#184 July 7, 2022

Mercedes-Benz Tech Innovation, with Sabine Wolz

Hosts: Craig Box

Why does a car manufacturer own an IT company? How did that IT company end up running 900 Kubernetes clusters, starting at version 0.9? Craig asks these questions and more of Sabine Wolz, Product Manager at Mercedes-Benz Tech Innovation.

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 your host, Craig Box.


CRAIG BOX: Depending on how you want to define the word, you could consider me a part-time journalist. I do produce a weekly news bulletin, and I want to thank you again for choosing to tune in to listen to it.

One of the good things about the space I cover is that it rarely changes in real time. As I speak these words, I'm somewhat distracted by the chaos in Her Majesty's Government. I considered recording two versions of this so I could correctly substitute in at time of release, whether the British Prime Minister has actually resigned or not.

When the dam first started to break here on Tuesday night, I was having dinner with two old colleagues, now great friends. It was interesting to hear how hard they find it to recruit good technical talent, especially in the Kubernetes space. I must say, knowing many of the people whose names are on the commits, and talking to talented people for this show every week, I just assumed that everyone was now an experienced Kubernetes person.

By listening to the show, I think you signal you're at least aware of the space. And so perhaps you can now use it as a shibboleth when you're interviewing. Yes, I know about Kubernetes. Did you hear that great interview with Justin Santa Barbara about configuration as data?

As a New Zealander traveling around the world, it's always been interesting to hear the international touchpoints. For a while it was Lord of the Rings, then Flight of the Conchords. And, recently, it's been Jacinda Ardern. It's nice to have two Prime Ministers to choose between. You can pick one for the occasion.

With a new TV series coming out, it might be Lord of the Rings again for a while. Only for the first series, though. If they make a second series, it has been announced that the production is moving to the UK. I do hope that it works out for them. Let's get to the news.


CRAIG BOX: A new GKE Cost Allocation feature has launched. As many customers advance and grow the use of GKE, being able to operate multi-tenant clusters while still doing accurate chargeback and showback for costs associated with utilization is critically important. This new feature lets users easily and natively view and manage the cost of a GKE cluster by name, namespace, or pod labels. The feature is available in public preview.

The CNCF has welcomed CubeFS, with a C, as an incubating project. CubeFS is a distributed file system designed to natively support large-scale container platforms. It separates computing storage with elastic and scalable metadata services and can be accessed with S3, HDFS, or POSIX interfaces. The project initially entered the CNCF in December 2019 under the name ChubaoFS. The current lead maintainer is the consumer electronics company Oppo, who also joined the CNCF this week as a Gold member.

Amazon has announced bare metal deployments for EKS Anywhere, their Kubernetes runtime for outside their cloud. Using open-source Tinkerbell by Equinix Metal, which we covered in episode 142, you can deploy onto servers that aren't running VMware. Management is either through a connector back to the EKS control plane or in a disconnected mode.

We are not paying Amazon anything, but you're on your own for tooling. If you're interested, check out Cubernetes, also with a C, where, in episode 20, guest Justin Garrison shows how to build a home lab to test EKS Anywhere for no less than $6,000.

New features and some of the other commercial distributions this week: Red Hat introduced version 2.2 of OpenShift Service Mesh, bringing the latest Istio and Kiali versions, and Tanzu Mission Control added continuous delivery, bringing the Flux CD project to allow controlling clusters through GitOps.

The Pixie project, contributed to the CNCF by New Relic, has built a new plugin system. The initial version addresses pixie's data storage limitations, wherein data was only available for the past 24 hours. Developers can now export their Pixie data in OpenTelemetry format for long-term retention. This allows you to look at trends over time and also to join Pixie data with other relevant insights.

Good blog of the week: a post by Google engineer Taahir Ahmed, who explains the concept of Kubernetes's bound service account tokens. The new OIDC-based authentication tokens are the default as of Kubernetes 1.21. As most of the change to adopt these is abstracted away by client libraries, there's very little chance you will need to change anything as a result of the new behavior. But it's an excellent piece of technical writing, which will explain the concept of client authorization and authentication to you.

