Method and apparatus for determining a predicted failure rate

• US 7,539,907 B1
• Filed: 05/05/2006
• Issued: 05/26/2009
• Est. Priority Date: 05/05/2006
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Inventors: Huse, Lars Paul, Hodoba, Predrag, Opheim, Hans Ludwig, Johnsen, Bjørn Dag
• Primary Examiner(s): Beausoliel; Robert
• Assistant Examiner(s): Ehne; Charles
• Application Number:
  US11/418,796
• Time in Patent Office: 1,117 Days
• Field of Search: 714/26, 714/47
• US Class Current: 714/47.2
• CPC Class Codes: G06F 11/008 Reliability or availability analysis

Abstract

A method and apparatus for determining predicted failure rates for computational resources provided by a system comprising multiple components. The method comprises storing failure rate information about individual components and determining a current configuration of components within the system. The method further comprises identifying from the current configuration dependencies between components within the system. The stored failure rate information and the identified current configuration dependencies can then be used to generate predicted failure rates for the computational resources in the current configuration of the system.

Claims

1. A method of determining a predicted failure rate in a system comprising multiple components, wherein said system provides one or more computational resources, the method comprising:

storing failure rate information about individual components;

determining a current configuration of components within the system;

identifying from the current configuration dependencies between components within the system; and

using the stored failure rate information and the identified current configuration dependencies to generate a predicted overall failure rate for each of said one or more computational resources provided by the system in its current configuration, wherein the overall failure rate for a computational resource represents a combined probability of failure for both the components within the computational resource and the components in devices that are not within the computational resource, but provide support for the computational resource.

2. The method of claim 1, wherein said predicted failure rates are generated by a box manager associated with the system.

3. The method of claim 2, further comprising reporting said predicted failure rates from the box manager to a resource manager, wherein the resource manager is responsible for allocating work to said one or more computational resources provided by the system.

4. The method of claim 2, wherein failure rate information relating to a field replaceable unit (FRU) is stored within the FRU itself, and is reported from the FRU to the box manager.

5. The method of claim 2, wherein at least some of the stored failure rate information about individual components is located remotely from the system and is retrieved by the box manager over a network.

6. The method of claim 1, further comprising using the identified configuration dependencies to determine one or more fault zones within the system, wherein no single point of failure can bring down more than one fault zone, and wherein each of said one or more computational resources is assigned to a corresponding fault zone.

7. The method of claim 6, wherein said faults zones are determined by a box manager associated with the system and are reported by the box manager along with the predicted failure rates to a resource manager.

8. The method of claim 1, further comprising:

detecting a change in the configuration of the system; and

updating the predicted failure rates in accordance with the changed configuration.

9. The method of claim 1, further comprising monitoring and recording parameters relating to the operational history of the system.

10. The method of claim 9, further comprising reporting the recorded parameters to a log site.

11. The method of claim 10, wherein said reporting is performed in response to a failure in the system.

12. The method of claim 9, further comprising using the recorded parameters to generate the predicted failure rates.

13. The method of claim 1, further comprising receiving over a network an updated model for generating the predicted failure rates.

14. A system comprising multiple components, said system providing one or more computational resources, the system further comprising:

a configuration manager for determining a current configuration of components within the system and for identifying from the current configuration dependencies between components within the system; and

a box manager for accessing stored failure rate information about individual components within the system and for using the stored failure rate information and the identified current configuration dependencies to generate a predicted overall failure rate for each of said one or more computational resources for the system in its current configuration, wherein the overall failure rate for a computational resource represents a combined probability of failure for both the components within the computational resource and the components in devices that are not within the computational resource, but provide support for the computational resource.

15. The system of claim 14, further comprising a resource manager responsible for allocating work to said one or more computational resources provided by the system, wherein the box manager is configured to report said predicted failure rates to the resource manager.

16. The system of claim 15, wherein failure rate information relating to a field replaceable unit (FRU) in the system is stored within the FRU itself for reporting from the FRU to the box manager.

17. The system of claim 14, wherein the box manager is configured to use the identified configuration dependencies to determine one or more fault zones within the system, wherein no single point of failure can bring down more than one fault zone, and wherein each of said one or more computational resources is assigned to a corresponding fault zone.

18. The system of claim 14, wherein the box manager is configured, in response to a detected change in the configuration of the system, to update the predicted failure rates in accordance with the changed configuration.

19. The system of claim 14, further comprising a monitor for recording parameters relating to the operational history of the system.

20. The system of claim 19, further comprising a communications facility for reporting the recorded parameters to a log site.

21. The system of claim 14, further comprising a communications interface for receiving an updated model for generating the predicted failure rates.

22. A method of operating a server comprising:

receiving communications over a network from a population of multiple systems, wherein each system monitors parameters relating to the operational history of that system, and wherein said communications report the monitored parameters to the server;

using the reported parameters to update a failure rate model for said population of multiple systems; and

distributing from the server over the network to the population of multiple systems the updated failure rate model for use by a system belonging to the population of multiple systems in predicting failure rates for that system.

23. The method of claim 22, wherein said failure rate model is sensitive to the operational history of a system.