Insight
The Cloud Migration 6 Rs: A Practical Framework for Enterprise Workloads
Rehost, replatform, refactor, repurchase, retire, retain — a working decision framework for choosing the right migration path per workload, with cost and risk trade-offs.
Migration failures rarely come from bad engineering. They come from choosing the wrong strategy per workload. The 6 Rs framework — popularised by AWS and Gartner — remains the cleanest way to sequence a portfolio migration.
The 6 Rs at a glance
- Rehost — "lift-and-shift" the VM as-is. Fastest, cheapest to start; highest run-cost tail.
- Replatform — small optimisations (managed DB, autoscaling) without changing the app.
- Refactor — re-architect to cloud-native (containers, serverless, event-driven).
- Repurchase — drop and replace with SaaS (e.g. CRM, HRMS, ITSM).
- Retire — decommission workloads no one uses.
- Retain — keep on-prem for compliance, latency, or licensing reasons.
A decision matrix that actually works
| Signal | Likely R |
|---|---|
| End-of-life OS, minimal change tolerance | Rehost, then modernise |
| Stateful monolith, high change frequency | Replatform |
| High traffic variability, need elasticity | Refactor |
| Undifferentiated commodity function | Repurchase |
| No owner, no traffic, no docs | Retire |
| Regulatory data residency, sub-5 ms latency | Retain (hybrid) |
Sequencing matters more than the choice
Migrate retire and repurchase workloads first — they shrink the estate and free budget. Then rehost the long tail while a modernisation squad refactors the crown-jewel systems in parallel.
Cost signals to watch
- Egress costs dominate multi-cloud designs — model them before commit.
- Reserved instances / savings plans only pay off after 8–10 months of steady state — don't lock in during discovery.
- Observability spend grows non-linearly; standardise on OpenTelemetry early.
Where teams lose time
- Missing a proper application discovery pass (dependencies, licences, data flows).
- Skipping a landing zone — every migration then re-invents IAM, networking, and guardrails.
- Underestimating change-management — cutovers fail on people, not tech.
How Solvin Globaltech helps
Our Cloud Infrastructure practice runs migrations end-to-end — discovery, landing zone, waves, FinOps, and post-migration modernisation. Pair with Managed IT Services for steady-state operations.
Next step: Request a 6 Rs portfolio assessment — we categorise up to 50 workloads and produce a sequenced migration wave plan.
/ FAQ
Frequently asked questions
- What are the 6 Rs of cloud migration?
- Rehost, Replatform, Repurchase, Refactor, Retain, and Retire. Each strategy trades off migration speed, cost, and long-term cloud-native benefit.
- Which R should we start with?
- Most enterprises begin with Rehost (lift-and-shift) for low-risk workloads to build momentum, then Replatform or Refactor for systems where cloud-native gains justify the engineering investment.
- How long does a typical cloud migration take?
- A mid-sized enterprise portfolio (200–500 workloads) typically takes 12–24 months when sequenced by the 6 Rs, with early wave-one workloads live in 60–90 days.