AWS Mainframe Modernization supports the migration and operation of mainframe workloads on AWS. Its capabilities cover application assessment, code transformation, development, data movement, functional testing, and runtime operations. A modernization project still requires application discovery, dependency mapping, target architecture design, and validation of business behavior; moving source code alone does not migrate an operational workload.
The service supports two principal modernization approaches. Automated refactoring uses AWS Transform for mainframe capabilities to convert legacy applications into modern languages and frameworks, including Java-based applications. Replatforming uses Rocket Software, formerly Micro Focus, technology to run compatible mainframe applications on AWS while retaining much of the existing source code. Select an approach after assessing language support, transaction dependencies, batch processing, data formats, and the amount of application change the team can support.
Access restrictions are important when selecting a target architecture. The managed runtime environment experience is no longer open to new customers. New customer access to the self-managed experience closes effective June 30, 2026. Existing customers can continue using the respective experiences, but AWS does not plan new features for the self-managed experience while continuing security and availability improvements. Confirm eligibility and regional availability before designing a deployment around either experience.
For existing managed runtime customers, an application represents a migrated workload and is associated with a runtime environment. The application specifies an engine appropriate to the modernization approach. Engineers deploy application versions, configure supporting resources, import data sets, submit batch jobs, and monitor execution. High-availability configurations and runtime sizing must reflect transaction throughput, batch windows, recovery objectives, and engine-specific requirements rather than mainframe capacity measurements alone.
Data migration requires explicit treatment of mainframe encodings and record structures. File Transfer with BMC can discover mainframe data sets and transfer selected data to AWS, with code page conversion options or binary transfer when downstream tools require the original representation. Managed runtime data sets can also be exported to Amazon S3, optionally with encryption. Validate record counts, packed decimal fields, character conversion, and application-level totals before accepting a migration.
Application Testing supports automated functional equivalence and regression testing. Build representative test cases for online transactions and batch processing, then compare results against established mainframe behavior. Differences in sorting, numeric precision, date handling, and file encoding deserve targeted checks. For supported refactored applications, restarting a batch job at a specific step can assist recovery, but operators must still assess previously committed database changes and external side effects before rerunning work.
Production architectures normally combine private networking, least-privilege IAM roles, encrypted storage, centralized logging, and repeatable deployment definitions. Separate development, testing, and production resources, and keep deployment artifacts distinct from production data. Budget for runtime capacity, storage, databases, data transfer, optional modernization tools, and testing rather than treating the service as a single migration charge.