Root Cause Candidate Determination in Process Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Determining the root cause of problems in complex data processing environments with multiple interrelated processes is challenging due to the complexity of dependencies and runtime metric values.
Innovation Solution
A method that uses a directed graph to identify forward and reverse dependencies between processes and analyzes runtime metric values to determine root cause candidates by tracing paths through the graph and identifying processes with non-desired metric values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual analysis of process dependencies and runtime metrics is used to identify root causes, then accuracy can be maintained through expert judgment, but the time required to identify root causes increases significantly
Solution Approach 1:
The patent segments the complex problem of root cause identification into distinct components: (1) constructing a directed graph representing process dependencies, (2) collecting runtime metric values from multiple processes, (3) analyzing the graph to identify affected processes, and (4) correlating metrics with affected processes to determine root causes. This segmentation allows automated processing of each component while maintaining overall accuracy.
Solution Approach 2:
The patent introduces an intermediary automated analysis system that acts as a mediator between the complex process environment and the root cause identification goal. This intermediary automatically processes dependency graphs and runtime metrics, performing the complex analysis work that would be time-consuming for manual expert judgment while maintaining consistency and accuracy.
2Loss of time
If automated analysis of all processes and dependencies is implemented to speed up root cause identification, then time loss is reduced, but the complexity of the analysis system increases
Solution Approach 1:
The patent applies preliminary action by constructing the directed graph of process dependencies in advance, before root cause analysis is needed. This graph structure pre-organizes all process relationships, so that when a root cause analysis is triggered, the system can quickly query pre-established relationships rather than analyzing all dependencies from scratch. Runtime metrics are also collected and prepared in advance for rapid analysis.
Solution Approach 2:
The analysis system is segmented into distinct functional modules: a graph construction module that builds dependency representations, a metric collection module that gathers runtime data, an analysis module that identifies affected processes, and a root cause determination module that correlates findings. This modular segmentation reduces overall system complexity by making each component's responsibility clear and manageable.
3Measurement precision
If comprehensive monitoring of all runtime metric values is performed to ensure accurate root cause identification, then measurement precision is improved, but the amount of data to be processed increases
Solution Approach 1:
The patent extracts only the relevant runtime metric values needed for root cause analysis from the comprehensive set of all process data. After identifying affected processes through graph analysis, the system selectively extracts and analyzes only the runtime metrics associated with those specific processes, rather than processing all available data. This extraction approach maintains accuracy by focusing on relevant data while reducing the overall volume to be processed.
Data Source
AI summary
Root cause determination in multiple process systems is disclosed. It is determined that a first process encountered a problem at a first point in time. A directed graph that represents each process of a plurality of processes, including the first process, as a corresponding vertex of a plurality of vertices in the directed graph is accessed. A path through the directed graph from a start vertex to an end vertex is determined to identify a subset of processes that correspond to the vertices in the path. Corresponding runtime metric values are accessed to determine whether a runtime metric value is a non-desired value for at least one of the processes in the subset of processes. A process in the subset of processes is identified as a root cause candidate of the problem encountered by the first process.