The 6Rs of migration: which R is yours?
The AWS 6Rs framework (Retain, Retire, Rehost, Replatform, Refactor, Repurchase) gives you a vocabulary for migration strategy. Here's when each fits.
Retire (15-25% of apps typically)
Nobody uses it. Check usage logs. If fewer than 5 users access an internal tool in 30 days, retire it. Savings are 100% of the app's cost forever.
Retain (5-15%)
Regulatory, licensing, or technical constraint means it stays on-prem. Common: old SAP, regulatory archives, proprietary hardware dependencies. Decision: schedule a 12-month revisit to see if constraints have changed.
Rehost / Lift-and-shift (40-60%)
Move as-is. Tools like AWS MGN, Azure Migrate, or CloudEndure do 80% of the work. Fastest path. Lowest risk. Lowest savings. Good for: apps you want off-prem quickly with no appetite for refactoring.
Replatform / Lift-and-adjust (10-25%)
Minor changes on the way over. Swap self-managed MySQL for RDS. Move from VMs to ECS/containers. Keep the app code mostly the same. Strongest balance of effort and reward.
Refactor / Re-architect (5-15%)
Major rework to take advantage of cloud-native patterns. Serverless. Microservices. Event-driven. Highest savings potential, highest cost, highest risk. Only for strategic apps.
Repurchase / Drop and shop (5-15%)
Replace with SaaS. Old Exchange → Google Workspace. On-prem CRM → Salesforce. Home-grown ticketing → Zendesk. The fastest way to eliminate infrastructure is to stop running infrastructure.
The honest workflow
For each app, score it with the Lift-and-Shift vs Refactor Decider. The output maps directly to the 6Rs. Then total the effort across apps to get a realistic timeline.