Approaches
Lift and Shift
- Quick—change of hosting environment
- Minimal advantage—not utilizing cloud benefits
Improve and Move
- Modernize during migration—take advantage of cloud capabilities
- Requires new skills and code changes
Rip and Replace
- Remove technical debt—take full advantage of cloud features
- Requires most time/skill investment
Strangler Pattern
- Lift and shift monolith, then break apart into microservices over time
Migrate for Compute Engine
- Migrate VMs to Compute Engine
- Lift and shift
- Test-clones—pre-migration validation
- Replicates instance data to Google Cloud in the background—zero interruptions
- Source environments:
- VMware vSphere
- AWS EC2
- Azure VMs
- Physical servers with supported OSes (e.g. CentOS, Ubuntu, Debian, Windows Server)
Phases
| Phase | Description |
|---|
| 1. Onboard | Choose source VM |
| 2. Replicate | Snapshot taken and replicated, incremental snapshot every two hours |
| 3. Configure | Set details for landing zone |
| 4. Test-clone | (Optional) Clone VM to Google Cloud for testing in sandbox environment |
| 5. Cut-over | Stop source, create new Compute Engine instance, roll-back manually if required |
| 6. Finalize | Remove replication data, delete test VMs |