Workload inventory and assessment
Every server, database and scheduled job catalogued with owner, dependencies, data volume, licence terms and a treatment: rehost, replatform, refactor, retire or retain.
Service · Cloud
Commission us to move running systems from data centres, colocation or another provider into the cloud, one planned wave at a time. Each move has a rehearsed cutover, a rollback route and a named owner, so the business keeps operating while infrastructure changes underneath it.
Dependency map before any move
Rehearsed cutover for every wave
Rollback route agreed in advance
Deliverables
Every server, database and scheduled job catalogued with owner, dependencies, data volume, licence terms and a treatment: rehost, replatform, refactor, retire or retain.
Workloads grouped into waves by dependency and risk, with lower-stakes systems first so the process is proven before critical ones move.
Destination accounts, networks and connectivity back to systems staying on-premises, ready before the first wave starts.
Step-by-step plans per wave covering data sync, freeze windows, DNS changes, smoke tests, go/no-go criteria and who makes the call.
Confirmation that source systems are shut down, retention obligations met and licences or contracts ended, so you stop paying twice.
Early decisions
Rehosting is fastest but carries old inefficiencies across; replatforming onto managed databases or containers costs more effort but removes upkeep. We decide per system, not per estate.
Native replication, change data capture or offline transfer, chosen by data size, acceptable downtime and available bandwidth.
The longest acceptable outage per system, agreed with the business, which decides between a weekend window and near-continuous replication.
Whether existing database and operating system licences can be brought across, or whether licence-included cloud services work out cheaper.
VPN or dedicated links such as Direct Connect, ExpressRoute or Cloud Interconnect, sized for replication traffic and for systems left behind.
Engagement
Agent-based or network-flow discovery reveals which systems talk to each other, including undocumented connections that break cutovers.
A small, low-risk group moves first, testing tooling, runbook format and the communication plan with real users.
Each wave follows rehearsal, freeze, sync, cutover and hypercare, with lessons from the previous wave written into the next runbook.
Once workloads settle, we rightsize against observed usage, retire the source environment and close with a before-and-after cost comparison.
Fit
A fixed exit date makes sequencing essential, because rushed final waves carry the most risk.
Where nobody is sure which applications share a database, discovery and wave planning prevent outages.
Migrating with minimal change, then improving workloads once in the cloud, spreads risk across two smaller steps.
Alternatives
If the current application is near the end of its life, building fresh on managed AWS services can cost less than migrating it.
Open registerWithout landing zones and networking in place, build the foundation first; the migration follows.
Open registerWhen working practices need rethinking too, a transformation programme tackles both together.
Open registerEach workload can be rehosted, replatformed, refactored, repurchased as SaaS, retained or retired. Assigning a treatment per system, with the reason recorded, keeps the programme honest about effort.
Retiring unused systems before migration is often the cheapest saving on offer.
Configuration databases rarely match reality. A reporting job that reads a production database nightly, or a file share mounted by three applications, will break if moved in the wrong order. Discovery tooling and network flow data give us a dependency map that waves are planned against.
Every cutover is rehearsed against a copy of production data before the real window, with timings recorded. Rehearsal finds the slow export, the missing firewall rule and the hard-coded IP address while there is still time to fix them.
After each cutover, a hypercare period watches error rates, performance and user reports closely. Once agreed criteria are met, the source is decommissioned. Leaving it running for months doubles cost and blurs which system holds the truth.
FAQ
Almost never. A single big cutover concentrates every risk into one window. Waves let you learn from early moves and keep trading if one system has to roll back.
We replicate continuously into the target using native replication or a service such as AWS Database Migration Service, then switch during a short window once the replica has caught up.
Each runbook has go/no-go checkpoints and a rollback route returning traffic to the source, which stays intact until the new environment has run cleanly through hypercare.
Not automatically. Rehosted servers sized like their on-premises predecessors often cost more. Savings come from rightsizing, managed services and switching off idle capacity, which we address after the move.
Related services
Controls checked on the destination before sensitive workloads arrive.
Open registerContainerising applications that are being replatformed during the move.
Open registerMoving commercial databases onto managed PostgreSQL as part of replatforming.
Open registerPrimary sources
Dated official documentation, standards, and research that support the factual claims on this page.
How to read this page
Scope, cost and timeline figures are planning guidance until confirmed in a written proposal. Diagrams and screens are illustrations, not client deliverables, and source-code rights and handover follow the signed agreement. See the delivery process and the full disclaimer.
Next decision
Define outcomes, constraints, evidence, rights and handover before delivery begins.
Commercial rights, repositories, environments, documentation, acceptance and handover remain contract-defined.