Celebrating Ubuntu’s 20th Anniversary with Canonical | KubeCon SLC 2024

This is Textron tv. Hey, everybody, welcome back. You know, we're just wrapping things up here.

Uh, we're here ready to do our last conversation. Last interview, Alan has taken off at the airport, and, uh, he, he asked me to put the CTO back on. And even though I'm, uh, head of, uh, of the, the VP of Practice lead for DevOps and app dev at, uh, Futurum, so kind of back in the old chair here doing the last interview.

So we kinda saved the best last, you know, it's like on the voice, the last singer, you know, it's always gonna be really good. So I think we've got that, that case here. So I'm joined by Cedric Ko, who is with, uh, VP of Product with Canonical.

Welcome. Exactly. Thank you very much.

Thank you for welcoming me. You bet. Great to have you here.

So tell me, I mean, so, so you have your eyes on everything I canonical and what's happening. What, what's the, what's at the spotlight here at Kon for you? Well, First, you know, it's always a pleasure to be at Kon, right?

Because we, you know, this is where we met our clone native fans, you know, people that already used Ubuntu for forever, I will say. Mm-Hmm. You know, that it's the 20 years, 20 years anniversary this year for canonical.

Wow. You know, we created Ubuntu in 2004. Really?

So for us, it's a big moment. So being able to share that, celebrate that we are, you know, Ubuntu lovers, and there are many, many of them. So it's, it's actually a great pleasure.

Right. Very cool. Well, congratulations.

20 years. That's, gosh, it doesn't seem like it's been that long. I remember when TU was new, I was like, ah, check this out.

This is really great. It really caught fire quickly. So, you know, we're talking Kubernetes, we're talking containers, we're talking security, sort of the whole ecosystem around these Kubernetes OB observability of course, but, um, container security.

Let's talk about that. I know that's a big thing for canonical. Yeah, it's a, it's, it's obviously a big thing in Q con and in the cloud native space, obviously, but, uh, you know, the, the, the good thing is that for every Ubuntu users, right?

It's easy to use the canonical product on top of it. It's like, uh, the, it's, it's already part, part of the ecosystem so they can use it easily, right? The, the, what we see that a lot of developers in the past ask us also to support not only the Dian package, but also the containers themself, and to have a, an opportunity for them to have the same level of support that they have with Linux distribution, with their own containers, right?

Mm-Hmm. So this is something we are proposing, and I mean, what was very interesting during this, uh, this week at CubeCon is that, uh, our offering is working very well. I mean, the, our proposal value proposition, which is to basically build the containers for our customers, maintaining it, doing all the security, the upgrade, and, and just providing this as a service for them, self-service for them, is actually really addressing the challenge, right?

That they have today is becoming more and more complex, man, maintaining the security of all the components that you have in your containers, uh, ensuring that you have the last version that you manage all the dependency graph Mm-hmm. Is, is really becoming a challenge for all the customers, all the people here. And, uh, it was very refreshing to see that, uh, what we envisioned a few years ago is actually, uh, becoming a real, a reality here.

I did some, uh, few years back, some papers around containerized everything, right? It's a transport package. It's the packaging that gets your software wherever, wherever you want to take it.

Once you've done that, I mean the, the hyperscalers the, the cloud pro providers have been, you know, building around your technology for one, do you see that same thing happening in the enterprise, them moving to an Ubuntu or Debian kind of platform to be able to containerize? And is that more of a platform thing or development or production? Where, where is it happening the most?

I think what is, what is happening is that now, you know, the, the infrastructure is becoming, you know, uh, mature. Even the cloud native infrastructure is becoming mature, mature. I will not say it's a commodity because there is a lot of innovation in this space, and obviously, uh, people want to rely on it, but people are more caring now about how they can deploy easily their containers and, and how they can deploy the infrastructure there, sorry, the application on different type of infrastructure and, uh, and they want the, the below to be abstracted, right?

And, um, so, so being, really having the ability when you de you develop your containers to deploy on every cloud the same manner, the same experience is actually key, right? And this is what we're providing with Ubuntu, because Ubuntu is bylaw by far the most distributed Linux distribution, right? And it's really the solution if you want to do multi-cloud, this is how it was, uh, designed, uh, at, at day one.

