Fault diagnostic system

• US 5,127,006 A
• Filed: 07/15/1991
• Issued: 06/30/1992
• Est. Priority Date: 05/01/1989
• Status: Expired due to Term

• Current Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Original Assignee: Digital Equipment Corporation (HP Inc.)
• Inventors: Subramanian, Krishnan, Billmers, Meyer A.
• Primary Examiner(s): BEAUSOLIEL JR, ROBERT W
• Application Number:
  US07/729,901
• Time in Patent Office: 351 Days
• Field of Search: 371/23, 371/29.1, 371/22.1, 371/19, 371/15.1, 371/16.5, 371/16.1, 371/49.1, 364/513, 364/200, 364/900
• US Class Current: 714/35
• CPC Class Codes: G06F 11/2273 Test methods

Abstract

A fault diagnosis system which detects possible faulty components in a complex system. The complex system is represented by a representation comprising a plurality of nodes interconnected by links defining a flow among nodes corresponding to flows in the complex system. At least one of the nodes is a checkpoint node, which generates a report if the flow received by it is not correct. A fault diagnosis processor performing upstream search operation to repeatedly (such as by recursion of iteration) step in upstream path, that is, opposite to the direction of flow in the complex system, from node to node starting with the reporting checkpoint node. At each step, the fault diagnosis processor indicts components represented by both nodes and links in the representation.

Claims

1. A method for using a fault diagnosis system to locate a fault in a complex system including a plurality of components that receive input signals and transmit output signals, wherein the fault diagnosis system includes a representation store for storing a representation of the complex system as a plurality of interconnected nodes, including checkpoint and outjunction nodes, wherein the representation store includes a plurality of data structures corresponding to the nodes, wherein a data structure corresponding to a node identifies as upstream nodes any nodes from which the node receives input signals and identifies as downstream nodes any nodes to which the node transmits output signals, wherein a checkpoint node detects errors in input signals and reports whether an error has occurred, wherein an outjunction node transmits output signals to a plurality of downstream nodes, and wherein the fault diagnosis system includes a suspect store for identifying nodes that are likely to be causing errors, and wherein the method performed by the fault diagnosis system comprises the following steps:

selecting a checkpoint node that is reporting an error as a current upstream node, to begin an upstream search;

selecting as a current upstream node, after the upstream search begins, a node that is identified as an upstream node in the data structure corresponding to a prior upstream node, wherein the prior upstream node is a node that was most recently selected as the current upstream node, and wherein the step of selecting the current upstream node is repeated to continue the upstream search;

indicting the current upstream node, after its selection, by identifying the current upstream node in the suspect store as a node that is likely to be causing errors;

exonerating the nodes that are downstream from the current upstream node, at times when the current upstream node is a checkpoint node that is reporting an error, by not identifying the nodes that are downstream from the current upstream node in the suspect store as nodes that are likely to be causing errors, after the current upstream node is selected;

selecting as a current downstream node, at times when the current upstream node is an outjunction node, a node that is identified as a downstream node in the data structure corresponding to the current upstream node, to begin a downstream search;

selecting as a current downstream node, after the downstream search begins, a node that is identified as a downstream node in the data structure corresponding to a prior downstream node, wherein the prior downstream node is a node that was most recently selected as the current downstream node, and wherein the step of selecting the current downstream node is repeated to continue the downstream search;

indicting the nodes that are downstream from the current upstream node and are not downstream from a checkpoint node that is reporting an error, at times when the current downstream node is a checkpoint node that is not reporting an error, by identifying the nodes that are downstream from the current upstream node and are not downstream from a checkpoint node that is reporting an error in the suspect store as nodes that are likely to be causing errors; and

exonerating the nodes that are downstream from the current upstream node, at times when the current downstream node is a checkpoint node that is reporting an error, by not identifying the nodes that are downstream from the current upstream node in the suspect store as nodes that are likely to be causing errors, after the current downstream node is selected.

2. A method in accordance with claim 1, in which the step of selecting the current downstream node is not performed at times when the prior downstream node is a checkpoint node.

