Method and apparatus for failure classification

Patent metadata

Field	Value
Publication number	US-10402249-B2
Application number	US-201715460682-A
Country	US
Kind code	B2
Filing date	Mar 16, 2017
Priority date	Mar 16, 2016
Publication date	Sep 3, 2019
Grant date	Sep 3, 2019

Abstract

Embodiments of the present disclosure provide a method and apparatus for classifying a failure. The method comprises obtaining a log associated with a failure and identifying a key event in the log, and the key event indicates a predetermined situation. The method further comprises determining a similarity between the failure and each of a set of historical failures based on the key event, and the set of historical failures have been classified into at least one category. The method further comprises classifying the failure based at least partly on the similarity. Embodiments of the present disclosure enable classifying the failure fast and accurately by determining the similarity associated with key event between failures.

First claim

We claim:

1. A method for classifying a failure, comprising:
obtaining a log associated with a failure;
identifying a key event in the log, the key event indicating a predetermined situation;
determining a similarity between the failure and each of a set of historical failures based on the key event, the set of historical failures having been classified into at least one category; and
classifying the failure based at least partly on the similarity, the classifying of the failure including (i) determining a subset of the set of historical failures based on the similarity, (ii) determining another similarity associated with a predetermined failure characteristic between the failure and each of the subset of historical failures, the predetermined failure characteristic including at least one of a recency value, a state, a release number, a version number, and a duplication value, and (iii) classifying the failure based on the similarity and the other similarity.

2. The method according to claim 1 , wherein the identifying a key event in the log comprises:
determining whether at least one of the one or more predetermined key events exists in the log; and
in response to determining the at least one predetermined key event exists in the log, determining the at least one predetermined key event as the key event in the log.

3. The method according to claim 1 , wherein the determining a similarity between the failure and each of a set of historical failures comprises:
determining, based on an event status list indicating whether one or more predetermined key events occurs, the similarity between the failure and each of the set of historical failures.

4. The method according to claim 1 , wherein the determining a similarity between the failure and each of a set of historical failures comprises:
determining a first event status list of the failure according to one or more predetermined key events;
determining a second event status list of each of the set of historical failures according to the one or more predetermined key events; and
determining the similarity between the failure and each of the historical failures by comparing the first and second event status lists, the first and second event status lists indicating whether the one or more predetermined key event occurs.

5. The method according to claim 4 , wherein the determining a similarity between the failure and each of a set of historical failures further comprises:
setting respective weight for individual key event in the one or more predetermined key events; and
determining the similarity between the failure and each of the historical failures based at least partly on the respective weight.

6. The method according to claim 1 , wherein the classifying the failure comprises:
weighting the similarity and, the other similarity by a first weight and a second weight respectively;
determining a weighted similarity between the failure and each of the subset of historical failures based on the similarity, the other similarity and the first and second weights; and
classifying the failure based on the weighted similarity.

7. The method according to claim 1 , wherein the determining another similarity comprises at least one of the following:
determining a first similarity associated with the recency value by comparing timestamp information of the failure and each of the subset of historical failures;
determining a second similarity associated with the state by determining a current state of each of the subset of historical failures;
determining a third similarity associated with the release number by comparing release numbers associated with failed applications in which the failure and each of the subset of historical failures occur respectively;
determining a fourth similarity associated with the version number by comparing version numbers associated with failed applications in which the failure and each of the subset of historical failures occur respectively; and
determining a fifth similarity associated with the duplication value by determining a number of times that the failure repeats in the set of historical failures.

8. The method according to claim 6 , wherein the classifying the failure comprises:
selecting, from the subset of historical failures, a predetermined number of historical failures based on the weighted similarity;
analyzing statistically categories of the predetermined number of historical failures; and
selecting, from the categories, a category to which the failure is to be classified.

9. A system, comprising:
a data storage system; and
computer-executable program logic encoded in memory of one or more computers enabled for classifying a failure, wherein the computer-executable program logic is configured for the execution of:
obtaining a log associated with a failure;
identifying a key event in the log, the key event indicating a predetermined situation;
determining a similarity between the failure and each of a set of historical failures based on the key event, the set of historical failures having been classified into at least one category; and
classifying the failure based at least partly on the similarity, the classifying of the failure including (i) determining a subset of the set of historical failures based on the similarity, (ii) determining another similarity associated with a predetermined failure characteristic between the failure and each of the subset of historical failures, the predetermined failure characteristic including at least one of a recency value, a state, a release number, a version number, and a duplication value, and (iii) classifying the failure based on the similarity and the other similarity.

10. The system of claim 9 , wherein the identifying a key event in the log comprises:
determining whether at least one of the one or more predetermined key events exists in the log; and
in response to determining the at least one predetermined key event exists in the log, determining the at least one predetermined key event as the key event in the log.

11. The system of claim 9 , wherein the determining a similarity between the failure and each of a set of historical failures comprises:
determining, based on an event status list indicating whether one or more predetermined key events occurs, the similarity between the failure and each of the set of historical failures.

12. The system of claim 9 , wherein the determining a similarity between the failure and each of a set of historical failures comprises:
determining a first event status list of the failure according to one or more predetermined key events;
determining a second event status list of each of the set of historical failures according to the one or more predetermined key events; and
determining the similarity between the failure and each of the historical failures by comparing the first and second event status lists, the first and second event status lists indicating whether the one or more predetermined key event occurs.

13. The system of claim 12 , wherein the determining a similarity between the failure and each of a set of historical failures further comprises:
setting respective weight for individual key event in the one or more predetermined key events; and
determining the similarity between the failure and each of the historical failures based, at least partly on the respective weight.

14. The system of claim 9 , wherein the classifying the failure comprises:
weighting the similarity and the other similarity by a first weight and a second weight respectively;
determining a weighted similarity between the failure and each of the subset

Assignees

• Emc Ip Holding Co Llc

Inventors

• Xiao Huibing
• Gao Jian
• Gao Hongpo
• Han Geng
• Dong Jibing
• Gong Shaoqin

Classifications

• G06F16/24578

  using ranking · CPC title

• G06F16/285

  Clustering or classification · CPC title

• G06F11/3072 Primary

  where the reporting involves data filtering, e.g. pattern matching, time or event triggered, adaptive or policy-based reporting · CPC title

• G06F11/0766 Primary

  Error or fault reporting or storing · CPC title

• G06F11/3476

  Data logging (G06F11/14, G06F11/2205 take precedence) · CPC title