KR-20260065458-A - METHOD AND SYSTEM FOR SUPPORTING MICROSERVICE LOOSE COUPLING BASED ON GENERATIVE AI LEARNING

KR20260065458AKR 20260065458 AKR20260065458 AKR 20260065458AKR-20260065458-A

Abstract

A method and system for supporting loosely coupled microservices based on generative AI learning are disclosed. A method for supporting loosely coupled microservices based on generative AI learning according to an embodiment of the present invention may include: a step of defining a microservice specification associated with a microservice; a step of proposing a method for classifying microservice specifications based on the details of the defined microservice specification; and a step of constructing a learned generative AI model by proposing a generative AI learning process that applies the microservice specification and the method for classifying microservice specifications.

Inventors

염근혁
조대영
박준석
정수민

Assignees

부산대학교 산학협력단

Dates

Publication Date: 20260508
Application Date: 20250110
Priority Date: 20241029

Claims (15)

A step of defining microservice specifications associated with microservices; A step of presenting a method for classifying microservice specifications based on the details of the defined microservice specifications; and A step of building a trained generative AI model by proposing a generative AI learning process that applies the above microservice specifications and the above microservice specification classification method. A method for supporting loosely coupled microservices based on generative AI learning, including
In paragraph 1, The step of defining the above microservice specifications is, A step of collecting microservice description documents for said microservices from a microservice repository; A step of generating microservice function specifications using the above microservice description document; A step of generating a microservice loose coupling specification using the above microservice function specification; and A step of defining a microservice specification, including the above microservice function specification and the above microservice loosely coupled specification. A method for supporting loosely coupled microservices based on generative AI learning, including
In paragraph 2, The step of generating the above microservice function specifications is, A step of dividing the above microservice description document into sections according to content and separating the microservice functions into functional sections; A step of refining the microservice functions by performing prompt engineering that processes natural language and formalizes the separated microservice functions using a generative AI model; and A step of generating a microservice function specification by adding image descriptions collected for the microservice from an image repository to the refined microservice function. A method for supporting loosely coupled microservices based on generative AI learning, including
In paragraph 2, The step of generating the above microservices weak coupling specification is, A step of collecting microservice deployment specifications from the microservice repository based on the service name of the microservice function specification; In the above microservice deployment specification, a step of selecting only the container orchestrator specification; and A step of generating the microservices loosely coupled specification by extracting loosely coupled specification elements from the selected container orchestrator specification. A method for supporting loosely coupled microservices based on generative AI learning, including
In paragraph 1, The step of presenting the above-mentioned microservice specification classification method is, A step of presenting a method for classifying the above-mentioned microservice specifications by deriving microservice answer pattern types according to the BCE pattern microservice answer criteria based on the details of the above-mentioned microservice specifications. A method for supporting loosely coupled microservices based on generative AI learning, including
In paragraph 1, The step of building the above-mentioned trained generative AI model is, A step of creating a fine-tuning dataset in which the above microservice specification is a question and the above microservice specification classification method is an answer; and A step of constructing the fine-tuned model as the trained generative AI model by proposing the generative AI training process of uploading the above fine-tuning dataset to the generative AI model to fine-tune the generative AI model. A method for supporting loosely coupled microservices based on generative AI learning, including
In paragraph 1, For microservices applications requiring pattern type identification, A step of inputting a microservice specification defined in the above microservice application as a question to the above-mentioned generative AI model to obtain a microservice answer pattern type from the above-mentioned generative AI model; and A step of identifying pattern types for the microservice application by processing the microservice answer pattern type and displaying it in a table form corresponding to the name of each microservice included in the microservice specification and the identified BCE pattern type. A method for supporting loosely coupled microservices based on generative AI learning, further including
A processing unit associated with microservices that defines microservice specifications; A presentation unit that suggests a method for classifying microservice specifications based on the details of the defined microservice specifications; and A model training unit that builds a trained generative AI model by proposing a generative AI learning process that applies the above microservice specifications and the above microservice specification classification method. A microservices loosely coupled support system based on generative AI learning, including
In paragraph 8, The above processing unit is, Collect microservice description documents for the said microservice from the microservice repository, and Using the above microservice description document, generate microservice function specifications, and Using the above microservice function specifications, generate a microservice loose coupling specification, and Defining a microservice specification, including the above microservice function specification and the above microservice loose coupling specification, Microservices loosely coupled support system based on generative AI learning.
In Paragraph 9, The above processing unit is, Divide the above microservice description document into sections according to content, separate the microservice functions divided into functional sections, and For the above separated microservice functions, prompt engineering is performed using a generative AI model to process and formalize natural language, thereby refining the microservice functions, and Generating the microservice function specification by adding image descriptions collected for the microservice from the image repository to the refined microservice function, Microservices loosely coupled support system based on generative AI learning.
In Paragraph 9, The above processing unit is, Based on the service name of the above microservice function specification, collect the microservice deployment specification from the above microservice repository, and From the above microservice deployment specification, only the container orchestrator specification is selected, and Generating the microservices loosely coupled specification by extracting loosely coupled specification elements from the above-mentioned selected container orchestrator specification, Microservices loosely coupled support system based on generative AI learning.
In paragraph 8, The above presentation section is, By deriving microservice answer pattern types according to the BCE pattern microservice answer criteria based on the details of the above microservice specification, the method for classifying the above microservice specification, Microservices loosely coupled support system based on generative AI learning.
In paragraph 8, The above model learning unit is, Create a fine-tuning dataset in which the above microservice specifications are the questions and the above microservice specification classification methods are the answers, and By proposing a generative AI training process that uploads the above fine-tuning dataset to a generative AI model to fine-tune the generative AI model, the fine-tuned model is constructed as the trained generative AI model. Microservices loosely coupled support system based on generative AI learning.
In paragraph 8, For microservices applications requiring pattern type identification, A model utilization unit that identifies pattern types for the microservice application by inputting the microservice specification defined in the microservice application as a question to the trained generative AI model, obtaining microservice answer pattern types from the trained generative AI model, processing the microservice answer pattern types, and displaying them in a table form corresponding to the names of each microservice included in the microservice specification and the identified BCE pattern types. A microservices loosely coupled support system based on generative AI learning, further including
A computer-readable recording medium having a program for executing any one of the methods of paragraphs 1 through 7.

