Accelerating Mainframe Modernization - Modernizing Legacy COBOL in Months with AWS Transform
This article explains mainframe modernization with AWS Transform for mainframe. It introduces automated analysis of COBOL code, conversion to Java, and phased migration strategies.
Challenges of Mainframe Modernization
Many financial institutions, insurance companies, and government agencies have run their core business operations on IBM z/OS mainframes for decades. Codebases of millions of lines written in COBOL, CICS transaction processing, IMS and DB2 databases, and batch jobs driven by JCL are intricately intertwined, and the number of engineers who can grasp the whole picture is shrinking as they retire. In traditional modernization, manual code analysis took months and conversion took years, and prolonged projects and cost overruns had become the norm. AWS Transform for mainframe is a service that takes on this challenge head-on by making use of agentic AI. It became GA in May 2025, and AI agents dedicated to mainframe modernization automate the work from code analysis to conversion and test generation.
Code Analysis and Dependency Visualization
AWS Transform automatically scans the mainframe codebase and classifies components such as COBOL programs, copybooks, JCL, BMS maps, and CICS transaction definitions. It visualizes the call relationships between programs, data flows, and the dependencies of shared copybooks as a graph, and clarifies the scope of impact of the conversion. Through cyclomatic complexity analysis, you can identify in advance the high-complexity programs that are difficult to convert and narrow down the places that require manual intervention. It also has a feature for detecting components with the same name and duplicate program IDs, so you can resolve the naming conflicts that tend to occur in large codebases before the migration. With the import / export feature for file classifications, sharing analysis results between teams and integration with external tools are also possible.
Automated Conversion from COBOL to Java and Reimagine
Once code analysis is complete, AI agents automatically convert COBOL programs to Java. Rather than a simple syntax conversion, they semantically convert COBOL-specific data types (COMP-3, PIC clauses), file I/O, CICS commands, and DB2 SQL into the corresponding Java constructs. Automated tests are generated for the converted code and verify behavioral equivalence with the original COBOL programs. The Reimagine feature proposes designs that decompose monolithic mainframe applications into microservices. It defines service boundaries based on business domains and shows a migration path to an API-based, loosely coupled architecture. The service recommends a wave approach in which conversion proceeds in stages starting from modules with low business risk instead of all at once, and each wave runs a conversion / test / deployment cycle.
Transform Pricing
Pricing for AWS Transform for Mainframe is based on the number of lines of code to be converted. The specific unit price is individually quoted with the AWS account team, but an investment on the scale of tens of millions of yen is typical for converting a COBOL codebase of millions of lines. However, compared with the annual license costs of a mainframe (on the scale of hundreds of millions of yen), in many cases the investment can be recovered in 1 to 2 years. Post-conversion AWS infrastructure costs are expected to be 50-70% lower than mainframe operating costs.
Benefits and Limits of Automated Conversion
Because mainframe modernization deals with a huge amount of code accumulated over many years, making use of automated conversion tools produces large benefits. By analyzing legacy code such as COBOL and mechanically converting it to a modern language, you can greatly shorten the migration period compared with rewriting it by hand. However, automated conversion is not a cure-all. It is necessary to thoroughly verify that the converted code returns the same results as the original system. For complex business logic and undocumented specifications, human review is indispensable. An attitude of raising efficiency through automation while performing verification carefully is the precondition for a successful modernization.
Post-Migration Maintainability and Skills
One of the important benefits that mainframe modernization brings is improved maintainability. Engineers who can work with legacy languages such as COBOL are decreasing year by year, and securing the people who can keep the systems running has become a business risk. If you migrate to a widely used modern language, the pool of engineers who can handle it becomes deeper, and hiring and handovers become easier. You also become able to make use of modern development tools and cloud services, and the efficiency of adding features and of operations rises. The essential aim of modernization is not merely to eliminate obsolescence but to put the system in a state where it can be maintained / evolved into the future.
Summary - Guidelines for Mainframe Modernization
AWS Transform for mainframe shortens mainframe modernization from years to months with code analysis / automated conversion / test generation by agentic AI. The key to success is the approach of first grasping the whole picture with code analysis and then advancing the conversion in stages starting from modules with low complexity. If you make use of the Reimagine feature, rebuilding into a cloud-native architecture, rather than a mere rehost, also comes into view.
References (Official AWS Resources)
The primary sources for this page are the official AWS website and documentation. Check the official pages below for the latest specifications and pricing.
- AWS Transform official page
- AWS Database Migration Service official page
- AWS Step Functions official page
- AWS Documentation (official)
If this page and the official documentation disagree, treat the official documentation as authoritative.