Controversial blog of the week is a post from Jeremy Brown, self-proclaimed former Kubernetes evangelist, who now claims the software is a red flag, signaling premature optimization. Hacker News was split down the middle on this one, with some people singing the virtues of Kubernetes simply as a deployment pattern, and others calling it resume padding.

Finally, if you just can't get enough eBPF, the third annual eBPF summit has been announced. The event will be held on the 28th and 29th of September, and a call for papers is open until July 22. Bee there or bee square (that's bee with two Es).

And that's the news.


CRAIG BOX: Sabine Wolz is a product manager at Mercedes-Benz Tech Innovation. Welcome to the show, Sabine.

SABINE WOLZ: Hi. Thank you very much for having me.

CRAIG BOX: By some accounts, Mercedes-Benz is the oldest car company in the world. You work for a subsidiary of the group called Mercedes-Benz Tech Innovation. Car companies are famous for their use of a large supply chain. Why does a car company own an IT company?

SABINE WOLZ: IT is very important nowadays for cars, and especially for car companies. But we are not around for a very short time. We are around for more than 25 years. And the importance of IT has grown over that time.

And I would say it has always been strategic for Mercedes-Benz to have IT close to them, as close as possible, because, formerly, it was important to support them and everything they did with IT. And, nowadays, it's very important to have us as close as possible because IT is such an important part of cars.

CRAIG BOX: You mentioned over 25 years of history there. What is the history of MBTI?

SABINE WOLZ: We were founded in 1998. Our name there was Technical Sales Support. And I think that was exactly the aim, why we were formed, because Mercedes-Benz — or back then, Daimler also — they needed technical sales support. So what we did is we supported them in everything they did.

So this still is our role, except for we are not supporting, but we are a major part. We have just recently been renamed. So we were called Daimler TSS formerly. And now we are Mercedes-Benz Tech Innovation.

CRAIG BOX: That renaming came as part of the larger group changing its name from Daimler to Mercedes-Benz. Was there anything in particular about the change that has refocused what your group does?

SABINE WOLZ: The split between Daimler and Mercedes was that we have Daimler trucks now. And we have Mercedes-Benz cars. And we are closer related to everything that's concerning cars. So it just made sense to put us back to the cars again. We were connected to the cars anyhow. And now it makes sense to be part of the whole family.

I still see us as a family, to be honest, because I think it's difficult to do the switch in the head. But the focus is on cars. It has always been, and it still is. And, therefore, it makes sense to be a Mercedes-Benz part.

CRAIG BOX: The focus of Mercedes-Benz itself is obviously moving towards electric cars. We also see the rise of autonomous vehicles. Do you see car companies becoming computer companies? Are cars just becoming computers that have wheels as peripherals instead of printers?

SABINE WOLZ: One comparison that is very often done is a comparison between cars and smartphones. And I can see many, many things which are kind of similar. We do have apps on our mobile phones.

And people are getting used to it because we have our mobile phone with us every day, all day. We know what we can do with apps. They support us in our daily life. And they can also support us in a car.

So what I can see is that cars do have many apps nowadays. And they support us, and they support us on the road. And we want them there. We kind of want this type of technology, which we are so used to in our private life, we also want that in the car. And, in that sense, I can see that cars become closer to computers, to mobile phones, to everything which is digitalized.

CRAIG BOX: What other places do cars and IT intersect that we might not normally think of?

SABINE WOLZ: As an end user, we tend to only look at the end product. We can see a car, and we can see a car moving around on the street. But, for me, being a part of this whole family and being part of Mercedes-Benz Tech Innovation, IT is important at every step of the supply chain of creating and developing cars.

In the beginning, we have planning departments. They do planning. They try to get the right material there in the right time. And this is, of course, IT-supported nowadays.

We also have IT in production. It's not only that IT is doing the production, but it's also supporting those people who work in production and in any kind of shop floor. And so IT is very important there. It helps to predict. It helps to mitigate mistakes, to support everyone who is involved in production.

And then we also have it in marketing and sales. So marketing and sales without IT would not be possible anymore nowadays. And, of course, we have the IT in the car. We have it on the street.

So it's kind of the whole supply chain. And, sometimes, I have the feeling that people tend to only see one piece of it, the IT in the car. But I would say it's way more.

