Model based reasoning system for network fault diagnosis

• US 5,428,619 A
• Filed: 11/14/1994
• Issued: 06/27/1995
• Est. Priority Date: 02/15/1990
• Status: Expired due to Term

• Current Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Original Assignee: Digital Equipment Corporation (HP Inc.)
• Inventors: Schwartz, Steven H., Billmers, Meyer A., Carifo, Michael G., Adler, Mark R.
• Primary Examiner(s): Baker, Stephen M.
• Application Number:
  US08/337,903
• Time in Patent Office: 225 Days
• Field of Search: 371/11.2, 371/15.1, 371/20.1, 395/916
• US Class Current: 714/712
• CPC Class Codes: G06F 11/2257 using expert systems
• CPC Class Codes: H04L 41/06 Management of faults, events, alarms or notifications
• CPC Class Codes: H04L 43/00 Arrangements for monitoring or testing data switching networks
• CPC Class Codes: Y10S 706/916 Electronic or computer, internal or network, circuit

Abstract

A model-based reasoning system for diagnosing faults in a computer network has an information base that stores information about the computer network. This information includes the topology of the network, a class hierarchy containing a component library and diagnostic expertise, and a diagnosis episode. A diagnosis engine receives a pointer to a situation. Each situation includes a focus role and a most general class for the focus role. Based upon this situation, the diagnosis engine creates a diagnosis episode, using the information stored in the information base.

Claims

1. A computer-based method utilizing model-based reasoning for performing fault diagnosis in a network having components and subcomponents, the method comprising the steps of:

providing a component library that stores, as a hierarchical structure, possible components and subcomponents for the network, information related to components and subcomponents, and interrelationships between components and subcomponents;

providing a network topology information base that models the components and subcomponents of the network;

providing a plurality of situations, each situation being an assertion about a single component of the network or a set of components of the network, each situation including an inclination that specifies whether the assertion is to be proved or refuted, each situation optionally including an action that is associated with procedural code to perform a fault diagnostic function, each situation optionally including one or more links to related situations;

receiving a report of a problem in a component of the network;

instigating a root situation representing the problem reported, the root situation being one of the plurality of situations;

setting the inclination of the root situation to prove; and

performing a hierarchical diagnosis to diagnose the problem reported by;

(i) performing the action for the root situation if the root situation has an action that supports the inclination of the root situation,

(ii) if the root situation does not have an action that supports the inclination for the root situation, utilizing the component library to select other situations that will support the inclination of the root situation,

(iii) setting the inclination for each selected situation,

(iv) performing the action for each selected situation if the selected situation has an action that supports the inclination of the selected situation,

(v) if the selected situation does not have an action that supports the inclination for the selected situation, utilizing the component library to select other situations linked to the selected situation that will support the inclination of the selected situation, and

(vi) recursively repeating substeps (iii) to (v) for each selected situation until the fault is diagnosed.

2. The method of claim 1 wherein the component library, the network topology information base and the plurality of situations are stored in a memory device.

3. The method of claim 1 wherein the component library, the network topology information base and the plurality of situations are stored in an information base.

4. The method of claim 1 wherein an action for a situation establishes a truth value for the situation.

5. The method of claim 1 wherein the network is a local area network.

6. The method of claim 5 wherein the local area network comprises components made by more than one manufacturer.

7. The method of claim 1 wherein each one of the plurality of situations includes roles that bind each situation to components and subcomponents stored in the component library.

8. The method of claim 7 further comprising the step of utilizing the network topology information base to bind each role to a relevant component or subcomponent of the network.

9. A computer-based method utilizing model-based reasoning for performing fault diagnosis in a network having components, the method comprising the steps of:

storing in a component library, as a hierarchical structure, possible components for the network, information related to components, and interrelationships between components;

storing a plurality of situations, each situation being an assertion about one or more components of the network, each situation including an inclination that specifies whether the assertion is to be proved or refuted, each situation optionally including an action to perform a diagnostic function, each situation optionally including one or more links to other situations;

receiving a report of a fault in a component of the network;

instigating a root situation representing the problem reported, the root situation being one of the plurality of situations;

creating a diagnosis tree by;

(i) adding the action for the root situation to an agenda,

(ii) if the root situation does not have an action, utilizing the component library to select other situations that will support the inclination of the root situation,

(iii) adding the action for each selected situation to the agenda if the selected situation has an action,

(iv) if the selected situation does not have an action, utilizing the component library to select other situations that will support the inclination of the selected situation, and

(v) recursively repeating substeps (iii) and (iv) for each selected situation;

ranking each action on the agenda according to a preselected scheme;

performing, in turn, each action on the agenda until the fault in the network is identified.

10. The method of claim 9 wherein each action is associated with procedural code that performs a diagnostic function.

11. The method of claim 9 wherein performing each action of a situation establishes a truth value for the situation.

12. The method of claim 9 wherein a component includes a set of subcomponents.

13. The method of claim 9 further comprising the step of setting the inclination of the root situation to prove.

14. The method of claim 9 wherein the component library, the network topology information base and the plurality of situations are stored in a memory device.

15. The method of claim 9 wherein the network is a local area network.

16. The method of claim 9 wherein each one of the plurality of situations includes one or more roles, each role binding each situation to a component stored in the component library.

17. The method of claim 16 further comprising the step of utilizing the network topology information base to bind each role to a relevant component of the network.

