Result: Automated Refactoring of Monolithic Applications to Cloud-Native Containers: Application Modernization using GenAI and Agentic Frameworks
Further Information
Modernizing monolithic applications into microservices is critical for scalability and maintainability, but manual refactoring is complex and resource intensive. This paper presents a fully automated AI-driven approach using Generative AI (GenAI) and multi-agent frameworks to refactor monoliths into containerized microservices with minimal human intervention. Our system orchestrates multiple specialized AI agents, each performing tasks such as code analysis, service decomposition, automated code transformation, containerization, orchestration, and CI/CD integration. By leveraging LLM-powered agents, the system not only identifies microservice boundaries but also modifies code, generates infrastructure configurations, and validates functionality through iterative AI-driven testing. We implement this approach on a representative enterprise monolith and achieve a successful decomposition that preserves functionality while improving modularity and deployment readiness. The results demonstrate that our AI-driven system can accelerate application modernization, reduce engineering effort, and enhance software quality. We discuss challenges such as LLM limitations, data consistency, and security concerns, and propose future directions for improving automation, scalability, and adaptability.