Method and apparatus for predicting remaining useful life for a computer system

• US 20080140362A1
• Filed: 12/06/2006
• Published: 06/12/2008
• Est. Priority Date: 12/06/2006
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Oracle America, Inc. (Oracle Corporation)
• Inventors: Gross, Kenny C., Urmanov, Aleksey M., Whisnant, Keith A.
• Granted Patent: US 7,702,485 B2
• Time in Patent Office: Days
• US Class Current: 703/2
• CPC Class Codes: G06F 11/004 Error avoidance G06F11/07 and subgroups take precedence
• CPC Class Codes: G06F 11/008 Reliability or availability analysis

Abstract

One embodiment of the present invention provides a system for predicting a remaining useful life (RUL) for a computer system. The system starts by collecting values for at least one telemetry metric from the computer system while the computer system is operating. The system then uses the collected values to generate a RUL prediction for the computer system or a component within the computer system

Claims

1. A method for predicting a remaining useful life (RUL) for a computer system, comprising:

collecting values of at least one telemetry metric from the computer system while the computer system is operating; and

using the collected values to generate a RUL prediction for the computer system or a component within the computer system.

2. The method of claim 1, wherein collecting the values of the at least one telemetry metric involves repeatedly measuring at least one computer system parameter at a predetermined time interval, wherein the computer system parameter includes at least one of:

usage rating, an idle time measure, a throughput, a latency within the computer system, a load in the computer system, a result returned from a software-based monitor, or another software parameter that indicates a current operating state of the computer system.

3. The method of claim 1, wherein collecting the values of the at least one telemetry metric involves repeatedly measuring at least one computer system parameter at a predetermined time interval, wherein the computer system parameter includes at least one of:

a temperature, a voltage, a current, or another physical parameter that indicates a current operating state of the computer system.

4. The method of claim 1, wherein using the collected values to generate the RUL prediction involves:

using a non-linear, non-parametric regression model to predict a value for a current value;

calculating a residual value between the current value and the predicted value;

checking the residual value using a sequential probability ratio test (SPRT), wherein the SPRT generates an alarm based upon the residual value and a prior sequence of residual values; and

determining a rate at which the SPRT is generating alarms.

5. The method of claim 4 wherein using the non-linear, non-parametric regression model involves using a multivariate state estimation technique (MSET).

6. The method of claim 4, wherein generating the RUL prediction involves:

inputting the rate at which the SPRT is generating alarms into a mathematical model for generating a RUL obtained from prior failure data; and

returning a RUL prediction calculated using the mathematical model.

7. The method of claim 6, wherein the mathematical model is a linear logistic regression model.

8. An apparatus for predicting a remaining useful life (RUL) for a computer system, comprising:

a collection mechanism coupled to the computer system, wherein the collection mechanism is configured to collect values of at least one telemetry metric from the computer system while the computer system is operating; and

a RUL prediction mechanism coupled to the collection mechanism, wherein the RUL prediction mechanism is configured to use the collected values to generate a RUL prediction for the computer system or a component within the computer system.

9. The apparatus of claim 8, wherein the collection mechanism is configured to collect the values of the at least one telemetry metric by repeatedly measuring at least one computer system parameter at a predetermined time interval, wherein the computer system parameter includes at least one of:

a usage rating, an idle time measure, a throughput, a latency within the computer system, a load in the computer system, a result returned from a software-based monitor, or another software parameter that indicates a current operating state of the computer system.

10. The apparatus of claim 8, wherein the collection mechanism is configured to collect the values of the at least one telemetry metric by repeatedly measuring at least one computer system parameter at a predetermined time interval, wherein the computer system parameter includes at least one of:

a temperature, a voltage, a current, or another physical parameter that indicates a current operating state of the computer system.

11. The apparatus of claim 8, wherein the RUL prediction mechanism is configured to use the collected values to generate the RUL prediction by:

using a non-linear, non-parametric regression model to predict a value for a current value;

calculating a residual value between the current value and the predicted value;

checking the residual value using a sequential probability ratio test (SPRT), wherein the SPRT generates an alarm based upon the residual value and a prior sequence of residual values; and

determining a rate at which the SPRT is generating alarms.

12. The apparatus of claim 11 wherein the RUL prediction mechanism is configured to use a multivariate state estimation technique (MSET) as the non-linear, non-parametric regression model.

13. The method of claim 11, wherein the RUL prediction mechanism is configured to generate the RUL prediction by:

inputting the rate at which the SPRT is generating alarms into a mathematical model for generating a RUL obtained from prior failure data; and

returning a RUL prediction calculated using the mathematical model.

14. The method of claim 13, wherein the RUL prediction mechanism is configured so that the mathematical model is a linear logistic regression model.

15. A monitoring system for predicting a remaining useful life (RUL) for a computer system, comprising:

a collection mechanism coupled to the computer system, wherein the collection mechanism is configured to collect values of at least one telemetry metric from the computer system while the computer system is operating;

a RUL prediction mechanism coupled to the collection mechanism, wherein the RUL prediction mechanism is configured to use the collected values to generate a RUL prediction for the computer system or a component within the computer system; and

an interface device coupled to the RUL prediction mechanism, wherein the interface device is configured to deliver the RUL prediction to a human user or to another system.

16. The monitoring system of claim 15, wherein the collection mechanism is configured to collect the values of the at least one telemetry metric by repeatedly measuring at least one computer system parameter at a predetermined time interval, wherein the computer system parameter includes at least one of:

a usage rating, an idle time measure, a throughput, a latency within the computer system, a load in the computer system, a result returned from a software-based monitor, or another software parameter that indicates a current operating state of the computer system.

17. The monitoring system of claim 15, wherein the collection mechanism is configured to collect the values of the at least one telemetry metric by repeatedly measuring at least one computer system parameter at a predetermined time interval, wherein the computer system parameter includes at least one of:

a temperature, a voltage, a current, or another physical parameter that indicates a current operating state of the computer system.

18. The monitoring system of claim 15, wherein the RUL prediction mechanism is configured to use the collected values to generate the RUL prediction by:

using a non-linear, non-parametric regression model to predict a value for a current value;

calculating a residual value between the current value and the predicted value;

checking the residual value using a sequential probability ratio test (SPRT), wherein the SPRT generates an alarm based upon the residual value and a prior sequence of residual values; and

determining a rate at which the SPRT is generating alarms.

19. The monitoring system of claim 18 wherein the RUL prediction mechanism is configured to use a multivariate state estimation technique (MSET) as the non-linear, non-parametric regression model.

20. The monitoring system of claim 18, wherein the RUL prediction mechanism is configured to generate the RUL prediction by:

inputting the rate at which the SPRT is generating alarms into a mathematical model for generating a RUL obtained from prior failure data; and

returning a RUL prediction calculated using the mathematical model.

21. The monitoring system of claim 20, wherein the RUL prediction mechanism is configured so that the mathematical model is a linear logistic regression model.