Hierarchical Network Diagnosis for Fast Root Cause Isolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large-scale networks face challenges in rapidly and efficiently identifying the root cause of performance disruptions, leading to prolonged network disruptions due to the high processing power required for traditional root cause analysis methods.
Innovation Solution
A hierarchical evaluation approach that divides network portions iteratively to identify failing subportions, using a combination of division and symptom-source analysis to pinpoint the root cause, reducing processing requirements to O(log(n)) compared to O(n).
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional root cause analysis methods are used to identify the source of performance disruptions in large-scale networks, then comprehensive analysis can be performed, but the processing power required becomes excessively high and identification time increases
Solution Approach 1:
The network is divided into multiple hierarchical levels or segments, allowing the root cause analysis to be performed on smaller portions rather than the entire network at once. This segmentation reduces the processing complexity from O(n) to O(log(n)) while maintaining identification accuracy through systematic evaluation of divided network sections.
Solution Approach 2:
The approach introduces a hierarchical dimension to the analysis process, organizing network evaluation into multiple levels. This dimensional change allows parallel processing at different hierarchy levels, significantly reducing the time required to identify root causes while maintaining comprehensive coverage of the network.
2Reliability
If traditional root cause analysis methods are used to evaluate the entire network, then complete coverage is achieved, but the processing power required becomes excessively high
Solution Approach 1:
The network monitoring system is segmented into hierarchical levels where only relevant portions are evaluated at each level. This reduces the overall processing power consumption from O(n) to O(log(n)) while maintaining reliable detection of performance issues through systematic evaluation of network segments.
Solution Approach 2:
Instead of performing exhaustive analysis on the entire network simultaneously, the system performs partial analysis on selected network portions at each hierarchical level. This partial action approach reduces processing power consumption while still achieving reliable root cause identification through iterative evaluation.
3Productivity
If the network is divided into smaller portions for evaluation, then processing efficiency improves, but the complexity of managing the division and reevaluation process increases
Solution Approach 1:
The network evaluation process uses a nested hierarchical structure where smaller network portions are nested within larger ones. This nesting approach simplifies the management of divisions by organizing them in a systematic hierarchy, improving identification speed while controlling process complexity through structured reevaluation at each level.
Solution Approach 2:
The evaluation process is made dynamic by adaptively reevaluating only the portions of the network that changed or show symptoms of failure. This dynamic approach improves productivity by avoiding redundant evaluations of stable network portions while managing complexity through intelligent selection of reevaluation targets.
Data Source
AI summary
A system and method provide for identification, diagnosis, and suppression of performance requirement failure in a network. Portions of the network are evaluated to detect a collective failure of the portion to meet a predefined performance requirement. The portion can then be iteratively divided and each subportion similarly evaluated, until a root cause portion of the initially-detected collective failure is identified. Alternatively, upon detection of a failure to meet a predefined performance requirement, a series of cause performance requirements are iteratively evaluated until a root cause performance requirement is identified. In either instance, the root cause can then be suppressed.