CRAIG BOX: You mentioned being able to see the car on the street. As the cars become electric, you can't hear them anymore because they don't have that motor sound. Will your team be responsible for putting an app in the car to make noise so it alerts people passing by?

SABINE WOLZ: So, actually, nobody from my team is doing that. But I know that there are sound engineers. And they try to put the sound in the car.

And I know many people who are really fond of cars. And they usually say a car without a sound, that's not the same feeling, especially people who like sport cars. So, of course, there is a way to pretend those sounds.

And you know what? Kids, if you see a kid playing with a car, they usually do the sound. So what will they do in future if an electric car doesn't give a sound? The only sound I can hear is if they drive backwards, it's doing this "beep, beep, beep" sound. But that's no fun to play. So sound and car is kind of connected.

CRAIG BOX: What you need to do, then, is record a kid going "brrm," and then just play that as the car drives down the street.


CRAIG BOX: Many opportunities now. Mercedes-Benz is headquartered in Stuttgart. But MBTI is based in Ulm. I hear you have a very nice church there. Is it a deliberate decision to be further away?

SABINE WOLZ: So you're right. Our headquarter is based in Ulm. But we also have offices located in Stuttgart, in Karlsruhe, and Berlin. And they are very strategic points for Mercedes-Benz because we have factories there. We have other important locations of Mercedes-Benz there.

All of those locations are growing. By the way, to be able to grow, we need more colleagues. So, in case anyone's listening and thinking, hey, that sounds really great, we are hiring. Just mentioning.

CRAIG BOX: Excellent.

SABINE WOLZ: I know. Our location in Stuttgart currently is as big as the one in Ulm. So, to be honest, I don't really know why we started in Ulm, because that was way before the time I even thought about working. We were founded in 1998. That was when I was 10 years old. So I didn't think about locations back then.

But, nowadays, we also have them at good points close to Mercedes-Benz. And, as you mentioned, Ulm has a very great church tower. To be honest, I like Stuttgart a lot as well. They only have a TV tower there. So, if you want to visit Stuttgart, it's worth it as well. But you should also visit Ulm and see the Minster tower.

CRAIG BOX: We're talking today because you and your colleagues told your Kubernetes story in a keynote at the recent KubeCon in Europe. In it, you mentioned how MBTI has been on a transformation journey over the last eight years. What can you tell me about that journey?

SABINE WOLZ: So, in general, what I can say is that the journey is a lot about using open-source technology. So the transformation is more towards open-source. Formerly, we were quite — especially Germans — we were kind of curious to use open-source because there is no real contract behind it.

So don't get me wrong. Of course there's a license behind it, and there are regulations to use it. But the contract gives you safety. And it tells you what you're allowed to do and what you can expect out of the license you just have bought.

I would say the management has figured out that open-source has more freedom, and it gives us more freedom. Freedom gives you ability, and it enables you to do things which otherwise would not be possible. And I can see a very good shift into the direction of open-source.

CRAIG BOX: You said you were running Kubernetes 0.9 in production. The paint probably wasn't dry, and we hadn't put the leather protector on the seats by that point. What was so compelling that you would run something before it was marked 1.0?

SABINE WOLZ: One thing I would like to mention if you ask this is that we have a very great mindset at Mercedes-Benz Tech Innovation. And we actually have the freedom to use technology in a very early stage and to just learn together with the technology. I know that other companies sometimes don't give this freedom because it's kind of a risk in it.

But, for us, we have put tech innovation into our name for a reason, because we like to be innovative. And we like to learn with technology and experience with technology. And, therefore, it made sense for us to work with a new technology which is out there, which was out there very new.

We have communities of practice where we have the freedom and the time to learn and to experience new technology. And I think this is what we did with Kubernetes 0.9.

CRAIG BOX: Was that shift driven by something you saw internally where you had a problem that you needed to solve? Or were you observing the market and saw that Kubernetes was out there and figured that it could apply to things within your business?

SABINE WOLZ: I would say it was a bit of both. So, on the one hand, it was that we had a — problem is maybe the wrong word — but there were challenges, and we had to face them. And Kubernetes was one way to figure out if those challenges can be solved by that.

The other way was also that we are very curious about new technologies. We want to go with the flow. We want to be innovative, as I've said. And I think, therefore, there was the curiosity to learn and to figure out new stuff.

