 Welcome back to theCUBE's live coverage here in Boston, Massachusetts for Red Hat Summit 2023. I'm John Ford, Paul Gillan, kicking off day two, live streaming, wall-to-wall coverage. Christian Uber, who's the CTO of Utah, subsidiary of Bosch is here, customer of Red Hat in a cutting edge automotive industry where Edge is the big store. Christian, welcome to theCUBE. Thanks for having me. It's a big keynote, automotive, big market. You can't get any Edge computers on cars with software. What's the key challenge in automotive? Because that's an area that's obvious. We see all the cars, electronic, more gadgets, more software. Yes. So it's a very diverse landscape. I think the biggest challenges with regard to software we currently see with the traditional big OEMs and for them the biggest challenge is development and integration, speed and stability. So it's not unusual for them to have week-long or even month-long integration cycles. And we know from other industries that the faster you iterate, the faster you deploy to production, the more efficient your software development process is. And this is a new experience for automotive. We have coming from safety, coming from manufacturing processes which rotates a little or iterate a little slower and getting through that, bringing the new technology into automotive while maintaining a high safety level that is the biggest challenge right now. The theme here this day is scalability, efficiency, obviously Edge, industrial Edge. Automotive is a great use case because it represents the most tactical Edge you could imagine. It's moving, it gets telemetry, different towers. Where's the signal coming from? What is the operating system behind? What makes it all work? It's so fascinating, the automotive. Well I think currently the operating system is not yet there. We're getting there. So we have a range of highly optimized special solution like real time operating system for microcontrollers. We have operating systems for high safety on vehicle computers. We have new players like Red Hat entering the vehicle which is really exciting. And they're catering different user needs right now. And this is all coming together. What we all want eventually maybe at the end of the decade or longer is that deploying onto our car feels more like deploying on a cloud environment. You don't have to know the specific hardware it's running on. You have some attributes that describe the requirements of the automotive function. You deploy it either in the cloud or on the vehicle and there's much less friction in between. A lot of customers say the next generation of software defined vehicles, they don't think as a car in the cloud but they try to think the vehicle from the cloud perspective. And making a lot of design decisions inside the vehicle from that perspective and this is really exciting. What are you specifically doing with Red Hat technology in the vehicles? So we came together when we got notice of Red Hat's ambitions to provide a safety certified version of Linux. And there's a huge market demand for that right now. Because there's not so many options right now in the market to choose from if that is what you need. And what is interesting for us with regard to Red Hat's competencies is they did for two decades now a great job of balancing high stability and keeping the systems up to date incrementally over time completely hassle free. And this is something that also automotive companies want nowadays. They don't want to think in like we have a four year cycle then a start of production and you start from scratch with every cycle but they want a continuous incremental update over decades eventually for the vehicles and then you need competencies like Red Hat and because of that Red Hat is really an interesting partner for us. What do you guys do specifically with the technology? Can you share your companies core competency? What do you guys deliver? Yes, so our whole portfolio is retail operating systems, middleware, development tooling, measurement hardware. It's all the things that an automotive developer needs to not only develop but road release automotive functions including highest degrees of safety and so on. This is our core expertise and we're developing this further right now in the direction of how do you do large scale data driven development. This is coming from ADAS and AD where we have like multi gigabit per second sensor erase a lot of data coming in and huge challenges to make these systems safe for road release. You sometimes have for example situations that you get in front of your cameras or sensors every 100,000 miles and this was your chance and what you need in such a situation is a very deterministically behaving operating system in middleware so that you can replay the same data over and over again until you have fixed it, put it into your CI infrastructure for regression tests and move on from there and this is something which is currently in the vehicle which is a very distributed place with many projects running in parallel, not happening and this is where we help, where we bring the newest technology which was developed in Bosch automated driving to the market via ADAS and now also in the future together with Red Hat. What is the status of fully autonomous vehicles at this point? It seems like five years ago there was an expectation that we wouldn't have to drive anymore by 2030. I don't hear anyone saying that anymore. Well, the normal answer is always in about two years it's going to be here and in about two years it's going to be here. This is the year of... But to be honest, it's