Apparatus and method for isolating and predicting errors in a local area network

• US 5,036,514 A
• Filed: 11/09/1989
• Issued: 07/30/1991
• Est. Priority Date: 11/09/1989
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Smith, Wilson E., Suciu, Ronald E., Wilson, Kenneth T., Downes, Robert W.
• Primary Examiner(s): Baker, Stephen M.
• Application Number:
  US07/433,930
• Time in Patent Office: 628 Days
• Field of Search: 371/5.1, 371/5.5, 371/11.2, 371/20.1, 371/20.2, 371/20.6, 371/4, 371/29.1, 364/200 MS File
• US Class Current: 714/704
• CPC Class Codes: H04L 43/0823 Errors, e.g. transmission errors
• CPC Class Codes: H04L 43/16 Threshold monitoring

Abstract

A method for correcting error weights in a communication system in which at least one error condition can result from more than one cause. Stations on the system send error reports to an error monitor station which maintains a count of the number of stations reporting an error having multiple causes and periodically corrects the error weights assigned to each station as a function of the maximum number of errors reported by any station only when the number of stations reporting a multiple cause error exceeds a predetermined value.

Claims

1. In a communication network including a plurality of data terminals, each of which detects one or more error conditions, at least one of which can result from more than one cause, and reports the identity and number of the detected errors to at least one error monitor station which accumulates weighted error counts for at least some of the reporting stations, a method for adjusting the weighted errors as a function of a cause of the at least one error having more than one cause comprising the steps at the error monitor station of:

storing a first limit and a plurality of scalar factors, one for each of the error conditions reported by the stations;

counting the number of stations within a predetermined time period reporting errors related to an error condition which can have more than one cause;

recording the largest reported error count which can have more than one cause;

comparing said reporting station count to said limit;

generating a correction factor when the count exceeds the limit by combining the largest reported error count and the scalar factor related to the error; and

applying the correction factor to the weighted error for each reporting station for which weighted error counts are maintained.

2. The method set forth in claim 1 in which the count of the number of stations reporting errors related to an error condition which can have more than one cause is not allowed to exceed said first limit plus one.

3. The method set forth in claim 1 in which the number of errors related to an error condition which can have more than one cause is limited to not exceed a predetermined value.

4. The method set forth in any one of claims 1-3 in which the correction factor is the product of the largest reported error and the scalar factor and the correction factor is used to reduce the weighted error for all reporting stations for which weighted errors are maintained.

5. In a communication network including a plurality of data terminals, each of which detects one or more error conditions, at least one of which can result from more than one cause, and reports the identity and number of the detected errors to at least one error monitor station which accumulates weighted error counts for at least some of the reporting stations, a method of adjusting the weighted errors as a function of a cause of the at least one error having more than one cause comprising the steps at the error station of:

examining each of the received error reports during a time period to determine if it includes error counts which can result from more than one cause and pre-filtering the error count related thereto;

weighting and summing the pre-filtered and unfiltered errors;

counting the number of stations reporting an error which can result from more than one cause; and

at the expiration of the said time period, filtering the weighted errors of the reporting stations when the count of stations reporting an error which can result from more than one cause exceeds a predetermined value.

6. The method set forth in claim 5 in which the filtering step includes modification of the weighted errors as a function of the maximum number of errors which can result from more than one cause reported by any station.

7. The method set forth in claim 6 in which the maximum number of errors is not allowed to exceed a predetermined value and is multiplied by a predetermined scalar and then subtracted from the weighted error counts.