CRAIG BOX: Your group built everything itself. I know that there weren't a lot of providers at that time. But you obviously mentioned an interest in open-source and being able to build and maintain things yourself. Was it a hard decision at the time to go for something where there wasn't a provider behind it? Or was that just part of that open-source ethos that you described?

SABINE WOLZ: I would say, at that time, it was just the right thing to do because there wasn't everything built up already. And there was a lot of potential in it. So it was the right thing to do to go into it and invent things newly.

If we were to have the same decisions today, maybe we would take different ones because there is much more on the market. But, back then, I think it was the right decision. And, therefore, the decision was kind of easy to make, yeah, because it just felt right to do it that way.

CRAIG BOX: What can you tell me about your Kubernetes environment today?

SABINE WOLZ: Nowadays, we have more than 900 Kubernetes clusters. And we have them on three continents. They are running on-prem. So they are running in Mercedes-Benz data centers. So the three continents, this is due to the locations of the data centers of Mercedes-Benz.

We operate 24/7. So even in the middle of the night, if something's going wrong, there's someone getting up and fixing the problem, which is, by the way, not that usual, because all of our colleagues sit in Germany. And many companies who do that have colleagues somewhere abroad, have them, let's say, in India or nearshore. So I think that is quite impressive that our colleagues are doing that.

We are evaluating going off-prem with our solution. So we are having a look at all the famous off-prem cloud providers. But, currently, we are only on-prem.

In general, we have five platform teams who are working in this area. So two of the teams, they offer the Kubernetes as a service. We have a logging monitoring and database as a service team, which is very popular, which is very well working.

We also have a container security service team. By the way, this is the only team who is using commercial tools at the moment. The rest uses the whole CNCF landscape, let's say. And we also have a team who is building Golden Path as they are especially well-working with IAM, identity and access management regulations of Mercedes-Benz. And they are building Helm Charts for that.

CRAIG BOX: Nine hundred sounds like a lot of clusters to me. Why so many, and what is your approach to managing them at that kind of scale?

SABINE WOLZ: Yeah, I think it's impressive, too. So my colleagues are doing a very, very great job there. Why so many? Because there are so many customers, and there's so much need in the whole Mercedes-Benz family.

And there's demand for it. And we are trying to fulfill this demand. And we are doing a great job.

So two of my colleagues had a great talk at KubeCon, as well. They were talking about how they are doing it, about working with cluster API, actually. This is the way they are doing it. And I could tell a little bit about it. But, to be honest, we could fill a whole talk only with this topic because it's such an interesting and deep topic.

And I have to say that my two colleagues did such a great job in their talk on KubeCon. I would suggest going to YouTube and watching it there because it's worth to watch what they have told about it.

CRAIG BOX: And, of course, you'll find a link to that talk in the show notes. With those teams and with the clusters that they get given, they have you-build-it, you-run-it software model, as I understand. Did that exist for those teams before Kubernetes? Or did Kubernetes enable you to extend those privileges to your application teams?

SABINE WOLZ: Kubernetes, of course, helped to support this. But I can see that this is a general trend in IT. So DevOps is the thing to do at the moment. It's not modern enough any more now to have development and operations separated.

For me, this is also something which has to do with empowerment, empowerment of your colleagues. You not only give them the ability to develop something, and afterwards, they are not allowed to take the responsibility. But you enable them to say, you can take the full responsibility. You can build it the way that you think it is worth to build it. And, afterwards, you can also operate it and you can still feel connected to the product.

And I think that is a great thing because it gives the people also, as I said, it empowers them and it gives them the responsibility they want to have. I think it helps for the developer happiness. And I would even say, if you know that you have to take the responsibility, even afterwards, you will build it completely differently because you know that this is your baby, and you are in charge for it.

I like to compare with game theory, where everyone knows that, in the last run of the game, you would behave differently because you know the game is over afterwards. And I think it's the same with development. If you know afterwards, someone else will deal with it, you will develop differently.

CRAIG BOX: Yes. I definitely think that incentivizes developers to put in debugging and logging features in applications that they might not have considered when that function was done by a completely separate team.

SABINE WOLZ: Absolutely. Yeah.

