Canonical founder Mark Shuttleworth's keynote at the second of this year's Openstack Summits was considerably less controversial than the last, where he explicitly compared the price between Red Hat and Ubuntu's services in Vancouver. However he still had plenty to say when Computerworld UK caught up with the hobby astronaut and entrepreneur on the show floor.
When asked to expand on his thoughts about IBM's recent acquisition bid for Red Hat, Shuttleworth wished them "every happiness in their marriage", with some caveats.
"It depends how forcefully they can steer RedHat Enterprise Linux (REL) on-prem users to the IBM cloud," he said. "At the end of the day, that's the only way for them to get a premium on REL itself. They've paid a very high premium, so they have to generate a premium, and the only way to do that, would be to use on-prem REL as a forcing function to get workloads onto the IBM cloud," he said.
"If they can do that, then yeah, it does make sense. IBM can be a real force in the public cloud, they have a long history of operating other people's data centres very profitably for IBM, so if they can extend that into public cloud and use REL as a lever for that then I think it would make sense for them."
Red Hat's Nick Barcet earlier this work confirmed to Computerworld UK that the company would commit to supporting Openstack for another 10 years. Shuttleworth made the same claim in his morning keynote on the second day.
"It's pretty straight forward, we published [Ubuntu LTS] 18.04, we've had a ton of enterprise engagement, and as we get deeper and deeper into the more mission critical type workloads that don't change so often, when people are saying: 'how long are you willing to support it?', that's always been sort of a balancing act.
"But there are now enough customers across some of those long-term sectors like finance, telco and industrial that it's good business sense for us to offer that 10-year support commitment. And, you know, I caucused widely within the engineering team and they're committed to doing a really good job of it as they're doing today, so everyone's excited."
The end of the big tent?
Openstack is arguably going through a reframing of its core branding, and indeed from now on the Openstack Summits will be called the Open Infrastructure Summits - perhaps signalling a pivot to a wider portfolio, rather than pushing the various flavours (and many projects) of Openstack at the core.
This comprises Kata containers, the distributed edge project StarlingX, automated provisioning toolset Airship, to name just a few. There are some comparisons to be made to the organisation's prior 'big tent' approach, where it tried to be all things to all people and consequently saw a lot of vendors joining the fray - and then leaving - once it got back to its core offering of open source private cloud.
Shuttleworth was at that time one of the biggest critics of the big tent approach - so Computerworld UK was curious to hear his thoughts on this diversification into wider infrastructure layers.
"I think the idea of open infrastructure is exactly right," he commented. "That lovely expression, 'the future is already here, it's just not evenly distributed' - if you look inside a Google, it's open infrastructure, so that's the future. That's the future for everybody, it's just not evenly distributed.
"So I absolutely believe in that, it's what I'm committed to delivering, when we help companies build their new data centres, their clouds, their new application management systems and so on, it's all moving them towards open infrastructure.
"If I have a concern about it, it's simply the loss of focus on Openstack itself. At the end of the day Openstack is a big complicated thing, it can easily become unnecessarily complicated, I think we've been a force for simplifying Openstack and focusing people's attention on the piece that's actually necessary.
"I hope that the changes that you see unfolding at the Foundation don't result in yet another big tent where it's completely unclear as to what you're supposed to be using or how you're supposed to be using it, just because there's a new bunch of vendors that have showed up at the Foundation because of a new mission. What matters is, do you help people get something done? Focusing on what people want and helping them get that done is, I think, the most important thing."
In general though, Shuttleworth believes the project is headed in the right direction - what he calls the core mission of VMs, discs and networks on demand - so that "core piece continues to get better and better".
And he also believes that the general perception of Openstack is over a hump of negativity (although some others in the wider open source community might disagree). "People started to understand that Openstack well delivered, well operated as a managed service or as part of a community like the Canonical Openstack community delivers a lot of value and gets the job done, so that's great," he said.
However he added one area he would like the Foundation to help steer the project is in the hiring of full-time project team leads [PTLs] to "get rid of a lot of the politics and posturing and ballot stuffing" that happens when you "invite people to get elected for various important roles". A model closer to the Linux Foundation, perhaps, where contributors trust Linus Torvalds to appoint the right people for the job.
Future use cases
Many of the next major use cases will relate to the form of distributed computing closer to what is colloquially referred to as 'edge', and the Foundation has stressed this for some time now too. Especially when 5G gets up and running, delivering compute closer to the user will become a very real proposition and will ultimately underscore the importance of telcos as network providers.
Canonical is in a number of pilots with those telcos and companies - Softbank, NTT Docomo, AT&T and Deutsche Telekom to name a few.
"They all have slightly different ideas about how they want to do that," said Shuttleworth. "But the key problems they will have to solve are physical provisioning.
"If you've got a server across the country, you don't want to drive a truck there to re-flash the operating system, so [Ubuntu cloud style provisioning for physical servers tool] MaaS solves that problem very nicely. MaaS also gives them the flexibility to take that rack and put VMware on it, Ubuntu on it, Windows on it - different machines can have different operating systems as they want it, and it leaves more possibilities open to how they might use that rack of compute, because people haven't figured out everything they want to do there."
Finally, in his keynote, Shuttleworth pointed out that Ubuntu widely supports many types of security cameras, including those with facial recognition technologies. Sporting a new beard that would be ideal for villainously running his hands through, we asked how he felt about carving out a niche in worldwide techno-dystopia - doing "his bit for 1984" as he put it.
"It's super twisty," he said. "Look, it's really tricky. I get the same feeling about reading about Ubuntu being used in rocket launchers, that's pretty ugly.
"But I figure my job is to make things - to make open source and enable basically, build a bridge between the people who are innovating and creating open source and the people who are consuming it. I don't want to do the 'guns don't kill people' routine because I think that's pretty gross.
"But I think open source has a huge number of really positive outcomes, so when I'm putting time into making Ubuntu amazing and getting the best open source software in everybody's hands, I feel like I'm doing that for all the people who are going to do wonderful things with it... and if there are people doing shitty things with it, I'll find some other way to hold them accountable."