Crash Dump Similarity via Simplified Stack Frames

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for determining similarity between crash dumps in computing devices are inefficient due to noise and complexity, leading to increased computational costs and loss of contextual information, making it difficult to identify similar failures and avoid repetitive analysis.

Innovation Solution

A method that generates simplified stack frames based on function and module name information, removing noise and preserving the original sequence, allowing for efficient similarity determination between crash dumps using a mapping table and Hash algorithm.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a complete text of stacks for crash dump is used for similarity determination, then individual words can be taken into account, but much noise information is introduced and calling order information is lost

Engineering Contradiction:
Improvesimilarity determination accuracyVSAvoidcalling order information
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent extracts only the essential function name information from stack frames, removing noise information such as variable parts of stack text, addresses, and other non-essential details. This extraction process preserves the calling order information while eliminating elements that contribute to noise in similarity determination.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the stack text into individual function name elements while maintaining their sequential order. By dividing the complete stack text into discrete function name components, the system can process and compare stacks efficiently without losing the calling sequence information, resolving the contradiction between noise reduction and information preservation.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If a customized full-text search engine is constructed to determine similarity between crash dumps, then individual words can be analyzed, but the device complexity increases significantly

Engineering Contradiction:
Improvesimilarity determination accuracyVSAvoidsearch engine complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

Instead of constructing a complex customized full-text search engine, the patent creates a simplified copy or representation of stack information by extracting only function names. This copying approach maintains the essential information needed for similarity determination while avoiding the complexity of a full-text search engine infrastructure.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The patent changes the parameter representation from complete stack text to extracted function name sequences. This parameter transformation simplifies the data structure and eliminates the need for complex search engine mechanisms, reducing device complexity while preserving similarity determination capability.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If complete stack text is compared with all existing stack files to determine similarity, then comprehensive analysis can be performed, but computational cost increases significantly

Engineering Contradiction:
Improvesimilarity determination accuracyVSAvoidcomputational cost
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent extracts only the essential function name information from stack frames, removing noise information such as variable parts of stack text, addresses, and other non-essential details. This extraction process preserves the calling order information while eliminating elements that contribute to noise in similarity determination.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the stack text into individual function name elements while maintaining their sequential order. By dividing the complete stack text into discrete function name components, the system can process and compare stacks efficiently without losing the calling sequence information, resolving the contradiction between noise reduction and information preservation.

Inventive Principle:
Principle #1Segmentation

4Measurement precision

If a customized full-text search engine is constructed for similarity determination, then individual words can be analyzed, but implementation difficulty increases

Engineering Contradiction:
Improvesimilarity determination accuracyVSAvoidimplementation ease
Core Design Contradiction:
Measurement precisionVSEase of manufacture

Solution Approach 1:

Instead of constructing a complex customized full-text search engine, the patent creates a simplified copy or representation of stack information by extracting only function names. This copying approach maintains the essential information needed for similarity determination while avoiding the complexity of a full-text search engine infrastructure.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The patent changes the parameter representation from complete stack text to extracted function name sequences. This parameter transformation simplifies the data structure and eliminates the need for complex search engine mechanisms, reducing device complexity while preserving similarity determination capability.

Inventive Principle:
Principle #35Parameter changes

5Loss of information

If complete stack text is used for similarity determination, then comprehensive context information is available, but the analysis results cannot provide simple understanding of failure-related problems

Engineering Contradiction:
Improvecontext informationVSAvoidresult understandability
Core Design Contradiction:
Loss of informationVSEase of operation

Solution Approach 1:

The patent extracts only the essential function name information from stack frames, removing noise information such as variable parts of stack text, addresses, and other non-essential details. This extraction process preserves the calling order information while eliminating elements that contribute to noise in similarity determination.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the stack text into individual function name elements while maintaining their sequential order. By dividing the complete stack text into discrete function name components, the system can process and compare stacks efficiently without losing the calling sequence information, resolving the contradiction between noise reduction and information preservation.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS10235234B2Method and apparatus for determining failure similarity in computing device
Publication Date: 2019.03.19 EMC IP HLDG CO LLC
  • US10235234B2 patent drawing
  • US10235234B2 patent drawing
  • US10235234B2 patent drawing

AI summary

Embodiments of the present disclosure provide a method and an apparatus for a computing device. The computing device may generate stacks for crash dump in response to failures, each of the stacks may include a plurality of stack frames from bottom to top, and each of the stack frames may include function information associated with a corresponding failure. The method may include: extracting corresponding function name information from the stack frames in the stacks; generating simplified stack frames based on the corresponding function name information to obtain simplified stacks for the stacks; and determining a similarity between the failures based on a similarity between the simplified stacks of the failures.