Method and system for generating a power consumption model of at least one server

• US 9,495,272 B2
• Filed: 06/11/2009
• Issued: 11/15/2016
• Est. Priority Date: 06/11/2009
• Status: Active Grant

• Current Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Original Assignee: Oracle America, Inc. (Oracle Corporation)
• Inventors: Elting, David Brian, Vaidyanathan, Kalyanaraman, Gross, Kenny C.
• Primary Examiner(s): Huynh, Kim
• Assistant Examiner(s): KINSEY, BRANDON MICHAEL
• Application Number:
  US12/482,925
• Publication Number:
  US 20100318828A1
• Time in Patent Office: 2,714 Days
• Field of Search: 713/300, 340/540, 347/101, 702/60, 702/127, 702/130, 702/132, 702/136, 374/57, 374/102, 374/103, 374/133, 374/137
• US Class Current: 1/1
• CPC Class Codes: G06F 11/3062 where the monitored property is the power consumption power management in a computing system G06F1/3203
• CPC Class Codes: G06F 11/3096 wherein the means or processing minimize the use of computing system or of computing system component resources, e.g. non-intrusive monitoring which minimizes the probe effect: sniffing, intercepting, indirectly deriving the monitored data from other directly available data

Abstract

A system for generating a power consumption model of at least one server includes one or more computers configured to obtain n time series telemetry signals indicative of operating parameters of the at least one server, obtain a time series power signal indicative of power consumed by the at least one server, and correlate each of the n time series telemetry signals with the time series power signal. The one or more computers are further configured to select a set of the n time series telemetry signals having an overall correlation with the time series power signal greater than a predetermined threshold, and generate a power consumption model of the at least one server based on at least the set of the n time series telemetry signals.

Claims

1. A method for generating a power consumption model of at least one server, the method comprising:

by at least one processor,

subtracting an ambient temperature signal from each of n temperature signals indicative of localized temperatures within the at least one server to normalize the n temperature signals,

correlating each of the normalized n temperature signals with a power signal indicative of power consumed by the at least one server,

generating a combination correlation coefficient for each combination of normalized n−

1 temperature signals if an overall correlation coefficient for the normalized n temperature signals is less than a predetermined threshold,

selecting a set of the normalized n temperature signals defined by one of the combinations of normalized n−

1 temperature signals having a corresponding combination correlation coefficient greater than the predetermined threshold, and

regressing at least the set of the normalized n temperature signals with the power signal to generate a power consumption model of the at least one server.

2. The method of claim 1 wherein correlating each of the normalized n temperature signals with the power signal includes generating a correlation coefficient associated with each of the normalized n temperature signals.

3. The method of claim 2 further comprising comparing the overall correlation coefficient with the predetermined threshold.

4. The method of claim 1 further comprising comparing each of the combination correlation coefficients with the predetermined threshold.

5. A system for generating a power consumption model of at least one server, the system comprising:

one or more computers configured to (i) subtract an ambient time series temperature signal from each of n time series temperature signals indicative of localized temperatures within the at least one server to normalize the n time series temperature signals, (ii) correlate each of the normalized n time series temperature signals with a time series power signal indicative of power consumed by the at least one server, (iii) generate a combination correlation coefficient for each combination of normalized n−

1 time series temperature signals if an overall correlation coefficient for the normalized n time series temperature signals is less than a predetermined threshold, (iv) select a set of the normalized n time series temperature signals defined by one of the combinations of normalized n−

1 time series temperature signals having a corresponding combination correlation coefficient greater than the predetermined threshold, and (v) generate a power consumption model of the at least one server based on at least the set of the normalized n time series temperature signals.

6. The system of claim 5 wherein correlating each of the normalized n time series temperature signals with the time series power signal includes generating a correlation coefficient associated with each of the normalized n time series temperature signals.

7. The system of claim 6 wherein the one or more computers are further configured to compare the overall correlation coefficient with the predetermined threshold.

8. The system of claim 5 wherein the one or more computers are further configured to compare each of the combination correlation coefficients with the predetermined threshold.

9. A non-transitory computer readable storage medium having information stored thereon for directing one or more computers to (i) subtract an ambient time series temperature signal from each of n time series temperature signals indicative of localized temperatures within at least one server to normalize the n time series temperature signals, (ii) correlate each of the normalized n time series temperature signals with a time series power signal indicative of power consumed by the at least one server, (iii) generate a combination correlation coefficient for each combination of normalized n−

1 time series temperature signals if an overall correlation coefficient for the normalized n time series temperature signals is less than a predetermined threshold, (iv) select a set of the normalized n time series temperature signals defined by one of the combinations of normalized n−

1 time series temperature signals having a corresponding combination correlation coefficient greater than the predetermined threshold, and (v) generate a power consumption model of the at least one server based on at least the set of the normalized n time series temperature signals.

10. The storage medium of claim 9 wherein correlating each of the normalized n time series temperature signals with the time series power signal includes generating a correlation coefficient associated with each of the normalized n time series temperature signals.

11. The storage medium of claim 10 having information stored thereon for further directing the one or more computers to compare the overall correlation coefficient with the predetermined threshold.

12. The storage medium of claim 9 having information stored thereon for further directing the one or more computers to compare each of the combination correlation coefficients with the predetermined threshold.