Mainframe workload lift and shift

This page provides architectural guidance and emulation strategies for migrating legacy mainframe workloads (such as IBM z/OS or IBM i) to Google Cloud with minimal code modifications.

If you are a cloud architect, mainframe engineer, or infrastructure specialist, use the information on this page to evaluate emulation runtimes, re-platform mainframe workloads to Compute Engine, reduce mainframe MIPS licensing costs, and stream batch data to BigQuery by using Mainframe Connector.

Before you read this page, you should understand your mainframe workload MIPS consumption, application dependencies, and target emulation runtime requirements.

Mainframe lift and shift architecture

A mainframe lift-and-shift migration replaces physical mainframe hardware with cloud virtual infrastructure, running legacy COBOL or PL/I programs in an emulation runtime environment on Compute Engine.

Alongside compute emulation, you can use Mainframe Connector to offload CPU-intensive reporting workloads and reduce mainframe MIPS consumption. Mainframe Connector lets you complete the following tasks:

  • Transcode EBCDIC datasets (QSAM and VSAM) into ORC, JSON, or CSV formats.
  • Move data to Cloud Storage and BigQuery.
  • Submit BigQuery jobs directly from Job Control Language (JCL).

The following diagram illustrates the mainframe lift-and-shift architecture:

Architecture diagram showing legacy COBOL and CICS programs running in a partner emulation runtime environment on Compute Engine with data ingestion into BigQuery and Cloud Storage.

Key components

A mainframe lift-and-shift architecture includes the following components:

  • Compute emulation runtimes: deploy partner emulation engines, such as Micro Focus, NTT DATA, or LzLabs, on Compute Engine to run COBOL or CICS binaries without rewriting business logic.
  • Dataset transcoding: transcode EBCDIC datasets into ORC, JSON, or CSV formats and ingest high-volume mainframe batch files directly into BigQuery or Cloud Storage by using Mainframe Connector.
  • Database emulation: re-platform DB2 or IMS data to managed cloud relational databases such as Cloud SQL or Spanner.

What's next