Kubernetes and Cloud-Native Operations

Alan Shimel and David Booth, vice president, cloud-native applications at Canonical, discuss the findings of Canonical’s recent “Kubernetes and Cloud-Native Operations Report.” The video is below, followed by a transcript of the conversation. You can download the report here.

 

 

Alan Shimel: Hey, everyone. Welcome to another segment here on TechStrong TV. My guest for this segment is David Booth. David is with the folks at Canonical, makers of Ubuntu and much more. Hey, David, welcome to TechStrong TV.

David Booth: Happy to be here.

Shimel: We’re so happy to have you. Real good to have you here. David, as we spoke off camera, our audience knows Ubuntu, they know Canonical, one of the leaders in the Linux opensource market, cloud native. A lot going on there. You know, distributions and so forth. But they may not know David Booth, believe it or not.

Booth: Sure.

Shimel: Well, some of us may.

Booth: There may be one or two out there who do, but for the vast majority –

Shimel: Relatives, relatives of yours.

[Laughter]

But beyond that, why don’t you share with our audience a little bit?

Booth: Sure. I’ve been in the tech space for going on about 20 years or so now. I spent a long time in Prague with a company called JetBrains that I think a few people may know about. After that I joined ZeroTurnaround to become the CEO there, where we built a tool called JRebel for the Java community. That went over really well, raised a bunch of venture capital, our company grew. The company was sold.

It went in a good direction. And then I built another startup, sold that one off, and kind of started thinking what’s really next for me? What do I really care about? What do I want to learn about and who do I think I could learn the most from? And I was in conversations with Mark Shuttleworth over at Canonical and he started talking about something that they’ve been working on for a while, their vision for the future, and I just couldn’t say no. So I joined Canonical in January as the VP of cloud native applications.

Shimel: Very cool. Very, very good, and they’re lucky to have you, man. Good stuff. So before we jump into the news that we want to talk about, let’s talk – you know, as I said, most people know Canonical, makers of Ubuntu. It’s a Linux distribution. When we talk about Canonical cloud applications, is it cloud applications that Canonical is developing and distributing or selling or is it more allowing Canonical to be used as the backend, as the OS for ISV’s cloud apps?

Booth: So it’s a really place to be where this is this operating system layer that everybody knows about and that helps to power a lot of the clouds that are out there. A lot of people’s applications and infrastructure are based off of it, whether on prem or on a public cloud. Canonical is looking at tackling other problems in the cloud-native space, so we see Kubernetes where we have Charmed Kubernetes and MicroK8s, a few different distributions that are helping people to solve different problems that they have with Kubernetes, particularly around getting started. And from there, we’re starting to look at moving beyond initial deployments of Kubernetes and cloud applications into the day-2 operations, the management of those applications.

One of the things that we actually saw in this survey was that the Kubernetes and cloud native space is still somewhat in its infancy. It’s really early. You’ll see a lot of venture capital dollars getting invested into the different companies in the space, but adoption is, although growing quickly, still at a stage where people are trying to figure out if we can move to Kubernetes, how are we going to do the deployments of our applications? Are we going to have to restructure our entire application from microservices to put it into containers to run it in Kubernetes or how are we going to do this?

Canonical is already starting to think about let’s assume that a whole bunch of people do. What happens next? What other problems can we solve? So those are some of the things that we’re thinking about in the company, and I’m happy to go into those a little bit later.

Shimel: Absolutely. It’s great stuff, David. So I wanted to talk to you today though – I think you mentioned it in passing – about a new report based on some survey data and analysis that the folks at Canonical had done. Why don’t you share with our audience, kind of give us some background, some highlights, and we can jump in.

Booth: Sure. So this is the Kubernetes and cloud native operations report, and what we did here was we looked at a lot of the other reports in this space, and I think like you mentioned before, there’s a lot of vendors out there who want to have reports and give people some different data to look at. I’ve run reports and looked at these for decades now. I often see maybe a bit of vendor bias in these types of things. They may have a point solution and they’re trying to just get that message across and trying to get people to go there, whether it’s a security tool and they’re trying to say, “Look, the cybersecurity space is terrifying, everybody needs to go and buy our tool.”

I thought, well maybe we should approach it from just a pure data point of view. Where is everybody? So we asked a whole group of people, some who were in some way using Ubuntu and others who were just at KubeCon about 50 questions, and we got them to go through about a half hour worth of questions to provide us with their insight and share their experiences. And then from there I thought rather than make this about Canonical and our opinions let’s invite in some industry experts and have them share their opinions. So we talked to folks like Kelsey Hightower, who a lot of people know, James Strachan, who’s over at CloudBees working Jenkins project.

