Data Center Expansion: Build or Buy?

4 comments

To build or buy your next data center is an age-old question for many business executives. It is sometimes driven by hard number-crunching business analytics, while in some cases it is controlled by human emotion and ego, or simply habit. Some issues can be clearly defined, such as cost and availability of capital and time to market. Other factors, such as market reputation and perceived strength, are a little fuzzier. This article series is designed to help C level business executives better understand the strategic implications and questions to be asked of your direct reports when considering whether to build, buy or lease your next data center.

There is an emerging trend toward outsourcing the data center as many organizations re-evaluate the need to build, own and operate their own data centers. They are either outsourcing it in whole or inpart, in an effort to concentrate their resources on their core business.

While building, owning and operating a data center could be seen by some as a normal part of any business with a P/L statement, it differs significantly for an enterprise customer vs. a colocation, hosting, managed services provider or cloud provider.

For enterprise data centers, the equation may not be as clear cut as an item on their normal P/L statement, since the “Profit” related to the data center is somewhat subjective, however reliable computing is critical to the operation of the business, while the “Loss” represents a potentially large financial risk factor should the data center experience an unplanned outage.

For colocation, hosting or managed services providers, the P/L somewhat resembles a hybrid of a traditional commercial real-estate operation, mixed with the seating revenue, economics and fuel efficiency issues of an airline. These are clearly understood and definable OpEx numbers to those organizations. However, even those organizations may not have the all in-house expertise to cost effectively build their data centers, and have outsourced the building and operations of their facilities.

Depending on your own organization’s type and size, performing an honest self-evaluation of core strengths and the level of design-build expertise and experience of internal resources is sometimes difficult. But it’s a necessary task in order to help make the best Build vs. Buy decision.

However, for all types of users the designing, building and operating of a data center is somewhat different than other buildings and requires a highly specialized level of expertise and experience. Even very experienced data center operators may not have the design and build resources necessary to undertake a major building project. The nuances and subtleties of a sophisticated data center design are somewhat similar to some other types of specialized facilities such as an oil refinery or chemical plant, all of which require highly specialized design skills and experience and may not be part of your internal team’s core skill set.

Over the next few weeks this article series examines the  following issues in more detail:

These are just some of the issues to understand in your diligence in making the Data Center Build vs Buy decision. You can download the complete Data Center Knowledge – Data Center Build vs Buy report, which includes all the articles in the series plus additional sidebars and analysis, compliments of Digital Realty Trust.

Julius Neudorfer is the CTO and founder of North American Access Technologies, Inc. and writes for Data Center Knowledge on issues and strategies relevant to seniors business executives.

About the Author

Add Your Comments

  • (will not be published)

4 Comments

  1. One of the factors that goes into such a decision, and that I didn't see mentioned in this article, is that in building a data center you can cut a lot of costs if you are willing to go through the used market for equipment. My company, Quantum Technology, buys and sells used data center equipment - often for a fraction of the cost of new equipment. Much of it has logged very few hours and is in great condition. Just another option to consider.

  2. I recently wrote a blog about this same topic from the point of view of "stranded" capital (http://www.switchscribe.com/?p=114). The true is that there are many unscientific factors like egos/myths that Julius hints at. Great stuff.

  3. I also wrote more specifically about the upfront/capital costs of building a data center in this article, Key Benefits of Leasing vs. Building a Data Center: http://resource.onlinetech.com/key-benefits-of-leasing-vs-building-a-data-center/

  4. This is always a lively topic. The company I work for, VISI, is hosting a seminar next week on this exact topic. We'll be discussing considerations such as as overall ROI, operational flexibility, future growth and virtualization. If any of your readership is going to be in Minnesota next week, you should check out the event at: http://www.visi.com/resources/events/data-center-decisions-build-versus-buy.aspx.