04 / Migration

Every move needs a way back.

We plan a platform move around the running business process. Applications, data and interfaces are considered together, critical assumptions tested in advance, and switchover and recovery prepared. Modernisation takes place where it makes sense for the target environment or subsequent operation.

Request an architecture discussion

Clarify the starting point and objective

We first clarify which problem the move should solve: infrastructure nearing end of life, insufficient scaling, difficult releases or changed security requirements. We record applications, data storage, interfaces and the user journeys that must function during the transition. A dependency overview reveals which components should move together and where separation is possible. Technical prerequisites are distinguished from organisational approvals. This produces migration groups, a sensible sequence and a realistic pilot scope.

Choose the right degree of change

Not every migration needs the same depth of intervention. A largely unchanged move can reduce time pressure, but often carries existing operational problems forward. Adapting the runtime or data storage may simplify target operations, but initially increases testing and change scope. We assess these options per application. Refactoring is aligned with a concrete benefit, such as replacing an unsustainable dependency or enabling reproducible deployment. It remains clear which work is necessary for the move and which modernisation can be planned independently.

Test data and dependencies

Testing must include inconspicuous connections: DNS resolution, certificates, identities, outbound connections and scheduled processing. For data, we clarify synchronisation, consistency and the period during which changes may occur at both source and target. A trial migration provides measurements for transfer, processing and validation. We check data volumes, business samples and access from the target application. Results determine the switchover window and reveal necessary preparation or intermediate states.

Prepare cutover and the return path

Switchover receives a process with owners, decision points and measurable approval criteria. Write access, DNS changes and the start of downstream processing are deliberately sequenced. Observation and communication belong in the same plan as technical commands. A return path depends particularly on data already written and schema changes. We state the point until which a simple switchback is possible and the procedure beyond that point. Where a conventional rollback is not viable, another recovery path must be clarified before approval.

Continue testing after the move

After cutover, we observe the application under actual usage. Response times, error rates, background processing and data reconciliation reveal whether the new environment meets expectations. Capacity, permissions and cost development are also compared against the underlying assumptions. The old environment is decommissioned only after agreed evidence is available. Until then, backups, access and responsibilities remain clear. Operational documentation and runbooks are updated to the target state so the migration project's completion leaves functioning operations behind.

Example: moving a web application and database separately

A staged move can reduce simultaneous changes. If the database initially remains in the data centre, however, the cloud application must handle additional latency and possible connection interruptions. Many small database calls may make this approach unsuitable. We test the user journey under realistic load and assess consistency, timeouts and retries. Results may favour moving components together or a limited interim phase with adapted interfaces. The interim architecture receives a clear purpose and a defined end.

Acceptance based on specific user journeys

A running container does not prove a successful migration. We agree the workflows that must function from the perspective of users, editors or connected systems. These may include login, search, order entry, processing and a complete data export. For each critical path, expected results, test data and verification methods are recorded. Operational processes such as deployment, alerting and recovery are checked as well. Acceptance thus connects functional correctness with the ability to continue operating the application after the move.

A migration plan with clear decision boundaries

The plan distinguishes preparation, trial runs, approval, switchover and stabilisation. It shows which evidence must exist before the next step and who decides. Scope changes remain visible through their consequences for testing and operational risk. The final sequence, time required and open actions are derived from the trial run. A go or no-go rests on this evidence. Unclear prerequisites do not become unplanned decisions only once the switchover window has begun.

Frequently asked questions

Does an application need modernising before migration?

Only as far as the target environment or an agreed objective requires. A simultaneous platform change and extensive redesign increase the testing scope. We distinguish necessary adaptations from further improvements and decide based on compatibility, operational effort and the risk of a longer transitional phase.

How do you handle a limited switchover window?

We measure transfer and validation in a trial migration and investigate preparatory synchronisation options. The plan accounts for write access and business consistency. Whether the required window is achievable must emerge from these tests; data volume alone cannot determine it.

Take the task forward with purpose

Cloud Consulting

Assess architectural options, cost assumptions and the operating model together. An assessment provides the basis for the next technical decision.

Clarify cloud architecture

Cloud Operations

Define responsibility for changes, incidents and recovery. Prepare for an operational handover or develop your team's work further.

Develop an operating model

Discuss your project

Describe your application, environment and desired result. From this, we define a technical investigation or an actionable first step.

Discuss your plans

Which dependency makes your move difficult?

Describe your source and target environments, critical data and desired switchover window. We define an investigable migration path and the evidence it requires.