Disciplines · 05 · Migration

Nobody loses Monday morning.

A migration is a move, and moves are where businesses get hurt. Email down for two days. A VM that came up without its data disk. A tenant cutover where nobody had the rights to change an MX record. The work isn't the tool — the tool is usually fine. The work is planning, batching, rollback, and knowing who actually holds the keys before anyone touches a button.

The discipline, not the tool

Migration isn't one thing. The discipline is.

Email from Exchange to Microsoft 365. Mailboxes out of a hosted provider. A full Google Workspace tenant folding into M365. A VMware estate being re-platformed onto Hyper-V because Broadcom's licensing changed the math overnight. Servers lifted and shifted out of a datacenter that's being evacuated. File shares pulled into SharePoint. A tenant-to-tenant move after an acquisition.

Different source, different target, different tooling — and yet the failure modes are almost always the same. The migration itself is the discipline. Do that well and the specific product choices mostly take care of themselves.

The signature visual

A pre-flight checklist, not a battle plan.

Good migrations look boring from the outside. That's because most of the thinking is done before anyone touches a button. Five things have to be true before a migration window is a safe bet.

Pre-flight — five pillars

Before anyone touches a button

  1. 01

    Planning & batching

    Who moves in which window, in what order, with what verification between batches. A full tenant on one weekend is a story; batched waves with clear gates is a plan.

  2. 02

    Rollback plan

    Every cutover step has a corresponding "how we back out" step, with the time budget and decision owner written down. If rollback isn't mapped, you don't have a migration plan — you have a hope.

  3. 03

    Architectural strategy

    The target state is designed before the move, not discovered during it. What's getting re-platformed, what's being consolidated, what stays where it is. The migration is a tool for the strategy, not the other way around.

  4. 04

    Communication

    The business needs to know what state the migration is in so it can plan its day. Users, team leads, help desk, executives — each one gets the right level of detail at the right moment.

  5. 05

    Asset control

    Confirmation — on paper — that the business actually controls every asset the migration will touch. DNS, tenant admin, physical access, vendor portals, break-glass accounts.

Zoom on pillar 05

Do you actually have the keys?

This is the pillar that catches people out, so it's worth saying clearly. Before the migration window opens, we need documented confirmation — not assurances, not "probably" — that someone on the team can, right now:

  • Change a DNS recordat the authoritative provider, including MX and TXT
  • Sign into the tenantwith admin rights, and isn't blocked by someone who's left the company
  • Get to the hypervisorphysically or via out-of-band, in case a VM goes sideways
  • Open the vendor portalfor every SaaS tool in scope — especially the identity provider
  • Reach the break-glass accountwith the password and MFA method both verified this week
  • Escalate to the incumbent MSPif one is still technically on the hook during cutover
Field note The worst time to find out you can't change an MX record is the morning the mailboxes are supposed to stop flowing to the old system. Ask the question weeks before, on paper, and watch someone actually log in.
What goes wrong without it

Without a plan, migration looks reactionary.

Without a plan of attack and without experience, the move turns into an exercise in responding to surprises. You're mid-cutover and someone asks for the password to the DNS console — nobody in the room has it. You're about to kick off a VM migration and realize nobody confirmed there was bare-metal access to the source hypervisor. You've scheduled a rollback window but never actually tested whether rollback is even possible for the workload in question.

Each of those is a sudden pivot, and sudden pivots are where migrations break. The goal isn't to eliminate surprises — there will always be some — it's to make sure the migration itself isn't the thing generating them.

Where clients usually start

The three moments migration comes up.

Migration is rarely the thing that gets initiated for its own sake. It's almost always downstream of something else — a price shock, a lease expiry, a hardware refresh that turns into a cloud conversation.

Moment 01

The Broadcom moment

VMware licensing changed the math. Teams that were comfortable for years are suddenly re-costing the estate. Hyper-V on Azure Stack HCI, Azure Local, or a straight lift-and-shift to Azure all enter the conversation. We help businesses work out what's actually cheaper once the full picture is in.

Moment 02

The hardware refresh moment

The server's five years old and a refresh quote just landed. That's the cheapest time to look hard at whether another box on the floor is the right answer — or whether a subset of the workload can move to cloud or managed infrastructure instead.

Moment 03

The datacenter evacuation

The colo is closing, the lease is up, or the parent company has consolidated real estate. Whatever the reason, there's a date on the wall and a rack that has to be somewhere else by then. This one is pure logistics — but it's also where the preflight checklist earns its keep.

What an engagement looks like

The discovery that finds what the business forgot.

Every engagement begins with an assessment — and an assessment that's just an inventory is half a job. The other half is business discovery. That's where the migration stops being a technical exercise and starts being a plan the business itself can defend.

Half one

The business discovery

Team-lead interviews across the departments that actually touch the systems being moved. How do they use email? What do they share, with whom, on what cadence? Which files must be live on day one, and which ones can stagger in?

This is where we usually find things the business itself had forgotten — a shared mailbox nobody owns, a service account that quietly runs payroll, a VM that reboots every night because someone set it up that way in 2017 and the reason is lost to history.

Half two

The technical discovery

Critical systems, dependencies, integration points, licensing posture, identity provider, DNS ownership, firewall rules, backup posture, and the policies that govern all of them. Tooling helps here — Azure Migrate, Exchange assessments, tenant-to-tenant discovery reports — but the judgement is in knowing what matters.

The output is a target architecture, a migration waves plan, a rollback playbook, and a communication cadence — tied together in one document the business can sign off on before any cutover happens.

"We avoid sudden pivots — the ones where someone realizes mid-cutover that nobody in the room can change an MX record, or that nobody checked for bare-metal access to the source hypervisor. Preventing those moments is the whole job." Why the checklist exists

Book a call

Migration on the calendar?

If there's a date on the wall — a VMware renewal, a datacenter lease, a tenant cutover — the conversation worth having is the one before the plan gets drawn. We can sit with you, walk the pre-flight, and tell you honestly what you already have and what's missing.

Or reach us directly: info@fouronesixit.ca · (647) 371-0400