Making Robust Data Center Design Decisions
April 24th, 2013 By: Industry Perspectives
Bruno Raeymaekers is a data centre consultant at ARCADIS, a global project management, consultancy and engineering company, which has delivered over 600,000 square meters of net IT space in the past 15 years.BRUNO RAEYMAEKERS
Over the years, our industry has come up with a veritable buffet of different technological solutions. Whether you’re placing many eggs in one basket and have chosen for a Tier IV facility, or your multi-site IT strategy allows for a number of decentralized Tier II locations – during design you’ll be hit with sheet after sheet of technical information, operational benefits, reliability impact assessments, and so on. Apart from those topics, cost will quite likely take a high position in the overall comparison of design choices. Many whitepapers have analyzed, down to the last nut and bolt, all the pros and cons of what’s available on the market.
This often leads to the assumption that standard designs should be readily available. And yes – whoever has worked in the industry for some time has encountered many technical solutions, has investigated many more, and has formed an opinion regarding preferred solutions, based on experience in construction and operation of mission critical facilities.
But do these preferred solutions stand fast for different markets? Can your off-the-shelf design withstand not only the expected reliability litmus test for your financial/banking client, but likewise the high expected EBITA/Return On Investment of your pharmaceutical client?
Venturing into the design process and costing exercise of your soon-to-be crown jewel, an evaluation into Total Cost of Ownership (TCO or however you wish to call it) will give you a bottom line comparison of the different alternatives. Be it your cooling system, the choice for a certain type of UPS or generator, or the facility as a whole: after incorporating investment, operational and replacement/End of Life (EOL) costs, you’ll get a pretty clear view on where to head.
Depending on your company situation and business case, you’ll have to account for different parameters in the calculations. What is your current and predicted electricity cost? Where do government incentives come in? What is your expected uptake profile over the next couple of years?
TCO expressed as Total Present Value: choose your parameters wisely…
Below table summarizes just a few high/low values for parameters which take part in any TCO analysis, and which we’ve encountered over the years:
Above figures combine a wide variety of clients – and locations. Building a facility in midlands US, Western EU or in the emerging markets will quickly tip any assessment in another direction. Most projects present a healthy mix of minimum and maximum values for that specific client, time and place.
Entering those parameter sets in design evaluation studies, will yield significantly different results in overall TCO values. Adding other weighing criteria such as required resiliency, modularity, and scalability, will show you that different design choices fit different projects.
An Example of Efficiency
In past years, the fight for energy efficiency has mostly been won on the cooling infrastructure of the facilities. Assuming you’re set for an air cooling system, and well into air flow management (cold or hot aisle containment and everything that goes with it), an air-side economizing system might prove to be your best bet.
But does it also make sense for your economic situation? Company 1, wants a 28°C supply air condition, has a business case for a 8% IRR, plans to take up the full 500m²/1000kW space within half a year, and pays 0,19 €/kWh in electricity costs.
Company 2, with its legacy equipment, doesn’t want to risk anything beyond 20°C (yes – they do still exist), yet expects a 15% IRR, will be building a site expected to be filled up over 3 years’ time, and is paying only 0,08 €/kWh.
Both have their minds set on a Tier III resilient facility.
After crunching the numbers, you might find that for company 1, the airside economiser solution would seem to make perfect sense, considering the impact above conditions have on CapEx and OpEx.
But company 2, with the same solution, will discover that its 15% IRR target is nowhere to be reached: the financial savings from the more efficient installation are in large part negated by the low electrical cost and longer uptake delay (even with a scalable build scheme). Furthermore, the 20°C expected supply air conditions directly hurt efficiency, but also require an additional chiller plant, compared to case 1 which is nearing the compressor-free tipping point, thus reducing significantly CapEx and maintenance costs.
So that’s that – or is it?
Looking closer into the company 1 analysis, the preferred design choice is just a few percentage points away in overall TCO from another option. You’ll thus want to finally make sure your solution is robust: what if your uptake delay is hit with a yet unforeseen increase due to some divestments in 3 years’ time? Or your energy costs turn out to increase at double the rate? Such changes can quickly take care of a couple of percents, and will impact your design options balance – but does the bottom line, all things considered, remain defendable?
Performing a sensitivity analysis will allow you to identify those risks and unforeseen factors, and to at least start thinking about planning for some alternative scenarios should you need to adjust your concept along the way.
Correctly assessing these topics with a broad and independent view on the technology market, and discussing them with your facility, IT, and financial managers, might set you back a couple of weeks in the early stages of your projects – but will avoid those unpleasant “Why on earth did you?” fights down the road. Or at least create a clear and written understanding on assumptions for those evaluations in 5 or 10 years time – when no one remembers who said what.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.
These comments are right on point. These same methodologies extend to risks, such as earthquake, forest fire, or proximity to a runway. I’m sure I’m preaching to the choir, but with the approach you are espousing above, one can quantify and rank different risks in financial terms that can then drive decision making in objective, fact-based manner. Sensitivity analysis and subsequent development work, to reduce variability and risk, is most lacking in our industry from my vantage point.
Glad to see you advocating for this rigor in the decision making process.