Method and apparatus for generating a model for an electronic prognostics system

• US 8,706,451 B1
• Filed: 12/15/2006
• Issued: 04/22/2014
• Est. Priority Date: 12/15/2006
• Status: Active Grant

• Current Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Original Assignee: Oracle America, Inc. (Oracle Corporation)
• Inventors: Gross, Kenny C., Whisnant, Keith A.
• Primary Examiner(s): Shah, Kamini S
• Assistant Examiner(s): Guill, Russ
• Application Number:
  US11/640,017
• Time in Patent Office: 2,685 Days
• Field of Search: 703/2
• US Class Current: 703/2
• CPC Class Codes: G06F 11/0706 the processing taking place on a specific hardware platform or in a specific software environment
• CPC Class Codes: G06F 11/0751 Error or fault detection not based on redundancy power supply failures G06F1/30; network fault management H04L41/06
• CPC Class Codes: G06F 11/3055 Monitoring arrangements for monitoring the status of the computing system or of the computing system component, e.g. monitoring if the computing system is on, off, available, not available error or fault processing without redundancy G06F11/0703; error detection or correction by redundancy in data representation G06F11/08; error detection or correction by redundancy in operation G06F11/14; error detection or correction by redundancy in hardware G06F11/16
• CPC Class Codes: G06F 11/3093 Configuration details thereof, e.g. installation, enabling, spatial arrangement of the probes

Abstract

A system that generates a model for an electronic prognostics system which is used to monitor a computer system while the computer system is operating correctly. During operation, the system receives telemetry signals generated by the computer system. Next, the system builds the model based on the telemetry signals, wherein the model uses correlations between different telemetry signals to identify anomalies in the telemetry signals. The system then validates the model against telemetry signals from the computer system.

Claims

1. A method for generating a model for an electronic prognostics system which is used to monitor a computer system, comprising:

receiving telemetry signals generated by the computer system while the computer system is operating correctly;

building the model based on the telemetry signals, wherein the model uses correlations between different telemetry signals to identify anomalies in the telemetry signals; and

performing a validation operation using the model, wherein performing the validation operation comprises validating the model against monitored telemetry signals from the computer system, which involves;

determining false alarms probabilities (FAPs) for the monitored telemetry signals;

determining missed alarm probabilities (MAPs) for the monitored telemetry signals, wherein determining the MAPs comprises observing alarms generated by applying the model to telemetry signals generated when a computer system is degraded;

comparing the FAPs and the MAPs to target values for the FAPs and the MAPs; and

if the FAPs or the MAPs exceed the target values,

adjusting sensitivity parameters for the model; and

rebuilding the model based on the adjusted sensitivity parameters and the telemetry signals; and

repeating the validation operation using the rebuilt model until FAPS and MAPs for the rebuilt model are within the target values.

2. The method of claim 1, wherein prior to building the model, the method further comprises sanitizing the telemetry signals which involves at least one of:

replacing missing values in the telemetry signals; and

removing values from the telemetry signals that are deemed invalid.

3. The method of claim 2, wherein after sanitizing the telemetry signals, the method further comprises resampling the sanitized telemetry signals to generate a uniform sampling interval for the sanitized telemetry signals.

4. The method of claim 2, wherein either sanitizing the telemetry signals or building the model involves eliminating duplicate telemetry signals.

5. The method of claim 2, wherein after sanitizing the telemetry signals, the method further comprises processing the sanitized telemetry signals to undo effects of quantization on the sanitized telemetry signals.

6. The method of claim 5, wherein processing the sanitized telemetry signals to undo the effects of quantization comprises:

receiving a time series containing a low-resolution quantized signal which is sampled from a high-resolution signal;

performing a spectral analysis on the time series to obtain a frequency series for the low-resolution quantized signal;

selecting a subset of frequency terms from the frequency series which have the largest amplitudes; and

reconstructing the high-resolution signal from the low-resolution quantized signal by performing an inverse spectral analysis on the subset of the frequency terms, wherein reconstructing the high-resolution signal facilitates removing the quantization effects from the low-resolution quantized signal.

7. The method of claim 1, wherein prior to building the model, the method further comprises selecting a subset of the telemetry signals to be used to build the model.

8. The method of claim 7, wherein rebuilding the model based on the adjusted sensitivity parameters and the telemetry signals comprises using a different subset of the telemetry signals to rebuild the model, wherein the different subset comprises at least one of the telemetry signals that is not in the subset that was used to build the model.

9. The method of claim 1, wherein building the model involves using a pattern recognition technique which includes at least one of:

a non linear, non parametric regression technique;

a neural network;

an adaptive Kalman filter; and

an autoassociative kernel regression technique.

10. The method of claim 1, wherein determining FAPs for the monitored telemetry signals involves:

observing alarms generated by applying the model to telemetry signals generated when a computer system is not degraded; and

normalizing the observed alarms to generate FAPs for the monitored telemetry signals.

11. The method of claim 1, wherein determining MAPs for the monitored telemetry signals

involves comparing the alarms to alarms observed from the computer system.

