Method for isolating a fault from error messages

• US 7,124,060 B1
• Filed: 11/03/2000
• Issued: 10/17/2006
• Est. Priority Date: 11/03/1999
• Status: Expired due to Fees

• Current Assignee: ABB Schweiz AG (ABB Limited)
• Original Assignee: ABB AB (ABB Limited)
• Inventors: Larsson, Magnus, Eriksson, Peter
• Primary Examiner(s): WACHSMAN, HAL D
• Application Number:
  US10/129,188
• Time in Patent Office: 2,174 Days
• Field of Search: 702/185, 702/58, 702/59, 702/122, 702/125, 702/176, 702/177, 702/183, 702/182, 702/184, 702/186, 702/187, 702/188, 714 45- 49, 714/57, 714/25, 714/26, 714/37, 714/2
• US Class Current: 702/185
• CPC Class Codes: G06F 11/0718 in an object-oriented system
• CPC Class Codes: G06F 11/0766 Error or fault reporting or storing
• CPC Class Codes: G06F 11/079 Root cause analysis, i.e. error or fault diagnosis in a hardware test environment G06F11/22; in a software test environment G06F11/36

Abstract

A method, a use of the method, a system, a use of the system, a computer program code element, and a computer readable medium for automatically isolating primary faults out from a system log including actual error messages in a system controlled by an object oriented program. Messages are isolated through clustering.

Claims

1. A method for isolating a fault from a plurality of error messages in a system log of a process controlled by an object oriented program, the method comprising:

creating a plurality of predefined signatures comprising a predefined signature for each error message of said plurality of error messages;

forming a time period to frame from said plurality of error messages the error messages belonging to a fault scenario of said fault;

clustering said error messages that fall within said time period;

forming from said predefined signatures of said clustered error messages a base graph, which is a formal representation of said fault scenario, carrying cause-effect relation information of said clustered error messages;

forming an explanation model from an underlying assumption of communication between system entities needed to carry out a specific task, the system entities comprising objects, packages and threads, and having said predefined signatures defined on a base of said explanation model;

analyzing said base graph to isolate said fault; and

reporting information concerning said fault.

2. The method according to claim 1, wherein each of said signatures is brought to contain information on a complaint and a complainer.

3. The method according to claim 1, wherein said reported information comprises a reliability estimate of said fault.

4. The method according to claim 1, wherein the information concerning the fault is reported to a control means, which provides measures for resolving the fault.

5. The method according to claim 1, wherein the process comprises operation of an industrial robot.

6. A method for isolating a fault from a plurality of error messages in a system log of a process controlled by an object oriented program, the method comprising:

creating a plurality of predefined signatures comprising a predefined signature for each error message of said plurality of error messages;

forming a time period to frame from said plurality of error messages the error messages belonging to a fault scenario of said fault, wherein said time period starts with the first error message in said fault scenario and ends with a stop message;

clustering said error messages that fall within said time period;

forming from said predefined signatures of said clustered error messages a base graph, which is a formal representation of said fault scenario, carrying cause-effect relation information of said clustered error messages;

analyzing said base graph to isolate said fault; and

reporting information concerning said fault.

7. A method for isolating a fault from a plurality of error messages in a system log of a process controlled by an object oriented program the method comprising:

creating a plurality of predefined signatures comprising a predefined signature for each error message of said plurality of error messages;

forming a time period to frame from said plurality of error messages the error messages belonging to a fault scenario of said fault;

clustering said error messages that fall within said time period;

forming from said predefined signatures of said clustered error messages a base graph, which is a formal representation of said fault scenario, carrying cause-effect relation information of said clustered error messages;

analyzing said base graph to isolate said fault, wherein if the base graph is inconclusive said analyzing further comprises extending said base graph with the help of a system model including information on the dependency relation between system entities, and analyzing said extended base graph with a set of predetermined explanation rules to isolate said fault; and

reporting information concerning said fault.

8. A system for isolating a fault from a plurality of error messages in a system log of a process controlled by an object oriented program, the system comprising:

a plurality of predefined signatures comprising a predefined signature for each error message of said plurality of error messages, wherein said predefined signatures comprise information on complaint, complainer and complainee, and wherein said predefined signatures are formed from an explanation model comprising an underlying assumption of communication between system entities needed to carry out a specific task, said system entities comprise objects, packages and threads; and

an interface located between the object oriented program and a means for process control, said interface comprising

means for clustering error messages belonging to a fault scenario of said fault,

means for forming from said signature of said clustered error messages a formal representation of said fault scenario,

means for analyzing said formal representation to isolate said fault, and

means for reporting information concerning said fault.

9. The system according to claim 8, wherein said means for reporting reports the information concerning the fault to said means for process control, which provides measures for resolving the fault.

10. The system according to claim 8, wherein said formal representation comprises a base graph.

11. The system according to claim 8, further comprising:

means for establishing a reliability estimate of said fault, wherein said fault information contains said reliability estimate.

12. The system according to claim 8, wherein the process is carried out by an industrial robot.

13. A system for isolating a fault from a plurality of error messages in a system log of a process controlled by an object oriented program, the system comprising:

a plurality of predefined signatures comprising a predefined signature for each error message of said plurality of error messages;

a time period starts with a first of said plurality of error messages in said fault scenario and ends with a stop message; and

an interface located between the object oriented program and a means for process control, said interface comprising

means for clustering error messages belonging to a fault scenario of said fault,

means for forming from said signature of said clustered error messages a formal representation of said fault scenario,

means for analyzing said formal representation to isolate said fault, and

means for reporting information concerning said fault.

14. A system for isolating a fault from a plurality of error messages in a system log of a process controlled by an object oriented program, the system comprising:

a plurality of predefined signatures comprising a predefined signature for each error message of said plurality of error messages; and

an interface located between the object oriented program and a means for process control, said interface comprising

means for clustering error messages belonging to a fault scenario of said fault,

means for forming from said signature of said clustered error messages a formal representation of said fault scenario, wherein said formal representation comprises a base graph,

means for analyzing said formal representation to isolate said fault, wherein if said base graph is inconclusive the system further comprises means for extending said base graph with the help of a system model containing information on a dependency relation between system entities and means for isolating said fault from said extended base graph by using predetermined explanation rules, and

means for reporting information concerning said fault.