[AusNOG] Data Centre Design was Internode goes Carbon Neutral

Bill Walker Bill.Walker at staff.snap.net.nz
Thu Nov 19 14:29:57 EST 2009


Obviously the biggest cost is in power. Emerson have done some research on energy use in data centres that showed for every watt your equipment used, the support services used 1.84 watts.  The general rule of thumb though is one watt of consumed power requires one watt of cooling. Based on that 50% of your power bill is cooling.  So you would need to do the numbers. If you are building based on 1MW for equipment, you would need an additional 1MW for cooling. But if you could use Free air cooling, Variable capacity cooling, cold aisle containment, more efficient PSU's etc you could significantly reduce that figure.

The key I think is to find a heating/cooling design engineer who knows his stuff.

-----Original Message-----
From: ausnog-bounces at lists.ausnog.net [mailto:ausnog-bounces at lists.ausnog.net] On Behalf Of Daniel Hooper
Sent: Thursday, 19 November 2009 3:21 p.m.
To: ausnog at ausnog.net
Subject: Re: [AusNOG] Data Centre Design was Internode goes Carbon Neutral

I'd like to see idea's on the most ideal place in .AU to build a green DC.

I'm still pondering if the location would be based on cheap cooling or cheap power (forgetting about current carrier & power infrastructure)

Ie, you could build somewhere in SA and take advantage of geothermal power, or you could build in Tasmania where (correct me if I'm wrong) the average summer temp is 22 degrees.

-Dan

-----Original Message-----
From: ausnog-bounces at lists.ausnog.net [mailto:ausnog-bounces at lists.ausnog.net] On Behalf Of lists
Sent: Thursday, 19 November 2009 10:38 AM
To: Matt Carter; 'Curtis Bayne'; Mark Prior
Cc: ausnog at ausnog.net
Subject: Re: [AusNOG] Data Centre Design was Internode goes Carbon Neutral

Thanks for everyones input.  There has been some good points raised in this 
discussion

I am like some others, and am one of those that is investing in technolgies 
that reduce the amount of power I use.  For those in that camp there seems 
to be a lot of focus on designing data centres to suit the existing building 
designs.  Correct me if I am wrong, but many data centres are located in 
existing buildings in capital cities.  These are generally the larger of the 
data centres, and as such these are where the major data centre vendors are 
focusing thier attention and solutions.

Some have mentioned good data centre design, which raises the question, what 
is good data centre design?

If we were to build a new data centre from scratch and we wanted to build it 
to be enviromentally friendly and power efficient then how would we do it?

Leaving aside having the data centre located as near as possible to 2 power 
grids and good access to carrier fibre and those issues and looking at it 
from a purely efficient power and cooling design point

Do data centres need to be located in capital cities?
Do we need physicall access to the data centres or do we need some capable 
hands and legs.?
When it comes to building design which material is best for maintaining low 
ingress or egress of heat and cold?
Is 240 volts the most efficient voltage to power such data centre equipment? 
perhaps 48volts DC may address some heat issues
Do we need raised floors?
etc etc

Personally I think there may be a case to do away with raised floors, there 
may also be a case for extracting heat from the racks and reusing it for 
alternative uses.  It may be possible to locate a data centre outside of a 
metro area and provide redundant paths back to metro areas.  There are many 
possibilities, and most of them in one way or another relate to the 4 or 5 
questions above.

While I commend Internode for taking the carbon neutral approach, I would 
like to see some discussion regarding what is "needed" to enable a data 
center to maintain uptime and avoid equipment failure.   Many of the vendors 
are now saying their equipments working temperature range is 18 to 28 
degrees C or even higher.  Perhaps 22 degrees is not longer what needs to be 
achieved.

My analysis shows that equipment will survive at higher temps, and I am 
always puzzled why we are using 240 volts instead of 48 Vdc.  The excess 
heat generated is substantial.   I did an install of some equipment in 2004, 
which had a 240 volt power supply that did POE at 48 volts to the equipment. 
Due to the location we wanted to solar power it.  We installed it on solar 
power and it worked fine using the manufactures power supply and an 
inverter.  The issue was that we needed to but 16 solar panels and the 
associated battery capacity to make it reliable.  We then decided to ask 
"why", we found the equipment itself used 0.133 amps at 48 volts,  yet it 
used 2.2 amps at 48v when using the manufactures power supply, most of which 
was disapated as heat.   We built a new power supply did away with the 
inverter and manufactures power supply, used 4 solar panels and a much 
smaller battery bank and 5 years later it is still working like it was in 
2004 ( with higher uptime than mains power), which is why I raised the 
questions above.  Each of us on our own are not big enough purchases of 
equipment to influence vendors, however collectively over time we may be 
able to drag the vendors along.  We just need to work out what in fact is 
actually needed as opposed to the status quo.

Regards

Tim McCullagh

HaleNET
 

_______________________________________________
AusNOG mailing list
AusNOG at lists.ausnog.net
http://lists.ausnog.net/mailman/listinfo/ausnog
_______________________________________________
AusNOG mailing list
AusNOG at lists.ausnog.net
http://lists.ausnog.net/mailman/listinfo/ausnog

No virus found in this incoming message.
Checked by AVG - www.avg.com 
Version: 8.5.425 / Virus Database: 270.14.72/2511 - Release Date: 11/18/09 07:50:00



More information about the AusNOG mailing list