12. The method of claim 1, wherein determining MAPs for the monitored telemetry signals involves:

simulating degradation mechanisms within the computer system by superimposing faults onto the telemetry signals to produce simulated faulty telemetry signals;

applying the model to the simulated faulty telemetry signals to generate simulated alarms; and

correlating the superimposed faults with corresponding simulated alarms to generate MAPs.

13. The method of claim 1, wherein receiving the telemetry signals involves receiving:

real time telemetry signals;

or

archived telemetry signals.

14. The method of claim 13, wherein receiving the telemetry signals comprises receiving the archived telemetry signals, and wherein the method further comprises:

periodically measuring a number of performance parameters in the computer system;

storing data representing values for the measured performance parameters in a first storage space;

recurrently compressing data stored in the first storage space, wherein compressing data allows additional data representing newly collected performance information to be stored in the first storage space;

assigning a second storage space for storing real-time data carrying performance information of the computer system; and

periodically consolidating the real-time data stored in the second storage space and storing the consolidated real-time data in the first storage space.

15. The method of claim 1, wherein receiving the telemetry signals comprises receiving the telemetry signals on a bus that comprises:

data for the telemetry signals from sensors, and

alarms generated by using the data.

16. A non-transitory computer-readable storage medium storing instructions that when executed by a computer cause the computer to perform a method for generating a model for an electronic prognostics system which is used to monitor a computer system, wherein the method comprises:

receiving telemetry signals generated by the computer system while the computer system is operating correctly;

building the model based on the telemetry signals, wherein the model uses correlations between different telemetry signals to identify anomalies in the telemetry signals; and

performing a validation operation using the model, wherein performing the validation operation comprises validating the model against monitored telemetry signals from the computer system, which involves;

determining false alarms probabilities (FAPs) for the monitored telemetry signals;

determining missed alarm probabilities (MAPs) for the monitored telemetry signals, wherein determining the MAPs comprises observing alarms generated by applying the model to telemetry signals generated when a computer system is degraded;

comparing the FAPs and the MAPs to target values for the FAPs and the MAPs; and

if the FAPs or the MAPs exceed the target values,

adjusting sensitivity parameters for the model;

rebuilding the model based on the adjusted sensitivity parameters and the telemetry signals; and

repeating the validation operation using the rebuilt model until FAPS and MAPs for the rebuilt model are within the target values.

17. The computer-readable storage medium of claim 16, wherein prior to building the model of the computer system, the method further comprises sanitizing the telemetry signals which involves at least one of:

replacing missing values in the telemetry signals; and

removing values from the telemetry signals that are deemed invalid.

18. The computer-readable storage medium of claim 16, wherein building the model involves using a pattern recognition technique which includes at least one of:

a non linear, non parametric regression technique;

a neural network;

an adaptive Kalman filter; and

an autoassociative kernel regression technique.

19. The computer-readable storage medium of claim 16, wherein determining FAPs for the monitored telemetry signals involves:

observing alarms generated by applying the model to telemetry signals generated when a computer system is not degraded; and

normalizing the observed alarms to generate FAPs for the monitored telemetry signals.

20. The computer-readable storage medium of claim 16, wherein determining MAPs for the monitored telemetry signals involves

comparing the alarms to alarms observed from the computer system.

21. The computer-readable storage medium of claim 16, wherein determining MAPs for the monitored telemetry signals involves:

simulating degradation mechanisms within the computer system by superimposing faults onto the telemetry signals to produce simulated faulty telemetry signals;

applying the model to the simulated faulty telemetry signals to generate simulated alarms; and

correlating the superimposed faults with corresponding simulated alarms to generate MAPs.

22. The computer-readable storage medium of claim 16, wherein receiving the telemetry signals involves receiving:

real time telemetry signals;

or

archived telemetry signals.

23. An apparatus that generates a model for an electronic prognostics system which is used to monitor a computer system, comprising:

a processor;

memory attached to the processor;

a receiving mechanism configured to receive telemetry signals generated by the computer system while the computer system is operating correctly;

a model building mechanism configured to build the model based on the telemetry signals, wherein the model uses correlations between different telemetry signals to identify anomalies in the telemetry signals; and

a validation mechanism configured to perform a validation operation using the model, wherein while performing the validation operation, the validation mechanism is configured to validate the model against monitored telemetry signals from the computer system, which involves;

determining false alarms probabilities (FAPs) for the monitored telemetry signals;

determining missed alarm probabilities (MAPs) for the monitored telemetry signals, wherein determining the MAPs comprises observing alarms generated by applying the model to telemetry signals generated when a computer system is degraded;

comparing the FAPs and the MAPs to target values for the FAPs and the MAPs; and

if the FAPs or the MAPs exceed the target values,

adjusting sensitivity parameters for the model;

rebuilding the model based on the adjusted sensitivity parameters and the telemetry signals; and

repeating the validation operation using the rebuilt model until FAPS and MAPs for the rebuilt model are within the target values,

wherein the processor comprises at least one of the receiving, model building, and validation mechanisms.