[LINK] How many Newspapers does it take to build Ayers Rock?

David Lochrin dlochrin at d2.net.au
Wed May 13 11:24:17 EST 2009


On Tuesday 12 May 2009 21:04, Rick Welykochy wrote:
>> I thought linkers mught enjoy my latest Koltai reveletion in 
>> relation to ICT vs traditional media distribution methods.
>> 
>> 47,742 tons per year - 2.83 times the volume of Ayers Rock.
> 
> That would be 2.83 times the mass of Ayers Rock, no?

At standard temperature and pressure (0 degC & 100 KPa) the density of carbon dioxide is around 1.98 kg/m³ (Wikipedia), which is 0.00198 g/cc.  Ayers Rock, which is sandstone, would have a density of say 2.3 g/cc.

So compressing CO2 at STP and the volume or Ayers Rock to the same density as Ayers Rock would reduce its volume to (0.00198 / 2.3) or 0.000861 that of the Rock.  Much of it is below ground but it's still 348 metres above, so our hypothetical CO2 would only be about 30 cm high!  I haven't calculated the temperature rise involved, but it would be substantial.

On Wednesday 13 May 2009 10:24, Glen Turner wrote:

> As far as carbon generated by equipment, that's a trickier thing. The 
> ASICs and CPUs used take 3-4 years to develop [...]
> 
> As for larger equipment, there's already a huge focus on the maximum
> energy usage of those items as ISPs desperately want to be able to run
> them from typical power supplies, rather than install bespoke systems
> for the equipment. There's still a lot that can be done to lower the
> average energy use of the equipment. [...]

Thanks Glen, that was a very interesting little tutorial.  I wonder how much energy could be saved with top-down energy savings rather than bottom-up savings achieved by redesigning chips.  Here are some examples which come to mind:

o   offices using X-windows displays on the desk and large centrally-managed servers rather than putting a system on every desk (I've never been able to understand why this is necessary);

o   servers with provision for DC input from a common source, which must surely be more efficient - is this what you mean by "run them from typical power supplies, rather than install bespoke systems"?

o   servers without console terminals;

o   servers which can be automatically powered up only when demand requires it - I wonder how much of the IT greenhouse load is generated by systems which are doing nothing productive at all;

o   and I wonder about the relative greenhouse efficiency of various O/S, taking into account the extent to which they support features such as those above.

David



More information about the Link mailing list