A data center.
Image: Sashkin/Adobe Stock

While the cloud remains a relatively small segment of corporate IT, its growth is staggering. Much of that growth stems from greenfield applications that are born and live in the cloud. However, enterprises are also moving their existing legacy workloads to the cloud at a rapidly increasing rate to increase agility and save money.

SEE: Checklist: Building and decommissioning data centers (TechRepublic Premium)

With this growing cloud migration demand, there’s a consequent demand for services and advice on how to migrate on-premises, data center applications to the cloud. AWS, Microsoft Azure and Google Cloud are some of the big cloud players that recognize this need and advertise their expert assistance.

Shifting from on-premises to the cloud can come with a variety of challenges, including budget overruns and poor capacity planning. Enterprises might also suffer if they don’t go far enough into a modernization project, merely moving applications to the cloud without fully benefiting from all the cloud can offer.

In this guide, I’ll walk through some simple steps you can follow to maximize your chances of success with your data center migration project.

Jump to:

When to plan for a data center migration

One rule of thumb for timing a data center migration is to move once your hardware is at least three years old, which tends to be when enterprises consider replacing it. However, this instruction is too simplistic and fails to account for an even more important factor: the will to migrate. That motivation needs to be top-down, even if it will ultimately be deployed bottom-up.

SEE: Data governance checklist for your organization (TechRepublic Premium)

When I worked at AWS, then-CEO Andy Jassy would stress that enterprises serious about the cloud need to think in terms of transformation rather than incremental improvements. Such transformation can’t happen without enthusiastic and consistent executive support.

As such, the easiest way to answer when a company should plan to migrate its data center to the cloud is when there’s sufficient executive support.

Common challenges in migrating a data center

This brings us to the biggest challenge that companies face during data center migration. As with so much in IT, it’s not about the tech — it’s about people.

As Jassy once noted, enterprises need to think about the changing experience their customers will demand over time. He concluded that these experiences “usually [will require] a pretty big change or transformation.” Such transformations involve a fair amount of pain and effort to deliver, and executive support is critical to weathering any short-term setbacks.

For many enterprises, it’s so daunting to consider transformational change that some simply default to rehosting applications in the cloud — often called “lift-and-shift” — rather than rearchitecting them for the cloud. The upside of a lift-and-shift approach is that it’s relatively straightforward; the downside is that it may simply migrate old ways of operating to new infrastructure.

Still, as much as an organization may wish to migrate, the talent and expertise to do so may be lacking. Even if the plan is transformation, it can pay to start with a more incremental approach. Migrating non-mission critical applications, for example, is a way to build cloud migration muscle with the right processes and knowledge.

SEE: Hiring kit: Cloud engineer (TechRepublic Premium)

All of this points to what arguably should come second only to will/motivation, and that’s planning. Enterprises might, for example, neglect to note seemingly minor steps in their migration plans. After all, they’re so obvious that they don’t need to be written down, right? Unfortunately, that’s a poor assumption that can lead to all sorts of issues as the company tries to piece together the moving parts necessary to effectuate a migration.

A poor plan often starts with a muddled strategy, while the most successful migrations begin with a clear, broadly supported one. It’s important to ask questions like:

  • Why are we doing this?
  • What do we hope to achieve?
  • How will we get there and when?

Using these questions to frame your thinking should be directly followed by a detailed project plan.

Steps to a successful data center migration

Create a phased plan and business case for migration

After ensuring sufficient executive support for delivering a data-center-to-cloud migration, the first step is to plan. While the ultimate goal is business transformation, you need to establish a phased approach to getting there. This might involve various migration strategies for different classes of applications: Rehosting, replatforming, or refactoring and rearchitecting depending on how central they are to your customer experience.

SEE: Cloud data warehouse guide and checklist (TechRepublic Premium)

The approach that fits a given application best will often become obvious when building out the business case for migration. You’ll want to ask yourself: What is the business case for the move, in terms of time to market and revenue? Determine which approach will best deliver the expected outcome.

Inventory current assets and map to provider infrastructure

Part of that planning should also include taking inventory not only of your applications and their requirements but also cataloging the services your preferred cloud provider offers.

You’ll need to map out the migration path for each, including mapping your servers to the cloud’s machine types or different database services. For example, you may choose to move from a self-managed instance of MySQL to a fully managed database service to support your application.

Additionally, you’ll need to plan out the cloud infrastructure. If you’re using Azure, for example, Microsoft offers landing zones as an environment for hosting your migrated workloads.

Institute checkpoints to measure progress and support continued testing

You should also institute checkpoints from the start and constantly assess progress against goals. This assessment phase should include constant, continuous testing to optimize performance and resolve any blockers.

No matter how solid the migration plan, there’s nearly a 100% chance that some element of the plan won’t work as predicted. As a result, it pays to periodically check progress so you can retool your approach as necessary.

Assess migration successes and optimizations post-migration

Finally, even after the migration is complete, smart strategy demands continued optimization. This might include increasing automation to lower your operational burden, improving observability tooling and more. In other words, migrating to the cloud isn’t the end of a journey, but rather the beginning of optimizing your data and other business operations for data center and/or cloud infrastructure.

Disclosure: I work for MongoDB, but the views expressed herein are mine.

Read next: Top cloud and application migration tools (TechRepublic)