Graph-Based Cloud Traffic Prediction for Nested API Trace Logs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud applications generate complex performance trace logs with nested API calls that are difficult to store and parse for accurate prediction of future API traffic, hindering effective cloud resource management.
Innovation Solution
A system that retrieves performance stack trace logs, parses them into a graph database with parent/child relationships, transforms the data for spatial and temporal information, and trains a transformer model to predict future traffic.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If performance stack trace logs are stored as plain text, then storage is simple, but parsing and generating intelligent prediction is difficult
Solution Approach 1:
The patent introduces a graph database as an intermediary structure between the raw performance stack trace logs and the prediction model. The graph database transforms the complex nested API call data into a structured representation with nodes (API calls) and edges (call relationships), making the data both easily storable and readily parseable for intelligent prediction. This mediator resolves the contradiction by providing a structured intermediate format that bridges simple storage and complex analysis needs.
2Measurement precision
If nested API calls are stored in a relational database, then structured storage is achieved, but the tree-like semi-structured data is not easily stored
Solution Approach 1:
The patent replaces the traditional relational database mechanism with a graph database mechanism. Instead of forcing tree-like semi-structured API call data into rigid relational tables, the graph database uses flexible node-edge representations that naturally accommodate the hierarchical nested structure. This substitution of the storage mechanism allows accurate representation of complex API call relationships while maintaining ease of storage and querying.
3Reliability
If cloud resources are manually adjusted, then resource optimization is achieved, but automation and efficiency are reduced
Solution Approach 1:
The patent implements a feedback-driven automated resource adjustment system. The transformer model continuously analyzes performance stack trace logs and generates predictions about future API traffic patterns. These predictions serve as feedback that automatically triggers resource allocation adjustments, creating a closed-loop system where resource optimization is both reliable and automated. The system learns from historical data and continuously improves its prediction accuracy, enabling autonomous resource management.
Data Source
AI summary
Some embodiments may be associated with a cloud computing environment. A computer processor of a traffic prediction server may retrieve performance stack trace logs from a traffic performance stack trace log repository that stores traffic information of the cloud computing environment. The traffic prediction server parses the performance stack trace logs as an objects list including parent/child object relationships and stores the parsed objects list in a graph database. The traffic prediction server may then transform the graph database into training data including spatial and temporal information and use the transformed training data to train a transformer model. According to some embodiments, the traffic prediction server also provides previous traffic input data to the transformer model when generates predicted traffic output data based on the previous traffic input data (e.g., to facilitate cloud load management).