Wait and Hope: A Strategy for Data Center Overcooling

This audio was created using Microsoft Azure Speech Services

“…the sum of all human wisdom will be contained in these words: Wait and hope.” Or so wrote Alexandre Dumas, in his novel The Count of Monte Christo. Once the most famous novelist and arguably the most famous Frenchman in the world, no doubt Dumas’ pronouncement appeared to be sound advice for the denizens of the mid-nineteenth century.

But in today’s rational world where the empirical method, nailed-on certainty and instant results are the currency that counts, I wonder whether “wait and hope” is a credible strategy. Would, for example, Dumas’ advice be of any use to anybody who’s been mulling over a potential role DCIM could play in their data centers today. Especially if they were waiting for some clarity.

Probably, though, the real axis for discussion – the nub of the matter – has been whether or not the cost and effort of software deployment can be demonstrated to have a proven ROI attached. Proving a return on investment has been no mean feat it seems. The guys that run the data centers so often do not have a view of the energy cost or any savings; ROI has been “soft”.

The effort that has been put into measuring and monitoring still requires someone to take the data and action it. Assuming the appetite is there to set, e.g., efficiency on an even par with availability. Or even to lay aside that very human thing of keeping the data center so much colder than it needs to be. Overcooling is a classic waste of resources, adding significantly to the cost of data center operations as well as carbon footprint. But it’s a tough habit to break.

The problem, perhaps, lies in overcooling providing the illusion of a safety blanket – some sort of comfort that if things go wrong the very air itself within the data center is on tap as redundant cooling capacity. But it is an illusion – you may not actually have the redundancy you expected. And it tends to be much worse in legacy data centers where there is tangible resistance to raising inlet temperatures. These types of data centers are uncomfortably cool, but despite being highly wasteful you get the sense that being cold is also a feel-good factor for the facility manager.

Cooling Fan

ASHRAE have done a lot work and provided a lot of well evidenced advice about the efficiency benefits of warming things up in the data centers. The Green Grid support their conclusions as do Uptime and practically every other serious advisor in the marketplace. What’s been missing to date, has been technology to put real control of the cooling system into the hands of facility managers, while maintaining an effective safety net.

As unlike me as it is, I’m writing this post to draw your attention to a breakthrough that we’ve made in this area through a new integration between Vigilent and Schneider Electric DCIM. The important outcome of this technology partnership is dynamic cooling provision for any data center – even those with legacy cooling infrastructure from any manufacturer. What’s more, there is a hard ROI because cooling efficiency and saved operational expenses can be seen almost from day one. Once implemented, this new dynamic control can ensure that cooling capacity can be kept in step with IT load requirements.

So if your strategy for DCIM implementation has been to wait and hope, then I think this new announcement is perhaps the answer you’ve been waiting for. Please look out for more technical details on the blog in the coming weeks and months. – or drop me a line via the comments.

Tags: , , , , ,