18. A computer-based method utilizing model-based reasoning for performing fault diagnosis in a network having components, the method comprising the steps of:

providing a class hierarchy, the class hierarchy including a component library and diagnostic expertise, the component library storing possible components for the network and interrelationships between components;

providing a plurality of situations, each situation being an assertion about one or more components of the network, each situation including an inclination that specifies whether the assertion is to be proved or refuted, each situation optionally including an action to perform a diagnostic function,

receiving a report of a fault in a component of the network;

instigating a root situation representing the problem reported, the root situation being one of the plurality of situations; and

creating a diagnosis tree by;

(i) adding the action for the root situation to an agenda,

(ii) if the root situation does not have an action, utilizing the class hierarchy to select other situations that will support the inclination of the root situation,

(iii) adding the action for each selected situation to the agenda if the selected situation has an action,

(iv) if the selected situation does not have an action, utilizing the class hierarchy to select other situations that will support the inclination of the selected situation,

(v) recursively repeating substeps (iii) and (iv) for each selected situation, and

(vi) performing each action on the agenda to elicit information to prove or refute a selected situation until the fault in the network is identified.

19. The method of claim 18 further comprising the step of modifying the diagnosis tree as a result of each action performed.

20. A computer-implement system for performing fault diagnosis in a computer network having components and subcomponents, the system comprising:

a component library storing, as a hierarchical structure, possible components and subcomponents for the network, information related to components and subcomponents, and interrelationships between components and subcomponents;

a network topology information base modelling the components and subcomponents of the network;

a plurality of situations, each situation being an assertion about a single component of the network or a set of components of the network, each situation including an inclination that specifies whether the assertion is to be proved or refuted, each situation optionally including an action that is associated with procedural code to perform a fault diagnostic function, each situation optionally including one or more links to related situations;

means for receiving a report of a problem in a component of the network;

means for instigating a root situation representing the problem reported, the root situation being one of the plurality of situations; and

means for performing a hierarchical diagnosis, said means including;

(i) means for performing the action for the root situation if the root situation has an action that supports the inclination of the root situation,

(ii) means for utilizing the component library to select other situations linked to the root situation that will support the inclination of the root situation if the root situation does not have an action that supports the inclination for the root situation,

(iii) means for setting the inclination for each selected situation,

(iv) means for performing the action for each selected situation if the selected situation has an action that supports the inclination of the selected situation, and

(v) means for utilizing the component library to select other situations linked to a selected situation that will support the inclination of the selected situation if the selected situation does not have an action that supports the inclination for the selected situation.

21. The system of claim 20 wherein the component library, the network topology information base and the plurality of situations are stored in a memory device.

22. The system of claim 20 wherein the network is a local area network.

23. The system of claim 20 wherein each one of the plurality of situations includes roles that bind each situation to components and subcomponents stored in the component library.

24. The system of claim 23 further comprising means for utilizing the network topology information base to bind each role to a relevant component or subcomponent of the network.

25. A computer-implemented system for performing fault diagnosis in a network having components, the system comprising:

a component library storing, as a hierarchical structure, possible components for the network and interrelationships between components;

a plurality of situations, each situation being an assertion about one or more components of the network, each situation including an inclination that specifies whether the assertion is to be proved or refuted, each situation optionally including an action to perform a diagnostic function, each situation optionally including one or more links to related situations;

means for receiving a report of a fault in a component of the network;

means for instigating a root situation representing the problem reported, the root situation being one of the plurality of situations;

means for creating a diagnosis tree, said means including

means for adding the action for the root situation to an agenda,

means for utilizing the component library to select other situations that will support the inclination of the root situation if the root situation does not have an action,

means for adding the action for each selected situation to the agenda if the selected situation has an action, and

means for utilizing the component library to select other situations that will support the inclination of the selected situation if the selected situation does not have an action;

means for ranking each action on the agenda according to a preselected scheme; and

means for performing, in turn, each action on the agenda until the fault in the network is identified.

26. The system of claim 25 wherein each action is associated with procedural code that performs a diagnostic function.

27. The system of claim 25 wherein performing each action of a situation establishes a truth value for the situation.

28. The system of claim 25 wherein a component includes a set of subcomponents.

29. The system of claim 25 further comprising means for setting the inclination of the root situation to prove.

30. The system of claim 25 wherein the component library, the network topology information base and the plurality of situations are stored in a memory device.

31. The system of claim 25 wherein each one of the plurality of situations includes one or more roles, each role binding each situation to a component stored in the component library.

32. The system of claim 31 further comprising means for utilizing the network topology information base to bind each role to a relevant component of the network.

33. A computer-implemented system utilizing a model-based reasoning for performing fault diagnosis in a computer network having components, the system comprising:

a class hierarchy including a component library and diagnostic expertise, the component library storing possible components for the network and interrelationships between components;

a plurality of situations, each situation being an assertion about one or more components of the network, each situation including an inclination that specifies whether the assertion is to be proved or refuted, each situation optionally including an action to perform a diagnostic function, each situation optionally including one or more links to related situations;

means for receiving a report of a fault in a component of the network;

means for instigating a root situation representing the problem reported, the root situation being one of the plurality of situations; and

means for creating a diagnosis tree, said means including;

means for adding the action for the root situation to an agenda,

means for utilizing the class hierarchy to select other situations that will support the inclination of the root situation if the root situation does not have an action,

means for adding the action for each selected situation to the agenda if the selected situation has an action,

means for utilizing the class hierarchy to select other situations that will support the inclination of the selected situation if the selected situation does not have an action, and

means for performing each action on the agenda to elicit information to prove or refute a situation until the fault in the network is identified.

34. The system of claim 33 further comprising means for modifying the diagnosis tree as a result of each action performed.

35. The system of claim 33 wherein the class hierarchy includes causal branches associated with classes in the class hierarchy.

36. The system of claim 35 wherein the class hierarchy is a multiple-inheritance class hierarchy.

37. The system of claim 36 wherein at least one situation has an associated action that establishes a truth of that situation or identifies a role player in the situation.

38. The system of claim 37 wherein at least one causal branch links two situations, said causal branch representing one of a plurality of causal relationships.