LLM-Trained Mock Servers for Microservices Integration Testing

Solution Overview

Problem

Integration testing of microservices is inefficient due to the need to launch and maintain full copies of system components, leading to time delays and additional costs.

Innovation Solution

Utilizing a Large Language Model (LLM) to generate training data for a mock server that imitates microservices interactions, allowing integration testing without running actual components, by recording requests and responses and using a mock server machine learning model trained with LLM-generated dependencies and configurations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If full copies of system components are launched and maintained for integration testing, then testing completeness is improved, but time consumption and cost increase

Engineering Contradiction:

Improvetesting completenessVSAvoidtime consumption

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent creates mock server components that replicate the behavior and interfaces of actual system components without requiring full copies. These mock servers are trained using recorded request-response pairs to simulate component interactions, enabling integration testing while avoiding the time and resource costs of launching complete system copies.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent performs preliminary recording of requests and responses during normal system operation, storing these interactions for later use in training mock servers. This advance preparation allows integration testing to proceed without needing to reconstruct or launch actual system components during the testing phase.

Inventive Principle:

Principle #10Preliminary action

2Reliability

If full copies of system components are launched and maintained for integration testing, then testing completeness is improved, but cost increases

Engineering Contradiction:

Improvetesting completenessVSAvoidcost

Core Design Contradiction:

ReliabilityVSLoss of energy

Solution Approach 1:

The patent creates mock server components that replicate the behavior and interfaces of actual system components without requiring full copies. These mock servers are trained using recorded request-response pairs to simulate component interactions, enabling integration testing while avoiding the time and resource costs of launching complete system copies.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent uses lightweight mock server instances that can be quickly created and discarded for testing purposes. These mock servers require minimal computational resources compared to full system components, reducing the cost and energy consumption of integration testing while maintaining adequate testing capability.

Inventive Principle:

Principle #27Cheap short-living objects (Disposable)

3Productivity

If LLM-generated training data is used to train mock server, then testing efficiency is improved, but dependency on LLM accuracy becomes critical

Engineering Contradiction:

Improvetesting efficiencyVSAvoidtesting accuracy

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent incorporates feedback mechanisms where the mock server's responses are validated against expected outcomes, and the system learns from discrepancies. This feedback loop helps correct inaccuracies in LLM-generated training data and improves the reliability of testing results over time.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent performs preliminary recording of requests and responses during normal system operation, storing these interactions for later use in training mock servers. This advance preparation allows integration testing to proceed without needing to reconstruct or launch actual system components during the testing phase.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20260044438A1Streamlining integration testing using large language models

Publication Date: 2026.02.12 SAP SE

AI summary

In an example embodiment, Large Language Models (LLMs) are leveraged to augment training data used to train a machine learning model to imitate responses to requests in a microservices system. A recorder is used to record requests from and responses to a microservice. This recorded information can then be used as context for an LLM prompt sent to an LLM. Based on this prompt, the LLM then generates dependencies, configurations, and integrations that can be used along with the recorded information itself as a training data set. The training data set is then used to train a mock server that is able to imitate an integration testing scenario, including replicating a setup procedure for the components and replicating responses and requests generated by those components, permitting integration testing without copies of actual components to be configured and run.