AWS migration
Move forward with a workable plan.
Understand dependencies, sequence changes and keep the service running through transition.

Change the estate without stopping the service
Most migrations are judged by whether the service kept running and whether the old system was actually switched off. Both depend on understanding what the current estate does, including the parts nobody wrote down.
We plan and deliver migrations and modernisation around service dependencies: what moves, in what order, how the old and new coexist, how cutover is rehearsed, and what has to be true before the old system is retired.
When to bring us in
- A data centre exit, hosting contract or end-of-support date is forcing a move.
- Earlier moves to AWS brought cost and complexity without the expected benefit.
- A migration plan exists, but nobody is confident in the dependencies or the cutover.
What we take on, and what stays with your team
We take on: discovery, the target architecture, sequencing, the migration engineering itself, cutover rehearsal and the evidence for each go/no-go decision.
Your team keeps: the decisions on scope and risk, the communication with users and suppliers, the go/no-go call, and the retirement of the old service.
Engineering scope
Estate discovery
Inventory applications, data, interfaces, scheduled jobs and manual steps through tooling and interviews. Record what each component depends on and who owns it.
Options and sequencing
Choose per component: retire, retain, rehost, replatform or refactor. Sequence migration waves by dependency and risk, not by the order on a spreadsheet.
Data migration and coexistence
Plan data movement (for databases, AWS Database Migration Service is often part of it), reconciliation checks, and how old and new run side by side.
Cutover and rollback
Rehearse the cutover with timings, define the point beyond which rollback is no longer practical, and agree the decision makers and criteria in advance.
Decommissioning and exit
Define exit criteria for the old system: data retained or deleted, contracts ended, access removed. Assign retirement obligations to named owners.
An illustrative project
Illustrative example, not a client engagement
Starting state. A service runs on ageing virtual machines in a hosting contract that ends in nine months. Its interfaces to two other systems are only partly documented.
Agreed work. Map dependencies and data flows, agree a target design on AWS, migrate in two waves, rehearse cutover twice and agree retirement obligations for the old environment.
Deliverables. A dependency map, the target architecture, migration runbooks, data reconciliation checks, rehearsal records and a decommissioning plan with owners.
Acceptance. Rehearsal evidence supports the go/no-go decision, service behaviour is accepted after cutover by the service owner, and retirement obligations for the old environment are assigned and dated.
What we need from you
- A service owner and someone who knows how the current system really behaves.
- Access to the current environment and to the people who run it, including outgoing suppliers.
- Agreed constraints: dates, permitted downtime, data retention rules and budget for dual running.
Costs, risks and trade-offs
Rehosting is fast but carries old sizing and habits into the new estate; refactoring costs more upfront but can change the running cost. Migrations often run two estates at once, so plan the dual-running period, the retirement date and who pays for the overlap.
FinOps and cloud cost engineering covers how running costs are made attributable and managed.
Related
Read What is still running when the programme closes? and Modernising an AWS estate for cost control.
See anonymised experience, or use the supplier transition checklist. Compare ways to work together.
A sensible first step
Bring the service, the deadline and what you know about its dependencies. We will discuss whether a transition plan, a migration build or embedded support is the right start.
A first conversation does not commit either side to anything, and we will not ask for access to your systems until a scope is agreed.