3. A method in accordance with claim 2, in which the step of selecting the current downstream node is not performed at times when the data structure corresponding to the prior downstream node does not identify any nodes to which the prior downstream node transmits output signals.

4. A method in accordance with claim 3, in which the step of selecting the current upstream node is not performed at times when the prior upstream node is a checkpoint node that is not reporting an error.

5. A method in accordance with claim 4, in which the step of selecting the current upstream node is not performed at times when the data structure corresponding to the prior upstream node does not identify any nodes from which the prior upstream node receives input signals.

6. A method in accordance with claim 5, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicting a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

7. A method in accordance with claim 4, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicting a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

8. A method in accordance with claim 3, in which the step of selecting the current upstream node is not performed at times when the data structure corresponding to the prior upstream node does not identify any nodes from which the prior upstream node receives input signals.

9. A method in accordance with claim 8, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicting a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

10. A method in accordance with claim 3, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicting a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

11. A method in accordance with claim 2, in which the step of selecting the current upstream node is not performed at times when the prior upstream node is a checkpoint node that is not reporting an error.

12. A method in accordance with claim 11, in which the step of selecting the current upstream node is not performed at times when the data structure corresponding to the prior upstream node does not identify any nodes from which the prior upstream node receives input signals.

13. A method in accordance with claim 12, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicting a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

14. A method in accordance with claim 11, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicting a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

15. A method in accordance with claim 14, in which the complex system is a digital data processing system, in which the input and output signals include data and parity signals, and in which each of the checkpoint nodes checks the input signals received by the checkpoint node for correct parity, and in which the step of selecting a checkpoint node comprises selecting a checkpoint node that is reporting a parity error.

16. A method in accordance with claim 11, in which the complex system is a digital data processing system, in which the input and output signals include data and parity signals, and in which each of the checkpoint nodes checks the input signals received by the checkpoint node for correct parity, and in which the step of selecting a checkpoint node comprises selecting a checkpoint node that is reporting a parity error.

17. A method in accordance with claim 2, in which the step of selecting the current upstream node is not performed at times when the data structure corresponding to the prior upstream node does not identify any nodes from which the prior upstream node receives input signals.

18. A method in accordance with claim 17, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicating a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

19. A method in accordance with claim 2, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicating a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

20. A method in accordance with claim 19, in which complex system is a digital data processing system, in which the input and output signals include data and parity signals, and in which each of the checkpoint nodes checks the input signals received by the checkpoint node for correct parity, and in which the step of selecting a checkpoint node comprises selecting a checkpoint node that is reporting a parity error.

21. A method in accordance with claim 2, in which the complex system is a digital data processing system, in which the input and output signals include data and parity signals, and in which each of the checkpoints nodes checks the input signals received by the checkpoint node for correct parity, and in which the step of selecting a checkpoint node comprises selecting a checkpoint node that is reporting a parity error.

22. A method in accordance with claim 1, in which the step of selecting the current downstream node is not performed at times when the data structure corresponding to the prior downstream node does not identify any nodes to which the prior downstreams node transmits output signals.

27. A method in accordance with claim 22, in which the step of selecting the current upstream node is not performed at times when the data structure corresponding to the prior upstream node does not identify any nodes from which the prior upstream node receives input signals.

28. A method accordance with claim 27, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicting a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

29. A method in accordance with claim 22, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicting a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

30. A method in accordance with claim 1, in which the step of selecting the current upstream node is not performed at times when the prior upstream node is a checkpoint node that is not reporting an error.

31. A method in accordance with claim 30, in which the step of selecting the current upstream node is not performed at times when the data structure corresponding to the prior upstream node does not identify any nodes from which the prior upstream node receives input signals.

32. A method in accordance with claim 31, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicting a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

33. A method in accordance with claim 30, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicting a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

34. A method in accordance with claim 33, in which the complex system is a digital data processing system, in which the input and output signals include data and parity signals, and in which each of the checkpoint nodes checks the input signals received by the checkpoint node for correct parity, and in which the step of selecting a checkpoint node comprises selecting a checkpoint node that is reporting a parity error.