Shimel: Sure. Know them both.

Booth: With Tim Hockin as well who’s worked on Kubernetes.

Shimel: Google.

Booth: Michael Hausenblas over at AWS. And a whole bunch of other people who all – Ken Sipe, for example, who is working on – he’s on the board for the Operator Framework, the CNCF, a former Java champion, works with Edward Jones right now, I believe, and they – we thought that it would be more interesting to share their insight into the data than just our opinion. So we literally try to approach the report from a vendor neutral standpoint.

Shimel: That’s fresh. That’s a fresh look at things. Because, David, I think you’re right. Look, it used to be – and guilty – vendors would come to MediaOps here or the Accelerated Strategies Group or our analyst firm and sort of get a commissioned report based on survey data and stuff, and too often, even though as much as we would try to not do it, there’s lies, damn lies, and metrics, and every question could be framed a certain way to give it the answer you desire. And I call it when you’re a hammer everything looks like a nail.

If you want everything to look like a nail because you’re a hammer, you’re going to make your report kind of show that. And so there’s two elements to it. One is the questions that we ask and how they’re asked, and then two is who does the analysis? So when you have a Kelsey or a James or Tim from Google, you have these guys giving you their analysis based on what they’re seeing from data and what they’re seeing in their lives and their career, their business world, it brings a certain amount of validity, of impartiality, of real –

Booth: As a group.

Shimel: Right.

Booth: I mean, individually, sure, everyone may still have their biases and have their points of view.

Shimel: Sure.

Booth: But as a group it’s nice to see that they don’t always agree.

Shimel: No.

Booth: So that’s a little bit nice to see.

Shimel: I agree.

Booth: And so you’ll see some of that in the report as well.

Shimel: Excellent. So let’s talk a little bit about what the report’s findings are.

Booth: Yeah. So hybrid cloud and multi cloud. We first wanted to take a look at use cases there and see what’s really happening here. Is there a broad adoption? How are people using multi cloud or hybrid cloud? What are their use cases? And it turns out, yeah, there’s a pretty broad adoption.

We were seeing about 78 percent of respondents are saying that they use at least one hybrid cloud or multi cloud use case in production in their organization. Frankly, we think 78 percent is a bit low, because if you were going to take, say, their SaaS or their third-party managed services that they may have, those are going to push those numbers up. But under their own management 78 percent of people are saying that they’ve got at least one use case, and those are often spread across a mix of – like 30 percent of people are saying that they use both bare metal VMs and Kubernetes.

Shimel: Really?

Booth: So there’s only 15 percent of people said that everything they do in production is in Kubernetes. Most people have a mix.

Shimel: Wait, wait. Run that by me again. 15 percent of people said that…?

Booth: Everything’s on Kubernetes.

Shimel: Oh, so they’re 100 percent Kubernetes.

Booth: That’s right. So 15 percent said that. 30 percent said it’s a mix of bare metal VMs and Kubernetes. Another, say, 28 percent said bare metal and VMs, but they’re either evaluating Kubernetes for dev or planning migrations. So this kind of starts to lead us down a path where we say this is still a nascent industry. It’s still really early –

Shimel: It sure is.

Booth: – of figuring things out.

Shimel: I was a participant on a session with Jonah Kowall from Logz.io. He’s a former Gartner guy. It was funny. One of the questions from the audience was do you think Kubernetes will stick around? How long do you think Kubernetes will be dominant for or will we see change? Will something else replace Kubernetes?

Dave, you and I have been around this business a long time. Of course something else will replace Kubernetes. Something else will replace you and I, right?

Booth: Rate of technology.

Shimel: Right. The question is when. I’ve learned to operate in technology where I might have – it gets fuzzy out there, but three years is probably as far as I can see. Right? You might see mega trends 5, 10 years and say they’re going to happen, but that’s throwing darts. For the next three years I do see Kubernetes becoming and remaining dominant. I also don’t think – and I think these numbers bear it out – it hasn’t crested yet.

Booth: Yeah.

Shimel: We haven’t hit the high point of Kubernetes adoption and penetration. Just like cloud, right? For the longest time – you know, 2005 I remember was a big year for cloud. It was on everybody’s lips. Everybody was thinking. But yet as recently as I bet 2015, 2012, you know, 80 or 90 percent of Global 2000 had some infrastructure or some assets in the cloud, but less than 10 percent of their mission critical stuff was in the cloud.

