Method and apparatus for isolating faults in a network having serially connected links

• US 5,299,201 A
• Filed: 08/26/1993
• Issued: 03/29/1994
• Est. Priority Date: 09/04/1990
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Tendolkar, Nandakamar N., Moffitt, Gerald T., Carusone, Anthony Jr., Garrigan, Albert W., Hunsinger, Wayne, Taylor, Jordan M.
• Primary Examiner(s): Beausoliel, Jr., Robert W.
• Assistant Examiner(s): Snyder, Glenn
• Application Number:
  US08/112,173
• Time in Patent Office: 215 Days
• Field of Search: 371/5.1, 371/11.2, 371/20.1
• US Class Current: 714/704
• CPC Class Codes: H04L 1/20 using signal quality detector

Abstract

A fault isolation system for use in a network in which bit errors resulting in detected code violation errors or disparity errors may be propagated from a first link to a second link of a network in which multiple serially coupled links may couple two devices. Errors in excess of a first threshold are detected at the receiving end of the first link, while errors in excess of a second threshold greater than the first threshold are detected at the receiving end of the second link. Errors are attributed to the first link in response to the detection of errors in excess of the thresholds on both links and are attributed to the link for which the threshold was exceeded in response to the detection of errors in excess of the threshold on only one of the links. If a substantial number of the bit errors resulting in the threshold being exceeded on the second link originated on the first link, the first threshold will also be exceeded with a high degree of probability. This prevents the erroneous identification of the second link as being the source of the fault where the fault actually originated from a condition on the first link.

Claims

1. In a system having a plurality of processing units and at least one control unit and at least one intermediate unit for coupling the other units via fiber optic links between each of the units, and wherein

each of said units has a fiber optic link adapter interface which may be actuated to form either a static or dynamic connection between the link adapter interfaces at the ends of each link units for establishing bidirectional communication between the units over the links, and wherein a processor or a control unit can function either as an originating unit or a destination unit in bidirectional communication between units, and in which an originating unit transmits data to a destination unit via an intermediate unit, said originating unit being coupled to said intermediate unit via a first link, said intermediate unit being coupled to said destination unit via a second link, said intermediate unit propagating at least some errors originating on said first link to said second link, with either said first link or said second link possibly having a fault and where faults generate error fault signals which may originate from several link adapters of the system from the plural links of the system simultaneously, and where errors occurring on a source link rating below a threshold are propagated to the destination link such that from the standpoint of the destination unit they are indistinguishable from errors arising on the destination link, and wherein said system has a fault analyzer for receiving a fault signal originating from said link adapters for isolating a link on which a fault originates, said fault analyzer providing a correct attribution of errors to a failing link when errors may occur either on a source link or a destination link without mistaking detected threshold crossings caused by cumulative errors occurring over a link, by a method of isolating errors originating on said links including the steps of:

(a) detecting the occurrence of errors on said first link in excess of a first threshold;

(b) detecting the occurrence of errors on said second link in excess of a second threshold greater than said first threshold; and

(c) attributing said errors to said links on the basis of the detection of said errors in excess of said thresholds, by

(1) attributing errors to said first link when errors have been detected in both of said links in excess of the thresholds on both of said links, and

(2) attributing errors to the link having an error detected when errors have been detected in only one of said links in excess of the threshold for the link.

2. A method as in claim 1 in which said errors are attributed by said fault analyzer to said first link in response to the detection of errors originating from the adapter interface for said first link in excess of the thresholds on both of said links.

3. A method as in claim 1 in which said errors are attributed by said fault analyzer to the link for which the threshold was exceeded in response to the detection of errors originating from the adapter interface for said first link in excess of the threshold on only one of said links.

4. A method as in claim 1 in which said transmitted data consists of a stream of characters from which certain characters are disallowed, each of said detecting steps including the step of detecting the occurrence of a disallowed character on the corresponding link.

5. A method as in claim 1 in which said transmitted data consists of a stream of characters having respective disparities associated therewith, each of said detecting steps including the step of determining the cumulative disparity of the characters transmitted on the corresponding link.

6. A method as in claim 1 in which each of said detecting steps includes the step of counting the number of errors occurring within a predetermined period.

7. A method as in claim 6 in which each of said detecting steps includes the step of counting the number of intervals of predetermined duration shorter than the corresponding period within which one or more errors occur.

8. A method as in claim 1 in which said first step includes the step of maintaining a running count of the number of errors occurring on said first link within a predetermined period of the present.

