#110 July 1, 2020

Mirantis, with Adrian Ionel

Hosts: Craig Box, Adam Glick

Over the past 20 years, Mirantis has grown from an outsourcing company for semiconductor engineers to a product company that is the new home of Docker Enterprise. Past and present CEO and “co-founder” Adrian Ionel oversaw Mirantis’s adoption of OpenStack and purchase of Docker’s enterprise business, and he joins the show to discuss them both.

Do you have something cool to share? Some questions? Let us know:

Chatter of the week

News of the week

CRAIG BOX: Hi, and welcome to the Kubernetes Podcast from Google. I'm Craig Box.

ADAM GLICK: And I'm Adam Glick.


My weekend was completely thrown upside down as I was listening to a podcast and heard the shocking news that Hello Kitty is not, in fact, a cat. For those of you who have spent any time or had any friends who are big fans of Sanrio, you're probably familiar with the iconic little non-cat, I guess I would say.

But apparently, in 2014, Sanrio had stated that Hello Kitty is not a cat. And at that point, I had the slight amount of existential dread trying to figure out what indeed we truly know in this world if I did not know that.

CRAIG BOX: The clue was in the name.

ADAM GLICK: Did you know Hello Kitty wasn't a cat?

CRAIG BOX: First of all, aren't we coming a little bit late to this news if it was first announced in 2014? How far behind are you on your podcast?

ADAM GLICK: Well, the podcast just came out this week on it, but they'd referenced this, which there is a surprising amount of detail, I have to say, about Hello Kitty out there, including Hello Kitty's full name, Hello Kitty's sister, Hello Kitty's blood type. And these are all official from the company. So I was very impressed at the depth to which they went on the details.

CRAIG BOX: The backstory. You can imagine the showrunners' Bible in the writers' room as to how to be Hello Kitty. In answer to your question, I did know that. It's not something I stored in my memory but it jogged a memory for me because there was a TV show on Netflix a year or two ago, which was called "The Toys That Made Us."

It was a six or eight part documentary which talked about various toys of your youth, assuming you're from the Western world from the '50s to the '80s, or in the case of Japan, the Hello Kitty. Mentioned Sanrio, the company who make them, talked about the rise and fall of this. And it covered all of those things. A good watch if you're interested in nostalgia, which I'm sure we all are in these tough times.

ADAM GLICK: Indeed. Other fun things that I did discover this weekend, for those of you who are gamers like myself, a bunch of the digital board games had gone on sale. And so I tried out a new one called Istanbul.

CRAIG BOX: Not Constantinople?

ADAM GLICK: No, no, it's Istanbul. You know, a traditional euro game style game. But it's a fun one. And as we were finding our way through it this weekend, trying it out for the first time, I was enjoying the different ways they've set up the game in order to give a bunch of different strategies for people to win in a game that doesn't take too terribly long. So that was my fun discovery for the week.

CRAIG BOX: You and I have never played a game together. We should look for something that's play-by-mail, where you can play across very long distances and report back to everyone on how that goes.

ADAM GLICK: Sure. We'll give it a shot.

CRAIG BOX: You might be too competitive for me. I don't know.

ADAM GLICK: Let's get to the news.


CRAIG BOX: Earlier this year, the CNCF's Technical Oversight Committee made changes to the way it adds projects to the sandbox, moving from a GitHub PR and review offline to filling in a form and having it discussed at the meeting. The first meeting has now happened, and a backlog of 11 new projects have been added to the sandbox, taking the total to 31.

The new projects are KUDO, a declarative toolkit for writing Kubernetes operators by D2iQ that we covered in episode 78.

ADAM GLICK: Crossplane, a bounds control plane for managing cloud infrastructure with Kubernetes.

CRAIG BOX: CNI-Genie, a plugin that enables Kubernetes to use different CNI plugins at runtime, contributed by Huawei.

ADAM GLICK: Keptn, with a K, a control plane for continuous delivery, built by Dynatrace.

CRAIG BOX: Cloud Custodian, a rules engine for policy definition and enforcement, originally built at Capital One.

