While reading NetworkWorld’s interview with Marius Haas, senior vice president and general manager of HP’s ProCurve Networking business — now part of HP’s $45-billion Technology Solutions Group — I got the distinct feeling that HP might consider a core-switching acquisition if private and public market valuations of logical target companies hadn’t run ahead of economic reality during the past few months.
HP still might make an acquisition of an emerging data-center switch vendor. Options would include those mentioned by Alan Leinwand earlier this year: Arista Networks, Blade Network Technologies, and Force 10/Turin Networks. If it chose to be more ambitious, HP could take a run at Juniper Networks, which has its own data-center switch offerings as well as its Blade-supplied EX2500 top-of-rack 10G Ethernet data center switch.
Juniper is Cisco’s primary competition in the router market, but it isn’t apparent that HP wants to play in that market. In fact, HP probably will conclude that it can derive more value from a smaller acquisition of one of the private data-center startups listed above.
But the timing for such an acquisition probably isn’t right. Arista is self-funded by its founders and presumably isn’t in a hurry to make an exit until it has developed a bigger installed base and greater market momentum. Blade has just raised a new round of financing, in which Juniper participated, and has set a valuation of approximately $230 million. Meanwhile, the reconstituted Force10 Networks, merged with Turin Networks, has been the recipient of prodigious funding and might be looking for a payout that HP would be reluctant to provide.
That’s probably why Haas has taken what he describes as a “pragmatic” stance in relation to HP’s core gap in data-center networking:
You can assume I’m looking at everything. I view the window of opportunity for us as now. The customer base is saying it wants us to step up now. So we’re looking at everything – build, partner, buy. What’s the best investment profile from a return standpoint, not just for our share holders but for our customers? So you should assume that’s the pragmatic approach I’m taking.
From the interview, as well as from everything HP has said and done during the past few months, it is obvious that Haas and his colleagues view Cisco as their main threat in data-center convergence. With regard to Cisco’s Unified Computing System (USC) and the vision that underlies it, Haas said the following:
But we’re not doing it the way they’re doing it. We’re not saying customers need to move into a converged network/storage/compute fabric with a proprietary stack that locks you into a 10 year+ architecture with an as yet undefined TCO model.
We’re saying we’ve been doing this for years. We know what the compute fabric looks like, the storage fabric, and we’re absolutely investing hard in the network fabric. And we believe that establishing your data center on a holistic and integrated infrastructure which includes servers, networks, storage and management, using common architectural building blocks, is the right way to do it.
So the marketing rhetoric they’re putting out there is interesting, but we decompose it and see it’s forcing some trends customers don’t want. It is a closed architecture with proprietary Cisco compute technologies. It doesn’t scale to the capabilities we have, and it doesn’t have overall data center management from an orchestration and automation standpoint, including identity management and policy management. Those are huge gaps they’re not addressing.
It’s an interesting argument, using Cisco’s relative inexperience in data-center computing, storage, and policy management as wedge issues to shake customer confidence in Cisco’s slick marketing pitch.
The rant against “proprietary Cisco technologies” might prove less persuasive. Most customers know that all vendors offer “open” products and technologies until they have the opportunity to invoke some form of proprietary lock-in. It’s in vendors’ DNA. Smart customers know it’s in their interests not to buy into a Manichean morality tale in which one vendor represents good and another represents evil. Save the schlock for Hollywood.
To the extent that HP can focus on its expertise, focus on lower total cost of ownership (TCO), focus on the benefits of being able to do a better job than Cisco of pulling the pieces together into a whole that is greater than the sum of its parts, it will have a good chance to maintain control of the accounts it already owns and of capturing new clientele.
The plan looks good, if not great. The ultimate outcome will depend on how well it is executed.