Not too long ago, data centers were unexciting facilities. Their performance and corporate value was a mystery to business executives. The same applied to the language that described these facilities: for example, energy consumption, capacity and comparative performance. It was uninspiring.
These abstract operational concepts have now emerged from obscurity. They have entered the strategic vernacular of business stakeholders. This change is partly due to the financial sums at play. Data centers command extraordinary amounts of money during their lifespan. Companies have invested hundreds of billions of dollars in these facilities since the microcomputer boom of the 1980s.
As is often the case with technology, data centers have moved up the agenda of the CFO and CIO. The long-term performance of a data center starts with a multidepartmental planning process—identifying an appropriate location from a financial, operational and CSR perspective.
It might seem obvious to avoid anywhere with severe weather conditions or a location far away from utility providers or energy sources, yet what is surprising is how often these factors are still inadequately taken into account.
Getting a Good Deal
Where to start? With the capital cost of land. If a company has a location in mind and it ticks the well-known risk-factor boxes, those managing the build should be aware of the premium often placed on land being sold for data center use. Several years ago, Microsoft paid $6 million for a tract of land in Iowa that would have cost just $1 million had it been sold for agricultural use. These numbers have grown dramatically since. Facebook’s new 110-acre data center in Dallas commanded a much larger price.
This expenditure must be considered in the context of the entire project, as costs are unrecoverable. Given that Microsoft’s budget was $1.1 billion, the cost of the land in this example was a small proportion of the overall spending, yet for smaller organizations any wrong decisions could derail a facility’s future profitability and business value.
Environmental regulations must be considered before any investment. That means carrying out an environmental-impact assessment to evaluate and predict a data center’s effect on the surrounding environment. Although a data center project is unlikely to pose a contaminant or chemical risk, these regulations also cover water resources—an important component of a facility’s cooling system.
Given the reliance of modern data centers on water for cooling, any plans must meet with water-conservation regulations. The availability of water at the location will need to be studied, provisions made for water storage, and sustainable water-usage practices considered. Apple’s Nevada-based data center, for example, is built underground to take advantage of the site’s submerged water resources. This design also protects the facility from natural disasters.
Other companies have taken advantage of cooler climates, which help to reduce energy costs and negate the need for expensive mechanical cooling systems. That said, “free cooling” is not always true to its name. When assessed using analytics applications, some initiatives actually end up costing more than prospective savings.
Another reason to be cautious of arid locations is that although the cold climate of say, Iceland, appears to offer cost benefits, it can be too remote a location. Microsoft apparently decided that the country was too far away from large population centers, risking higher data latency. Network connectivity is therefore a critical consideration when deciding on a data center location, especially for data center hubs in cross-continental “landing zones,” such as the Goonhilly Earth Station.
Alongside temperature considerations, risks in relation to climate and local conditions must be assessed in the short and long term. Having access to historical climate data, therefore, is extremely powerful in the context of site selection—i.e., understanding the operational impact of fluctuating temperatures, renewable power resources and the ability to use ambient cooling.
Where planning procedures are concerned, the wrong decision can be massively costly, as Apple found out. The company’s new $940 million Irish data center suffered delays when locals lodged complaints of environmental insensitivity. In Apple’s case, construction delays were due to an independent planning group tasked with evaluating the company’s appeal case. Understanding any potential planning objections and applicable regulations before investment avoids delays and unnecessary costs.
Energy will be one of the facility’s highest operating costs, and site location can have an impact on its price depending on whether energy source is grid power, carbon or sustainable sources such as hydro, wind and solar. In the U.S., energy rates vary from state to state, but compared with international energy they are generally low.
The other consideration is availability of energy to the site: if the site is remote or is being newly developed, utility services will have to be extended, adding to the cost and potentially causing delays. Tax on energy should be part of analytics calculations too, so checking local rates and identifying any incentives is crucial. Being able to simulate prospective tax increases or reductions is an extremely valuable capability for both the financial and operational department.
This process is crucial for data center service providers. Margins are being squeezed, and the profitability of a facility stems directly from maximizing available space, energy and capacity. If the data center design can’t handle its maximum possible workload, the investment will never reach its potential. Any small change in energy costs can have a dramatic impact on future financial success and commercial governance.
Energy costs are a real threat. After all, they are always changing. Norway announced data center electricity tax cuts in the 2016 state budget, which highlighted the country’s determination to win more data center business. The proposed cuts, which reduce rates by a massive 96% per kilowatt-hour for data centers with more than 5MW of power usage, have the potential to make a major difference to the annual electricity bills of enterprises investing in the country. It’s a policy change that is gaining momentum in the Nordics—large-scale data centers already benefit from lower energy-tax rates in Finland, while the Swedish government is now considering a 97% tax cut for data centers and data-service providers.
In the U.S., more than half of the country’s 52 states and other jurisdictions have introduced changes to their tax structures in order to encourage data center operators to invest locally. Understanding which locations offer the best tax deals, and measuring them against other site-selection criteria, is important.
Due diligence requires that all of the different considerations relating to a site are taken into account. But achieving a competent and accountable evaluation requires the analysis of different scenarios and the ability to accurately predict costs and/or performance levels.
Crucially, these simulations must not only forecast the current situation; they must also consider changes such as energy-cost increases and tax decreases, which have a longer-term impact. Doing so is possible only with a solution that is built to model, simulate and predict data center investment and operational performance.
About the Author
Richard Jenkins is Senior Vice President of Marketing & Strategic Partnerships for Romonet. Richard’s career in sales and marketing includes roles ranging from enterprise-systems management to Internet of Things (IoT)–enabled smart footwear. On leaving the Royal Navy in 1990, he worked for UK-based IT resellers developing channels across EMEA, building and selling a Y2K IT-contractor firm and providing IT consulting services to London-based banks.
Richard has held management roles with Tivoli/IBM and Crystal Decisions (later Business Objects) and senior executive positions with Corporate Radar, Kyoto Planet, Plantiga and RF Code. In 2004, he launched his own company, Performedia International, which published strategy books for the IT industry, provided video marketing solutions and established global sales and marketing divisions for growing technology companies.