What Is Data Center Migration? Strategies & Best Practices

Data center migration vs. data center relocation

Relocation usually describes the physical move of equipment from one site to another. Migration is a broader term that also covers the logical side: moving workloads, re-addressing networks and cutting over services. Many projects are both at once. A company may relocate racks to a new colocation suite and migrate some workloads to new hardware at the same time.

We focus on the physical side, including planning, de-install, packing, transport and re-rack, and we coordinate with your teams on the cutover. Our data center migration services page describes how that division of work usually looks.

Common types of data center migration

Most projects fall into one of a few patterns:

• Physical relocation: moving equipment to a new room, building or city
• Colocation migration: moving between facilities, or from an office server room into colocation
• Consolidation: merging several rooms or sites into fewer locations, covered on our data center consolidation page
• Cloud migration: moving workloads to a cloud provider, often with hardware retired afterward
• Hardware refresh: moving to new equipment and retiring the old

Data center migration strategy: the main approaches

The strategy is the central choice in the project, and it is driven by how much downtime the business can accept and how much risk the team can carry.

Lift and shift moves equipment largely as it is, often with racks shipped loaded to cut re-cabling time. It is simple to explain and works well when the destination matches the source. A phased migration moves the estate in waves, so a problem in one wave does not stop everything. It takes longer but contains risk. A swing or parallel migration builds a temporary or new environment first, then moves workloads across the network before any hardware moves. It gives the shortest outage but needs extra equipment and capacity.

Many projects combine these. A team might move non-critical systems in a first wave to rehearse the process, then use a swing approach for the systems that cannot go offline.

Data center migration methodology

A sound methodology is repeatable. Whatever the size of the project, it tends to follow the same sequence: discovery, inventory, planning and scheduling, preparation of the destination, de-install, packing, transport, re-rack, testing and close-out. Each stage has an owner, a deliverable and a check before the next one starts. Our data center migration checklist lays out the data center migration steps in this order, and the project plan page shows how the phases fit together.

The key idea is that nothing moves until the previous stage is verified. The inventory is checked against the asset register before packing, and the delivery is checked against the inventory before power-on.

Best practices for a data center migration

• Start planning early. Lead times for carrier circuits, facility access and crating are measured in weeks
• Build the inventory from the floor, not only from a database, and reconcile the two
• Test your backups by restoring them, and write a rollback plan before you need one
• Label everything at both ends and photograph racks before and after
• Move in waves when the estate is large, and rehearse with lower-risk systems first
• Keep a custody log and verify the load at each handoff
• Decide what not to move. Retiring end-of-life equipment saves transport and rack space
• Name one decision maker and one owner per phase

Common risks and how to reduce them

Most failures are planning failures. Undocumented dependencies cause a system to start in the wrong order. Unlabeled cables turn re-racking into guesswork. Destinations that are not ready force the move to wait. Hardware that has not been powered down cleanly can fail to come back.

Reduce these risks by recording dependencies in the plan, labeling and photographing before de-install, confirming the destination in writing and following a documented shutdown and power-on order. If a part of the move can fail, plan what happens next.

Where equipment disposition fits

Every migration leaves something behind: aging servers, spare networking gear, retired racks. Treating this as part of the plan, rather than an afterthought, lowers the size of the move and can offset its cost. Our buy back program buys servers and storage, network gear, and racks, UPS and power equipment, paying cash or giving credit toward the move. We quote it per project. For retirements without a move, see data center decommissioning.

How to get started

The best first step is a short scoping conversation: what you are moving, where, by when and what depends on it. From there we can recommend a strategy, outline the phases and give you a written quote. Request a quote to start, or read our guide on what a move costs to understand the factors involved.