherein the step of adjusting delivery of power to at least one of said elements includes transitioning said element to a state selected from a group consisting of:

decreasing operating frequency of a processor in communication with said element, increasing operating frequency of a processor in communication with said element, halting a clock of a processor in communication with said element placing said element in suspend mode, hibernating said element, and powering off said element.

9. A system for managing power delivery to a partitionable element comprising:

two or more partitionable elements;

a collection manager adapted to collect data related to a service level agreement associated with a usage of said elements in response to a transaction adapted to employ at least one of said elements; and

an adjustment manager adapted to control delivery of power to at least one of said elements to ensure compliance with said service level agreement.

10. The system of claim 9, further comprising a response time manager adapted to collect response data of said computer system subsequent to an adjustment of at least one of said elements.

11. The system of claim 9, wherein said element is selected from a group consisting of:

multiple CECs, multiple blades, a processor housed within a CEC, a group of processors housed within a CEC, a processor housed within a blade, and a group of processors housed within a blade.

12. The system of claim 9, wherein said adjustment manager is responsive to work load selected from a group consisting of:

current work load, prediction of work load, and terms of said service level agreement.

13. The system of claim 9, further comprising a prediction manager adapted to analyze past performance associated with said service level agreement and to predict anticipated usage periods.

14. The system of claim 9, wherein said adjustment manager conducts an in-band call to an operating system in communication with said element.

15. The system of claim 9, wherein said adjustment manager conducts an out-of-band call to a service processor in communication with said element.

16. The system of claim 9, wherein said adjustment manager transitions delivery of power to at least one of said elements to a state selected from a group consisting of:

decreasing operating frequency of a processor in communication with said element, increasing operating frequency of a processor in communication with said element, halting a clock of a processor in communication with said element placing said element in suspend mode, hibernating said element, and powering off said element.

17. An article comprising:

a computer-readable signal-bearing medium;

means in the medium for collecting data pertaining to a service level agreement associated with usage of said computer system;

means in the medium for monitoring a transaction on said computer system; and

means in the medium for adjusting delivery of power to at least one of said elements to ensure compliance with said service level agreement.

18. The article of claim 17, wherein the medium is selected from a group consisting of:

a recordable data storage medium, and a modulated carrier signal.

19. The article of claim 17, further comprising means in the medium for monitoring response time of said multiple element computer system once power has been adjusted to at least one of said elements.

20. The article of claim 17, wherein said element is selected from a group consisting of:

multiple CECs, multiple blades, a processor housed within a CEC, a group of processors housed within a CEC, a processor housed within a blade, and a group of processors housed within a blade.

21. The article of claim 17, wherein said means for adjusting delivery of power to at least one of said elements is responsive to work load selected from a group consisting of:

current work load, prediction of work load, and terms of said service level agreement.

22. The article of claim 17, further comprising means in the medium for analyzing past performance associated with said service level agreement to predict anticipated usage periods.

23. The article of claim 17, wherein said means for adjusting delivery of power to at least one of said elements includes an in-band call to an operating system in communication with said element.

24. The article of claim 17, wherein said means for adjusting delivery of power to at least one of said elements includes an out-of-band call to a service processor in communication with said element.

25. The article of claim 17, wherein said means for adjusting delivery of power to at least one of said elements includes transitioning said element to a state selected from a group consisting of:

decreasing operating frequency of a processor in communication with said element, increasing operating frequency of a processor in communication with said element, halting a clock of a processor in communication with said element placing said element in suspend mode, hibernating said element, and powering off said element.