David Tebbutt, Teblog
People around the world are tearing their hair out trying to get to grips with environmental issues. well, some of them are, anyway. Should they replace their car or run it into the ground? Should they commute to the office or work from home? Should they buy power-efficient servers or stick with the old stuff?
They’ve heard that new stuff comes with an embodied environmental footprint, but how do they know what it is? And they know that disposing of stuff can cause environmental damage, so what do they do?
These question and many more, have been around for years but information is hard to come by. It’s not like a fridge or a washing machine that has a dead easy sticker on it. It almost doesn’t matter what the bars mean if you go for the shortest and greenest then you’re doing your bit for the environment.
When it comes to computing, you can find plenty of sources of information, but nothing as simple as a fridge sticker. Energy Star, EPEAT, RoHS and the Carbon Disclosure Project are just a few of the big names. Then you have the so called carbon calculators, device comparators and, in the data centre, computational fluid dynamics, among other things. It takes a lot of time to rummage this stuff and figure out the environmental implications of what you’re buying. And, of course, it’s constantly changing.
Welcome then, to the British Computer Society’s initiative with the Carbon Trust to produce a way of modelling the intricacies and interrelationships of the equipment in a data centre to churn out the cost and energy implications of your choices. The catchily-named BCS Data Centre Energy and Cost Simulator started beta trials last week and is expected to go live next March, but you can get a head start and buy your way into the beta programme for £25k (half price if you’re a non-profit). Assuming there are any places left in the 20-organisation trial.
The modeller looks at the following elements of the data centre: application workload, IT devices, PDUs, IPS, air flow, CRACs, outside climate, chiller, power cost energy use and carbon intensity. Variations in each of these are taken into account. For example, the climate varies by location and by time of day. The CRAC load varies by climate. The energy cost varies by thetime it is purchased. The carbon intensity varies by the source of energy.
You get the idea. The project is open source and extensible with new components and XML data. The outputs are clear – 3D wireframe graphs, bar charts and the like.
The modelling system restricts itself to the data centre but it seems to me that it would make a useful component in a broader system. One of the speakers mentioned that the CMA, now part of the BCS, is also working with the Carbon Trust on helping UK businesses make better use of ICT to significantly reduce carbon emissions, with the emphasis on the communications element, as you might expect.
If things go according to plan, you can expect more initiatives from this quarter. I’ll keep you posted.