Software Failure Influence Range Specification via Propagation Direction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods fail to accurately specify the influence range of a software failure in a network system, as they do not consider the difference in dependency propagation direction between software, leading to incorrect identification of affected devices.
Innovation Solution
A method that determines communication paths between devices with correlated resource usage, generates an influence propagation model based on co-occurrence probabilities from message logs, and traces the propagation direction to accurately identify influenced devices, using system configuration and message log data to specify the influence range of a software failure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the influence range of a software failure is specified without considering the propagation direction difference between software, then the specification process is simple, but the accuracy of identifying affected devices deteriorates
Solution Approach 1:
The patent segments the influence propagation analysis by creating separate influence propagation models for each software type. Each model captures the specific propagation direction characteristics of that software, allowing accurate identification of affected devices without oversimplification. This segmentation resolves the contradiction by maintaining high accuracy through software-specific models while managing complexity through systematic organization.
Solution Approach 2:
The patent applies local quality by tailoring the influence propagation model to each specific software's characteristics. Different software types have different propagation directions, and the patent creates localized models that reflect these differences. This ensures that the accuracy of influence range specification is maintained for each software type while the overall system manages complexity through modular, software-specific approaches.
2Measurement precision
If correlation analysis of resource usage is performed to determine communication paths, then the identification of influenced devices becomes more accurate, but the computational load increases
Solution Approach 1:
The patent performs correlation analysis of resource usage in advance to establish communication paths between devices before actual failure analysis is needed. This preliminary action creates a pre-computed map of device relationships that can be quickly applied when failures occur, reducing the computational load during actual incident response while maintaining high accuracy in identifying influenced devices.
3Measurement precision
If co-occurrence probability analysis of message logs is conducted to generate influence propagation models, then the accuracy of failure impact prediction improves, but the data processing time increases
Solution Approach 1:
The patent conducts co-occurrence probability analysis of message logs in advance to generate influence propagation models before failures occur. This preliminary processing creates ready-to-use models that capture the probabilistic relationships between software failures and device impacts. When actual failures occur, these pre-generated models can be applied immediately, reducing data processing time during critical incident response while maintaining high prediction accuracy.
Data Source
AI summary
A non-transitory computer-readable storage medium storing a program that causes a computer to execute a process, the process includes determining an information transmission path between two devices with a correlation in use status of resources based on system configuration information and use resource information; generating, for each of a plurality of software, an influence propagation model in which a direction from a first device to a second device is set as an influence propagation direction, when a co-occurrence probability that the second device outputs a message within a certain time after the first device outputs a message through execution of the software is equal to or greater than a threshold, based on a message log; acquiring failure information indicating a failed device; specifying an influenced device reached by tracing the information transmission path from the failed device in an influence propagation direction; and outputting information indicating the influenced device.