I've had some discussions on the greenness of Cloud Computing the last few weeks. Most people consider the Cloud to be Green IT, but there are skeptics. Their main argument against the greenness of the Cloud is usually componentisaton.
Componentisaton allows for ever greater speeds of evolution and innovation. It leads to more consumption thus you might have more efficient units but you'll end up consuming vastly more of them.
This might seem logical, but if you ask the inevitable question - what are the alternatives - most people I talk to come up shorthanded. We are inevitably in a maelstrom of ever increasing usage of computational units, cloud or no cloud, and as far as I can understand, the underlying business models of Cloud Computing of providing the most efficient and cost-effective platform for computational units, this really has to be greener than the waste we are seeing on desktops and in-house server farms.
New, and novel forms of consumption always arise with componentisaton they say
The first insight into this are seen today in the CI/test environment scenarios. Where we used to have a sparse set of environments on which to make do, most Cloud companies and projects do not post any limitation of how they utilize virtualization of such environments to beef-up productivity and quality. If we do a side-by-side greenness comparison today, the cloud win by large margins, but at some point this picture might reverse.
So, while I still believe the cloud to be green, we must watch the Cloud business models closely as they evolve to ensure that the cost of Cloud computation does not get to low to measure, which will lead to massive waste which will kill the idea of green clouds.
1 Comment
comments.show.hideSep 21, 2009
Thor Henning Hetland
A side note to this from Norwegian perspective.
It seems like the Norwegian Government is sponsoring a lot of new "Green Energy" plants these days, and while this is good, the lack of power-line capacity out of Norway, this will likely reduce the energy costs significantly the next few years (from an already low level) so we might hit a temporary "too cheap to measure" sweet spot here quite soon...