How HP's Dave Donatelli Plans to Win the Next Generation Data Center
HP's growth numbers are hard to miss and the company could drive massive industry change.
There are three founding elements to this impressive growth: One is that HP aggressively uses its own technology now, and the second is that HP, by having all three hardware legs, has been able to both better integrate and better reduce the number of redundant components in its integrated solutions. The third is that at the core of HP's solution are servers that have to compete in the most price-sensitive and lowest margin segment of the three, driving efficiencies that HP argues no one else can match.
HP has two exposures: It doesn't have telephony, which could become a problem as Cisco scales into its space over the next decade, and HP is light on software, which is why it hired Bill Veghte – one of the most highly-regarded Microsoft executives – and has Leo Apotheker as its CEO, partnered with Ray Lane as its new chairman of the board. Bill was largely responsible for Microsoft's most successful software launch since Windows 95, Windows 7.
Let's explore both the advantages and disadvantages.
Step back and take a look at what might be termed "the next generation of the glass house," which is being driven by three entities: Acadia, a joint project between EMC, Cisco and VMware building on a concept called "VBlocks" or packaged data centers; Microsoft, which is exploring the idea of packaged data centers; and HP, which is integrating storage, networking and servers into single solutions that both embrace Microsoft's solution and mirror Acadia's.
What appears to give HP a potential advantage is that all of its core components report up to the same executive, Dave Donatelli, and can be designed to work together as opposed to being fitted after the fact. This allows HP to do some interesting things with the components and things like cooling systems, transformers and even the way the systems are wired and designed to be interchangeable.
This is manufacturing 101 and one of the ways other high-volume industries, like automotive, keep costs down. For instance, a cooling assembly, rack or power supply can be used interchangeably for storage, networking and server rack mounted systems, which drives up the manufacturing volume of those components, lowers service and manufacturing inventories and thus lowers the manufacturing, logistics and support costs.
One of the most impressive aspects of this is that by integrating the components, not only can you interchange them in the rack, but what otherwise would be a shopping cart of wiring and networking gear, can now be cut down into two little boxes, which dramatically cut installation costs and both service cost and complexity.
On top of the hardware, HP has a management solution that blends storage, network and server management into a single console allowing the coordinated management of all three areas with drag-and-drop ease of use. As a result, HP, when bidding on a blended next generation data center solution comprising all three components, can argue savings of between 30 percent and 50 percent over alternatives. This savings is what creates its rapid growth and success. HP couldn't have even conceived of doing this if it hadn't first aggressively moved to deploy its own technology internally and there is a good lesson in that.
HP has two major competitors that it isn't ideally positioned for: Oracle and Cisco. Oracle can optimize its software for its hardware solutions and, in the long term, this likely represents HP's greatest risk at least partially because Oracle's CEO appears unreasonably focused on the hardware solutions. However, in building a solution, typically you'd start with a hardware foundation because Microsoft's strategy of staying away from hardware until the foundation is complete has proven vastly more successful, as referenced by Sun's failure.
It isn't even clear the market will allow a company that is as tightly integrated and controlling as IBM once was to exist and IBM is actually much closer to this goal than Oracle. HP is moving aggressively to contain this exposure but it will take years and its best path may remain partnering with companies like Microsoft, which can then better position itself against Oracle.
With Cisco, the exposure is telephony and the move to consolidate telephony into solutions like this. This has been attempted before by IBM with its acquisition of ROLM, and AT&T with its acquisition of NCR. Both failed, and as one of the analysts who did the post mortems stated, they failed because it was like mixing oil and water – it fundamentally couldn't work.
With VoIP technologies, the environments are becoming integrated. However, the networking market at the core of this is comparatively high margin and is exposed to aggressive price competition. In addition, there currently is a trend to abandon PBX-like technology and move to virtual cloud-based services tied to cell phones.
Given the speed of change, it is likely that this last trend will mitigate HP's exposure long before Cisco can complete its solution, and Acadia, where the greatest risk likely exists for HP, doesn't appear to be integrating telephony yet anyway.
Wrapping Up: Solid Advantages and Mitigated Risks
It is easy to get lost in the rhetoric between HP and Oracle and to lose track of the basics surrounding what appears to be the rebirth or recreation of the data center. HP seems to have thought this all the way through largely because it has tightly integrated its own needs for affordable data center solutions to its roadmap, and hired folks who could address the related complex problems.
I'm kind of surprised this level of integration hasn't happened sooner but I expect it will now as HP's growth numbers are hard to miss and the company could drive massive industry change. Oracle clearly should have picked an easier target.