ADAM GLICK: Dex, an OIDC and OAuth provider used to connect Kubernetes to external authentication services that dates back to CoreOS in 2015.

CRAIG BOX: Litmus, a Kubernetes chaos engineering tool from MayaData, who we spoke to in episode 56.

ADAM GLICK: Artifact Hub, a project originally built for the CNCF to run their own such hub but is now downloadable for everyone.

CRAIG BOX: Kuma, a service mesh and microservices control plane by Kong.

ADAM GLICK: Parsec, the platform abstraction for security built by Docker and ARM.

CRAIG BOX: And BFE, an open source version of the Baidu front end layer 7 load balancer. Five other projects remain in the pipeline for the next meeting.

ADAM GLICK: Last week, JFrog held their Swamp Up conference, really leaning into the "amphibian" theme. Chart Center is a new public Helm chart repository built on JFrog's Artifactory. It provides much of the same functionality of Helm Hub or Artifact Hub but with immutable version charts so that you can always install the chart you used last time or see the metadata for each change.

It also includes integrated vulnerability scanning so any known flaws that are in a chart are flagged for you. Chart Center is free to use and available today.

CRAIG BOX: The CNCF has announced the schedule for the virtual KubeCon EU event this August. The four-day event will kick off each day at 1:00 PM Europe time, which does make it a little less of an EU and a little more of a convenient overlap with the US event. Keynotes will be mid-late day, or as it's commonly known, 9:00 AM on the West Coast.

Presenters will be available for live Q&A. And if that gets too much for you, you can always go to the Chair Yoga channel. Sponsors will have demos available, and there are discounts on certification tests.

ADAM GLICK: Solo.io has released version 1.04 of their Gloo API gateway. The new version brings improved scalability, enhanced support for Kubernetes Ingress, support for Istio 1.06, as well as security and user experience enhancements. You can learn more about Solo.io and Gloo by checking out our interview with Idit Levine in episode 55.

CRAIG BOX: Jacob Tomlinson from NVIDIA and its RAPIDS AI project has launched Frigate, an autogeneration tool for documenting Helm charts. This is opposed to launching a frigate, which is generally only done by the Queen.

The tool creates output in HTML, markdown, or reStructuredText depending on where you intend to put the output. The templates used by the tool are extensible, using jinja2.

ADAM GLICK: Bridgecrew have announced their Checkov infrastructure script scanning tool can now scan Kubernetes manifests. The tool scans configuration scripts to look for vulnerable components. And Kubernetes joins support for Terraform and CloudFormation. Checkov is open source and currently available.

CRAIG BOX: Ingress controller Contour is now up to version 1.6. Aside from bug fixes, this version fully deprecates the IngressRoute object introduced in 0.6 in favor of HTTPProxy introduced in the 1.0 beta.

A fun fact-- Contour is the same age as this show, having been announced the week of and covered in episode 1. The change to remove Ingress route was first covered in episode 74. And this will therefore be a breaking change if you've not done anything for 36 weeks.

ADAM GLICK: Microsoft has announced the latest Edge version of Docker desktop for Windows. Allows you to use the Docker desktop tools to deploy a container into an Azure Container instance directly from the Docker command line interface. First announced last month, the tooling is now available to download, and also has an extension available for Microsoft's Visual Studio Code Editor.

CRAIG BOX: In other Microsoft news, gRPC Web is now available .NET. If you heard episode 94 with Richard Belleville, you'll know that gRPC Web is a cutdown version of the protocol, designed for working in the constraints of a browser. This release brings support for the protocol to ASP .NET and older .NET platforms that don't support HTTP/2, and thus cannot use the full gRPC.

ADAM GLICK: HBE has announced Ezmeral, a new umbrella brand for their digital transformation offering that seems to cover just about everything. Ezmeral, which derives from the Spanish word for emerald, includes, quote, "container orchestration and management, AI/ML and data analytics, cost control, IT automation and AI-driven operations, and security." If you're worried that edge and hybrid weren't mentioned, don't worry. Both of these buzzwords were also covered later in the release.

