Throwing down the gauntlet over power and cooling costs

EMC is in the spotlight this week, with a mammoth announcement that covers just about every product line, including Symmetrix, Clariion Celerra, and Disk Library.

It will take some time to digest and make sense of all the news. Moreover, some of the software novelties that EMC is introducing, such as assigning different performance levels to applications on a Clariion, won't be available until December.

Interestingly enough, I heard more from competitors such as Hitachi Data Systems -- or their agents -- than from EMC on those news bites, which tells me that the marketing machine from Hopkinton, Mass., must have bruised some noses.

I will be glad to host a debate on the EMC announcements, but let's do that on my blog, where the two vendors and all others are welcome to make their point.

As for what may have triggered EMC competitors' reaction, I have a few ideas. Very high on my list is that EMC is making power and cooling demands one of the selling points of its new high-end box, the Symmetrix DMX-3 950. In fact, EMC is claiming an astonishing "up to 70 percent better power efficiency" than both competitors' gear and its own previous models.

To be fair, I did not see the formulas that EMC used to back up those claims. Hopefully, EMC and other vendors will follow the example set by Brocade, which recently published a chart showing its director switch power efficiency compared with others from Cisco and McData. Those numbers are not necessarily gospel, but at least they're out there for anyone who wants to challenge or improve on them.

EMC's claims may need some peer review and are probably marketing-driven (no surprises there), but I have to applaud the vendor for bringing its weight to the storage power efficiency debate. I can't even begin to tell you how difficult it is to find some factual, recent data on that topic.

Power efficiency is a new theme for storage; as of about a year ago, it was rarely if ever mentioned. So it's quite possible that vendors' initial claims may need to be put in some context to be sure customers can compare apples to apples. 

For example, a simple efficiency/capacity ratio would not be much help to customers with limited floor space. Those customers may find it easier to compare solutions according to criteria such as watts per square foot that take space into consideration.

Moreover, customers with space floor constraints will probably prefer solutions that can be mounted in a rack, rather than alternatives with similar ratios but requiring dedicated floors space -- such as the Symmetrix line.

Compared with the simple server alternative of blade vs. traditional form factor, storage has a much more diverse range of devices that make it even more challenging to come up with just one or two criteria for measuring power and cooling efficiency. The fact that many arrays can mount either small SAS drives or large SATA drives creates another challenge if we use only capacity as denominator, due to the large divergence of efficiency ratios for the same device.