With many enterprises migrating critical workloads and services to cloud environments over the last two years we’re seeing an inflection point where IT leaders are now being forced to make a definitive decision about what they’ll do with all those legacy data centers. But not all service providers are equipped to deal with the almost traditional work of migrating the last bits of data and then tearing down data centers to the last brick. IT Leaders must adapt their cloud strategy to include a clear plan for what happens with their existing data centers and, crucially, assess which partners will be able to deliver the goods.

Even though enterprises began pushing services into cloud environments several years ago, legacy applications, regulations, and culture necessitated on-premise data centers to effectively continue as normal. However, with the maturation of cloud solutions across both public and private clouds, on-premise data centers are finally ending up on the scrap heap. Anyone with a modicum of experience in the space should understand that it’s not as simple as locking the doors and wandering off. Migration alone is resource intensive— and as demand for this type of work has increased smart professionals have chased it, meaning that the core skills and tools enterprises require to break down a data center are increasingly scarce.

Enterprise leaders must find the right partner with the experience to execute data center takedown activities from end-to-end, and in a complex market, they must pick carefully. As not all providers have nurtured core infrastructure management capabilities – instead chasing selective engagements in over-hyped digital technologies. Two providers, however, have recently talked us through how they are taking this growing, albeit less-glamorous, part of IT transformation.

DXC has gained some real-world experience from rationalizing its own sprawling estate

It’s fair to say that vendors look at projects a little differently when they’re internal, which gives DXC an interesting story to tell when it comes down to the end-to-end rationalization of sprawling DC estates. When two major IT services companies, HPE and CSC,  plus a raft of smaller acquisitions combined to create the leviathan firm, it faced a major challenge. With all of its constituent parts pushing their own private and public cloud narrative before the merger, the combined entity was saddled with a glut of capacity spread across the globe in almost 100 data centers. With the increasingly positive market perceptions of public cloud, particularly among the big hyperscale players, infrastructure services on the sell-side were already evolving.

So DXC had a simple choice, rationalize its sprawling estate or risk its spread of assets becoming an anchor that consumed resources that could be better spent elsewhere. Much unlike the migration conversations we’ve had over the past few years, DXC needed to scrap its data centers completely—not just modernize and migrate workloads. In a few short years, the firm reduced its sprawl from over 90 data centers to less than 60, with a blueprint for further consolidation still in the cards. According to firm leadership, this consolidation has created a unique pool of talent and expertise that the firm can offer clients that are facing similar challenges, managing the end-to-end process from app modernization to pulling the final bricks out of the server room.

Atos evolves traditional outsourcing capabilities to support data center takedown

Atos, too, has spotted the market’s growing need for service providers to enter client organizations with a clear vision and understanding of how to pick up, manage, and deconstruct a data center over a fixed period. For Atos, it becomes a matter of simple economics and a return to commercial models that have generally fallen out of fashion in the digital and IT services space.

It starts by flipping the commercial model on its head. According to Atos executives, there is a perception in the market that vendors are incentivized to take their time migrating from legacy data centers to cloud environments. And why not? They’re paid as long as they manage the assets, so they might as well pack deals with extensive and unnecessary migration and management services and keep the gravy train running as long as possible. This is where the opportunity is for Atos. To break this market perception, the firm builds a fixed timeline, which it tries to make much shorter than usual. The fixed project timelines wrapped into the commercials reassure clients that Atos won’t lock them into a pricey asset management gig—something Atos says is resonating well with the market.

Atos has constructed a model that works for all parties and that enables the company to accurately calculate cost benefits. Atos buys the asset and takes it off the client’s books, then runs and collapses the data center within a fixed timeline. Crucially, the savings are balanced across the engagement so that clients get immediate benefits and find it much easier to gain senior buy-in. Another interesting element to this story is Atos’ close work with its hyperscale cloud partners—for example, the firm advised that it’s working with Google on similar projects, where both firms cooperate and share the risk and investment.

The Bottom Line: As the cloud market matures, enterprise leaders will need to rethink how they approach the final termination of data center assets. Finding a partner with the expertise and the right model will be key.

There’s a story that’s often circulated about the shortsightedness of market forces. In this story, a market depression stopped investment in road infrastructure projects for several years. The only challenge was that when roadwork started up again, many of the companies that had serviced the space had gone bust or moved into other lines of work because of the market drought. It’s possible that we will see a similar, albeit less acute, theme emerging in the infrastructure and cloud services space. With service providers pushing selective offerings and steering clear of heavy-lifting infrastructure projects in favor of higher value digital engagements, we are losing the skills, talent, and capabilities in a section of the market that is likely to boom as the public cloud continues its onslaught against traditional services. The reality is, enterprises will need to do something with their creaking data centers in the future, once their cloud-first strategies mature. Finding a partner that can help migrate the last application and see the job through to the final bulldozing of the server rooms may be much harder to find.