So, so we see that the, the developers, this is what they're looking for, right? Mm-Hmm. They come to the booth and say, okay, I'm using Ubuntu.

I love Ubuntu. I want to see what I can do with on top of Ubuntu with canonical on many clouds. And in a way, I don't want to care about the infrastructure anymore, how you can help me, right?

This is basically the, the conversation. And Another paper I did about two years ago was thinking about ai and then, then it wasn't so much generative ai, it was all around, uh, kind of ML ops, but containerize your, your AI development environment with data engineers, you know, AI engineers, knowledge experts as you go through that cycle. 'cause they're eventually gonna put it into a DevOps or some kind of a deployment path.

And if everything's containerized, you sort of like already got the path to get there. Do you see anything like that happening in the AI world, adopting containers? Is that a trend that Oh, yeah, yeah, yeah.

I think the, you know, all the things that we discussed about few years, I mean, few years ago or 10 years ago, 15 years ago, about DevOps is now happening in the MOPS space, right? Mm-Hmm. So now we are not only speaking about containers, we're speaking about models and how we can manage those dependencies, right?

Mm-Hmm. And, um, and, uh, what is very important for us is first to be sure that, you know, we're using, we are building something which is composable one, and also it's veneer the vanilla open source, right? So that's why we're promoting tools such as Q Flow, er, um, uh, ML Flow, et cetera.

All those tools, which, uh, exist are quite mature now. And enable the AI developers to rely on something which is, you know, standard, a uniform platform standard, and on top of which they can build the application without being locked in, into specific platforms, right? That's very important for us.

But the main challenge for them is actually how can I maintain all these complexity, all those, the, the different containers, different models that are very, very, I will say, sticky with the hardware, right? The models needs to be optimized for the GPUs you have below, right? And the inference engine depend, depend on what you have below in term of hardware.

So it's much more complex than what we thought in term of DevOps in the past. And what we're proposing, uh, with our technology is to create this infrastructure for the developers to, as I said, to have something composable open source and, um, fully efficient in term of distribution of the, of the, of the workload, right? Of the workload that they are.

Is that Some more kind of instructions or metadata about where, what kind of hardware you want your container deployed? Yeah. You need, if you're using GPUs for inferencing or, or maybe you know, uh, neural processors for in, You know, that that's, that's, that's something that we, the asset we have, because we know the hardware below is thanks to the Ubuntu, we know how to optimize the, the Ubuntu on top of the hardware, how to, uh, signal or, or, or inform, uh, the application on top what is in below in term of hardware.

So I think the full end-to-end platform is actually very useful in the m lops, uh, space. Uh, because then the developers, you know, again, again, see that everything could be abstracted and they can just focus on the development of their, uh, AI applications, right? This is, I think that this was a discussion of the previous year and previous Q actually, now people understand the challenge now they are looking for, they're looking not only for forms of, for, for solutions, they're looking for an ecosystem, standard ecosystem that will stay for, for a long time.

Mm-Hmm. Because the main challenge in the AI is that it's moving so fast that if you are selecting one technology, maybe in six months from now, this technology will be replaced by another one. So it's very important to rely on, you know, the ecosystem, which is stable.

And I mentioned few, few, few components like Q Flow ER and stuff like that. We're very active in the community, in those communities, very active in promoting those tools, again, in the Van Air. So when you have really strong brand, I mean canonical, but especially with Ubuntu, sometimes that's what you're known for.

And it doesn't always, the other things always don't shine quite as brightly. The other things that you're doing. What do you think some of the most important work that you're doing or products that you're offering that maybe it's not the best kept secret, but, you know, it was really a lot happening with what, what you're doing in that area of canonical?

What is, I mean, your question is what is the most, I would say, valuable product or about the flagship products Place? Yeah. What do you wish people knew more about?

'cause you're doing some really great work. I think, I mean, people usually when they come to a booth, they say, oh, you have Ubuntu. I know, oh, you have, uh, you know, you are very involved in OpenStack or thing like that.