CRAIG BOX: One of the hardest parts of setting up an SRE function is getting management to sign off on the idea that you can hand the service back to the engineers. In the case with Google, for example, the engineers will quite often build and operate a service. And then an SRE team might take that on for them.

But, if it starts failing too much, they'll say, actually, you haven't built it well enough. And they'll hand that back to the developers, and they will be responsible again. When you have that model, with one team looking after developing and managing it in scale all the whole time, do you have any trouble getting management to buy into operating it this way rather than the previous model that they perhaps used to, where you have the wall that the code is thrown over?

SABINE WOLZ: In general, I think it's difficult to say. But I can talk for the management I am experiencing. And I have figured out the same as I've just mentioned, that it makes sense to give the full responsibility and that DevOps is the way to go. So my management, at least, is supporting us on that way. And they have the same opinion as I have, that it makes sense and that it's the way to go.

CRAIG BOX: You've talked about the teams that you have, some running the clusters, some providing shared services on top of it. I understand that you are in one of those shared services teams.

SABINE WOLZ: Yeah, absolutely. I'm the product manager of one of the shared service teams.

CRAIG BOX: So what are the products on the Kubernetes stack that you're personally responsible for?

SABINE WOLZ: We have monitoring and logging as a service. And we also have databases as a service. And, in general, our aim is to have services or to provide tools which enable other DevOps teams on their daily business so they can focus on their actual, let's say, application they are developing. And everything around it, they don't have to think about it too much, because it's there already and we are supporting, maintaining, developing it for them.

CRAIG BOX: Is that something where your team are going out and evaluating the space and figuring out what's available? Or are you taking direction from your application team, who are saying, hey, we've heard there's this Prometheus thing, we'd like you to offer it to us?

SABINE WOLZ: It's a bit of both. Of course, we try to listen to our users as much as possible. And they give input a lot.

But the loudest person is not always the one who knows best for the whole community. This sometimes is a problem because we have users who tell us constantly what they need. But that's the single use case.

And what we try to do is we try to listen to them, and on the same side, filter what is good for a bigger group for many use cases. And what we do there, of course, is we also have a look at the CNCF landscape. We get in touch with other end users. We try to figure out what can help many people because I think that's the sense of a shared service.

CRAIG BOX: What happens in the case where one of your users wants something that is outside the scope of what your team provides? Are you able to onboard something specific if someone needs a different database, for example, than the ones your team offer? Or are you, in general, able to guide them down the path of using the tools that are provided by the team?

SABINE WOLZ: Again, it depends. I know that's always a good answer. It depends on the demand.

We sometimes have users asking us for, let's say, different monitoring tools, or for an APM solution, stuff like that, for a specific tool which we can't offer. And what we can do is internally, in our Mercedes-Benz intranet internal environment, that it is kind of easy to open firewalls and to give them the access, which we try to — if there is a plugin we can install, we try to do that — and by that, offer them the possibility to use the tools they want to use.

But that's always in between boundaries, of course, because if you let them do everything they want to do, then you can't really manage it anymore. So it actually really depends on the request.

CRAIG BOX: And, likewise, if someone needed something that was perhaps specific hardware or cluster configuration, like someone wants to use a cloud machine learning service that isn't available to run on your on-premises hardware, do you work with those teams to find them ways to get access to external services that are still within those guidelines?

SABINE WOLZ: Not really. If we have knowledge about it, then we can answer their questions. But what we don't do is a kind of consulting to let them know where they can find the solution they are looking for. I don't think we have the capacity for that one.

But, of course, we don't say just no. But we listen to their question, and we try to see if we can do something for them. Because, sometimes, you have heard things in between the company, and they might have not heard of this service yet. And then it's, of course, a good thing to do to let them know, hey, there is something. And you can check it out, let's say, in the intranet or wherever.

CRAIG BOX: In 2021, Mercedes-Benz posted their FOSS manifesto, the set of core values demonstrating the commitment to free and open-source software. How did that come about?

SABINE WOLZ: So we love to call it the FOSS "manifossto". How did that come about? So, in the beginning, it was written for ourselves. There was a small competence group, and they were really fond of FOSS. And they have seen the potential in it.

