GCP Study Hub

Migration

Migration Strategies (6 Rs)

Framework for moving workloads to GCP — Round 2 focus area

AWS equivalent

Same 6 Rs framework — your AWS experience applies directly

MigrationRound 2 FocusStrategy

Architecture Diagram

The 6R Migration Framework

🔵
Rehost
Lift & Shift

Move VMs as-is. Fastest, least value.

Speed: FastEffort: Low
🟢
Replatform
Lift & Reshape

Minor changes. e.g. VM→Cloud SQL.

Speed: FastEffort: Low-Med
🟣
Refactor
Re-architect

Cloud-native redesign. Containers/microservices.

Speed: SlowEffort: High
🟡
Repurchase
Replace w/ SaaS

Switch to SaaS (Workspace, Salesforce).

Speed: MedEffort: Med
🔴
Retire
Decommission

Remove unused workloads. Reduces scope.

Speed: FastEffort: Very Low
⚪
Retain
Keep On-prem

Stay on-prem — compliance/latency/cost.

Speed: —Effort: None
💡 Interview answer: "I wouldn't apply one strategy to the entire estate. I'd start with a workload assessment using Migration Center, then map each workload to the right R-strategy based on complexity, risk, and business value."
🔄

AWS → GCP: Key Differences

  • ▸

    The 6 Rs framework is cloud-agnostic. Your AWS migration experience maps directly.

  • ▸

    GCP-specific nuances: Cloud Run makes 'replatform' easy (just containerize). BigQuery makes 'rearchitect' compelling for analytics.

📌

Key Concepts to Know

  • 1

    🔵 REHOST (Lift & Shift): Move VMs as-is. Fastest, lowest risk. Little cloud-native value initially.

  • 2

    🟢 REPLATFORM (Lift & Reshape): Minor changes. e.g., on-prem MySQL → Cloud SQL.

  • 3

    🟣 REFACTOR / RE-ARCHITECT: Redesign for cloud-native. e.g., monolith → microservices on Cloud Run.

  • 4

    🟡 REPURCHASE: Replace with SaaS. e.g., on-prem email → Google Workspace.

  • 5

    🔴 RETIRE: Decommission unused workloads. Always do a workload audit first.

  • 6

    ⚪ RETAIN: Keep on-prem for now. Compliance, latency, cost reasons.

💡

DCE Interview Tips

  • ★

    ALWAYS start with: 'I wouldn't recommend a single strategy for the whole estate. I'd do a workload assessment first, then apply the right R to each workload.'

  • ★

    Phasing rule: migrate dev/test first, then non-critical prod, then mission-critical prod.

  • ★

    Assessment tools: Migration Center (free), StratoZone (deeper analysis). Do this BEFORE architecture design.

⚠️

Common Gotchas

  • !

    Don't lead with re-architect — it takes longest. Start with rehost for speed, then refactor incrementally.

  • !

    Dependency mapping is often the hardest part — apps that look standalone often have 10+ dependencies.