Ezmeral appears to be a similar strategy to what VMware did with their Tensor platform. HBE currently lists two services that are part of the Ezmeral platform which are rebrands of their Container and ML Ops platforms.

CRAIG BOX: Finally, Codefresh has raised a $27 million Series C round to continue to grow their software delivery solution for Kubernetes. The round was led by Red Dot Capital partners, and Codefresh says the funding will go to investments in open source, as well as their CICD tools.

ADAM GLICK: And that's the news.


ADAM GLICK: Adrian Ionel is the CEO, co-founder, and chairman of Mirantis. He joined the company in 2009 and initiated its investment in OpenStack. He earned a master's of science in electrical engineering degree from the University of Erlangen-Nuremberg and started his career in our R&D and product management. Welcome to the show, Adrian.

ADRIAN IONEL: Happy to be here. How are you guys doing today?

CRAIG BOX: Very well, thank you. You're the co-founder of a company that was founded 10 years before you joined it. That sounds like it might be an interesting place to start?

ADRIAN IONEL: Yeah. I'm happy to share the story behind it. It is actually quite weird and interesting. First of all, I should say that I am a co-founder because the other co-founders felt that I should have this title, whatever that might mean. When I first got to Mirantis, it did something very, very different. It was an interesting company made up actually of 100 super smart engineers who are doing something quite different from what we do today.

And it was a good business, a profitable business, serving maybe five or six customers in the Bay Area, mostly companies in the semiconductor software design space. The specific skill set of those engineers was actually very difficult to find, which are people who understand data science and scalable algorithms which you need in chip design to solve all kinds of very difficult computational problems. And at the same time, people who could implement that in very efficient and scalable code.

So maybe simplified, people who understood math and data science extremely well and could also write very, very good code at scale. And companies like Mentor Graphics and Cadence and Synopsis and KLA-Tencor really ran out of engineers in the US with this particular type of skill set. It's far removed from the average software developer. And they felt they needed to tap new markets for that, and Mirantis really had an answer by giving them access to very, very capable engineers in Ukraine and Russia.

So that was the business at the time. And when my partner and co-founder Alex Freedland, who actually did found the company 10 years earlier, ran into me with this business, one of the challenges that he had is that he felt that this is not a business that can scale. It's not a company that can become really big and make a huge impact in the world because there are only so many companies that need this type of skill set. And it was a services business.

So the reason why Alex and their colleagues felt that I earned the title of co-founder of the company is because I uncovered something which I felt could leverage the existing talent and skillset and really turn into something very big. Where I felt the opportunity was was I very much felt that the way internet consumer companies were building and running software is going to move over into the enterprise.

I felt that internet consumer companies were building software and running software in very different ways than Wells Fargo, Bank of America, or Procter and Gamble, and that those ways will inevitably swap over. And I also felt that we had the skill set to capitalize on that.

Let's start, maybe-- we ultimately believe that internet consumer companies, like Google, for example, but many others, were building software much, much faster. They were able to create new features, deliver updates, literally many times a day, not because necessarily they wanted to but because they needed to. Because consumers are very demanding and very fickle, and they expect very rapid innovation.

And at the same time that's one unique threat. How can you build software with new updates every day much, much faster? And the other thing that these internet companies were doing very differently is they build internal infrastructures that were orders of magnitude more cost efficient than the traditional enterprise.

And they did that by using open source technology wherever possible-- so nobody paid for operating systems or for middleware or for databases in these internet companies-- and by creating very, very high levels of automation. Now if you were to contrast traditional enterprise software development with internet software development, you can, I think, arrive at a very clear distinction.

Enterprise software development means very slow release cycles, maybe one or two upgrades a year. Means infrastructure that's very heterogeneous, made up of commercial technologies, whether it's, Oracle or VMware or Red Hat or whatever you want to pick.

Fairly low levels of automation where there is a very big air gap between the developers and the infrastructure. And that air gap is really populated by system administrators and IT operations people, which end up deploying applications on top of infrastructure.

ADAM GLICK: Were those changes that you helped implement as part of your shift towards OpenStack?

