In my last posting on the death of the data center, I noted that most data centers are built in the wrong place. Let’s examine this idea. There are eight primary costs to the data center. They are:

1. Staff
2. Land
3. Architectural/Building
4. Core and shell costs
5. Power costs
6. Power distribution and cooling costs
7. Server hardware costs
8. Networking equipment & comms costs

So let’s just think about location.

First we have the fact that if cheap power won’t come to the data center, then the data center must go to cheap power.  Cheap power may turn out to be a wild variable in the coming years. Today’s cheap electricity may turn out to be tomorrow’s pricey power. The climate change dynamic has promoted significant investment in “green” power, which might possibly shift the variables here. It also opens up the possibility that if you build a data center, you might decide to build the power source too, rather than risk being beholden to an electricity company. A few windmills here and a field of solar cells and mirrors there, etc.

More likely, you just move the data center to a known reliable source of cheap electricity. The electricity generation options include: hydroelectric, oil, liquified natural gas (LNG), coal and nuclear. Fuel costs can move the price of oil, LNG and coal considerably. Hydroelectric is generally more expensive because the electricity has to travel further to its customers – losing considerably power in transit.  (I believe the ultimate limit is 500 miles, but that’s a figure from a while ago). Otherwise hydroelectric power would be a bargain – and it may be for a data center.

Anyway here is a set of criteria for location costs:

• Choice of Power Source: Cheap is great but cheap, highly reliable and fuel cost independent is much better. For medium to long term price stability, power companies will doubtless be willing to strike long term deals.
• Distance from source: There are significant advantages to locating very close to a power plant, because you can bargain for a lower cost supply on the basis of distance alone. Also power interruption from electrical storms can be far less likely if you have a private circuit direct to the power station.
• Land costs: The days of siting data centers in skyscrapers in Manhattan are over. Better to use cheap land with low property taxes.
• Architectural/Building/Core and Shell: A designed-entirely-as-a-data-center building is a must. Heat management is the overriding priority so building out will almost certainly make more sense than building up. Cool geographical areas may make more sense than hotter ones. There is also the issue of safety. Data centers need to be electrically safe, secure and fireproof.
• Staff: Although I’ve noted that staff costs are very low as a percentage of the whole, location in areas or even countries where staff costs are low can make a contribution to taking costs out of the data center.
• Data Laws: You may need to be be a little careful about data laws. Some countries require that certain kinds of data (particularly personal data) stays within the country.
• Investment incentives and taxation: There are many areas of the world including states in the US that welcome inward investment and will help to finance it with very generous tax exemptions and even cash incentives for job creation. It would be churlish not to take advantage of these things.

All in all, how much is this worth?

It varies, of course, but the answer is a lot. Staff costs may be as low as 3% but if you cut them in half, that’s significant. Investment incentives can often pay for wages for a whole decade, making the staff costs descend to zero or below, as well as slicing a huge amount from property taxes – which can make up most of ongoing building costs (which are about 7% of all costs). Siting next to a power plant can reduce electricity costs by 20% before you even start to negotiate a lower industrial price for the electricity. And as for building costs, data center costs per square foot can be in the range of $300 to $1000 in the US. Designing for purpose can take you to the bottom end of the range, or lower.

Location, location, location

Location is a huge factor in data center costs. Using the rough model I created (see Part 1), factors that contribute 33% of data center costs are affected by location. You can probably cut those costs in half by a good choice of location, and maybe you can do better than that. But let me emphasize what I’m talking about here. I’m talking about a cost model for a scaled out data center, not an expensive already existing data center paying high electricity prices, where staff costs are 33% or more of the total.

If you compare to that, then the price for these elements is being cut to ribbons.

See also:

The Death of the Data Center: The Model
The Death of the Data Center: Location, Location, Location
The Death of the Data Center: Power
The Death of the Data Center: Cooling
The Death of the Data Center: Networking
The Death of the Data Center: Server Hardware
The Death of the Data Center: The Software
The Death of the Data Center: Software Optimization