SAP reports that two of its data centres in Germany have been certified as energy efficient by TÜV Rheinland, a German technical services group. Apparently to date only 10 data centres from various companies have received this certification and of those, the SAP data centre in St. Leon-Rot, Germany, has set a new standard
The press release says that the SAP data centres were specifically recognised for the use of advanced technologies, which in 2009 saved around 55 million kilowatt hours of electricity or 14,000 tons of CO2 emissions. It helps towards the company’s 2009 target to reduce its carbon emissions back to the levels of 2000 by 2020, the equivalent of half its year-2007 peak.
Also on the subject of data centres, Greener Computing reports that Robin Johnson, Dell's CIO, and Dane Parker, the company’s Global Facilities Lead, said in the Fortune Brainstorm Green conference this week that Dell may never need to build a new data centre.
By virtualisation, removing underused assets and updating old equipment with the latest technology the company had halved the power demand of two of its data centres in Texas and could see the possibility of delaying the need for new data centres indefinitely. By replacing the oldest 25% of servers every year and replacing them with the latest models the company worked out that it could recover the cost from reduced energy use and delay data centre expansion.
I have a problem with both of these reports. Firstly, well done to SAP for checking out the efficiency of its data centres, but for most of us a certification by TÜV Rheinland doesn’t mean very much (and a quick look through their web site didn’t shed much light). For these sorts of measures to be meaningful they need to be widely known and available so that comparisons can be made. For example, the US EPA is expected to launch the Energy Star program for data centres in June, and Energy Star, the EU and others are collaborating with the Green Grid to develop global data centre energy efficiency metrics based on the widely quoted PUE measure.
As for Dell, replacing a quarter of all servers each year looks more like a cost-based move, rather than a sustainability benefit. Energy consumption may be reduced, but it doesn’t sound like the embedded energy in manufacture is an issue in the calculations (and as machines become more efficient it’s likely that the embedded energy will represent a greater proportion). Perhaps if it prevents a new data centre from being commissioned it will have a net benefit, but that pre-supposes that the technology in the data centre structure has not itself moved on.