Automated API Testing via Performance Doubles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Testing APIs that operate with database management system-implemented data models is challenging, especially in production environments, due to the difficulty in measuring performance against standards without impacting the system and maintaining compliance post-release.
Innovation Solution
An API test framework that operates automatically, generating performance doubles for database tables and view data structures, allowing for performance measurement against standards in a production environment without affecting the actual system, and enabling modifications to maintain performance compliance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If automated testing is implemented in production environment, then performance measurement capability is improved, but system disruption and complexity increase
Solution Approach 1:
The patent introduces performance doubles as intermediary objects that mediate between the testing framework and the actual production system. These doubles replicate the interface and data structure of real database tables and view data structures, allowing tests to execute against substitutes rather than the live system, thereby measuring performance without causing disruptions.
Solution Approach 2:
The patent creates copies of production database structures (tables and view data structures) called performance doubles. These copies contain representative data and replicate the schema of the original structures, enabling automated testing to execute against the copies instead of the actual production data, thus achieving measurement without impact.
2Measurement precision
If performance testing is conducted against actual production data, then measurement accuracy is improved, but system reliability deteriorates due to potential disruptions
Solution Approach 1:
Performance doubles serve as intermediaries that preserve the structural and interface characteristics of actual production data while isolating the testing process from the live system. This allows accurate measurement of API performance against representative data without the risks associated with testing against real production data.
3Productivity
If manual testing processes are used, then system complexity is reduced, but productivity and testing coverage decrease
Solution Approach 1:
The automated testing framework is designed to self-configure by automatically discovering the API interface, generating performance doubles that match the production data structures, and executing performance measurements without manual intervention. This automation significantly improves productivity and testing coverage while the framework manages its own complexity internally.
4Measurement precision
If production data is used for testing, then measurement realism is improved, but data security and system stability worsen
Solution Approach 1:
Instead of using actual production data, the patent creates performance doubles that are structural copies containing representative or synthetic data. These copies maintain the same schema, relationships, and interface characteristics as production data, providing realistic measurement conditions while eliminating security risks and stability concerns associated with using real production data.
Data Source
AI summary
Various examples are directed to systems and methods for evaluating an Application Program Interface (API) for interfacing an application to a database through a data model. A computing system may access a first view data structure associated with a first API call. The computing system may use a first view data structure to select a first table from the one or more tables, the first table being associated with the first API call, where the data model is described by a plurality of view data structures based on one or more tables at a database, including the first view data structure. The computing system may measure the API against a standard of performance for the API using a first performance double view data structure and a first performance double table.