Although having a laugh at so-called “enterprise clouds” is a respected pastime in some circles, there’s an argument to be made that they do serve a legitimate purpose. Large-scale public clouds such as Amazon Web Services, Microsoft Azure, and Google Compute Engine are cheap, easy and flexible, but a lot of companies looking to deploy applications on cloud architectures simply don’t need all of that all of the time.
So says Bill Hilf, the senior vice president of Helion (the company’s label for its cloud computing lineup) product management at HP. He came on the Structure Show podcast this week to discuss some recent changes in HP’s cloud product line and personnel, as well as where the company fits in the cloud computing ecosystem. Here are some highlights of the interview, but anyone interested in the details of HP’s cloud business and how its customers are thinking about the cloud really should listen to the whole thing.
Download This Episode
Subscribe in iTunes
Amazon matters . . . and so does everything else
“First and foremost, our commitment and focus and investment in OpenStack hasn’t changed or wavered at all,” Hilf said. “It’s only increased, frankly. We are fully committed to OpenStack as our core infrastructure-as-a-service platform.” HP has been a large backer of the open source project for years now, and was building out an OpenStack-based cloud platform exclusively before acquiring Eucalyptus and its Amazon-Web-Services-compatible cloud technology in September.
However, he added, “As we started working with customers around what they were looking for in their overall cloud environment, we did hear the signal loud and clear that the AWS design pattern is incredibly relevant to them.” Often times, he explained, that means either hoping to bring an application into a private cloud from Amazon or perhaps moving an application from a private cloud into Amazon.
Hilf thinks vendors targeting enterprise customers need to make sure they’re selling enterprise what they actually want and need, rather than what’s technologically awesome. “Our approach, from their feedback, is to take an application-down approach, rather than an infrastructure-up approach,” he said. “How do we think about a cloud environment that helps an application at all parts of its lifecycle, not just giving them the ability to spin up compute instances or virtual machines as fast as possible.”
Below is our post-Eucalyptus-acquisition podcast interview with Hilf, former Eucalyptus CEO Marten Mickos and HP CTO Martin Fink.
Enterprise applications might be boring, and that’s OK
Whatever HP’s initial promises were about challenging Amazon or Microsoft in the public cloud space, that vision is all but dead. HP still maintains a public cloud, Hilf explained, bu does so as much to learn from the experience of managing OpenStack at scale as it does to make any real money from it. “It not only teaches us, but allows us to build things for people who are going to run our own [private-cloud] products at scale,” he said.
But most of the time, he said, the companies that are looking to deploy OpenStack or a private cloud aren’t super-concerned with concepts such as “webscale,” so it’s not really in HP’s financial interests to go down that path:
“[W]e don’t have an intention to go spend billions and billions of dollars to build the infrastructure required for, let’s say, an AWS or an Azure. . . . . It’s not because ‘Oh, we don’t want to write a billion-dollar check,’ it’s because [with] the types of customers we’re going after, that’s not at the top of their priority list. They’re not looking for a hundred thousand servers spread across the globe. . . . Things like security are much higher on their list than the intergalactic scale of a public cloud.”
“What we typically hear day-to-day, honestly, is actually pretty unexciting and mundane from customers. They’re not all trying to stream the Olympics or to build Netflix. Like 99 percent of the enterprise in the world are doing boring things like server refreshes or their lease in a data center is expiring. It’s really boring stuff, but it matters to them.”
“If a customer came to me and said, ‘Hey I need to spin up a billion instances to do whatever,’” he said, “. . . I’d say, ‘Go talk to AWS or Azure.’”
Get over the talk about lock-in
Despite the fact that it’s pushing a lineup of Helion cloud products that’s based on the open source OpenStack technology, Hilf is remarkably realistic about the dreaded concept of vendor lock-in. Essentially, he acknowledged, HP, Amazon and everyone else building any sort of technology is going to make a management interface and experience that’s designed to work great with their particular technology, and customers are probably going to be running multiple platforms in different places.
Hilf thinks that’s a good thing and the nature of business, and it provides an opportunity for vendors (like HP, coincidentally) with tools to help companies get at least some view into what’s happening across all these different platforms.
“People often use the term ‘lock-in’ or ‘proprietary.’ I think the vendors get too wrapped up in this,” he said. “The enterprise is already through the looking glass. They all know they’re going to have some degree of lock-in, it’s just where.”
Why boring workloads trump intergalactic scale in HP’s cloud biz originally published by Gigaom, © copyright 2015.
Related research and analysis from Gigaom Research:
Subscriber content. Sign up for a free trial.
- What developers should know when choosing an MBaaS solution
- The top 10 cloud trends for 2014
- How the mobile-first world will transform the data center