And they have decided that they will just write it down because it makes it easier to get something as a baseline for a company if someone has written it down and there's a kind of a paper which you can go around with and discuss it with. So the aim was actually a cultural shift in our company. And, as it turned out, the cultural shift went much further. And it also went directly into Mercedes-Benz.

And, nowadays, the FOSS manifesto is live, and it's online. And everyone can have a look at it because it's so, I would say, even impressive, what my colleagues did there, because they just believed in it.

CRAIG BOX: And that "manifossto" — I'm only going to say that once, by the way — it talks also about inner-source. How do you define that?

SABINE WOLZ: Inner-source can be very much compared to open-source, except for it's in between the boundaries of our company. And the aim kind of is the same as for open-source. But we want to have the code which we are developing to be open for everyone to use.

So you can forget. You can open pull requests. You can use the code which is there. You can support in developing it.

And the reason is that it's always not a good idea to have knowledge at one team. But it's good to spread the knowledge. There's a saying, "If Mercedes would know what Mercedes knows." And I think inner-source can cope with this problem because it helps other people or other projects because they don't have to invent the VLU. But they can share code and share knowledge that way.

CRAIG BOX: When you are developing things internally, do you also have an eye to how they can be released upstream?

SABINE WOLZ: I think it depends, in between boundaries again. Upstream, yes, but still, there is the competition between the car companies and between all the companies. So, of course, we have an eye on it if we can put it upstream or not. I wouldn't say that it's in general forbidden or in general a thing that we do. Again, it depends on the use case and on the topic.

CRAIG BOX: Let's take, then, Kubernetes as the topic. Mercedes-Benz are one of the top 20 corporate contributors to Kubernetes. Is this through work that you've done customizing the platform to do what it needs to do? Because it doesn't sound like your use case is substantially different to what other companies might need. But it might just be that you're ahead of the curve and doing it early.

SABINE WOLZ: I would say it's because we have very, very good colleagues and very talented developers. If you experience that something is missing, one thing you can do is you can complain. And the other thing is you can try to contribute and change.

And I think that is just our mindset, that we try to improve things. And, as I said, the colleagues, they just have the ability to do so. And the other thing is they have the freedom to do so because the management lets them do it. And they say, hey, if it's not the way you need it, why don't you just contribute there and see if you can change it in a way that it makes sense?

And that, again, is kind of taking responsibility, and, yeah, getting into action. And I think we are very, very good at that, although our use case is not that different from others.

CRAIG BOX: You were one of the first presenters in the first keynote at the KubeCon recently. How was the rest of the event?

SABINE WOLZ: So I enjoyed it very much. It was a great event. I have to say it was one of the first in-person events in a very, very long time. So it was kind of overwhelming to see all those people and to actually speak to people in real life. That was great.

I've met so many other end users. What I think was great is that we shared our story on stage. And, afterwards, people asked us about our story. But, on the other hand, they told us their story.

And that was very, very great for me because I've learned that we kind of have different stories, we are struggling with the same things, and that it's worth it to talk to others because they might have solved your problem already.

CRAIG BOX: Do people recognize you and stop you in the halls?

SABINE WOLZ: [LAUGHING] At the beginning, after my talk they did. And I was very surprised because I thought I looked like every second woman in the world, and they won't recognize me. And then I was very surprised that they did. And I still had fame in Valencia.

But it stopped at the airport because I think nobody expected me there anymore. So I saw many people with KubeCon T-shirts, and nobody recognized me anymore. And when I had to switch the plane in Amsterdam, fame was over. So that's the story.

CRAIG BOX: Fame is a fickle mistress, 15 minutes. My advice to you is do a podcast because that way, people don't generally know what you look like. And then, if you don't talk, no one knows who you are.

SABINE WOLZ: [LAUGHING] That is good. Yeah, good advice.

CRAIG BOX: What is it like being part of the end user community at the CNCF event?

SABINE WOLZ: I think it's great to be part of the end user community because, for me, I'm a very social person. For me, it's always great to be part of a community and not being an individual, being by myself. And, as I have mentioned, the end user community, we all have similar problems. We all have similar use cases.

And it's just great to know that there are other people around who try to solve the same problems. And you can talk to them. And not all of them are, let's say, on the different side.