ADRIAN IONEL: Yes. That's exactly right. So we believe that we can bring to the enterprise a way of building and running software that was nearly identical to the way that Amazon is doing it or Google is doing it or Facebook or Twitter are doing it by leveraging open source technology and helping enterprises implement an internal cloud infrastructure and software development processes that would yield the exact same outcomes that internet companies have. So that was a shift.

ADAM GLICK: Yeah. If we go through a time machine a little bit, we think about things in 2020, like it is right now, but what was the state of OpenStack and OpenStack's ecosystem at the time when you're making this change?

ADRIAN IONEL: It was extremely early. Pretty much no one had OpenStack in any form of commercial use. And in fact, we built our very first OpenStack cloud in a Cisco data center, which Prem Jain, one of the senior executives at Cisco-- was one of the founders of UCS-- had loaned us to be able to do. But pretty much nobody had OpenStack in any form of commercial or production deployment back in 2010. It's very, very early days.

CRAIG BOX: For the benefit of some of our younger listeners, it might be worth explaining a little of the history of OpenStack. I remember hearing that it came out of some work from NASA originally?

ADRIAN IONEL: That's exactly right. NASA was looking to build an internal cloud infrastructure for very similar reasons we talked about before. They wanted to have much, much faster software development cycles, and they wanted to operate at very, very large scale. So they created a scheduler to build an internal cloud infrastructure that's fully automated.

And Rackspace had developed some storage capabilities, similar to S3 AWS. And between the two-- NASA and Rackspace-- they decided to create an open source project and donate this into a community that could then pick it up and run with it.

ADAM GLICK: What attracted you to the open source side of these conversations? Some similar technology existed that was proprietary at the time. If I think about it, it still exists today. But you made the investment towards the open side when it was still very nascent.

ADRIAN IONEL: Well, we felt that the open source approach has a dramatically better opportunity for innovation and scale. And also-- this connects now to the core DNA of the company-- our engineers, the 100 engineers that we had at the time, were extremely proficient with open source infrastructure technology.

In fact, they were using and consuming the same building blocks that make up OpenStack today day in, day out, to build stack for the semiconductor software companies. So we felt open source was absolutely the right way to go.

CRAIG BOX: So you've been helping these companies adopt software that was built somewhere else. And your own engineering team started then enhancing that software. What was some of the work that Mirantis did to improve and build on top of OpenStack?

ADRIAN IONEL: We focused on a few areas which we considered absolutely critical for customer adoption. We made OpenStack dramatically easier to deploy and configure and integrate into an enterprise setting. So we pioneered a piece of technology called Fuel, that's actually still widely used today. Many, many companies all around the world.

That turns a very complicated system integration project into pretty much instant deployment where you can configure your cloud visually on a GUI, and you can automagically, very, very quickly, get to a production grade OpenStack cloud. So that's one area where we focused.

Second area where we focused is making OpenStack more scalable and resilient and highly available. So we ended up creating the most robust high availability architecture for OpenStack. And then, in a fairly short period of time, we ended up scaling to more than 100 open source contributors who ended up contributing major, major portions of the entire OpenStack code base, mostly around the core.

ADAM GLICK: After about five years at Mirantis, you chose to depart in 2015. Why did you depart, and what did you start at that point?

ADRIAN IONEL: I did not choose to depart. The board fired me. Or actually, to be more complete, the board felt that I should become the chairman and maybe go and build other parts of our business because we were already, at the time, very geographically distributed. We had presence all over the world, including China.

So they floated an idea of me going to other parts of the world and building up the business there. And at the same time, they felt that maybe it'd be much, much better for the company to hire an outside professional CEO to scale the company to the next level. We had grown by that time to about 500 or so people in a fairly short period of time. And our business had grown probably 50x or so, our cloud business, from the time we started it.

So that was actually a pretty big fight in the board. It was a pretty big argument that narrowly fell what I believe on the wrong side or wrong decision. But at the same time, I had other ideas, and I had a good number of investors who very much believed into what I could do. And I had also my own capital available, so I felt that that would be fine. And there are some other things that I could go do.