35. A method in accordance with claim 30, in which the complex system is a digital data processing system, in which the input and output signals include data and parity signals, and in which each of the checkpoint nodes checks the input signals received by the checkpoint node for correct parity, and in which the step of selecting a checkpoint node comprises selecting a checkpoint node that is reporting a parity error.

36. A method in accordance with claim 1, in which the step of selecting the current upstream node is not performed at times when the data structure corresponding to the prior upstream node does not identify any nodes from which the prior upstream node receives input signals.

37. A method in accordance with claim 36, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicting a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

38. A method in accordance with claim 1, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicting a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

39. A method in accordance with claim 38, in which the fault diagnosis system includes a current upstream node pointer for identifying one of the nodes during an upstream search operation as the current upstream node, and in which the step of selecting the current upstream node includes changing the current upstream node pointer.

40. A method in accordance with claim 39, in which the fault diagnosis system includes a current downstream node pointer for identifying one of the nodes during a downstream search operation as the current downstream node, and in which the step of selecting the current downstream node includes changing the current downstream node pointer.

41. A method in accordance with claim 38, in which the fault diagnosis system includes a current downstream node pointer for identifying one of the nodes during a downstream search operation as the current downstream node, and in which the step of selecting the current downstream node includes changing the current downstream node pointer.

42. A method in accordance with claim 38, in which the complex system is a digital data processing system, in which the input and output signals include data and parity signals, and in which each of the checkpoint nodes checks the input signals received by the checkpoint node for correct parity, and in which the step of selecting a checkpoint node comprises selecting a checkpoint node that is reporting a parity error.

43. A method in accordance with claim 1, in which the fault diagnosis system includes a current upstream node pointer for identifying one of the nodes during an upstream search operation as the current upstream node, and in which the step of selecting the current upstream node includes changing the current upstream node pointer.

44. A method in accordance with claim 43, in which the fault diagnosis system includes a current downstream node pointer for identifying one of the nodes during a downstream search operation as the current downstream node, and in which the step of selecting the current downstream node includes changing the current downstream node pointer.

45. A method in accordance with claim 43, in which the complex system is a digital data processing system, in which the input and output signals include data and parity signals, and in which each of the checkpoint nodes checks the input signals received by the checkpoint node for correct parity, and in which the step of selecting a checkpoint node comprises selecting a checkpoint node that is reporting a parity error.

46. A method in accordance with claim 1, in which the fault diagnosis system includes a current downstream node pointer for identifying one of the nodes during a downstream search operation as the current downstream node, and in which the step of selecting the current downstream node includes changing the current downstream node pointer.

47. A method in accordance with claim 46, in which the complex system is digital data processing system, in which the input and output signals include data and parity signals, and in which each of the checkpoint nodes checks the input signals received by the checkpoint node for correct parity, and in which the step of selecting a checkpoint node comprises selecting a checkpoint node that is reporting a parity error.

48. A method in accordance with claim 1, in which the complex system is a digital data processing system, in which the input and output signals include data and parity signals, and in which each of the checkpoint nodes checks the input signals received by the checkpoint node for correct parity, and in which the step of selecting a checkpoint node coomprises selecting a checkpoint node that is reporting a parity error.

23. A method in accordance with clain 22, in which the step of selecting the current upstream node is not performed at times when the prior upstream node is a checkpoint node that is not reporting an error.

24. A method in accordance with claim 23, in which the step of selecting the current upstream node is not performed at times when the data structure corresponding to the prior upstream nodes does not identify any nodes from which the prior upstream node receives input signals.

25. A method in accordance with claim 24, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicating a node includes increasing the suspicion value corresponding to the node in the suspect store; and

the step exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.

26. A method in accordance with claim 23, in which the suspect store in the fault diagnosis system includes a suspicion value corresponding to each node that may be causing errors, and in which:

the step of indicating a node includes increasing the suspicion value corresponding to the nodes in the suspect store; and

the step of exonerating a node includes reducing the suspicion value corresponding to the node in the suspect store.