But they also want to support you. You know they want to learn from you, but they also want to give their knowledge to you. And I think that that's a great thing, and that's kind of how community works.

What I have seen there is that, still, there's a little amount of women in the community. There were women, of course. But I think we should have more of them in there. This is something I would wish for the community because I think a diverse community can reach better things.

CRAIG BOX: One area that I noticed at KubeCon, obviously, is the vendors and the people building the tools, they are obviously the people who are in the showcase hall in the most part. And they are a lot of the people who are doing the presentations.

The end user group, however, is probably the group that's growing the most. There are more vendors coming online every day. But there are probably 10 times as many people who are using Kubernetes. What do you get out of going to KubeCon? And what were you able to take back to the rest of your team?

SABINA WALTZ: There were many vendors. And sometimes it felt like there are more vendors than end users. But for me, in my role as a product owner of a shared service, it was quite interesting because I got the chance to talk to several vendors. I have learned a lot about new database technology, maybe even also logging and monitoring solutions, new ones.

It doesn't necessarily mean that we will change our portfolio. But it was just a good way to get in touch in a very quick time, in a very short time, with different vendors and even compare them directly. Of course, they had great souvenirs there as well. I love my new sunglasses and my new bottle and stuff like that.

That was the fun part about talking to them. Sometimes I didn't even want to get the stuff. But then they said, hey, take a T-shirt. And I thought, OK, I'll take the T-shirt. And I'll take the information, which I can bring back to my team and discuss with them if it makes sense to put it into our portfolio.

CRAIG BOX: There were a lot of times where I'd see about 100 people hanging around a booth. And then, a minute later, one of them would win a drone, and then they'd all go away. It was drones and the Nintendo Switch. I think they were the two top prizes of the conference. You didn't win a drone or a Switch, I take it?

SABINE WOLZ: No, I didn't win anything. But, to be honest, I didn't compete in anything. They sometimes ask you to take a picture and put it somewhere in social media. To be honest, I didn't do that. I didn't have the time to do that.

And they also had LEGO. They had a lot of LEGO prizes there.

CRAIG BOX: What, then, about the talks, as an end user? Do you feel that the talks are aimed at you? Or do you feel, a lot of times, the maintainers are simply talking to themselves?

SABINE WOLZ: There were so many different talks. Some talks really caught me. And, with some, I really had the feeling that the aim was to get the knowledge to the end users. Of course, there were some where I had the feeling that it's more of a sales show. But, in general, I would say it was such a diversity in the talks that there was something in it for everyone.

CRAIG BOX: Finally, I can't miss the opportunity to ask someone who works — indirectly, even — for a car company. What do you think we will have first, fully self-driving cars or flying cars?

SABINE WOLZ: So, from my point of view, it will be self-driving, autonomous cars. I have to say that because some of my colleagues work in this area, and none of them are working in the area of flying cars. So, of course, I assume that autonomous driving will be there first.

CRAIG BOX: I'm disappointed in that, Sabine. If I'm honest, we were promised flying cars as far back as the 1950s. And they still are not here today.

SABINE WOLZ: I know. Life can be disappointing. But my feeling is that people who are down to earth, they are much more light. So it's always difficult to talk to people who have their nose to high up in the air. Maybe that's the reason why we have decided to stay on Earth.

I know we all love Peter Pan, and I know we all want to fly around and have fun. But maybe it makes sense to just remember where we have to stay, and that's on the Earth.

CRAIG BOX: That is good advice. Thank you very much for joining us today, Sabine.

SABINE WOLZ: Thank you very much for your time.

CRAIG BOX: You can find links to Sabine on LinkedIn, Mercedes-Benz Tech Innovation, and the KubeCon talks linked in the show notes.


CRAIG BOX: And, with that, we are, once again, at the end of the show. If you enjoyed it, please help us spread the word by telling a friend. Or tell someone who's interviewing you.

If you have feedback for us, you can find us on Twitter at @KubernetesPod or reach us by email at kubernetespodcast.com. You can also check out the website at "Kubernetes Podcast," where you will find transcripts and show notes as well as links to subscribe.

Please consider making a recording of yourself going "brrm, brrm, brrm," and we'll forward the best ones on to the sound engineers at Mercedes. Thanks for listening, and we'll see you next week.