Could a return to an integrated technology 'stack' turn out to be the perfect way to build and grow private clouds for the enterprise? HP thinks so - and has some strong arguments to back up its claims, too. CBR reports.
In April last year, Executive Vice President and general manager of Enterprise Servers, Storage and Networking at HP Dave Donatelli was quoted as saying that, "When you're the largest buyer of disk drives in the world, the largest buyer of memory in the world, the largest buyer of Intel products in the world - chances are you're going to get the best prices". The "you" in his sentence was his new employer, HP. And that set of superlatives is one way of saying that HP thinks it's changing the game when it comes to enterprise infrastructure.
Donatelli's remarks need, in fact, to be seen in the context of the ongoing serious, long-term commitment by his company to basically rebuild its entire systems architecture offerings: to fuse together the server, the storage capacity and the networking functions of the stack in a way that we haven't seen in the industry since the mainframe.
A year ago, HP set just such a course with its announcement of a new set of products and services that included, for example, an expanded CloudSystem portfolio, its VirtualSystem for Microsoft and VirtualSystem for Superdome 2/HP-UX, as well as other new hardware and software offerings it positioned as offering "highly optimised, fully tested virtualised application environments".
Though targeted at the data centre - the October 2011 announcement specifically promised among other benefits for such facilities a chance to cut such lifetime costs by up to 37% - it's now becoming clear that what Donatelli and his team really have in mind is something much bigger: the Cloud. In effect, HP says it's building the ideal platform for massive private enterprise Clouds - and that it's really the only serious game in town for this type of operation, too.
Pay to play
HP is unabashed about the fact it's had to partly buy its way into this position. Insiders point to three significant acquisitions that bear out how serious it is in trying to forge an integrated Cloud-ready set of technologies. The first was its November 2009 $2.7bn swoop on network equipment firm 3Com; the second, its buyout of storage leader 3PAR in September 2010 for $2.4bn; and the final piece of the jigsaw - though not perhaps immediately obvious in terms of infrastructure - was its October 2011 $11bn purchase of UK enterprise software firm Autonomy. That's a combined outlay of the best part of $16bn in the past three years alone, on top of which the company's significant R&D investment and other spot technology purchases has to be factored in.
Of course, in a way getting storage industry veteran Donatelli on board in April 2009 was also part of this programme - after 22 years at EMC, executives like him don't come to companies that really had no convincing story in his area of expertise without commitments he was coming to build something real. What does the company - which is going through significant restructurings and cost-cutting exercises in parallel - think it's getting back from this expenditure?
"We've become used to the idea that servers, storage and networks are all separate, that they are different silos of IT activity," says Chris Johnson, the firm's head of all things Storage in Europe. "In our own past, HP was more known for servers and not really in the other two areas at all. But we are convinced that the next way forward for customers is to be able to get rid of these distinctions and move to a converged architecture, which we think is the best combination to support Cloud. By our acquisitions and by infusing our own IP into the new product set, too, we our going to offer the market a truly unique, tightly-integrated stack that no-one else in the enterprise market will be able to compete with, too."
Back to 'the bad old days?
For some CIOs, that ambition may have set off some alarm bells. Isn't that kind of 'one computer architecture to rule them all' vision what Big Iron was all about - the monolithic, mainframe mindset that allowed certain vendors to dominate the glasshouse (the data centre of its day) and customer budgets for decades? The whole idea of the mini revolution of the 1970s and open systems in the 80s, after all, was to break the cycle and get more choice into the market?
Johnson counters that even if that were true, times have changed and CIOs have to approach the problems they are dealing with now with a new mindset. "IT teams are exhausted by having to act as systems integrators," he told Computer Business Review. "They're kept so busy determinedly seeking best of breed for all parts of the system and then spending all their time gluing it all together that they're not left with enough resources to do what they really are there for - helping support the organisation drive growth and control cost. They are ready for a combined approach once again that will save them time and also money; without proper integration and convergence, they're just going to spend their time watching their budget get eaten away by buying more and more storage and servers to throw at the problem."
For Donatelli, the problem is also about the fact that lots of resource gets put into data centres without the kind of results back clients should really be getting. "We've gone through a building boom, like crazy, of next-generation data centres," he has said. "Depending on the size of your company, you're spending hundreds of millions of dollars, to in some cases billions, on them... and as soon as you're done, you always have regrets."
The way forward is to still build big, is the idea - but in the Cloud. That's a domain where size and scale soon starts to dwarf conventional CIO frames of reference, says HP. Again, Donatelli: "We have a social media customer who buys 40,000 servers at a time. Most of our enterprise customers don't have 40,000 servers total."
Cloud, yes, perhaps; though as we all know, not everyone is convinced as yet, the data still does suggest that demand is moving "up and to the right," as they say in terms of customer interest in storage delivered this way. Thus last October market-watchers IDC predicted that overall spending by public Cloud service providers on storage hardware, software, and professional services will grow at a compound annual growth rate of 23.6% from 2010 to 2015 - while enterprise spending on storage for the such resources is expected to increase at a CAGR of 28.9%. In other words, within three years, public and private Cloud storage could be in the order of $23bn per year, worldwide.
Patently, HP thinks it will be a significant player in just such a market, via the advantages it sees accruing from this converged architecture move. "We really are the only vendor that can offer all three elements," says Johnson. "Other players are very strong in the server side, or the storage side, the networks side, perhaps in some cases the server and the networking side. But the HP approach is the only one where all three are being brought together in a unified, efficient way - and also with a design principle of making this the perfect large-scale Cloud platform. Plus, the addition of our Cloud services and consulting propositions to the matrix means we are the best option for companies serious about taking advantage of all that private Cloud can offer."
The end of the 'internal IT company'?
The promise, then: before long HP will be able to give the CIO that famous "single pane of glass" that would mean they can with a few points and clicks consume and deliver all IT - infrastructure to software - as a service, knowing that the underlying plumbing was optimised to deliver that the way that's cheapest for him. It's a compelling vision, one that it might indeed take a company as big and focused as HP to make a reality.
It's also going to be a play that doesn't only concentrate on hardware and boxes, by the way. We mentioned the Autonomy purchase earlier as key to this programme - something that might seem odd, given that for all its benefits, that software doesn't seem at first glance to be that connected to storage. But Johnson points out that with that clever box of tricks from Cambridge, HP can start delving into information flows right across a customer organisation, linking data in its many guises (structured to unstructured, emails to SharePoint files) and routing their electronic storage to the best place in terms of security, access and archival.
What, then, is the message to the CBR CIO reader in all this potent combination of technologies and product sets? Johnson says that in actuality, it's really rather simple, at the end of the day. "What's keeping him awake at night is the fact that he knows his storage capacity is growing at something like 50% a year. That's simply unsustainable, it's costing him a fortune and it's also inflexible and slow to implement anyway. He needs to start working toward a way of dealing with all the data in his organisation in a way that's simple to manage and much more cost-effective.
"At the same time, he also knows he can finally stop linking things together at the back end all the time and start concentrating on the front office again at last. He needs to stop being an SI and get back to being a CIO. With the right Cloud platform, he really can do that."