Service

System migration: once, and done right

A server replacement, the end of support, a move to the cloud: migrations are the projects where the most can go wrong – and which you should do right exactly once. Since 1992 we have completed over 1,200 migration projects, from single servers to the changeover of several sites.

Over 1,200 completed projects Parallel operation instead of a hard cutover A way back always prepared

Triggers

When a migration is due

Rarely by choice. These six triggers bring companies to us – the first three with a deadline looming.

  • A version reaches end of support. Without security updates, a system that runs reliably becomes a risk. The date is fixed and cannot be negotiated.
  • The hardware has reached the end of its service life. No more maintenance contract, no spare parts. The replacement can be planned – or it happens unplanned on a Monday morning.
  • A contract is expiring. Hosting provider, data centre, licensing model. If you start too late, you end up renewing on worse terms because you have no choice.
  • The business has grown. What was built for twenty workstations no longer copes with sixty – usually the first thing you notice is the speed.
  • Working from home and multiple sites. An environment designed for one building has to be set up differently when half the staff access it from outside.
  • Compliance requirements arrive. Clients, certifications or NIS2 require separate permissions and documented processes – that is the moment to clean up structures that have grown over time.

Scope

What we migrate

The most common projects – on their own or as part of a larger overhaul.

Windows Server

From an older version to a current one, from physical to virtual hardware or into a data centre. This includes Active Directory, file services, print services and the line-of-business applications that run on them.

  • Active Directory and permissions
  • File and print services
  • Line-of-business applications and their databases

Exchange Server

Since 14 October 2025, Exchange Server 2016 and 2019 no longer receive security updates. There are three options: move to Microsoft 365, switch to Exchange Server Subscription Edition, or run a hybrid setup.

  • Mailboxes, calendars, public folders
  • Distribution lists and permissions
  • Rules and forwarding

On-premises to cloud

File storage to SharePoint, mailboxes to Exchange Online, servers to Azure. Not everything at once, but in the order that disrupts your business least.

  • Microsoft 365 and Azure
  • File storage with permissions
  • Step by step instead of a hard cutover

Hybrid environments

Part stays on premises, part moves to the cloud – often the right answer when a line-of-business application cannot run in the cloud or data should not leave the building.

  • Directory synchronisation and sign-in
  • Hybrid Exchange connection
  • Access from both sides

Databases and line-of-business applications

ERP systems, accounting, industry-specific software. The hardest part is rarely the database itself, but coordinating with the vendor and deciding what happens to the interfaces.

  • SQL Server and other databases
  • Coordination with the vendor
  • Interfaces taken into account

Relocation and data centre

Moving premises, consolidating after an acquisition, or moving from your own server room into a data centre. What counts most here is planning the cutover day.

  • Site consolidation
  • From server room to data centre
  • Lines and fallback

Four routes

The migrations we carry out most often

Each route has its own pitfalls. What they all have in common: the effort lies not in the changeover but in the preparation.

On-premises Exchange to Microsoft 365

The most common case. Mailboxes, public folders, distribution lists, rules, signatures and the devices that send via the mail server – printers with scan-to-email, ERP systems, time recording. The devices are almost always overlooked and are noticed the day after.

  • Mailboxes run in parallel beforehand
  • Identify devices that send email beforehand
  • The old environment stays in place for now

Windows Server to a new version

When a server version reaches end of support, there are no more security updates – regardless of how reliably the system is still running. The change is rarely a simple switch; it is a rebuild with data transfer.

  • Check line-of-business applications are supported
  • Set permissions afresh instead of carrying them over
  • Old server stays as the way back

On-premises systems to the cloud

Virtual machines from Hyper-V or VMware to Azure, file storage to SharePoint. The technical move is usually the easier part; what matters is bandwidth, latency and whether the line-of-business application supports running in the cloud at all.

  • Measure data volumes and connection beforehand
  • Get the vendor's approval
  • Fallback for the internet connection

Change of data centre or provider

A hosting provider is changed, a contract expires, a site relocates. What counts most here is completeness: logins, certificates, DNS records, licences and contracts that are registered with the old provider.

  • Record all logins and DNS entries
  • Cutover with a quick way back
  • Dismantle the old environment only after sign-off

Approach

How we run a migration

The principle is always the same: the new system is running before the old one is switched off. A hard cutover with no way back is the most common reason migrations fail.

  1. Assessment

    What runs where, what depends on what, which interfaces exist, which licences and contracts are affected. The effort put in here decides how the whole project goes – surprises almost always surface at this step, or not in time at all.

  2. Target design and sequence

    What moves where, in what order, and what deliberately stays where it is. Plus the awkward question: what happens if a step fails? The way back is defined in advance, not improvised.