Description

Method and System for Supporting Microservices Loosely Coupled Based on Generative AI Learning The present invention relates to a method and system for supporting loose coupling of microservices based on Generative Artificial Intelligence (GAI) learning, which performs support for loose coupling of microservice architecture by learning Generative Artificial Intelligence. Registration No. 10-2456386 (2022.10.14) "Microservice Batch Device and Microservice Batch Method" Registration No. 10-2456386 discloses a microservice deployment device comprising: a resource monitoring control unit for identifying resource requirements of microservice units constituting an application service; a workload profiling control unit for profiling the application service and the workload for the application service; a microservice deployment control unit for deploying the microservice to a cloud environment according to resource requirements per microservice using the profiling results obtained from the workload profiling control unit; and a cloud control unit for controlling the cloud environment. Registration No. 10-2456386 proposes a microservice deployment decision system that performs heuristic deployment by applying a greedy algorithm with workloads such as resource requirements and load at the microservice level as input. Registration No. 10-1996840(2019.07.01) "Microservice Store Operating System" Registration No. 10-1996840 comprises: a microservice store operating server for operating a microservice store for selling microservices; a developer terminal for connecting to the microservice store operating server via a network to register the microservice developed by a developer as a product in the microservice store; a buyer terminal for connecting to the microservice store operating server via the network to search for the microservice sold in the microservice store and to purchase the microservice; and, in cases where the developer grants the buyer the right to modify the source code, the right to distribute, or the right to resell when registering the microservice in the microservice store, an adapter terminal for connecting to the microservice store operating server via the network to modify the purchased microservice and register the modified microservice as a product in the microservice store. The microservice includes a curator terminal for developing a new application by combining multiple microservices purchased via the network to the microservice store operation server, and registering and selling the developed application in the microservice store. The microservice store comprises: a community module that provides communication channels to buyers, developers, adapters, and curators; a microservice app store module that guarantees the usage rights of copyright holders and buyers by managing license rights and providing authentication functions, wherein the entire functionality is provided via a web screen, the curator reviews microservices registered by developers to register them as catalog services and provides customized catalog curation services to buyers, and buyers can use the service by installing the product in their cluster operating environment after paying for the catalog using a payment platform and acquiring license rights; a microservice repository module that allows developers to freely register microservices they have created and provides copyright and configuration management functions for the services; a microservice SDK module that is installed in a local environment and provides functions for developers to develop, test, and register catalog services in a Web UI format; and a payment module that provides a payment interface for purchases and offers various payment method integration functions through blockchain-based cryptocurrency and PG integration. The store operating system is being launched. Registration No. 10-1996840 proposes a system that builds and operates a microservices store to reduce opportunity costs and ensure ease of use of microservices. Registration number 10-1996840 establishes a system that provides registration, sale, purchase, payment, and license registration for microservices, enabling organizations to obtain necessary microservices without developing microservices within the organization. Registration number 10-1996840 focuses only on the process for supporting the purchase of microservices. Microservices architecture is a structure that supports the development of cloud-native applications, where individual microservices are implemented as functional units capable of independent deployment to provide cloud services through loose coupling. One way to structure a microservices architecture is to classify the types of microservices by applying the Boundary-Control-Entity Pattern (BCE), a software design pattern. Currently, the implementation of microservices architecture utilizes containers, which are lightweight virtualization systems, and their lifecycles. To achiev