CRAIG BOX: Did anyone call John Sculley and see if he was available to be your new CEO?

ADRIAN IONEL: [LAUGHS] I don't know. But I do want to say, before getting to what I did afterwards, is that maybe in a way that's quite unique and different for how these things usually go down, this was super friendly all around. And what that means is there was no bad blood of any kind between any of the parties, with the other co-founders, or the investors, with the board members. It was fine.

So I got it, why they felt they wanted to go that direction. And at the same time, I was pretty happy to do other things still.

CRAIG BOX: So what were those other things?

ADRIAN IONEL: The other thing that I was very excited about doing is finding a way to democratize and surface the true capabilities of technical talent in a way that this technical talent can monetize it. And I know this sounds pretty esoteric, but I ended up finding a company called Dorsal, which I invested in, and then a number of other people also invested, like Two Sigma, like Foundation Capital, like Jerry Murdock who is the co-founder of Insight Venture Partners, the largest outside investor in Mirantis.

And what Dorsal does-- it's very much focused on open source but on data science and databases-- is it has a very, very unique approach to providing open source support. One of the challenges with adoption of open source software is that enterprises actually need support and indemnification. Many, many companies would use the completely free version of a piece of software if they were able to get support.

And today, the only way you can get support is by buying the subscription. But subscription is kind of like an old business model where you have to pay in advance. And maybe you have just three or four support cases a year but they are very important once. But you have to buy a whole subscription in advance.

So the entire concept behind Dorsal is that instead of you having to buy a support subscription for your database or for Elasticsearch, for example, you could just enter your support case on a website, and we would parse that support case, understand what kind of expertise would be required to solve it.

And then we would very, very quickly identify the best people in the world who can solve that support case and then connect them with the customer. And we would process the entire support case from top to bottom.

ADAM GLICK: You didn't stay away long, though. You were honest about not having a lot of bad blood there, and you rejoined as CEO of Mirantis in 2018. How did that process happen?

ADRIAN IONEL: Well, it happened because of, I would say, very much the good relationships I'd left behind. Some time in 2017, the original co-founder of the company, Alex Freedland, came over to my home and we sat down in my kitchen and living room for close to a day, talking about the challenges that the company was experiencing. And he invited me to come back in and maybe help the company with some of these challenges-- formulate its vision and its strategy forward, maybe look at some of the internal challenges, and provide some suggestions.

So it turned into kind of an advisory role for a number of months. At the same time, the board felt that they needed to make a CEO change. And they had a search open. And then they asked me if I wanted to be considered. So I was one of the candidates who was considered, among others. And so, then they offered me the job, and I rejoined.

CRAIG BOX: While you were away, Kubernetes has happened. What did Mirantis do in those three years?

ADRIAN IONEL: Well, actually, Mirantis did do some wonderful things that we leverage even to this day. So Mirantis took, I would say, the next technological step in delivering an infrastructure as a code experience for our enterprise customers. So where the entire infrastructure could be modeled as code and changing the infrastructure would be as simple as changing a line of code. And that's an enormous step that I want to give credit for to Alex and our engineering team for making happen. Huge step forward.

Also, we continue to gain very, very large number of customers and very large customers around the world, whether it's somebody like Reliance Jio, Pioneer in India, or AT&T on a much larger scale. So a lot of fantastic things happened for the company during those two years where I was away.

ADAM GLICK: People often talk about Kubernetes and OpenStack and kind of how they sit together and how they're different. You have the unique experience of being part of an organization that has been involved with both. What did your experience with OpenStack teach you that you were able to put into place as your company moved to its work on Kubernetes?

ADRIAN IONEL: That simplicity and the robustness are extremely important for enterprises. What we learned from OpenStack is that the vast majority of enterprises are not interested and not prepared for dealing with a complicated piece of infrastructure technology. They want something which just works and always works at scale for mission critical applications.

And what we also learned is that the developers in the enterprise don't want complexity. They want simplicity. They want ease of use. So that's extremely important. So a big part of the value that we add is turning a very complex and powerful piece of technology into something that's extremely easy to deploy and operate and use. And that's the key lesson learned.