9. A method as in claim 1 in which each of said links has a source end nearer to said source unit and a destination end nearer to said destination unit, said errors being detected at the destination ends of said links.

10. A method as in claim 1 in which said system contains at least one processing unit coupled to a device control unit via a switching unit, said processing unit and said switching unit and said switching unit and said control unit being interconnected by respective links.

11. In a system for bidirectional communication among digital units, comprising:

a plurality of processing units and at least one control unit and at least one intermediate unit for coupling the other units via fiber optic links between each of the units,

each of said units having a fiber optic link adapter interface which may be actuated to form either a static or dynamic connection between the link adapter interfaces at the ends of each link units for establishing bidirectional communication between the units over the links, and wherein a processor or a control unit can function either as an originating unit or a destination unit in bidirectional communication between units, and in which an originating unit transmits data to a destination unit via an intermediate unit, said originating unit being coupled to said intermediate unit via a first link, said intermediate unit being coupled to said destination unit via a second link, said intermediate unit propagating at least some errors originating on said first link to said second link, with either said first link or said second link possibly having a fault and where faults generate error fault signals which may originate from several link adapters of the system from the plural links of the system simultaneously, and where errors occurring on a source link rating below a threshold are propagated to the destination link such that from the standpoint of the destination unit they are indistinguishable from errors arising on the destination link, and wherein said system has a fault analyzer for receiving a fault signal originating form said link adapters for isolating a link on which a fault originates, said fault analyzer providing a correct attribution of errors to a failing link when errors may occur either on a source link or a destination link without mistaking detected threshold crossings caused by cumulative errors occurring over a link, said system including for isolating errors originating on said links;

(a) means for detecting the occurrence of errors on said first link in excess of a first threshold;

(b) means for detecting the occurrence of errors on said second link in excess of a second threshold greater than said first threshold; and

(c) means responsive to said detecting means for attributing said errors to said links after the detection of said errors in excess of said thresholds, by

(1) attributing errors to said first link when errors have been detected in both of said links in excess of the thresholds on both of said links, and

(b

     2) attributing errors to the link having an error detected when errors have been detected in only one of said links in excess of the threshold for the link.

12. Apparatus as in claim 11 in which each of said detecting means forms part of said link adapter interface which includes means counting the number of errors occurring within a predetermined period.

13. Apparatus as in claim 12 in which each of said detecting means includes means for counting the number of intervals of predetermined duration shorter than the corresponding period within which one or more errors occur.

14. Apparatus as in claim 11 in which each of said links has a source end nearer to said source unit and a destination end nearer to said destination unit, said detecting means being located at the destination ends of said links.

15. Apparatus as in claim 11 in which said system contains at least one processing unit coupled to a device control unit via a switching unit, said processing unit and said switching unit and said switching unit and said control unit being interconnected by respective links.

16. In a system having a plurality of processing units and at least one control unit and at least one intermediate unit coupling the other units via an interconnection switch of the intermediate unit and fiber optic links between each of the units, and wherein each of said units has a fiber optic link adapter interface for coupling the unit to a link, and wherein said intermediate unit includes an interconnection switch which may be actuated to from either a static or dynamic connection between the link adapter interfaces at the ends of each link units for establishing bidirectional communication between the units over the links, and in which an originating unit transmits data a destination unit via an intermediate unit, said originating unit being coupled to said intermediate unit via a first link, said intermediate unit being coupled to said destination unit via a second link, said intermediate unit propagating at least some errors originating on said first link to said second link, a method of isolating errors originating on said links including the steps of:

(a) detecting the occurrence of errors on said first link in excess of a first threshold;

(b) detecting the occurrence of errors on said second link in excess of a second threshold;

(c) attributing said errors to said first link in response to the detecting of errors in excess of the thresholds on both of said links; and

(d) attributing said errors to the link for which the threshold was exceeded in response to the detecting of errors in excess of the threshold on only one of said links.

17. A method as in claim 16 in which said transmitted data consists of a stream of characters from which certain characters are disallowed, each of said detecting steps including the step of detecting the occurrence of a disallowed character on the corresponding link.

18. A method as in claim 16 in which said transmitted data consists of a stream of characters having respective disparities associated therewith, each of said detecting steps including the step of determining the cumulative disparity of the characters transmitted on the corresponding link.

19. A method as in claim 16 in which each of said detecting steps includes the step of counting the number of errors occurring within a predetermined period.

20. A method as in claim 19 in which each of said detecting steps includes the step of counting the number of intervals of predetermined duration shorter than the corresponding period within which one or more errors occur.