Cloud Migration Strategy

The 6 Rs (popularized by Gartner and AWS) categorize how each application is migrated: Rehost ("lift and shift" — move as-is to cloud VMs), Replatform ("lift and reshape" — minor optimizations like a managed database), Repurchase (drop and replace with SaaS, e.g. move to Salesforce), Refactor/Rearchitect (redesign cloud-native, e.g. break a monolith into serverless), Retire (decommission what is no longer needed), and Retain (keep on-prem for now — too risky or costly to move). A migration assessment assigns each app one of these.

Lift and shift (rehosting) moves the app essentially unchanged onto cloud infrastructure. The appeal is speed and low engineering risk — you avoid rewriting code and can hit a data-center-exit deadline. The trade-off is that you bring the old architecture's problems with you: you may pay for always-on VMs instead of elastic services, miss managed-service benefits, and gain little scalability. Many organizations lift-and-shift first to exit a data center, then iterate toward replatforming or refactoring afterwards — a pragmatic "migrate then modernize" approach.

Large migrations are organized into waves — batches of related applications moved together. Wave planning sequences work to manage risk: early waves often contain low-risk, low-dependency apps to build the team's muscle and validate the landing zone; later waves tackle complex, interdependent, business-critical systems. Applications with tight dependencies are usually grouped into the same wave to avoid latency or connectivity issues spanning environments. The output of a migration assessment feeds a wave plan: which apps, in what order, using which of the 6 Rs.

A landing zone is the well-architected foundation you build before migrating workloads. It establishes the account/subscription structure, network topology (VPCs, subnets, connectivity back to on-prem), identity and access management, logging and monitoring, encryption defaults, and policy guardrails. The goal is that every workload lands in a consistent, secure, compliant environment rather than each team improvising. AWS Control Tower, Azure Landing Zones, and GCP's landing zone blueprints automate this baseline so migration waves have a ready target.

A cutover is the carefully planned moment production traffic moves from the source to the migrated target. A good cutover plan covers data synchronization (final delta sync), DNS/traffic switching, validation checks, a defined rollback path, and a maintenance window. The pilot light pattern keeps a minimal core of the system always running in the cloud (like a pilot flame), so it can be rapidly scaled up during cutover or disaster recovery — cheaper than a full warm standby but faster to activate than building from scratch.