ADAM GLICK: Some people have speculated that Kubernetes is on the same path as OpenStack, and some people have even said it could face the same fate. That doesn't seem to have happened, though, if we take a look at the trajectory of the two projects. What do you think was different about the Kubernetes journey versus the OpenStack journey?

ADRIAN IONEL: Oh, yeah. There are a few very specific differences. One is Kubernetes was much more, I think, designed with a developer in mind and is designed by developers. And I think that's a huge difference. The second very important difference that Kubernetes community I think has done much, much better than the OpenStack community is becoming very, very close and very interested in the end user, whereas the OpenStack community at some point developed this more self-centered view and focus more on enterprise IT decision makers and the needs of the vendors.

I feel that the Kubernetes community was never about the needs of the vendors or the enterprise executive leadership team. It's all very much about the end users. And then, finally, the last point that I think that the Kubernetes community is doing much better is much tighter control and purity of vision for what Kubernetes is aiming to achieve. So not have something that sprawls out and trying to solve so many problems for so many people so it ends up being good at actually nothing, right?

So these are some, I think, very, very big differences that helped the Kubernetes community. There is another huge difference, which is Kubernetes has been adopted as a standard in public clouds. And that's incredibly compelling to end users because now I can have the same public cloud experience that I can have on-prem and anywhere in between. So it is a standard that I can rely on. And that's incredibly important.

CRAIG BOX: Speaking of companies who are all about the developers, that brings us to Docker. Docker have a tools and desktop and hub business, and they also had an enterprise business. And in November, Mirantis purchased that enterprise business from Docker. Who called who?

ADRIAN IONEL: Yes. A very good friend of mine who invested $10 million of his own money in Docker called me.

CRAIG BOX: So was this that Docker were looking to move out of that business and looking for a good home for that part of it?


CRAIG BOX: And how does something like that go down?

ADRIAN IONEL: With a lot of drama.

CRAIG BOX: OK. Any of that that you're able to share with us?

ADRIAN IONEL: Well, as you can imagine, a deal like this will be also fraught with emotion, right? Docker has changed the world, without any question. They have pioneered the usage of containers for developers at massive scale. They literally changed the world. And the company definitely had an expectation of enormous commercial success. They thought that they would be the next VMware. And I think that their strategy was very much to become the VMware for containers.

And for a variety of reasons, that didn't work out exactly like that. Now, the reason this was very, very compelling for us is as follows. One is that their longer term vision in terms of the role that Kubernetes displays and containers will play in modern software development and our vision was identical. So we were very much marching towards the same path, towards the same vision.

We also found out very quickly that our product developers and our engineers and the Docker engineers spoke the exact same language and had similar principles for developing excellent products. So there was a very good chemistry there right away. So that was very, very helpful. The challenge was working on a deal that would be compelling for all shareholders-- for the Docker shareholders and for the Mirantis shareholders-- which took a bit of time.

But I think, as far as these things go, from mid-September when we had the first serious conversation until literally the evening of November 12 we were closed, it was quite a bit of work to do in between.

ADAM GLICK: At the time, Mirantis already had a Kubernetes as a service offering. When you brought in the Docker Enterprise technology, how did you integrate the two teams, and how did you view kind of what that would become for your customers?

ADRIAN IONEL: That was part of the premise for doing the deal in the first place-- how we would unify it. And we build the plans to do it before the acquisition. And what we decided to do is that the Docker container technology-- fundamental container technology, Docker Enterprise Engine, Docker Enterprise Cluster-- will become the foundation of the roadmap forward, but it would be re-engineered to become more Kubernetes-native, which is the expertise that we had and we have, and we would bring our as a service capabilities into play.

So the fundamental piece of technology that Mirantis adds to the table is continuous updates and the ability to spin up and manage a very large fleet of Kubernetes clusters, regardless of where they reside-- on public or private clouds. And these are technologies that we added to the Docker Enterprise platform.

