HP reinvents self as data center designer
Mission Critical Facilities
After a freeze in the construction of new data centers in the past couple of years, the market is starting to thaw a bit. And Hewlett-Packard wants to provide more than servers, storage, and networking. It also want to design and manage the construction of the candy coating that wraps around all that gear.
That was the idea when HP bought EYP Mission Critical Facilities, a designer of data centers with over 350 employees when the deal was done back in November 2007. HP has been gradually working its way up to productizing its data base construction management service. EYP was renamed Critical Facilities and tucked under the behemoth's Technology Services arm.
According to a recent survey performed by Gartner, 46 per cent of enterprise customers surveyed said they would need to build one or more new data centers in the next two years, and 54 per cent of CIOs said they would need to expand an existing data center over that same term.
Rick Einhorn, worldwide director of Critical Facilities services at HP, tells El Reg that based on data he has seen, the expectation is that for new data center construction worldwide will generate somewhere between $30bn to $40bn in 2012, which is "on par" with the levels set back in 2008, before the Great Recession froze a lot of projects.
Einhorn says that to figure out how to ramp the data center design and construction management practice, the company has done more than 30 engagements worldwide that it handled soup-to-nuts. Before and after being acquired by HP, the EYP staff has done thousands of data center designs, including both greenfield and retrofit projects. And it is now moving from reactive to proactive mode.
HP is obviously touting its own expertise with IT equipment and in running data centers, but Einhorn says HP is not just pushing its own hardware. Critical Facilities operates independently of the Enterprise Systems, Storage, and Networking division, which has created the Performance Optimized Datacenters - aka PODs - riff on containerized data centers.
The CFI service can be used to test and install PODs, but the factories down in Houston, Texas, where the servers and storage are built is where the PODs are manufactured. The "Butterfly" Flexible Data Center, launched last July, is a CFI product.
"Anybody can build a shell, but building a shell is not what people want," explains Einhorn. "They want to essentially build what is a giant computer from the ground up." And they have to do this now because, says Einhorn, any data center that is ten years old or older is "woefully out of date" in terms of floor styles, ceiling heights, and power and compute density. With around 75 per cent of the cost of the new construction of the data center coming from mechanical and electrical systems - and these items can change three or four times in the course of a 15-year lifespan of a data center - it is important to try to get it right the first time.
"By bringing the IT company that has all of this expertise internally into the beginning of the process, you can get a flexible and expandable data center and avoid over-provisioning and under-provisioning, both of which are disastrous."
A typical engagement with the new Critical Services Implementation service, which involves everything from design to construction management, takes somewhere between 20 and 30 people just for the design and then construction managers on top of that. Pricing for the services depends on the scope of the project and how involved HP is in the process. A data center might, in effect, be a giant server, but don't expect HP to treat it like a ProLiant and provide a price. Although that is precisely what HP should be doing, come to think of it. ®