* HP vs. IBM blade server efficiency
I have to start this newsletter by conceding the fact that the study I’m going to talk about was commissioned by HP. And it does make HP and its blade servers look good. But there is a lot of interesting information in the study that was done by Sine Nomine Associates. It’s all about power and cooling so I figure it’s worth at least a read-through.
Sine Nomine is a research and engineering firm based in Ashburn, Va. It conducted the study for HP earlier this year, comparing the power demands of HP’s BladeSystem servers with those of the BladeCenter systems from IBM.
Specifically, the engineering firm did tests on the HP BladeSystem c7000 enclosure with the ProLiant BL460c blades and the IBM BladeCenter-H with HS21 blades. Both blades are dual-core Xeon based.
The study also looked at 1U rack servers from Dell and IBM “to ascertain the relative power and airflow requirements per unit server in traditional vs. blade mechanical configurations.”
The main difference between the HP and IBM blades, according to HP, is the Thermal Logic technology that HP has architected into its systems. HP introduced the Thermal Logic technology in June, giving customers tools to better manage power usage. HP says the Thermal Logic technology makes its blades more energy efficient.
Sine Nomine seems to agree. Its study concluded that HP’s BladeSystem uses up to 27% less electricity than the IBM BladeCenter-H. It also found that because HP’s Thermal Logic technology lets customers manage power usage such as fan speed, it is able to reduce electricity demands by slowing or stopping fans, as appropriate. The IBM blades, on the other hand, would sometimes produce too much airflow representing “wasted effort by the fans in the IBM blade server, which is a component of the higher power draw of the hardware.”
IBM already has denied that HP’s Thermal Logic technology provides any kind of edge. And, in fact, Big Blue conducted its own internal tests in the fall showing that its servers use as much as 30% less power than HP’s BladeSystem. IBM did say at the time that it would be open to a third-party assessment of the energy efficiency of both systems.
You can take a look at the detailed Sine Nomine report here. Let me know what you think.
