AWS Mainframe Modernization
A service that refactors or replatforms mainframe COBOL/PL/I workloads on AWS, accelerating legacy system modernization
Overview
AWS Mainframe Modernization is a platform service for migrating legacy workloads running on mainframes - COBOL, PL/I, JCL, and more - to AWS. It provides two engines: Micro Focus (replatform) and Blu Age (refactor), offering the choice between running existing code as-is on AWS or automatically converting it to Java/.NET. It emulates or replaces mainframe-specific execution environments including CICS, IMS, and batch jobs in the cloud.
Selection Criteria for Replatform vs. Refactor
Replatform (Micro Focus engine) runs existing COBOL/PL/I code unchanged in a compatible execution environment on AWS. Since no code changes are required, migration risk is low and AWS migration can be completed in a short timeframe. However, the codebase remains legacy, limiting long-term maintainability improvements. Refactor (Blu Age engine) automatically converts COBOL code to Java, replacing CICS transactions with Spring Boot microservices and JCL batch jobs with Step Functions workflows. Converted code can be maintained with modern development toolchains (IDE, CI/CD, test frameworks) and gains cloud-native scalability. As selection criteria, refactoring suits business-critical systems intended for long-term operation, while replatforming suits systems approaching EOL (End of Life) or early phases of staged migration.
Migrating Batch Jobs and Transaction Processing
Mainframe batch processing (JCL jobs) is managed by Mainframe Modernization's batch scheduler. The replatform approach interprets and executes JCL as-is, while the refactor approach converts it to Step Functions and Lambda. VSAM file and DB2 database access are replaced with S3/DynamoDB and RDS/Aurora respectively. CICS online transactions run in Micro Focus's CICS emulator under the replatform approach, or are converted to Spring Boot services with REST APIs under the refactor approach. BMS maps (3270 terminal screens) are converted to HTML/React web UIs, enabling end users to access through modern browser interfaces. For data migration, conversion from mainframe EBCDIC encoding to ASCII/UTF-8, packed decimal processing, and fixed-length record parsing are handled automatically.
Testing Strategy and Staged Migration Patterns
In mainframe modernization, equivalence testing that guarantees the migrated system returns identical results to the original system is paramount. Mainframe Modernization's test environment provides a regression testing framework that feeds mainframe input data into the AWS environment and automatically compares output results with mainframe output. Tens of thousands of test cases run in parallel, with detailed reports on detected discrepancies. For staged migration patterns, the Strangler Fig pattern is effective. An API Gateway is placed at the front, routing requests for new features to modernized services on AWS while routing unmigrated feature requests to the mainframe via Direct Connect. Migration proceeds function by function, with the strategy of eventually reducing mainframe traffic to zero.