What we double down on also within the Docker Enterprise is all the security aspects. Docker Enterprise has absolutely fantastic security built in throughout the stack, which is why we have so many customers in the federal government, in National Labs, in the defense community, not just in the US, but also throughout the world in banks and insurance companies and so forth.

So we took the best capabilities of Docker Enterprise. We took that as a foundation. We re-engineered it to become even more Kubernetes-native while still retaining the Swarm capabilities. And we added the as a service experience with continuous updates, zero touch operations, and very large scale fleet management for Kubernetes to the picture. So this is how we unified it.

But day one, we had one product development team and one product roadmap. And on December 19, we rolled out a product roadmap to the customer base.

CRAIG BOX: As you mentioned there, there are two different engines that are part of the Docker Enterprise product. There's Docker's own Swarm, and then there is, of course, Kubernetes. Reading between the lines and some of the things that have been said, it might seem you were looking to bring Docker Swarm to an end of life but you got a lot of feedback from its users and its community that they still want that technology. How do those two communities relate to each other?

ADRIAN IONEL: Well, they mostly relate to each other in our customer base. We see Docker Swarm. And the more we spoke with our customers, we see that our customers love the simplicity of Swarm. And this goes back to also a question that you asked earlier, is it possible that Kubernetes will take the same path as OpenStack? We don't believe so, but we do believe that Kubernetes is still quite complex for many developers.

And for a number of use cases it's overkill. It's not necessary. And that's where Swarm does a fantastic job. If we just have a half dozen containers in a very simple multi-tier app and you want to deploy them on the platform and run them efficiently, Kubernetes is going to be overkill for you.

ADAM GLICK: You have picked up the parts of Docker Enterprise. There's also the Docker company that's focusing more on developer tooling as it currently exists. How did the two of you work in terms of the name Docker, the brand Docker, in almost every developer's mind?

ADRIAN IONEL: We use the Docker Enterprise brand name. And it's always Docker Enterprise. It's never Docker by itself. And we use it strictly for that purpose. Whereas Docker Inc. uses the Docker brand for developers and for any other purpose they actually want to. So this is how we work together to keep the brand distinction out there.

I also want to point out that in terms of our product strategy, we seek to be a very, very developer-centric platform. And we have excellent integrations with the Docker Engine, with Docker Compose. All of that works beautifully in a native way, as you might imagine.

CRAIG BOX: Another company that was building Kubernetes services is Kontena, a Finnish company. And they closed down last year. And then you acquired their team in February-- not the technology but the team. Tell us a little bit about that acquisition.

ADRIAN IONEL: We acquired the team because we feel they've done some spectacular work in the Kubernetes space to dramatically simplify the developer experience for people adopting Kubernetes. And while they're a small team, we felt that the quality of their work was incredible and that they had a tremendous passion for creating beautiful, compelling end user experience.

And we as a company, we believe that infrastructure needs to become a lot more developer-centric than it already is today. So we're marching in that direction. And we felt that a container team would add tremendous value to that. And we have a highly distributed engineering team. We're literally all over the world, from China to Poland, Czech Republic, Berlin, UK, several sites in the west.

So it wasn't difficult for us to pick up a team in Finland, and that was fantastic. And since, we also acquired what we feel is the central piece of unique technology that the Kontena team contributed, Lens. And we are integrating it in the product family.

CRAIG BOX: Another area of overlap with both OpenStack and Kubernetes is the 5G space. A lot of companies who are now looking at the services they need to operate next to their 5G antennas, for example. Mirantis joined an OpenStack project called Airship last year. What can you tell us about that collaboration?

ADRIAN IONEL: We joined the Airship because we feel that telecommunications infrastructure is undergoing a massive revolution which Kubernetes container technology and even OpenStack can play a huge role in. And the massive revolution is that the future of communications is all about software running on standardized, highly efficient infrastructure that's powered by open source.

So what Airship does, it uses Kubernetes, which we have a lot of expertise in, as the foundational layer to run OpenStack, as well as native 5G services on top. So everything is containerized, everything is standardized on top of Kubernetes as the foundational layer for the 5G infrastructure.

