Lift-and-Shift to Re-Platform Blueprint
The follow-up to lift-and-shift: progressively re-platform rehosted cloud VMs onto managed databases, containers, and serverless. Value-ranked waves use the strangler-fig pattern to cut cost and operational toil without big-bang refactors.
What and Why
Many cloud migrations stop at lift-and-shift: VMs rehosted unchanged, often costing as much as on-prem while gaining little. This blueprint is the second act, progressively re-platforming those rehosted workloads onto managed databases, container platforms, and serverless functions to capture real cloud value. It is cloud-agnostic; examples reference AWS but apply to Azure and GCP.
Phases
Baseline. Measure the current rehosted estate: per-app cost, utilization, patch burden, and operational toil. This quantifies the case for re-platforming and exposes the easy wins (idle VMs, self-managed databases).
Prioritization. Rank workloads by value-to-effort. Self-managed databases moving to RDS/Cloud SQL, web tiers moving to containers or PaaS, and batch jobs moving to serverless usually rank highest.
Re-platform waves. Tackle one pattern at a time. Move databases to managed services with native migration tooling. Containerize web tiers and run them on ECS, AKS, or GKE. Convert event-driven and scheduled jobs to functions. Use the strangler-fig pattern so legacy and new run side by side during cutover.
Validation. Test functionality, performance, and failover for each re-platformed component before retiring the rehosted version.
Optimization. After re-platforming, adopt autoscaling, commit-based discounts, and right-sizing. Tighten observability and cost allocation.
Key Risks and Mitigations
- Scope creep into full refactors. Keep each wave to one re-platform pattern with a defined exit.
- Cost overrun if managed services are misconfigured. Pilot, measure, then scale.
- Downtime during database moves. Use replication-based cutover with a short switch window.
- Skills gap on managed services. Pair with enablement and reference architectures.
Recommended Tooling
Native database migration services, Terraform for repeatable managed-service provisioning, container orchestration, serverless frameworks, and an observability and cost platform such as Datadog plus the cloud cost explorer.
Success Metrics
Track cost reduction per re-platformed workload, deployment frequency improvement, reduced operational overhead (patching, backups), and mean time to recovery.
Prerequisites
A completed lift-and-shift, a cost and utilization baseline, infrastructure as code in place, and a prioritized re-platform backlog with business sponsorship.