Visiting major data center events is always an education, as I was recently reminded when I visited the Gartner Data Center Conference 2012 in Las Vegas at the beginning of December. Without a bit of external and objective input, there can be a tendency to become a little introspective. So I was looking forward to exchanging the naturally chilled climate of my native Denmark at this time of year (well, any time of year), for the air conditioned “comfort” of Sin City.
To me, one of the outstanding insights from Gartner concerned the anticipated growth of enterprise and on-premise data centers. During the last couple of years the projected major growth in colocation has materialised big time, driven by the trends to cloud source and use managed services. I suspect that there’s a lot of organic growth buried in these stats too – after all, we’re all busy generating data like there’s no tomorrow. So perhaps it should have come as no surprise that companies would also be increasing their own resources.
Old habits die hard and no doubt uncertainty and fear of the unknown also play a role in constraining rates of adoption – especially in a market which is by nature conservative. Answering this, Garter quoted ADP, one of the largest payroll service providers in US. I think I remember correctly that they said that today as much as 60 percent of companies outsource this function. Yet ADP laid the foundations of success at a time when payroll was considered far too sensitive a process to outsource. Yesterday’s taboo has become a more than acceptable business habit today.
So if outsourcing and cloud are set to become more widely used, does this also mean that some of todays critical applications will be considered less so tomorrow? How do you decide what’s mission critical, when do you review this categorization (should you put a process in place to ensure this is done regularly), and what impact does this have upon on-premise data centers? We already see that outsourcing creates new pressures and questions for internal resources. For example, the high levels of availability and comparitively lower operational costs on offer from service providers can bring into sharp contrast the economics of data center ownership.
At Switch Communications’ SuperNap in Las Vegas they offer a 100 percent uptime guarantee. I’ve never seen such an SLA before – but would you offer it based on your existing infrastructure. More to the point, perhaps, how much would you need to spend on facility upgrades and operational expenses in order to be able comfortably offer something similar to your internal customers?
So if the growth in enterprise data centers is going to be realised, what lessons can we take from colocators? Because certainly, if the data center is the business (or at least central to the business), there must be something on offer from companies that have elevated ownership to a profitable business which also provides cost, convenience and reliability advantages to an increasing number of customers. If you’re running services or applications in your own data center as well as in outsourced services, I’d be interested to learn what lessons you can share about how one domain informs the other. Have you experiences to share?