Service Graph Generation with Order and Exclusive Relationships
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing service topology visualization techniques, such as those described in NPL 1, fail to express dependency relationships including order and exclusive relationships between components in microservice architectures, making it difficult to identify the root cause and failure range of service failures.
Innovation Solution
A service graph generation apparatus that acquires trace data, analyzes time data to estimate order and exclusive relationships between components, and creates a service graph expressing parent-child, order, and exclusive relationships, enabling the visualization of dependency relationships between components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If service topology visualization is performed using existing techniques, then parent-child relationship between components can be expressed, but dependency relationship including order relationship and exclusive relationship cannot be expressed
Solution Approach 1:
The patent segments the dependency relationship information into three distinct types: parent-child relationships, order relationships, and exclusive relationships. Each type is represented by different visual elements in the service graph (different line styles or connection types), allowing comprehensive information expression while maintaining clear visual distinction and avoiding overwhelming complexity.
Solution Approach 2:
The patent adds temporal and logical dimensions to the traditional service topology visualization. By incorporating time data from trace spans and analyzing execution sequences, the system elevates the visualization from static parent-child connections to dynamic multi-dimensional relationships that include temporal ordering and exclusivity constraints.
2Difficulty of detecting and measuring
If comprehensive dependency relationships are visualized, then root cause identification becomes easier, but the complexity of analyzing trace data increases
Solution Approach 1:
The patent performs preliminary analysis of trace data by automatically comparing time stamps of spans with identical parent components to pre-determine order and exclusive relationships. This preliminary processing extracts dependency patterns before service failures occur, so that when failures happen, the pre-analyzed service graph can immediately provide root cause information without requiring time-consuming real-time analysis.
Solution Approach 2:
The system makes the trace data self-analyzing by automatically extracting temporal relationships and dependency patterns without requiring manual intervention. The analysis unit autonomously processes span time data to infer order and exclusive relationships, reducing both manual analysis time and the burden on operators.
Data Source
AI summary
Provided is a service graph generation apparatus 10 for generating a service graph expressing a dependency relationship between components constituting a monitoring target service 50. The service graph generation apparatus 10 includes an acquisition unit 11 that acquires trace data including spans recording a parent-child relationship and time data of processing of components in a series of processing in a service, an analysis unit 13 that compares, for every piece of the trace data, time data between spans of the components having the same parent to estimate an order relationship or an exclusive relationship between the components, and a creation unit 14 that creates a service graph expressing all components formed of the monitoring target service 50, and the parent-child relationship, the order relationship, and the exclusive relationship between the components.