Year: 2025
Type: article
Abstract: Legacy COBOL and C systems are still widely used in industries, yet they are increasingly costly, insecure, and incompatible with modern platforms. Traditional modernization methods, although effective, often require significant time and resources and carry high risks of disruption. This paper proposes a scalable alternative that leverages Large Language Models (LLMs) within a structured multi- agent framework, guided by the “7Rs of Modernization.” The frame- work comprises three agents: an Analysis Agent that interprets and maps legacy code, a Coder Agent that generates modern equivalents, and a Review Agent that validates correctness, security, and compliance through iterative feedback. By automating much of the migration process, the proposed approach enables faster, more transparent, and less risky mod- ernization. It helps enterprises transition from outdated systems to modular, secure, and cloud-ready solutions, offering a cost- effective and future-proof pathway to digital transformation.
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4416135505