The Build or Buy Decision is All About Risk

Build or Buy?  The real answer is elusive.  Finding what is right for your enterprise is about understanding risk. A data center has a very specialized design. The build vs. buy decision isn’t about KW or cost per square foot. Sorry, I know a lot of dialog in this space goes right to looking for the one technical decision that will lead to a choice.

If you think of a data center as a function that includes services such as power distribution, HVAC, electrical systems, network, cabling, flooring, fire suppression, racking, compute/storage hardware and backup systems, then you begin to see it is a system that is more complex than the sum of its parts.

The technical complexity, along with funding and capital, are very much a part of the risk equation. Add to the equation that a computer room has to be maintained over time.

To assess risk, start with a few questions.

  • 1. Do you understand your business processes and know what level of availability those applications need? What is the cost of an outage over time?
  • 2. Do you know the level of availability you need to design into your computer room? There are four tiers of availability. Moving from one tier up to the next has both a financial and technical impact.
  • 3. Do you have the skills in-house to be a computer room management company? Once it is built, someone has to do the monthly tests, the analysis of capacity on all support systems and coordination of adds, moves and changes to all systems, including the compute, storage and network equipment.
  • 4. Do you understand the financial risk portfolio of your company and the preference for purchase for lease? Most CFO organizations can provide you the analysis they use to make those decisions.

The greater the requirement for availability and redundancy for your business critical processes, the higher the tier of computer room design you need. That equates to greater cost and complexity and a higher ongoing cost of support in both technology and trained personnel. That is the fundamental business risk question you need to ask.

Let me give some context to the risk issue. The question of build or buy comes down to who do you want to carry the burden of risk for specific elements of the data center. If you choose to buy, then the risk of providing a fully supported data center is on the shoulders of the provider, assuming you’ve signed a good contract.

If you decide to build, you are taking the burden of risk on your shoulders. That is not a technical decision; it is a business decision that has a technical component. To underscore how complex it is: Do you know whether the floor drains in your data center have check valves? If you say you don’t think you have floor drains, then that should raise a flag. A data center requires a very specific support infrastructure and supply chain. Fuel for your generators and potable water for staff are all part of the analysis.

In summary, the decision of build vs. buy is best made by taking a step back from the technical issues and putting it into a business context based on risk analysis.

Jerry Gentry writes for Data Center Knowledge on strategic issues relevant to enterprise class centers and is the VP, IT Program Management at Nemertes Research

Get Daily Email News from DCK!
Subscribe now and get our special report, "The World's Most Unique Data Centers."

Enter your email to receive messages about offerings by Penton, its brands, affiliates and/or third-party partners, consistent with Penton's Privacy Policy.

About the Author

Jerry Gentry is a research analyst for Nemertes Research.

Add Your Comments

  • (will not be published)


  1. Spot on Jerry. People love to get wow'ed by building things and start down paths too early in a desire to get their hands on technology. Building things is fun, but the hard part is up front, assessing risk, asking the hard questions on why applications need to be moved and the business processes that support these decisions. We work with our clients every day to pull them out of the technology and answer questions like yours. Else, you're bound to get the cart way before the horse.