Monitoring System Root Cause Analysis Time Width Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large-scale information processing systems face challenges in accurately analyzing the root cause of events due to varying event occurrence periods across different apparatuses, leading to increased analysis time widths that can result in noise inclusion and reduced accuracy, and the risk of missing important cause candidates with lower certainty.
Innovation Solution
A monitoring system performs cause analysis based on a rule, displaying change in certainty, sorting by an index value showing affected range, and calculating analysis time width based on this index value to improve accuracy and reduce noise inclusion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the analysis time width is increased to absorb the difference of occurring period of events in large-scale systems, then the coverage of event detection is improved, but the number of noise events included increases and analysis accuracy deteriorates
Solution Approach 1:
The patent segments the analysis process by dividing events into different time width groups based on their occurring periods. Instead of using a single uniform analysis time width for all events, the system categorizes events into multiple groups with different time widths, allowing each group to be analyzed with an appropriate time window that matches its characteristic occurrence period, thus avoiding noise inclusion while maintaining event detection coverage
Solution Approach 2:
The patent implements dynamic adjustment of analysis time width based on event characteristics. The system automatically determines appropriate analysis time widths for different events based on their occurring periods and other attributes, rather than using a fixed time width. This dynamic adaptation allows the system to optimize the balance between coverage and accuracy for each specific event type
2Measurement precision
If the analysis time width is decreased to reduce noise inclusion, then the analysis accuracy is improved, but the coverage of event detection is reduced and important causes may be missed
Solution Approach 1:
The patent applies local quality by assigning different analysis time widths to different events based on their specific characteristics and occurring periods. Each event is analyzed with a locally optimized time width that is appropriate for its nature, rather than applying a global uniform time width. This allows high accuracy for events with short occurring periods while maintaining adequate coverage for events with longer periods
3Ease of operation
If cause candidates are sorted and displayed only by certainty in descending order, then the display simplicity is improved, but the importance of cause candidates with lower certainty but large affected range is lost
Solution Approach 1:
The patent adds another dimension to the display by introducing multiple sorting criteria beyond just certainty. The system can sort cause candidates by multiple attributes including affected range, occurrence frequency, and time width, allowing users to view cause candidates from different perspectives. This multi-dimensional display approach preserves important information about cause candidates that would be hidden in a single-dimension sorted list
Data Source
AI summary
A monitoring system performs cause analysis of an event occurring in any of a plurality of monitoring-target objects to be monitored based on a rule. In this case, the monitoring system makes a detection during the analysis time width and determines a plurality of conclusions based on an event corresponding to the condition for determining the conclusion. Moreover, the monitoring system performs one or more of (A) displaying change of certainty to be used for determination of a conclusion in a case where the analysis time width is assumed to be changed, (B) performing sort display of the determined conclusion based on an index value showing an affected range and (C) calculating the analysis time width based on the index value.