ReleaseBytes
gcp Google Cloud Blog ·

Google Cloud AI for Mainframe Modernization

blogaigcpengineerfinanceenergygcp-bigquerygcp-cloud-sqlgcp-cloud-storagegcp-spanner
feature announcement

Google Cloud introduces an AI-powered approach to mainframe modernization, moving beyond simple code conversion to address complex dependencies and data formats. This strategy leverages Gemini models for code understanding and specialized AI agents to offer iterative modernization, aiming to reduce the risks associated with traditional 'big bang' migrations. The solution targets enterprises with legacy mainframe systems, providing tools for assessment, modernization, de-risking, and data migration.

  • →AI reverse-engineering for legacy application assessment
  • →AI-powered code transformation for modernization
  • →Dual Run for de-risking before go-live
  • →AI-driven mainframe modernization strategy
  • →AI agents codify proven modernization methodologies
Features (3) ›
  • AI reverse-engineering for legacy application assessment

    The Mainframe Assessment Tool (MAT) reverse-engineers legacy codebases at scale to provide explainability and form a foundation for modernization. It offers dependency visualization, automated business rule extraction, automated documentation generation, and domain/business function discovery.

  • AI-powered code transformation for modernization

    Google Cloud offers two main modernization patterns: 'Rewrite/Reimagine' for business logic innovation using AI agents and forward-engineering, and 'Deterministic modernization' for exact legacy application behavior preservation with AI-driven code-to-code modernization.

  • Dual Run for de-risking before go-live

    Google Cloud Dual Run processes real-world production workloads simultaneously on both mainframe and Google Cloud environments. It automatically captures mainframe transactions, runs them against modern applications, and compares outputs to ensure logic and data equivalence before go-live.

Notes (2) ›
  • AI-driven mainframe modernization strategy

    Google Cloud proposes an alternative to traditional mainframe modernization by leveraging AI and cloud agility for iterative and continuous updates. This approach acknowledges that modernization involves more than just code conversion, requiring updates to data models, handling dependencies, and ensuring functional equivalence.

  • AI agents codify proven modernization methodologies

    Specialized AI agents, built in partnership with Mainframe Modernization Professional Services, codify proven methodologies and field experience into structured, agentic modernization workflows. These agents are integrated with tools like MAT and Antigravity for an AI-accelerated development pipeline.

Read the original announcement →

https://cloud.google.com/blog/products/infrastructure-modernization/mainframe-migration-and-modernization-with-ai/

Related releases