AWS Transform for Mainframe: A new era for Legacy System Modernization

#aws #cloud #mainframe #modernization

Note: ✋ This post was originally published on my blog wiki-cloud.co

Introduction

A central component in many traditional business infrastructures has been the mainframe, a platform that for decades has served as the backbone of critical applications, especially in sectors such as banking, insurance, retail, government, among others.

Mainframe systems are known for their reliability, high performance, and ability to process large volumes of transactions. However, they represent a significant challenge today because they are expensive to maintain, difficult to integrate with modern technologies, and there is a growing shortage of professionals trained in languages such as COBOL, PL/I or assembler.

One of the biggest challenges facing organizations is precisely the migration of applications written in COBOL to modern programming languages. This difficulty responds in large part to a generational shift in the technology workforce, because younger developers are trained in contemporary languages such as Java code generation or Python, and many of the experienced professionals continue to master older technologies such as COBOL where most are already retiring. Despite this context, COBOL is not going away anytime soon and will continue tobe an essential component in many critical business systems.

💡 According to Reuters, 43 % of banking systems still rely on COBOL, and it is currently estimated that there are more than 220 billion lines of COBOL code active worldwide.

At the same time, the growing need to drive digital innovation and respond with agility to market changes has led many organizations to explore alternatives that are more flexible, scalable and aligned with cloud environments.

In this scenario, many organizations are beginning to consider the use of Generative Artificial Intelligence solutions as a key tool to accelerate and automate the modernization process of their Mainframe systems. These technologies can facilitate the transition to modern platforms without the need to completely redesign application functionality or alter the behavior of workloads currently running on Mainframe environments.

As part of its innovation plan, AWS made available AWS Transform for Mainframe, an innovative solution designed to modernize legacy systems. This was presented as “Amazon Q Developer Transform capability for Mainframe” during re:Invent 2024 and is now generally available. This service incorporates a specialized AI agent to significantly accelerate modernization processes that traditionally took several years, helping organizations complete their transformations and/or modernizations in a more agile and efficient manner.

What is AWS Transform for Mainframe?

AWS Transform for Mainframe is a service designed to accelerate the modernization of legacy applications in mainframe environments through the use of Generative AI (GenAI). This solution automates traditionally complex tasks such as COBOL code analysis, mainframe system documentation, business rule extraction, decomposition of monolithic applications into functional domains, and code refactoring.

By intelligently decomposing applications, AWS Transform allows you to plan and execute the migration sequence more efficiently, enabling the parallel transformation of business functions. This not only accelerates the modernization process but also improves decision-making, operational agility, and overall migration efficiency.

Source: AWS

💡 Note: Currently AWS Transform for Mainframe is only available in the regions of: USA (Northern Virginia) and Europe (Frankfurt).

AWS Transform Agents for Mainframe

As part of the Mainframe application transformation process, AWS Transform for Mainframe incorporates a series of specialized agents designed to facilitate and automate the modernization of legacy code to modern languages such as Java code generation. These agents enable critical analysis, refactoring and conversion tasks to be performed more quickly and accurately, optimizing each stage of the transformation process. The agents are described below:

Agent	Description	Key capabilities
Code Analysis Agent	This agent is responsible for analyzing legacy COBOL code to understand its structure, identify dependencies, detect missing artifacts and prepare the ground for automated Java code generation. Its in-depth analysis assesses the technical complexity of the codebase and automatically categorizes components.	- Cyclomatic complexity analysis to determine code stability. - Identification of internal and external dependencies. - Detection of missing or inconsistent artifacts. - Automatic classification of code according to functionality.
Documentation Generation Agent	This agent generates detailed technical and functional documentation from existing COBOL code. It can process millions of lines of code in minutes, identify the underlying business logic and provide clear descriptions tailored to the profile of each user (developers, analysts, business leaders). In addition, it can be integrated with knowledge bases to enrich the technical and domain context.	- Automatic expansion of customized acronyms. - Component and person-driven documentation generation. - Intelligent business logic summaries. - Context-sensitivity of code and functional domains. - Integration with knowledge bases.
Code Decomposition Agent	This agent identifies the boundaries between technical and business domains within monolithic code, making it easier to prepare for iterative, segmented migration to the AWS cloud. It is key to moving towards more modular and decoupled architectures.	- Detailed analysis of dependencies between code and data. - Logical separation of business and technical domains. - Improved detection of hidden dependencies. - Simplified initialization mechanism for code segmentation.
Migration Plan Agent	This agent assists in strategic migration planning, allowing users to define priorities, constraints and objectives. Based on these inputs, the agent generates optimized sequences of migration waves, providing different options to adapt the plan to business needs.	- Interactive and dynamic planning of migration waves. - Generation of multiple transformation scenarios based on customized criteria. - Prioritization based on technical complexity, business value, and dependencies.
Code Refactoring Agent	This agent has the ability to automatically refactor code, maintaining its functional equivalence and preparing it to run in a modern environment. It operates under human supervision, allowing to validate each step of the process to ensure that the new code is complete, maintainable and aligned with the defined functional domains.	- Automatic transformation of the code to modern languages (e.g. Java). - Guarantee of functional equivalence with respect to the original system. - Refactoring based on previously identified domains and components. - Horizontal scalability to handle large volumes of code simultaneously.

Source: AWS

Conclusions

Modernizing mainframe systems no longer has to be a lengthy, costly or risky process. With AWS Transform for Mainframe, organizations now have a comprehensive solution based on Agentic AI, capable of automating critical tasks such as COBOL code analysis, technical documentation, business rule extraction, migration wave planning, and refactoring to modern languages such as Java.

With its specialized agents, AWS Transform for Mainframe not only accelerates transformation and modernization projects, but significantly reduces the technical complexity and human effort required, allowing companies to continue to innovate without compromising the stability of their current operations. This hybrid approach between intelligent automation and human oversight opens the door to scalable, secure modernization aligned with business needs in today's digital world.