You are correct – it is woeful and pretty amazing when you think about it.
Equally amazing, in my mind, is the low level of awareness among data centre owners / operators, particularly corporates as opposed to colos and data centre companies. And the majority of data centres are owned by corporates.
Amazing, too, is the low level of instrumentation and monitoring deployed right now in data centres to address your headline question – “Where does the power go?” Until you can see where the power is going within your data centre, to a fairly granular level and in near real-time, then you cannot start to manage it.
The good news is that awareness is growing, a desire to do something about it is growing, and solutions are appearing.
Isn’t it easier to force improvement by cost? Follow my tangent here, if you would…
I’m unpopular because I tend to argue for countries such as USA to stop subsidizing oil products; their current £0.40/litre of petrol cannot last. I know that we’re accustomed to such prices, os we’re used to distant warehousing of supplies for logistical reasons among others. The thing is, it cannot last, and an easy way to bring about the proper change is to simply reduce subsidy or (gasp!) tax utilities.
As a counter-point, an old co-worker argues that polarbears losing ice-cap requires them to evolve into better swimmers; I argue with him that it’s happening too fast for Darwin to catch up. The key point here is that influencing change based on environment has to be done at a pace that could be matched by polarbears or by entrepreneurs.
The way this would appear in datacentres is by an increased cost of power, slowly over time, influencing more consideration of power as a more significant aspect of DC choice. That consideration should influence more creative ideas: although we’re not going to get back that inefficiency at generation (and remember, the petrol engine hovers at 17%-19% efficiency), we may save in distribution (slightly) and use kooky ideas such as geological/geographical aspects for cooling. Nuke plants near rivers are not just for the scenery; heat-exchangers to aquifers might become a consideration.
Much as it may come off as a Granola-inspired Green-hallucination, the object of that tangent is to express that price and taxation can influence the desired behavior, if leveraged over sufficient time for the reaction/players to keep up. That price change will eventually be passed on to consumers, but remaining competitive in the marketplace means keeping overhead low and net impact to consumers low.
[...] From the coal or oil delivered to the power station all the way through to the energy that drives our business applications the efficiency story is woeful. Read the entire blog entry here >> [...]
Post a Response
Follow The Hot Aisle
About the Author
Steve O’Donnell has over 3 decades of global high technology experience. He was CTO at Amlin, was Chairman & CEO at GreenBytes Inc, is Chair at Lanix Ltd and Preventia Ltd, and founder at Chalet Tech, an advanced Big Data Security firm. He serves on multiple advisory boards including Violin Memory, Rackwise, and DCPro.
He was CEO at MEEZA, a Qatar Foundation Joint Venture, Managing Director at ESG, Global Head of Data Centre Operations at BT and has held senior positions at Cable & Wireless, Lehman Brothers, and Deutsche Bank. He spends much of his time serving on fiduciary and advisory boards around the world.