Managing power consumption in a data center

• US 9,264,334 B1
• Filed: 11/12/2013
• Issued: 02/16/2016
• Est. Priority Date: 12/07/2009
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Brandwine, Eric Jason
• Primary Examiner(s): Ibrahim, Mohamed
• Application Number:
  US14/078,365
• Time in Patent Office: 826 Days
• Field of Search: None
• US Class Current: 1/1
• 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: G06F 9/5094 where the allocation takes into account power or heat criteria power management in computers in general G06F1/3203; thermal management in computers in general G06F1/206
• CPC Class Codes: H04L 43/0876 Network utilisation, e.g. volume of load or congestion level
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

Systems and methods are provided for managing resources. In one implementation, a method is provided in which a management server determines whether a condition related to one or more resources has occurred. The management server further determines at least one program instance to terminate. The at least one program instance executes on one of a plurality of servers. The management server further terminates the determined at least one program instance, which was used by an excess program execution capacity user.

Claims

1. A computer-implemented method comprising:

obtaining, by a management server computer, performance data related to operation of a plurality of computer-related resources;

modeling, by the management server computer and based at least in part on the obtained performance data, a predicted future condition related to the operation of the plurality of computer-related resources;

determining, by the management server computer and based at least in part on the predicted future condition, to terminate one or more executing program instances to reduce usage of the plurality of computer-related resources;

selecting, by the management server computer and in response to the determining, at least one executing program instance to terminate using priority data associated with the at least one program instance; and

terminating, by the management server computer and in response to the determining, the selected at least one program instance.

2. The computer-implemented method of claim 1 wherein the plurality of computer-related resources include one or more servers on which the at least one program instance is executing before the determining.

3. The computer-implemented method of claim 2 wherein the one or more servers are located in a data center, and wherein the obtained performance data includes a temperature within the data center.

4. The computer-implemented method of claim 2 wherein the one or more servers are located in a data center, and wherein the obtained performance data includes a temperature external to the data center.

5. The computer-implemented method of claim 2 wherein the obtained performance data includes a power consumption level for the one or more servers.

6. The computer-implemented method of claim 2 wherein the one or more servers are located in a data center, and wherein the obtained performance data includes an environmental condition related to the data center.

7. The computer-implemented method of claim 2 wherein the obtained performance data includes a processing load for the one or more servers.

8. The computer-implemented method of claim 2 wherein the one or more servers further execute one or more other program instances for one or more dedicated program execution capacity users who receive dedicated access of a specified amount of program execution capacity, and wherein the priority data further indicates that the at least one program instance has a lower priority than the one or more other program instances.

9. The computer-implemented method of claim 2 wherein the one or more servers further execute one or more other program instances for one or more on-demand variable program execution capacity users who do not use excess program execution capacity, and wherein the priority data further indicates that the at least one program instance has a lower priority than the one or more other program instances.

10. The computer-implemented method of claim 1 wherein the modeling of the predicted future condition includes modeling an anticipated increase in power utilization by one or more computer-related resources of the plurality of computer-related resources.

11. The computer-implemented method of claim 10 wherein the obtained performance data includes telemetry data indicating current power utilization by the one or more computer-related resources.

12. The computer-implemented method of claim 10 wherein the obtained performance data includes temperature data, and wherein the modeling of the predicted future condition further includes determining a trend in temperature changes that is predicted to result in the anticipated increase in power utilization.

13. The computer-implemented method of claim 10 wherein the modeling of the predicted future condition further includes determining a weather trend that is predicted to result in the anticipated increase in power utilization.

14. The computer-implemented method of claim 1 wherein the modeling of the predicted future condition further includes determining a weather trend, and determining the predicted future condition based on the determined weather trend.

15. The computer-implemented method of claim 1 wherein the at least one program instance is executed on behalf of one or more excess program execution capacity users who use excess program execution capacity, and wherein the priority data associated with the at least one program instance is a priority of the one or more excess program execution capacity users that is lower than one or more other priorities of one or more other users.

16. The computer-implemented method of claim 1 wherein the terminating of the selected at least one executing program instance includes powering down one or more of the plurality of computer-related resources that had been being used by the selected at least one executing program instance before the terminating.

17. The computer-implemented method of claim 1 wherein the management server computer is affiliated with a program execution service that executes program instances for a plurality of users, and wherein the at least one program instance is executed by the program execution service for at least one of the plurality of users.

18. A computer system comprising:

a processor for executing program instructions; and

a memory storing the program instructions, wherein the program instructions, upon execution by the processor, cause the system to;

obtain performance data related to operation of one or more computer-related resources;

model a predicted future condition related to the operation of the one or more computer-related resources based at least in part on the obtained performance data;

determine, based at least in part on the predicted future condition, to terminate one or more executing program instances to reduce usage of the one or more computer-related resources, and selecting in response at least one program instance to terminate; and

terminate, in response to the determining, the selected at least one program instance.

19. The computer system of claim 18 wherein the one or more computer-related resources include one or more servers on which the at least one program instance is executing before the determining, and wherein the obtained performance data indicates one or more current conditions of a data center in which the one or more servers are located.

20. The computer system of claim 19 wherein the one or more servers further execute one or more other program instances, and wherein the selecting of the at least one program instance to terminate is further based on priority data that indicates that the at least one program instance has a lower priority than the one or more other program instances.

21. The computer system of claim 18 wherein the modeling of the predicted future condition includes modeling an anticipated increase in power utilization by the one or more computer-related resources, and wherein the obtained performance data includes telemetry data indicating at least one of a group of data types including power utilization by the one or more computer-related resources, temperature data associated with the one or more computer-related resources, or a processing load for the one or more computer-related resources.

22. The computer system of claim 18 wherein the modeling of the predicted future condition further includes determining a weather trend, and determining the predicted future condition based on the determined weather trend.

23. A computer-implemented method comprising:

monitoring, by a management server computer, a plurality of computer-related resources for a data center to determine performance data related to operation of the plurality of computer-related resources;

modeling, by the management server computer and based at least in part on the performance data, a predicted future condition related to the operation of the plurality of computer-related resources;

determining, by the management server computer and based at least in part on the predicted future condition, at least one resource of the plurality to terminate or shut down using associated priority data to reduce usage of the plurality of computer-related resources; and

terminating or shutting down, by the management server computer, the determined at least one resource.

24. The computer-implemented method of claim 23 wherein the plurality of computer-related resources includes multiple servers and multiple program instances executing on the multiple servers, and wherein the determined performance data includes at least one of a group of data types including power utilization by the multiple servers, temperature data associated with the multiple servers, or a processing load for the multiple servers.

25. The computer-implemented method of claim 24 wherein the determining of the at least one resource to terminate or shut down includes selecting the at least one resource to be one of the one or more executing program instances in response to the determining, and wherein the terminating or shutting down of the at least one resource includes terminating execution of the selected one program instance or shutting down one of the multiple servers on which the selected one program instance is executing.