Some of the most prominent messaging that emerged from the recent SUSEcon event, held in Nashville last month, was that it was now the world's largest, independent open source company (particularly in wake of the Red Hat-IBM deal). But what does this spell for its relationship with OpenStack? We met with Alan Clark, a SUSE employee and chairman of the OpenStack Foundation board, to find out.
SUSE OpenStack Cloud 9, based on OpenStack Rocky, shipped this week. "SUSE has, has been, and will continue to ship OpenStack as part of their solution set," said Clark, speaking with Computerworld UK at the Open Infrastructure Summit in Denver, Colorado this week. "But if you look at SUSE's strategy - and particularly now that they're an independent open source company - their strategy is expanding."
He added that OpenStack remains a critical piece for SUSE, but that the company also recognises that its customers "aren't homogeneous".
"They're not just containers, they're not just virtual machines... they're not just bare metal machines, they're not just Windows and Linux - all kinds of stuff. And they need all that to work together," he said. "And so you need virtual machines to work with containers to work with bare metal services. The use of bare metal is dramatically expanding, we're seeing it driven by the use of machine learning and data analytics and so forth. So you need infrastructure that supports all these up-and-coming technologies."
That messaging is certainly in line with many of the speakers at this week's Open Infrastructure Summit. Previously known as the OpenStack Summit, this is its first iteration of the event with the new branding, which emphasises that, although OpenStack remains core, other projects that are emerging, such as Airship - the cloud provisioning and management tools designed to make Kubernetes and OpenStack play nicely - can operate independently of it. Everyone is talking about interoperability.
Clark says that although this is the first big public event as the Open Infrastructure Summit, the OpenStack Foundation had been putting in much of the groundwork to encourage cross-community participation years ago. For example, bringing the OpenStack and [open source network function] OPNFV boards together. And now that's evidenced by the fact that staff, community members, and engineers are actively collaborating with Kubernetes and vice versa.
"From the SUSE perspective, we recognise that our customers don't buy individual projects," said Clark. "They don't come in and say: I want OpenStack. They come in and say: here's my problem, I need to solve it, how do I do that... solutions are typically a combination of multiple open source projects, we as a distribution put those together, which means that we have engineers in all these different communities, with the goal of making sure that those technologies are enterprise ready."
The beta of SUSE OpenStack Cloud 10 will open in the coming months. Just how central will these additional 'Open Infrastructure' projects such as Airship feature in upcoming SUSE enterprise releases?
"We're actually containerising the whole thing [OpenStack Cloud 10]," said Clark. "We're using Airship to containerise the control plane. So there's a case where we're taking other project technologies like Airship, we're leveraging our containerisation technologies that we've gotten from the CNCF efforts, and the networking efforts, and pulling those together to build a whole new style of OpenStack.
"We recognise that others have containerised OpenStack in the past, but our approach is leveraging our experience from all these other groups - we think the way we're doing it is very unique. And the reason we're doing this - it's not a release, it's just early access, so that we can get feedback from our customers, and learn from that before we get to the seriousness of rolling out the release"
But is there a danger that these new projects emerging in the ecosystem might struggle from a sense of inertia, or find it difficult to sustain their momentum - is there a risk of zombie projects once the excitement and the sheen has worn off?
Jonathan Bryce, executive director at the OpenStack Foundation, told us that these new collection of tools and projects have emerged organically from the community to solve specific problems.
Clark is dismissive that the risk is there.
"My experience is that doesn't happen, people won't walk away from the project," he said, adding that one of the keynote bare metal demonstrations was based on code from one of the longest-running OpenStack projects, Nova. Or the Linux Kernel, today, or that the Kata containers launched under the auspices of Open Infra were a combination of hypervisors and Firecracker bundled into a new package.
Another recurrent topic this year was the role that open source foundations play in general. OpenStack Foundation COO Mark Collier, for example, was interviewed by TechCrunch journalist Frederic Lardinois on the subject.
When we interviewed the CDF earlier this year (a Linux effort along with Jenkins/CloudBees and Spinnaker) - they said that there was some early skepticism about yet another Foundation arriving on the scene.
Is there a tension between the inherently distributed and decentralised nature of active open source projects at the grassroots, and the role of foundations? Are strange new silos emerging?
"What we've been talking about at the leadership level is not to create these silos, not to do reinvention," says Clark. "That's why we're reaching out and working with Kubernetes and so forth - we don't want silos and nobody wants... particularly with my SUSE hat on, I don't want to see three projects doing the same thing. We don't have the time, energy and money to do that. We very much want this open collaboration to happen with adjacent communities pulling together.
"That's our aim, that's our goal, anything we're doing, we're trying to reach out and make sure that doesn't happen. There are always competing ideas. Even within CNCF itself, if you look at projects, there's competing projects.
"That's a natural part of business, and it's a natural part of open source as well. What you typically see is everybody eventually adopts one, or the ideas merge. I really like it when the ideas merge together, and things go forward."
Clark added that the role should be to encourage development and standard interfaces, to make them interoperable so that whatever project is being used doesn't destroy the rest of the stack.
"Keep that interoperability - let the communities decide and let people choose," he said. "We don't want to take over choice... we're trying to get rid of the tension, enable innovation, let's not reinvent the wheel where it's not needed. Allow innovation - sure there'll be competing ideas, there always is. Let those flourish, let those go - let's flesh out and figure out which one works best. There might be multiples that work better in certain markets or certain environments than others. That's fine."