So we know that what we did not know is that you are building a full portfolio of application cloud native application on top of it. And we can actually leverage on this portfolio with just one subscription, which is Ubuntu Pro subscription. So we already are Ubuntu Pro client customer, and we can enjoy all the portfolio of application you have, like I mentioned, Q Flow, but also Postgres, Kafka, open search, you know, all the com, those tools which are part ecosystem.

So, so they discover that actually they can use all the canonical port portfolio as a pass, basically a platform as a service, right? Mm-Hmm. So they can build their own platform as a service with a clinical portfolio.

That's basically the, usually the main, the main, uh, surprise, I will say surprise, that we, that we, we, we acknowledge and we witness on our booth and for the, I mean for the people who are open to user, but they did not know that we have those components right now in qm, specifically when they know canonical, they know portfolio. What was important is to promote our Kubernetes edition. And the fact that this core Kubernetes edition is boosted, I will say by the SNAP format, the juju on charms technologies that were, uh, putting around Kubernetes in order to simplify the lifecycle management to be able to have mechanisms to deploy at the age.

And so, and, uh, and having orchestration around it. And, uh, this is all what the Kubernetes, uh, the experts discover with us, that actually they can use a lot of cloud native tools around the Kubernetes, uh, tool toolkits that will really help them to deploy the application in different type of configuration, right? The edge, edge computing create another level of complexity in terms of management of the, of infrastructure.

So it's very important to rely on tools. That's the one we are providing, right? Like charms with juju, which are basically Kubernetes operators, right?

Plus, uh, the SNAP format, which helps them to, you know, to deploy securely their, um, their applications and the Kubernetes edition as well. Very good. com.

It's a new site. And, and part of what you're saying seems like that'd be very attractive to platform engineering teams, right? Exactly.

Exactly. You, you might argue that essentially they're trying to create different forms of a pass for them to be able to develop, deploy, test on, et cetera. Seems like that would be a, a good headstart if not a lot of the work done for you.

I think, you know, the, um, when people speak about platform engineering, the, the, you know, it's a very large domain where coming from, you know, platform as a service, self-service for, or with a self-service portal for developers down to cloud management platform, right? And we have so many actors here, even, even today, if you look and so many actors working in this space, that, uh, it's important for us, you know, to rely on this dynamic, this this innovation dynamic on the open source dynamic behind this, right? So, so we are not trying to create yet another platform engineering solution.

We're trying to provide components in order for others to actually build their platform engineering. Mm-Hmm. Right.

So I think you, you, you mentioned it completely correctly that we, we have very good interaction with platform engineering teams because they have all the different platform engineering teams. They have their own needs, right? And what they're looking, uh, for or looking at working with us is to have a toolkit, a set of toolkit to build their own platform, right?

And I think that's important because, you know, otherwise, you know, again, they may select a technology today and it's one year, two years from now, this technology, they will have to replace everything. Yeah. So it's important that they just rely on the, on the, on the frameworks and build their own platform instead of taking one technology.

Right. Excellent. Well, I think that's a great note to end on.

Yeah. So thank you very much. Thanks for chatting with you.

I told you we're saving the best for last, and I wasn't talking about me, I was talking about him. So, hey, thanks everybody for joining us at Kup Con here in Salt Lake City. It's, it's another fantastic Kup Con and we really enjoy being here.

There's some unique things that you get to do at Kku Con that you don't always get to do at other conferences. And a lot of it is just the, the community interaction, the ecosystem of all, not just the vendors, but the people contributing open source, using open source, uh, startup community, well established companies, the Canonicals of the world and others. So we hope that you've enjoyed our live stream, or if you're watching this recorded, we hope that you're getting a lot of out of the interviews that Alan's been doing, and this last one I got to do with, uh, Cedric.

So we thank you for joining us and we look forward to seeing you at our next livestream. Be sure in text, check out Techstrong TV for all kinds of recorded content as well as livestream. We have Techstrong TV show running every day.

We have, uh, the Textron gang, uh, starting off those shows Monday through Friday. And, uh, I'll be at the Open Text world next week, and Mike Vard will be there as well, doing some, uh, interviews. So, we'll, we'll be joining you that way.

Thanks again. com. I'm sure Alan and the team will be very happy if you did that.

And thanks again for watching and listening to, uh, what we have to say. Take care.