I think that’s changed. COVID has been an accelerator, digital transformation in general and all of that. But I think the same kind of thing is Kubernetes. 15 percent was the number you gave me, which is 100 percent Kubernetes.

Booth: Right.

Shimel: That could double or triple.

Booth: The people who we’re speaking with were saying – we’re asking them what – now that they’ve got this infrastructure set up – hybrid, multi cloud, some Kubernetes in there and some not – what are the main things that they’re doing on hybrid clouds and on multi clouds? And the use cases are really they want to use it to accelerate development, to increase automation, to expand their cloud backup options, disaster recovery, clustering and mission-critical databases. All of these things are stuff that you really need that second cloud for.

Either you’re a developer and you’re like, I need to do this faster and so you just want the extra power, or you need a backup and you can’t do that on the cloud that you’re currently on or the space that you’re currently on. There was a really small number, 5 percent or something like that we’re talking about, bursting into cloud, into a normal cloud use case. So whenever they have tremendous load then they would burst into a multi cloud. And this is one of those signs of the earliness of that industry still.

So they’re not taking advantage of the infrastructure that’s available to them because that’s a more advanced technique and they’re not quite there yet. So we’re still looking at the multi cloud, hybrid cloud space as an early industry. So between Kubernetes being early, the use cases being early, most people are focused it seems on can I deploy this? Can I deploy this now? Can I start using this technology and can I do this stuff?

A few, however, like that 15 percent who are all in on Kubernetes, are starting to say, “Okay, well now I’m doing it all the time so now I’ve got to manage my applications, I’ve got to upgrade my applications, I’ve got to destroy them, I’ve got to upgrade Kubernetes itself, I’ve got to go through all of these day-2 operations and these processes there. What do we have? We’re seeing people who are taking – kind of ops folks who are taking their previous mentality from the VM space and saying, “Okay, well maybe it’s Ansible, maybe it’s Terraform, maybe it’s something else that’s in there.”

Some people are starting to use Helm, although it looks like adoption’s really low, like really small adoption. Maybe that’s for testing or something. But we certainly see 30 percent of people are starting to look at operators now as a tool to extend Kubernetes and manage their day-2 operations. So this is early stage. The operator kind of space is not yet well formed. Canonical has some stuff in there, Red Hat has some stuff in there, but it’s still really early stage as well.

But the idea is that at some point software is going to be managing our software, we’re not going to need to do all the configuration management that everyone does right now where they need to know every line of every YAML file, and the software will know that. So operator seems to be a path towards it, and we’d like to think that that’s a path towards _____ driven operations, but we didn’t really ask about that in the report because we wanted to be vendor neutral.

Shimel: I hear you. Good stuff. You know what? Dave, we’re almost out of time, but where could – the report’s downloadable, I assume?

Booth: Yeah, that’s right. It’s on where Juju is. It’s in the learn section of the site under resources. I can send a link so that we can –

Shimel: Sure. Send us a link. We’ll try to put it in notes. For those who aren’t watching it on TechStrong TV but maybe off of our Digital Anarchist as a standalone, we’ll try to put that in the notes in there once you send it along, David.

Booth: Perfect.

Shimel: Good stuff. Hey man, this is great. You know what? We didn’t get into some other things I wanted to talk to, but that’s a good reason for you to come back and we’ll talk some more.

Booth: Happy to do it.

Shimel: Alright. David Booth –

Booth: Thanks a lot for having me.

Shimel: – here on TechStrong TV. Thank you.

Booth: Bye-bye.

Shimel: We’re going to take a break. We’ll be right back with another guest.

[End of Audio]

Alan Shimel

As Editor-in-chief of DevOps.com and Container Journal, Alan Shimel is attuned to the world of technology. Alan has founded and helped several technology ventures, including StillSecure, where he guided the company in bringing innovative and effective networking and security solutions to the marketplace. Shimel is an often-cited personality in the security and technology community and is a sought-after speaker at industry and government conferences and events. In addition to his writing on DevOps.com and Network World, his commentary about the state of technology is followed closely by many industry insiders via his blog and podcast, "Ashimmy, After All These Years" (www.ashimmy.com). Alan has helped build several successful technology companies by combining a strong business background with a deep knowledge of technology. His legal background, long experience in the field, and New York street smarts combine to form a unique personality.

Alan Shimel has 54 posts and counting. See all posts by Alan Shimel