HP isn't getting out of the public cloud business.
Bill Hilf, HP's senior vice president for cloud product management, made that painfully clear in a company blog posted on Monday, April 13, after being quoted in the New York Times last week as saying that it doesn't make sense for Hewlett-Packard to be trying to compete mano a mano with Amazon Web Services.
Under the headline, "HP Comes To Terms With Cloud," The Times' Quentin Hardy reported Hilf as saying: "We thought people would rent or buy computing from us. It turns out that it makes no sense for us to go head-to-head."
That was taken as a statement of HP defeat and retreat. Actually, any of Amazon's competitors could have said it as well, with the possible exception of Microsoft. The real problem with what Hilf said is the unsettled environment in which he said it.
For those who still claim to be competing with AWS, we will soon get a measure of just how desperate the race has become.
Amazon announces its first quarter results on April 23. It's slated to say at that time what AWS's revenues are, and its earnings, if any. Everyone's going to look to see whether AWS is making any money in cloud services. I'm going to see whether the ratio of public cloud infrastructure that Amazon enjoys over 14 of its competitors combined -- five to one -- might also apply to revenues. (None of Amazon's competitors are prepared to break out their revenues, if that's any clue.)
[Want to learn more about containers? See Containers Explained: 9 Essential You Need To Know.]
Hilf was trying to give voice to a new reality in HP's approach to cloud computing and, in part, explain why he was there in the public spotlight, and HP's senior vice president for cloud, Marten Mickos, was not.
"In the past week, a quote of mine in the media was interpreted as HP is exiting the public cloud, which is not the case. Our portfolio strategy to deliver on the vision of Hybrid IT continues strong," Hilf wrote in Monday's blog post.
What HP Can Offer
Hilf was saying HP must stick to the thing it can do, which is to provide an alternative data center and self-service provisioning to its regular customers. HP doesn't want to compete, let's say, in the ruthless price war that Amazon, Google, and Microsoft are waging with each other, or in the many features that Amazon is piling on top of its infrastructure platform.
Nevertheless, there's still a role for HP to play, if only HP could clarify what its role will be, since it's a trusted partner of long standing inside many companies. For those companies that don't want to go all in on the cloud, HP might be just the right place to send enough of their workloads so that they don't have to build another data center. It could be the place they go when a short period once a year generates more traffic for their customer-facing applications than they can handle. It could even be the outside data center they turn to for system image storage when they're improvising a disaster recovery site.
These are not the customers that will generate multimillion-dollar accounts with their cloud provider, and HP now knows it.
That HP wasn't competing on the same level has been clear for some time. When other cloud services announce price cuts, HP remains silent. When Amazon announces CodeDeploy and Amazon Machine Learning, HP remains silent. It inadvertently exposed itself when it acquired Eucalyptus Systems in September and described the high-profile Eucalyptus CEO Marten Mickos as general manager of its cloud business. Lately, Mickos has been absent from its cloud announcements. HP veteran Hilf, who knows the fiscal restraints HP must operate under and how they will play out in HP culture, has taken the helm.
Companies without deep pockets probably shouldn't commit to compete in cloud computing -- it takes a huge capital investment. Amazon has rejected the assertion that it costs $1 billion to create a large-scale modern cloud data center, but it hasn't said exactly how much it does cost. Let's say $250 to $350 million might be a good range to consider, if you're thinking of making a modest start. And of course you will need a chain of these data centers.
Hilf in his blog listed HP's strong positions in virtual private cloud and managed private cloud services, noting Forrester had named HP a leader in private cloud in China in the first quarter of 2015. HP's managed cloud services are likely to suffice for some, and to prove profitable. Since shifting its emphasis into managed services, at the expense of its public OpenStack cloud, Rackspace's revenues and stock have recovered nicely. Instead of trading near its 12-month low of $26.18, it doubled, and closed at $52.99 on the day Hilf's blog posted.
Hilf also wrote that HP is "one of the largest OpenStack-based public clouds," and would continue to be a leading figure in providing OpenStack from the public cloud arena. On that point, it seems the ice he's standing on just got a little thinner.
Some notable parties have adopted OpenStack for their private cloud build-outs, including Comcast, PayPal, and The Gap. Many other companies are considering it as the best open-source option. But implementations remain difficult and expensive, according to some who have tried. OpenStack's future is clouded, not so much by lack of confidence in what it will eventually produce, but by whether what it's producing will still be relevant by the time it gets there.
Many companies have conceived of their need for an OpenStack cloud as one of the few ways they could get to a more automated, agile infrastructure. They see a shimmering image of the architecture they wish they had, and it looks a lot like Amazon's or Google's. But the image they're pursuing is reflected in pool of oil, and Linux containers are taking an eggbeater and applying it to the surface.
Containers, Containers, Containers
Instead of moving to a full cloud infrastructure, companies are finding they can automate several of the steps they sought in the cloud through Docker containers.
Instead of frequent updates to applications through a uniform cloud infrastructure, they're finding they can frequently update separate layers in a container within their existing and varied server infrastructure. Instead of agility in a more uniform, x86 architecture, they're finding containers give them agility with what they've already got. Instead of needing cloud computing for elastic scalability, they're finding containers instantiate quickly and can be scaled through new systems much smaller than OpenStack, such Kubernetes, Docker Swarm, and open source Deis. (Deis was just acquired by PaaS provider Engine Yard.)
Containers are a new and rapidly expanding technology whose patterns of use have yet to be established. How much will they encroach on what we think of as cloud computing and change its practices, no one knows for sure at the moment. The lines are not just blurred, they're scrambled.
It was into that environment that Hilf made his statements, and it will be in that environment that HP will have to recover its lost momentum in the public cloud. Chances are, Hilf, the Helion OpenStack cloud, and HP will muddle through somehow. But Amazon Web Services is launching CodeDeploy, EC2 Container Service, and Amazon Machine Learning, while HP struggles once again to explain itself. It's not only not competing with Amazon head-to-head, it's also not picking up enough cloud users in its own customer base.
Instead of leading in cloud computing, it appears to be settling in to follow a trail of crumbs behind the pack. Yes, it's still offering public cloud computing, but its back-and-forth in an uncertain environment has left it with a handicap it still needs to overcome.
Attend Interop Las Vegas, the leading independent technology conference and expo series designed to inspire, inform, and connect the world's IT community. In 2015, look for all new programs, networking opportunities, and classes that will help you set your organization’s IT action plan. It happens April 27 to May 1. Register with Discount Code MPOIWK for $200 off Total Access & Conference Passes.