CIOs are facing a critical resource shortage and, if they don’t plan accordingly, will face escalating costs and end up squandering a large portion of their budgets. Yes, the looming data center shortage, which has already hit London hard, is fast approaching the United States as well.
With over 85 percent occupancy, data center lease costs have doubled in London in just the last 12 months. And with the Olympics coming in 2012, and a moratorium on building new “data centres” (as they spell it in merry old England) proclaimed by the power company that supplies London’s electricity, the situation is bleak.
Across the pond in New England, I’ve seen monthly lease prices for colocation data center space increase 675 percent over the last five years. The same provider that offered Tier 3-type space for $4 a square foot back then is now charging about $30. In the metro New York market, don’t be surprised to see prices in the $45 to $50 range.
And keep in mind that this is just for the facilities and the redundant infrastructure costs necessary for Tier 3 / N+1 availability. The increases are even greater when you add in the rising cost to power and cool modern severs.
Some people believe (incorrectly) that data center costs will drop because the bad economy is driving down real estate prices. Wrong. Real estate costs account for less than 10 percent of a leased or newly built data center. Even if the real estate were free, it wouldn’t change the situation much.
I see a perfect storm brewing, with three powerful forces colliding to drive up data center costs:
1. Shrinking Supply. Since 2005, the demand for high-quality colocation space has outpaced supply. Many facilities are filling up. Making a bad situation worse, data center providers DuPont Fabros, Savvis, Equinix, even Google and Microsoft, already have scaled back plans to build new data centers. And with the time-to-market for large data center projects taking 12 to 24 months, I don’t expect to see relief any time soon. With limited supply, prices will continue to rise.
2. Increasing Demand. Enterprise applications (ERP/BI) and infrastructure (storage and virtualization) are driving the resurgence of centralized IT and its increased demands on existing data centers. Plus, the growing need for regulatory compliance, and upgraded security and disaster recovery systems are all fueling the need for more resources by colocation customers. This expansion is causing ever more of the colocation facility’s resources to be consumed by its current customers.
3. Infrastructure Limitations. The migration to high density servers is causing data centers to bump into power and cooling limitations long before anyone anticipated they would, and forcing enterprises to make additional investments in order to increase the capacity of their existing facilities. Facilities built for 100 watts per square foot capacity are being upgraded to accommodate increased loads up to 400 watts per square foot. With current colocation customers locked into long term deals, these cost increases are being passed on to the facility’s new customers. Lucky them.
If you are not bumping up against limits of space, power or cooling in your current facility, you may be able to wait out the storm. On the other hand, if your data center lease is expiring over the next 24 months, you need a plan now or you’ll pay a big price later. It may, for example, make business sense to build your data center yourself. You can now build a 2,000 square foot facility with Tier 3 redundancy for around $2.5 million to $3.5 million, all inclusive (based on 300 kilowatts of IT equipment capacity). With equivalent leased space running about $750,000 per year, you’ll break even early in the third year — if, of course, you do it right.
In addition, a modern design using the latest energy efficient (green) strategies and technologies will also reduce your power consumption (and your monthly electric bill) significantly.
One last point: Don’t be tempted by a lease price that’s “too good to be true.” It probably is. Look beyond the building and ask yourself these questions:
1. Are there multiple power sources (i.e. fed by different substations)? There should be.
2. Are there multiple network provider options? Competition can keep down network costs. And do the network providers have redundant fiber trunks to assure high availability?
3. Is the location susceptible to disaster or disruption? And I’m not just talking about “acts of God” but equally likely accidents or terrorist attacks.
4. Who built and who operates the facility? How well funded is their business? And what are their plans to address growth and capacity. Do they have any?
If you’d like, I’d be happy to write more about data center site selection in future posts. As always, thank you for sending comments, tips and topic suggestions to me at CIOblog@TransitionalData.com.
Michael Bullock is the founder and CEO of Transitional Data Services (TDS), a consulting firm helping clients implement energy saving green data center solutions, data center relocations, web based enterprise applications and 24/7 technical operations.