CONCEPT OF SOFTWARE FOR MICROSERVICE ARCHITECTURE DESIGN WITH USAGE OF ARTIFICIAL INTELLIGENCE

Abstract

This paper presents a comprehensive concept of software designed for the holistic development of microservice architectures (MSA) empowered by artificial intelligence (AI). The proposed solution addresses major challenges in MSA by automating key stages of its lifecycle: reconstructing of an existing deployed system into a formal model, interpretation of user-defined requirements written in natural language, architectural model transformation, and automated generation of infrastructure artifacts compatible with DevOps practices. At the core of the approach lies a formal architecture model described using a modeling, capable of capturing service semantics, communication protocols, persistence strategies, scalability mechanisms, and service ownership. The system leverages AI technologies - particularly large language models (LLMs) and natural language processing (NLP) - to analyze contextual information and derive architectural decisions such as service decomposition, refinement, or restructuring. Unlike traditional diagram-based tools, the proposed concept ensures traceability of changes, semantic consistency, support for iterative updates, and seamless integration with version control and deployment pipelines. The solution also features an interactive graphical user interface based on the C4 model, enabling engineers to visualize the system at different abstraction levels and to inspect architectural alternatives, including rejected ones, which are stored for historical traceability. Existing AI-based tools for MSA design often remain at the proof-of-concept level and do not provide formal outputs. This work addresses these limitations by introducing a complete conceptual framework that integrates AI reasoning, formal modeling, and DevOps automation. Future work will focus on the development of the modeling language suitable for MSA, implementation of a software prototype, and experimental validation in industrial case studies.