So we felt that was just a natural thing for us to do. And AT&T has been a longtime partner and customer for us, along with Ericsson. Now we have the unique and compelling value to add there. So this is why we did it.

CRAIG BOX: You mentioned before the roadmap that you set out to for the new releases of Docker Enterprise. You've recently made the first release of Docker Enterprise as Mirantis. How much of what you released in this new version comes from Mirantis rather than what Docker was working on pre-acquisition?

ADRIAN IONEL: First of all, it has been a combined engineering team. So the Docker, or our containment development team has a very, very large portion of engineers from the Docker folks. Probably the majority are Docker engineers. But the work is all new. This is all work that has been done in the last six months. So none of this work, or very, very few pieces of this work-- probably none of the code-- existed prior to the acquisition.

So it's all built fresh and new. But it's built, to a large degree, also by the longtime Docker engineering team, combined with the prior Mirantis software development team. And actually, even the Kontena folks contributed enormously to this release.

ADAM GLICK: One of the things that I've been very impressed with about Mirantis is the focus that you've put on culture and the recognition that the company has received about being a great place to work. You've put focus on that personally. How did you create the culture, and what do you think defines a great culture?

ADRIAN IONEL: Yeah. I am very proud of it, although we can always do better. And I can tell you what role culture played in our growth so far, the evolution of the company so far. So what I'm aiming for is to create a transparent meritocracy which is focused on constant learning. That's what I'm aiming for.

So we have a very high level of transparency within the company. We share financial results. We share the bad and ugly news along with the good news. In fact, we had an all hands call just this morning, breakfast Q&A, where some very, very tough questions were asked. And we're very much focused on constant learning and progress.

And the idea here is that what excellent companies do is they adapt to change. A successful company, a successful business is all about recognizing, embracing, and capitalizing on change. So what I'd like us to do is to be extremely good at discovery, at learning, and at embracing and adapting to change. So I want us to be a vibrant, dynamic environment.

ADAM GLICK: When you think about change in culture, the Kubernetes community has certainly gone through this as it's grown from a small project to a very large project with many people involved in the community. Organizations and cultures often change as they grow. You talked about how Mirantis has grown since you've been there. How do you avoid cultural shift that happens as it gets larger? How do you scale culture?

ADRIAN IONEL: Yeah. I suspect what you're referring to is maybe cultural drift. It's very difficult. It's very, very difficult. I would say first and foremost by role modeling the right behaviors in a way that's very consistent. It's by being very clear and intentional as to what it means to have the culture we want. What behaviors do we support and encourage, and what behaviors are not acceptable?

It's very much about selecting, developing, evolving the right leadership team. And it's about making culture a constant topic of attention for everybody in the company. But it's very, very hard. And you're never going to get a 100% cultural integrity or adherence. There's always going to be pockets or differences. So it's always kind of a work in progress.

CRAIG BOX: Finally, at OpenStack conferences and other events that Mirantis were at in the past, there was quite famously someone in a bear suit up on stage with Boris, one of your other co-founders.


CRAIG BOX: Was that ever you inside the suit?

ADRIAN IONEL: No, I was never in the bear suit, though I can totally relate to the bear.

CRAIG BOX: Why is that?

ADRIAN IONEL: Why can I relate to the bear?


ADRIAN IONEL: Well, we wanted to be a company that's quite vigorous and that can be also quite determined in pursuing what we believe in. And we feel that the bear very much reflects that.

ADAM GLICK: Adrian, it's been wonderful having you on the show. Thanks for joining us.

ADRIAN IONEL: Yeah, my pleasure. All the best to you guys.

ADAM GLICK: You can find Adrian on Twitter at @adrianionel, and you can find Mirantis on the web at mirantis.com.


CRAIG BOX: Thanks for listening. As always, if you've enjoyed the show, please help us spread the word and tell a friend. If you have any feedback for us, you can find us on Twitter at @kubernetespod, or you can send it to us by email to kubernetespodcast@google.com.

ADAM GLICK: You can also check out our website at kubernetespodcast.com, where you'll find transcripts and show notes, as well as links to subscribe. Until next time, take care.

CRAIG BOX: See you next week.