  3. Trial run

    The migration is rehearsed on a small sample – one mailbox, one folder, one database. This reveals what the assessment missed, without any consequences.

  4. Pilot group

    Five to ten people work productively in the new system for a week, deliberately drawn from different departments. They are the ones who notice what gets in the way day to day – not us.

  5. Changeover in parallel operation

    The old system remains accessible while the new one takes over. For mailboxes and files, this means both are available for a while, and nobody is left without access.

  6. Follow-up and decommissioning

    For two to four weeks after the changeover, we fix whatever comes up in day-to-day use. Only then is the old system switched off – and even then it is initially just switched off, not deleted.

From experience

Why migrations actually fail

After over 1,200 projects, it is the same six points again and again – and rarely the technology.

  • The assessment was too thin. An interface nobody thought of. A device that works via an old sign-in method. A script that has been running every night for years and that nobody can put a name to.
  • Nobody asked the vendor of the line-of-business application. Some applications are not approved for running in the cloud or require a specific version. You clarify that before the project, not after.
  • There was no way back. If the old system is switched off on changeover day, every fault is an emergency. If it stays in place, a fault is merely a nuisance.
  • Permissions were carried over instead of reviewed. A migration is the best opportunity to clean up access rights that have grown over time. If you transfer them one to one, you take the mess with you into the new system.
  • Nobody talked to the users. Whatever changes in their daily work has to be explained beforehand – otherwise on day one it looks as if something is broken.
  • The timing was bad. Not just before the year-end close, not in peak season, not on a Friday. If there are things to fix, someone has to be there the next day.

Frequently asked questions

What we are often asked about this

How long will our business be down?

As a rule, not at all. Because we switch over in parallel operation, the old system keeps running while the new one takes over. There are short windows for individual steps – switching the mail flow, for example – usually outside business hours and lasting minutes.

Physically moving hardware is a different matter. For that we plan a weekend and keep a fallback ready in case of emergency.

We run Exchange 2016 on premises. What now?

Since 14 October 2025 it no longer receives security updates. An Exchange server without updates is reachable from the internet and therefore a favoured target – this is not a theoretical risk.

There are three options: move the mailboxes to Microsoft 365, switch to Exchange Server Subscription Edition, or run a hybrid setup. Which one fits depends on whether devices and applications on your premises still send mail via Exchange, and whether the data is allowed to leave the building. We clarify this in the assessment.

Can we migrate step by step, or does everything have to move at once?

Step by step is almost always the better route, and for mailboxes and files it is technically built in. Some users are already working in the new system while the rest are still in the old one – and both groups can see each other.

With databases and line-of-business applications this is less often possible; there is usually a real cutover point. That is exactly why it gets rehearsed.

What happens to the old data?

Initially it stays where it is. After the changeover the old system is switched off but not deleted straight away – usually it is kept for several weeks, often longer if retention periods apply.

In addition, every migration starts with a full backup whose restore we test beforehand. A backup you cannot rely on in an emergency is particularly expensive in a migration project.

What does a migration cost?

That depends on the number of mailboxes or workstations, the number and condition of the servers, the line-of-business applications involved, and how well the existing environment is documented. That last point regularly makes the biggest difference.

We quote a fixed price for the assessment and then a reliable estimate for the implementation. After the assessment you can stop without any obligation and take the results with you.

Do you also take over a migration someone else started?

Yes, and it happens more often than you might think – usually when a project has stalled halfway: half in the cloud, half on premises, and nobody dares to go further.

Even then, we start with an assessment. Taking over a half-finished state without having understood it regularly goes wrong.

Over 1,200 migrations – what counts towards that?

We count completed migration projects since the company was founded in 1992, not individual mailboxes or computers. A project can be the replacement of a single server or the changeover of several sites with hundreds of workstations.

Over 34 years, that is roughly three to four projects a month. By far the largest share are Windows Server replacements and Exchange migrations – the two tasks that come up again and again in small and medium-sized businesses.

Can you guarantee a migration without downtime?

Not as a blanket promise – and anyone who makes one has not looked at your environment. How much downtime there is depends on the volume of data, the line-of-business applications and whether parallel operation is technically possible.

What we do promise: before the changeover, you will know what interruption to expect, when it will happen and what the way back looks like. In many cases the interruption can be limited to a weekend or a few hours – but that is the result of planning, not a promise made in advance.

Free assessment

Before we migrate, we take stock.

We record servers, applications, interfaces and dependencies and tell you which route is right for your business – even if the answer is that you should not change anything for now.

+49 221 984300-0Switchboard and support hotline

[email protected]Reply within 4 hours on working days

Robert-Perthel-Straße 7250739 Köln – Bilderstöckchen

Mon–Fri 9 am–6 pmEmergency support outside these hours by arrangement