Managing the performance of a computer system

• US 7,890,298 B2
• Filed: 06/12/2008
• Issued: 02/15/2011
• Est. Priority Date: 06/12/2008
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Oracle America, Inc. (Oracle Corporation)
• Inventors: Coskun, Ayse K, Gross, Kenny C
• Primary Examiner(s): TSAI, CAROL S W
• Application Number:
  US12/137,977
• Publication Number:
  US 20090313623A1
• Time in Patent Office: 978 Days
• Field of Search: 702/186, 718/100, 718/102, 718/103, 718/108, 711/148, 711/151, 711/152, 711/153, 711/158, 713/300, 713/322, 713/324
• US Class Current: 702/186
• CPC Class Codes: G06F 2209/508 Monitor
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: G06F 9/5088 involving task migration
• 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: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

Some embodiments of the present invention provide a system that manages a performance of a computer system. During operation, a current expert policy in a set of expert policies is executed, wherein the expert policy manages one or more aspects of the performance of the computer system. Next, a set of performance parameters of the computer system is monitored during execution of the current expert policy. Then, a next expert policy in the set of expert policies is dynamically selected to manage the performance of the computer system, wherein the next expert policy is selected based on the monitored set of performance parameters to improve an operational metric of the computer system.

Claims

1. A method for managing a performance of a computer system, comprising:

executing a current expert policy in a set of expert policies that manage one or more aspects of the performance of the computer system;

monitoring a set of performance parameters of the computer system during execution of the current expert policy; and

using a loss function to dynamically select a next expert policy in the set of expert policies to manage the performance of the computer system, wherein the next expert policy is selected based on the monitored set of performance parameters to improve an operational metric of the computer system; and

wherein the loss function is based on performance parameters in the set of performance parameters.

2. The method of claim 1,

wherein monitoring the set of performance parameters of the computer system includes systematically monitoring and recording a set of performance parameters of the computer system;

and

wherein the recording process keeps track of the temporal relationships between events in different performance parameters.

3. The method of claim 1,

wherein dynamically selecting the next expert policy includes selecting a specialist from a set of specialists based on the set of performance parameters, wherein each specialist is associated with at least one expert policy in the set of expert policies;

and

wherein a specialist in the set of specialists may be associated with different expert policies based on performance parameters in the set of performance parameters.

4. The method of claim 3, wherein at least two specialists in the set of specialists are associated with the same expert policy.

5. The method of claim 1, further including:

generating a set of metrics based on the set of performance parameters, wherein the set of metrics is related to one or more of;

a hot spot metric, a thermal cycle metric, a spatial gradient metric, and a load average metric; and

determining the loss function based on metrics in the set of metrics, wherein determining the loss function includes calculating an updated loss function for the one or more specialists associated with the current expert policy.

6. The method of claim 1, wherein dynamically selecting the next expert policy includes using an online learning technique.

7. The method of claim 6, wherein the online learning technique includes a switching experts framework.

8. The method of claim 1, wherein the operational metric includes at least one of:

a computer system load average, a computer system reliability, and a computer system energy efficiency.

9. A non-transitory computer-readable storage medium storing instructions that when executed by a computer cause the computer to perform a method for managing a performance of a computer system, the method comprising:

executing a current expert policy in a set of expert policies that manage one or more aspects of the performance of the computer system;

monitoring a set of performance parameters of the computer system during execution of the current expert policy; and

using a loss function to dynamically select a next expert policy in the set of expert policies to manage the performance of the computer system, wherein the next expert policy is selected based on the monitored set of performance parameters to improve an operational metric of the computer system; and

wherein the loss function is based on performance parameters in the set of performance parameters.

10. The computer-readable storage medium of claim 9,

wherein monitoring the set of performance parameters of the computer system includes systematically monitoring and recording a set of performance parameters of the computer system;

and

wherein the recording process keeps track of the temporal relationships between events in different performance parameters.

11. The computer-readable storage medium of claim 9,

wherein dynamically selecting the next expert policy includes selecting a specialist from a set of specialists based on the set of performance parameters, wherein each specialist is associated with at least one expert policy in the set of expert policies;

and

wherein a specialist in the set of specialists may be associated with different expert policies based on performance parameters in the set of performance parameters.

12. The computer-readable storage medium of claim 11, wherein at least two specialists in the set of specialists are associated with the same expert policy.

13. The computer-readable storage medium of claim 9, further including:

generating a set of metrics based on the set of performance parameters, wherein the set of metrics is related to one or more of;

a hot spot metric, a thermal cycle metric, a spatial gradient metric, and a load average metric; and

determining the loss function based on metrics in the set of metrics, wherein determining the loss function includes calculating an updated loss function for the one or more specialists associated with the current expert policy.

14. The computer-readable storage medium of claim 9, wherein dynamically selecting the next expert policy includes using an online learning technique.

15. The computer-readable storage medium of claim 14, wherein the online learning technique includes a switching experts framework.

16. The computer-readable storage medium of claim 9, wherein the operational metric includes at least one of:

a computer system load average, a computer system reliability, and a computer system energy efficiency.

17. An apparatus for managing a performance of a computer system, comprising:

an execution mechanism configured to execute a current expert policy in a set of expert policies that manage one or more aspects of the performance of the computer system;

a monitoring mechanism configured to monitor a set of performance parameters of the computer system during execution of the current expert policy; and

a dynamic selecting mechanism configured to use a loss function to dynamically select a next expert policy in the set of expert policies to manage the performance of the computer system, wherein the next expert policy is selected based on the monitored set of performance parameters to improve an operational metric of the computer system; and

wherein the loss function is based on performance parameters in the set of performance parameters.

18. The apparatus of claim 17,

wherein the monitoring mechanism includes a mechanism configured to systematically monitor and record the set of performance parameters of the computer system;

and

wherein the recording process keeps track of the temporal relationships between events in different performance parameters.