Inferential power monitor without voltage/current transducers

• US 20070040582A1
• Filed: 08/17/2005
• Published: 02/22/2007
• Est. Priority Date: 08/17/2005
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Oracle America, Inc. (Oracle Corporation)
• Inventors: Vaidyanathan, Kalyanaraman, Gross, Kenny, Dhanekula, Ramakrishna
• Granted Patent: US 7,869,965 B2
• Time in Patent Office: Days
• US Class Current: 326/88
• CPC Class Codes: G06F 1/28 Supervision thereof, e.g. detecting power-supply failure by out of limits supervision
• CPC Class Codes: G06F 1/3209 Monitoring remote activity, e.g. over telephone lines or network connections
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 11/3452 Performance evaluation by statistical analysis
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

A system that facilitates estimating power consumption in a computer system by inferring the power consumption from instrumentation signals. During operation, the system monitors instrumentation signals within the computer system, wherein the instrumentation signals do not include corresponding current and voltage signals that can be used to directly compute power consumption. The system then estimates the power consumption for the computer system by inferring the power consumption from the instrumentation signals and from an inferential power model generated during a training phase.

Claims

1. A method for estimating power consumption in a computer system by inferring the power consumption from instrumentation signals, comprising:

monitoring instrumentation signals within the computer system, wherein the instrumentation signals do not include corresponding current and voltage signals that can be used to directly compute power consumption; and

estimating the power consumption for the computer system by inferring the power consumption from the instrumentation signals and from an inferential power model generated during a training phase.

2. The method of claim 1, wherein estimating the power consumption further involves:

pre-processing the instrumentation signals using an analytical re-sampling program; and

using a nonlinear, nonparametric regression technique to produce an inferential power estimate from the pre-processed instrumentation signals and the inferential power model.

3. The method of claim 2, wherein generating the inferential power model during the training phase involves:

monitoring the instrumentation signals within the computer system;

monitoring a corresponding power consumption using a hardware power monitor attached to the computer system;

pre-processing the monitored instrumentation signals and the corresponding monitored power consumption using an analytical re-sampling program; and

using a nonparametric regression technique to build an inferential power model from the pre-processed instrumentation signals and the corresponding monitored power consumption.

4. The method of claim 3, wherein a linear regression technique is used to produce the inferential power estimate from the pre-processed instrumentation signals and the inferential power model.

5. The method of claim 2, wherein the nonlinear, nonparametric regression technique can include a multivariate state estimation technique.

6. The method of claim 1, wherein the instrumentation signals can include hardware variables and/or software variables.

7. The method of claim 6, wherein the software variables include load metrics, CPU utilization, idle time, memory utilization, disk activity, transaction latencies, and other performance metrics reported by the operating system.

8. The method of claim 6, wherein the hardware variables include temperature and voltage.

9. The method of claim 6, wherein only software variables are used to estimate the power consumption of the computer system.

10. An apparatus for estimating power consumption in a computer system by inferring the power consumption from instrumentation signals, comprising:

a computer system; and

a power-estimation mechanism;

wherein the power-estimation mechanism is configured to;

monitor instrumentation signals within the computer system, wherein the instrumentation signals do not include corresponding current and voltage signals that can be used to directly compute power consumption; and

to

estimate the power consumption for the computer system by inferring the power consumption from the instrumentation signals and from an inferential power model generated during a training phase.

11. The apparatus of claim 10, wherein while estimating the power consumption, the power-estimation mechanism is configured to:

pre-process the instrumentation signals using an analytical re-sampling program; and

to

use a nonlinear, nonparametric regression technique to produce an inferential power estimate from the pre-processed instrumentation signals and the inferential power model.

12. The apparatus of claim 11, wherein while generating the inferential power model during the training phase, the power-estimation mechanism is configured to:

monitor the instrumentation signals within the computer system;

monitor a corresponding power consumption using a hardware power monitor attached to the computer system;

pre-process the monitored instrumentation signals and the corresponding monitored power consumption using the analytical re-sampling program; and

to

use a nonparametric regression technique to build the inferential power model from the pre-processed instrumentation signals and the corresponding monitored power consumption.

13. The method of claim 12, wherein a linear regression technique is used to produce an inferential power estimate from the pre-processed instrumentation signals and the inferential power model.

14. The apparatus of claim 11, wherein the nonlinear, nonparametric regression technique can include a multivariate state estimation technique.

15. The apparatus of claim 10, wherein the instrumentation signals can include hardware variables and/or software variables.

16. The apparatus of claim 15, wherein the software variables include load metrics, CPU utilization, idle time, memory utilization, disk activity, transaction latencies, and other performance metrics reported by the operating system.

17. The apparatus of claim 15, wherein the hardware variables include temperature and voltage.

18. The apparatus of claim 15, wherein only software variables are used to estimate the power consumption of the computer system.

19. A computer system for estimating power consumption in a computer system by inferring the power consumption from instrumentation signals, comprising:

a computer system; and

a power-estimation mechanism;

wherein the power-estimation mechanism is configured to;

monitor instrumentation signals within the computer system, wherein the instrumentation signals do not include corresponding current and voltage signals that can be used to directly compute power consumption; and

to

estimate the power consumption for the computer system by inferring the power consumption from the instrumentation signals and from an inferential power model generated during a training phase.

20. The computer system of claim 19, wherein while estimating the power consumption, the power-estimation mechanism is configured to:

pre-process the instrumentation signals using an analytical re-sampling program; and

to

use a nonlinear, nonparametric regression technique to produce an inferential power estimate from the pre-processed instrumentation signals and the inferential power model.

21. The computer system of claim 20, wherein while generating the inferential power model during the training phase, the power-estimation mechanism is configured to:

monitor the instrumentation signals within the computer system;

monitor a corresponding power consumption using a hardware power monitor attached to the computer system;

pre-process the monitored instrumentation signals and the corresponding monitored power consumption using an analytical re-sampling program; and

to

use a nonparametric regression technique to build an inferential power model from the pre-processed instrumentation signals and the corresponding monitored power consumption.