Skip to content
North Ark

Azure Migration

Migrate to Azure in waves that follow the dependencies, not the server list.

Most migrations stall on identity, DNS, shared services and network paths that nobody mapped. We plan the waves around those, build the landing zone before the first workload moves, and leave an environment that can be operated.

  • Fixed-fee phases
  • Microsoft Partner
  • AZ-104 and AZ-305 certified engineers
  • Senior engineers only
  • Brisbane-based, working across Australia
Start with an Infrastructure Review
Small server rack alone in a glass-walled room, with cloudy sky reflected in the glass.

Discovery that finds the real dependencies

  • Application and server dependency mapping from network flow data, not only CMDB records
  • Identity: what authenticates against what, service accounts, Kerberos and NTLM use, hybrid identity health
  • DNS and name resolution, including hard-coded addresses and split-horizon zones
  • Shared services: file, print, certificates, licensing, backup agents and monitoring
  • Data gravity: which databases and file shares set latency and cutover windows
  • Retire, retain, rehost, replatform or rebuild decisions per workload, with the reasoning written down

The landing zone comes first

Workloads land in a zone that already has management groups, policy, hub networking, private DNS, logging, identity boundaries and backup in place. Building it first means wave one does not become an accidental design.

Connectivity to the existing WAN, whether ExpressRoute or VPN, is proven with production-like traffic before cutover, including failover behaviour.

Waves, cutover and rollback

  1. Wave design groups workloads by dependency and business calendar, starting with a low-risk pilot wave that exercises the full pattern.
  2. Each cutover has a written runbook, a validation checklist agreed in advance and a defined point after which rollback stops being cheap.
  3. Replication, cutover and post-cutover validation are timed against data change rates and application owner availability.
  4. Decommissioning is part of the plan, so the old estate does not linger and keep costing money.

Data centre migration and exits

When a data centre lease, hardware refresh or colocation contract ends, the whole estate has to move by a fixed date. We plan the exit back from that date: every server, appliance and circuit is inventoried, each workload gets a destination in Azure, Azure Local or a smaller on-premises footprint, and the waves are sequenced so the facility can be emptied and handed back on time.

When Azure is not the answer for a workload

Some workloads are better on Azure Local or staying on-premises: latency-bound systems, licensing that changes the maths, hardware-attached applications. We say so per workload and design the hybrid path to match.

Handover to operations

Monitoring, alert routing, patching, backup and DR tests are live before the last wave completes. Runbooks, diagrams and decision records go to whoever operates the environment, which can be your team or North Ark under Managed Engineering.

Questions buyers ask

How long does an Azure migration take?

It depends on the number of workloads, their dependencies and how quickly application owners can test. Discovery produces a dated plan with waves.

Should we lift and shift or modernise?

Decided workload by workload. Some move as they are, some are replatformed onto Azure services, and some are retired. The reasoning is written down.

Do we need a landing zone first?

Yes. A minimum landing zone with identity, networking, DNS, logging and backup comes before the first production workload.

Do you account for licensing?

Yes. Windows Server and SQL Server licensing, including Azure Hybrid Benefit, is part of the plan and the cost model.

Can you move us out of our data centre?

Yes. A data centre exit is planned back from the lease or contract end date, with every workload given a destination and a wave, and decommissioning built into the plan.

Can you take over a migration that has stalled?

Yes. We assess what has moved, why it stopped and what needs to change, then restart with a dependency-led plan.

Who runs the environment afterwards?

Your team, with runbooks and code, or North Ark under Managed Engineering.

Related

Start a conversation

Plan the migration around what depends on what.

Tell us about the estate you are moving.

Talk to an Engineer