Thinking Strategically about Data Centers

29 Oct

Sloth“I’m sorry sir, but the decision to place an office in Thailand puts a serious strain on our current infrastructure.” What do you mean by a “serious strain”?  “Well,” I said, “our primary DCs are all in the Eastern US and our legacy applications will be severely impacted by the 250 milliseconds of additional latency.  We also don’t have any network agreements with providers in Asia, so it will take some time to get contracts in place and they will likely be costly.”

Locked in to infrastructure choices

Why would you want to lock yourself in to infrastructure choices that could lead to the above scenario?  What about needing quick regional capacity or access to cloud providers that can extend your reach globally? Unfortunately, the aforementioned situations, along with too many others to list, can have a serious impact on your company’s growth strategy.  In a previous blog, I talked about the data center as a “15 year business plan,” the basic concept being that there’s no such thing as a workable 15 year business plan. The future of IT in two words is “agility” and “innovation.”

Why would you build applications and infrastructure that can be modified, extended, upgraded or replaced quickly if your facilities don’t support that? Hosting modern applications designed for cloud in a facility that is bound by the skills and the resources available to your enterprise. What about being bound to a physical construct that takes over 18 months to complete and which remains on your books for 10-20 years? It just doesn’t make sense.Locked old building

Pragmatic decisions

All of us who have been in IT for any period of time have had to make pragmatic decisions. If you don’t think you have been, then you need to review your ability to measure success and weigh risk/value.  In many cases we make pragmatic decisions or fail to implement a reasonably decent solution because the long term risks of ownership cast dark clouds over our choices. In other cases, we make these same pragmatic decisions believing we have no workable alternative; consequently finding ourselves stuck with something we either can’t use or that shackles us and limits our options. The funny thing about these shackles is that most of us don’t even know we have them, as the manifestation of the shackles in our daily work is often higher levels of fire-fighting and or statements like “we can’t do that because…”.

As a former leader of infrastructure teams, I rarely asked myself many of the tough questions. Instead I focused on finding justification for building it myself.  After all, who in IT doesn’t like the idea of building an empire? Now I would ask why would I limit my options for speed, skills, and resources to that which can be provided only through use of my internal data center environment and its associated limitations of access, scale and performance?

The Empire of Today

The empires of yesterday were fine for the times, but as they say “times they are a changin’.”  If your boss were to ask you to build or buy a new application, would you buy legacy client server designs? Would you plan on building unique vertical infrastructure to support said application? No, you wouldn’t; any IT leader worth their salt today is looking at “cloud first,” and then finding an alternative if cloud absolutely won’t work.  So, if we’re smart enough to think cloud first, why would many of us still think “build a data center” and then force fit every choice we make into it?

The hybridized model for IT going forward requires an entirely new strategy from IT in the way resources, people, and infrastructure are acquired and applied to solve problems.  In this hybridized environment, you will likely be partnering with suppliers in a much deeper way than ever before, and you’ll be distributing your capabilities across several cloud providers, storage providers, and locations.  Again, assuming the hybridized model rings true, how will you support the new models adoption from the data center perspective? Will it be to try to find the most “pragmatic” way to build the “right” sized and scoped facilities in the “best possible” locations?  Assuming your decisions are still correct five years into the future, what will you do about the need to work in proximity to myriad providers? You won’t be able to do that in your internal DCs? What about the need for diverse cost effective, connectivity solutions? Can you really convince yourself that you can compete on network pricing and options against the best colocation business that is designed around enabling that very capability?

Design for Tomorrow

Most data centers being built today—even those by many colocation, hosting, and cloud companies—are testaments to the slow nature of change in the data center market. These data centers are often shining examples of what a great data center was in 2007. So when you go about the process of determining the best approach for enabling your business to grow well into the future, pick wisely and pick based on how you expect IT will be operating in two to five years.  The best approaches to IT operating models will include the ability to adopt, change, update, and replace solutions quickly and to use them only for as long as they are the best fit.  In other words, we’ll be moving away from treating IT solutions as possessions and moving towards thinking of them as consumable tools and resources.


Don’t get caught flat footed building a 10, 25, or 500 million dollar albatross for your business. Instead, find ways to make your organization more valuable by designing for rapid adoption and geographic elasticity.  The best solution is always the best solution. You can’t hope to be providing the best solution for your business if your primary focus is always “how do I build this myself?”.  Find your way to freedom, even if you have to crawl, the business will reward you, and your career will accelerate if you apply the “best” options; so leverage the people, tools, and resources available to increase your value to the organization.

Turtle on beach

Supporting Blogs:

Professionally copy edited by Kestine Thiele (@imkestinamarie)

Did you like this? Share it:

Tags: , , , , , , , , , , , , , , , , , , , , , , , , , , , ,

  1. George Cvetanovski

    October 30, 2013 at 4:06 am

    Cloud providers that aren’t offering flexible Private / Hybrid Cloud solutions on their clients’ land must start now.

    • Mark Thiele

      November 6, 2013 at 6:03 pm

      Hi George,

      I would agree. In fact, many companies (HP, VMware, Virtualstream, and others) are attempting to do just that.