an amazing heart to fix long tail problem. So it's very easy, you can go to a university and within three months you have something that drives, drives awesome, but to really cover the long tail like seven, eight, nine behind in your certainty that this vehicle is going to be safe under all conditions it is a really tough problem. And I think we will solve it in this decade. And I think the range of companies capable of doing this will be one or two. One or two? Yes. Will ETAs be one of them? ETAs is not doing level four systems. You're not right. Just for people building this. And I wouldn't wonder if it's really if you go by who's actually having this at scale on the road all the time in a commercially viable fashion not in the sense of a vehicle cost you 300,000 euros or dollars, but in the sense of, hey, this is here now and you see it everywhere. I wouldn't place my bets on this decade that this is coming in scale. Oh well. You said during your keynote that you think 2024 will be a turning point that we have a new generation, new architecture for software defined vehicles. What do you think that will look like? Well, I think some of the technologies in the vehicle especially for the highest safety levels they are here to stay, they're proven in use, they work, do their job but they're not suitable for highly iterative fast development. And this is what we talk about when we say software defined vehicle, software defined vehicle is not just a different software stack. It's an OEMs capability to continuously update the vehicle all the time in a very short cycles. And with regard to these stacks there are many, many options right now in the market. If you look at, for example, Eclipse, STV, AD members by now everyone is contributing their stuff and there is a lot of things to choose from and we will probably not get out of this with the traditional methods of the automotive industry like building committees, standardization or joint ventures and so on. This is all gonna take at least two years and in two years the manufacturers believe they will be gone from the market in attractive Asian markets or also versus new competitors in the US if they don't act now. And so there are a lot of interesting behind the doors discussions right now in the sense of, hey guys let's decide for one of the stacks that is good enough we'll put several million vehicles of volume behind it then we're going to say, hey this is our stack for the next couple of years we are going to support this who's going to take over what responsibility? And not in the sense of you, I give you something but really in the sense of safety, security supply chain security and so on and then you compile a stack of things and make a decision and I think this is going to happen within the next 18 months. Christian why the Red Hat in vehicle operating system and why is it a fit with Etos? Can you explain that? Yes, so first we have to give safety guarantees to the higher layers. So from the OS we need to understand very well what is the scope of safety guarantees I'm getting and what can I build on top of this? And so far there was no vendor in the Linux ecosystem who we could have really reliable discussions with or discussion within the sense of that we believe this is going to work. When I first talked to the Red Hat guys they had that impression. So the whole approach like not certifying once, certifying with a big focus on automation to certify it continuously with a big focus on very well defined scopes of operation. I immediately understood they are no, they know what they're talking about and this is why they are now an interesting partner for us. We get the competencies to make that work and we get also a continuously updated business model, delivery model that is well aligned with what our customers expect from us in the future. How is the cloud and the in vehicle operating systems converging, what role does the cloud play in managing these increasingly automated vehicles? That's an interesting question. So we have seen a large range of discussion. So the most extreme position was let's have something like Kubernetes in the vehicle. And we found out it solves a lot of problems you don't have in the vehicle but it doesn't solve problems you have in the vehicle especially you cannot scale out if you hit a resource limit. You have to live with that. You have to degrade, you have to kill processes and so on in a very well defined behavior if safety is important for you. And so I believe we need a low barrier between cloud deployment and in vehicle deployment but it doesn't have to be necessarily the same technology. So we're seeing interesting things emerging there right now that cater to the demands of resource constrained vehicle but use for example the same YAML description for the automotive functions that you deploy in the vehicle and you can, as they are you can immediately deploy them to the cloud and back and forth. And I think that is a very attractive option going forward. It's interesting YAML comes up a lot and there's a demo yesterday where there's no YAML in it. We were joking, hey you got to get your driver's license for the self of the driving car, pass the YAML test. Only kidding but as ease of use comes in with the car functionality, the interaction with the human is going to be a very interesting enablement. How do you view that when you look at the roadmap with Red Hat, you got a lot of open source movement with AI right now. How do you see that middleware enabling a user experience with a human, might even not be driving but maybe minimal interactions, what's