“Whether tis nobler in the mind to suffer the slings and arrows of outrageous [outages]. Or take arms against a sea of troubles, and by [colo’ing] end them?”
– William Shakespeare
Well maybe that quote is a little melodramatic, but for IT infrastructure managers who have been outraged by an in-house data center outage the words of the great bard hold water. Nearly every growing company must face a data center build vs. buy decision at some point. Too often the decision is made to invest in an in-house computer room for the production environment. Don’t get me wrong, there are a number of reasons to build an onsite data center: local or campus networking, test/dev, proof of concept, and so on. But the housing of production infrastructure is not one of them.
The rise of the large scale data center has made it possible for small and mid-size organizations to leverage world-class infrastructure that they wouldn’t be able to afford otherwise. This, in combination with the commoditization of metro Ethernet and other network products, has made it possible to cost-effectively remove critical infrastructure from the homespun computer room and move it into industrial-strength data centers.
We could examine nearly any aspect of a large scale data center in comparison to what is usually built into an on-site computer room and find stark inadequacies in the latter. Perhaps the discipline this is most prevalent in is electrical power. IT leadership understands the need for clean, always-on power however they often don’t have a complete picture of what it takes to provide 100% uptime. Take, for example, ServerCentral’s DF-CH1 data center constructed by Dupont Fabros. This facility has four, completely separate and path-diverse power feeds from multiple substations. It also has 33x 1.3 megawatt rotary UPS systems and 33x 2.25 megawatt diesel powered generators. These eye-popping figures are just a fraction of the systems and expertise needed to provide true, mission-critical power.
Better power, HVAC, networking, security, and economies of scale aren’t the only reasons a large data center is best suited for production infrastructure. Data center providers often offer 24/7 remote hands capabilities, further reducing strain on IT staff already stretched thin. ServerCentral has been providing interminable remote hands services for more than a decade, giving us keen insight into the pain points associated with running production equipment. Our Data Center Operations team can handle any physical plant need, freeing corporate IT resources to focus on the core of their job functions, which revolve around providing technology services to the business.
If you’re starting a build vs. buy data center project, make sure touring at least one large-scale data center is on your to-do list. Preferably tour three or more data centers to get a feel for the range available. Make a day of it! Get the decision makers and influencers together for a whirlwind, whistle stop, carpooled day of data center enlightenment. Take the provider’s quotes and compare against the 5-7 year costs of operating your own computer room. You’ll likely discover that not only can you achieve more uptime, but you can do it for about the same money, or even less.