What's Important in the Data Center Location Decision?
Energy reliability and costs weigh heavily on the data center location decision, but tax incentives may ultimately come into play as well.
Thomas Freeman, Managing Director, Jones Lang LaSalle and Jason Warner, Managing Director, Mission Critical Solutions Group, Jones Lang LaSalle (Spring 2011)
The data center sector is one of the fastest-growing property types in the United States and in many other countries. Unlike other property types, data centers generally don't need to be located in the center of urban or suburban business areas. The factors that drive office or retail site decisions - high visibility, proximity to large pools of employees or shoppers, access to highways and transit systems - are usually not important at all in a typical data center location decision. In fact, data centers don't even need to be near the people using the data.
So what is important? And how can companies in the market for data centers take advantage of their unique situation? Data center site searches are likely to focus on energy capacity and costs, fiberoptic connections, tax impacts, municipal incentives, and the availability of large land sites at low prices.
It's no surprise that demand for new data center space is rising. While office, retail, and other property sectors face high vacancy and low net absorption rates, demand for data center space is rising at a faster pace than new construction.
Computer servers have become increasingly compact, allowing more of them to fit into a smaller space. As a result, a data center built 10 years ago is unlikely to have the electrical capacity and physical characteristics needed to accommodate the increased server density, as well as to keep all those servers cool. Redeveloping a data center to meet the increased electrical load is often no less expensive than building an entirely new facility.
In the past few years, the credit crunch, combined with the larger economic crisis, has caused companies to put off capital-intensive projects in favor of leasing space on an incremental basis. In 2010, this situation eased somewhat with the emergence of private equity and REITs (real estate investment trusts) investing in data centers.
Also, compared to other property types, data centers with excess capacity have a limited opportunity to sublease space to other users. Therefore, despite the recent development of more multi-tenant data centers, the criticality of maintaining operations without interruption and the required separation of infrastructure make the market for data center space not as fluid as for other property types.
Reliability Is Job One
The overriding concern of data center site selection is the availability of reliable power. They are called mission-critical facilities because even a brief shutdown can be devastating to a business. Consider a bank that loses track of deposits and withdrawals or an airline that can't track planes or take reservations, even temporarily.
Ensuring zero downtime means that data centers are typically not located in areas prone to natural disasters such as hurricanes, earthquakes, or tornadoes. Nor are areas that experience frequent power outages appealing to data center users. Usually these areas will be eliminated from consideration early in the site selection process, unless there is a compelling reason for locating a center in such a place, such as the need to be close to customers.
In conducting location searches for data centers, companies typically look for fiberoptic connectivity to handle high-speed, high-volume data transmission. This information can generally be obtained from communities that want to attract these facilities. A location decision-maker can obtain a map of fiberoptic cables in the area, or ask the fiber providers to provide information about the potential to extend lines as needed.
And, while some data centers can be located hundreds of miles from a major city, in some cases it is important that servers not be too far removed from the people using them. Therefore, latency (the time it takes for data to travel between one place and another at the speed of light) may be an incredibly important site selection criterion as well.
For a stock trading system, for example, the milliseconds that it would take for buys and sells to travel through lines to and from a distant data center might be an unacceptably long wait. New Jersey and eastern Pennsylvania are popular areas for data centers for their proximity to New York and other East Coast cities for this very reason. In setting up criteria for an optimal location strategy, companies need to consider these issues not only in terms of distance, but also in terms of the number of data centers required to serve the needs of business across the country or around the world.
Energy and Sustainability
After reliability and other "must-have" factors, energy cost is one of the biggest drivers of data center location decisions. Currently, IT represents 30 to 40 percent of a typical company's energy consumption, and corporate data center operating costs are reportedly growing by 20 percent per year.
Since energy is one of the main costs of operating data centers, companies look for ways to minimize energy cost in their site selection. The most obvious way to do that is to find an area that naturally has low rates. It follows that climate also factors into this decision: An area that gets extreme heat through much of the year will require more energy to cool a data center than an area with cool to moderate temperatures throughout the year.
Sustainability, as related to energy, is another location factor, although it is not a primary consideration. The Environmental Protection Agency, which monitors energy consumption for its associated greenhouse gas emissions, has estimated that energy use in data centers may double over the next three years, and is already approaching the amount of energy used by the airlines and heavy manufacturing industries. Companies with large data center operations - such as Internet service providers, telecommunications providers, and financial institutions - have set goals to limit their carbon footprints by operating data centers in a more efficient manner and by locating in areas where utility providers have a "greener" generation mix. Sustainability is also taken into account in the design and day-to-day operations of these data centers as well as in the technology employed within.
Nevertheless, green energy from renewable sources such as solar and wind power is more likely to increase - rather than decrease - the cost of a kilowatt-hour. So an area serviced by a utility that provides renewable energy as part of its mix of power resources may lose out to an area with less costly energy that is not sustainable. That equation could change if nonrenewable energy costs rise, or if government mandates for renewable energy continue to increase. Currently, government or utility rebates for companies that implement renewable technologies are available in more than 30 states. Organizations considering data center location strategies should consider all the angles - but inevitably must deal with current realities.
Local incentives rarely drive facility location decisions, and data centers are no exception. As noted, a range of other factors must be satisfied before a data center user will seriously consider an area. However, when multiple communities can meet all the key criteria equally well, incentives can drive the final decision to locate in a particular community.
For example, there are a dozen states that abate sales taxes on certain construction materials as well as IT infrastructure and equipment. Since servers and some associated equipment are replaced every three to five years, these sales tax abatements can total tens of millions of dollars over the lifetime of the data center. When everything else is equal, data center users may favor communities with the flexibility to create an incentive package that furthers the user's specific goals.
Just as important as the size and shape of the incentive package is the knowledge and responsiveness of economic development agencies (EDAs) and local real estate interests that are vying for data center development. An EDA with a database of potential sites, and a working knowledge of competitive utility rates as well as climate and weather data, can play a role in helping local landowners and developers attract a data center development.
Ultimately, a company that performs its due diligence in the data center site search should find plenty of locations that meet its criteria. Understanding the specific need and taking the time to conduct a thorough process will ensure a positive outcome.