your vision there? Well I think, especially in a safety critical context, a set of well chosen constraints enable a lot of trying out things. So what you can do with middleware for example, you can secure that the system is always in a well defined state and you can trust it. And if you give that in a very high confidence level, you can enable users to try a lot of things. Like deploy your own AI model as an end user because you scan the video images coming in, you have some idea you want to monitor maybe birds or whatever. And I cannot foresee all of this but what I can make sure is that I build a safe platform that you can play on top on. And I make sure that you don't interfere with the safety state of the vehicle and that enables a lot of creativity facing the user. They can bring the customization, could be playlists for music, it could be some computer vision, whatever foundation model that they want to bring in but you have a hard top. Yes, secure. What we make sure is that we give you a safe deployment platform that you can try out new things even as an end user if the OEMs want to enable this. And but this is going to grow in these ecosystems. This is not what we pre-define. We just make sure that it's safe and operates as intended. How open have the automakers been to the idea of open source, of open sourcing their innovations as it develops the way the software industry has. We've seen an interesting development. I think a couple of years ago it was this new thing where no one is accountable and they didn't want it. And it developed over time. Then it was like purchasing organizations, seeing hey, this can be more cost effective. Can we have this open source and so on? It became a strategic discussion and nowadays it's changed again in the sense of hey guys, we have a problem. We have a fractured ecosystem. We're not delivering to our customers right now. It's too late for standardization. It's too late for joint ventures. Couldn't we work together more quickly, more efficiently, more effectively in an open source kind of environment? Like we don't have to sign off anything. We just put in our stuff. We need people who support us with that and then we move on. And this is the current state of open source discussions and I would really welcome that. Talk about the roadmap with Red Hat. As you look at the edge, it's going to develop very fast. If it develops as fast as AI is kind of coming down the road, so to speak, you're going to probably see a lot more speed, agility, and latency is a huge issue at the edge. And so all these things are coming together. What are the core issues? When you look at the edge as it develops, what are some of the core things that need to happen quickly to keep advancing and it gets into a position where it's continually reliable, stable, and programmable? Well, I think we need something with a container-like experience, but with the same level of assurances and guarantees you get in the classical ecosystems which iterate too slow. I do not believe it's necessary to be slow if you want to assure these things. There's a lot of potential we haven't tapped into yet. Especially, for example, the deployment model. In a container, you bring all your own dependencies. You don't have a complicated web of dependencies that you need to agree on with your colleagues, what library and what version. You just package it, you have an outside interface, and that's it. It doesn't get more complicated than that. We aren't using that in automotive so much yet, but we should. And especially also with the infrastructure. It's cloud native coming to the edge. What? It's cloud native coming to the edge. Yes, and also with regard to supply chain security, if you hide your dependencies, you may be open up new security angles. An old library hidden in some container, but this is all transparent. If you do it right with the modern stacks, you have versioned layers, you can see in your manifest, hey, this is outdated, I have to update it. You have transparent update mechanisms. There's a lot of potential we can tap into. So dependencies managed with the containers, software supply chain, transparent, visible. You know what you're dealing with. Is it declarative environment? Yes. Secure, hard and top. Yes. And then bring your own foundation model, bring your own AI to the car. Yes. Paul, we just got the whole roadmap right there. Yeah. Can't wait. Christian, thanks for coming on the keep, really appreciate it. Final question, what are some of the hallway conversations that you're having here as you give your talks and are talking about some cutting edge stuff? Literally edge. What are some of the conversations you're having here with folks here at Red Hat Summit? I think the interesting one is, everyone is familiar with the range of options now and everyone is trying to find out what's it going to be now? What is going to be the next stack? How much Reddit is going to be in there? How much of something else? Eclipse, SDB or whatever? And what's it going to be now? And I think it's a very interesting time. I think it will materialize very quickly. Christian, Uber, CTO of Etos, part of Bosch, really kind of building that operating system in the edge for cars, automotive, big sector. We all know what's happening. We see self-driving. We see electric. We see user experience changing. Bring your own AI into your car. I love that future. Soon they'll be flying cars in the world. It's theCUBE coverage day two, wall-to-wall live coverage. We'll be right back with our next guest. I'm John Furrier with Paul Gillin. We'll be right back.