Notes from the Consultant’s Jungle

Regular readers of this blog will know that we are strong advocates of container data centers as a step toward modular data center design and as a facility component for extremely dense data processing.  Earlier posts talked about 2008 as a breakout year for containers (it wasn’t) and containers used as facility components in cargo ship data centers (haven’t seen them either).

It seems that 2009 isn’t going to click for containers either.  We’ve seen reaffirmations from Microsoft and select others that containers are still a planned component in the construction of data centers for very high density data processing, but this too seems to be lagging.

…Where are the containers?  They’re coming.  When?  They’re coming.

In work that we did late last year for a Client who was interested in building a facility in the Southeastern US solely on container technology, we had the opportunity to speak to the sales organizations of container manufacturers.  That experience struck me as very interesting because even though I had a paying Client, I could not get a vendor to clearly articulate a price to me.  Yes there’s customization involved, but even still it was like pulling teeth to get even a rough order of magnitude understanding of what the costs associated with containers really are.  Whether behind Door #1 or Door #2, in the end I can tell you a dollar range of +/- 40% of what you can expect to pay for a container, but I still can’t tell you with certainty what drives the price up or down.

So what does all this mean for our use of containers as mission critical facility components?  Well, I think that it paints a niche market for containers, made largely of content and entertainment and certain Web 2.0 applications.  If there are no pressures to move the price point lower, we’re likely to not see them enter the general enterprise space, even though many are waiting for just this sort of solution.

As a stepping stone toward modular construction techniques for data center facilities, if the number of use cases is insufficient to throw off knowledge and experience with their use, it seems unlikely they’ll contribute to that tangent either.  In the mean time, we have many advances in cold aisle and exhaust heat containment that are enjoying open visibility in the data center world.  Maybe someday, unlike Godot, a modular construction technique for data centers will arrive.