About this talk
In this talk, Andy discusses his transition from automotive software development to his new role as an open-source advisor at Schwartz Digits. He highlights the importance of ecosystem visibility in the rapidly growing software-defined vehicle (SDV) landscape and shares insights from the automotive project governance at Eclipse. The speaker emphasizes the need for a structured overview of various projects, using the Cloud Native Computing Foundation (CNCF) as a successful example. Andy explains the current state of the SDV landscape project, detailing its interactive website that categorizes components based on their specific functionalities. He also encourages contributions to the automotive open-source collection, which serves as a repository for various automotive software projects. The session concludes with a discussion on the necessity of collaboration and clear communication in advancing the automotive open-source ecosystem.
Full transcript
Thank you so much. So I'm Andy. Um I have also a little bit of history in automotive. It's about 30 years and I have also some similar topics than Dana. So I do not want to continue now the safety talk but one thing um when I started at Bosch uh in 1997 so I was also working in the ABS team means in the anti-lock braking system we
do did not talk about safety not really so there was no IEC 61508 but we had a team it's called safety software so we did also a lot of safety stuff without the official regulation. So this is what I want to add here. So I have a new role. So I'm out of automotive more or less right now. So I started 3 weeks ago in the cloud
section. So at a German company called Schwartz Digits and um I'm now an open source adviser. But nevertheless, I'm still related to automotive with especially with Eclipse and I want to show you a little bit about the landscape project. So what I want to uh tell you a little bit is why ecosystem visibility is really important. Then an existing success story. I guess everyone of you is
uh uh every one of you knows CNCF. And then let's go a little bit deeper into the SDV landscape project. What do we have inside today? Who works on that? So Tilo is one of them. Just a spoiler. And how it works and what comes next. So why visit visibility matters? So the SDV ecosystem is growing rapidly. So I'm also a member of the automotive PMC. So
we took over the the governance of all these projects and I can tell you yeah it's fully right. So a lot of things happen there and it's really really fast um how they grow. So therefore you need a kind of overview means yeah you can can compare it with a a big box of Lego bricks. Currently, everyone is throwing things in there, different color, different sizes, sometimes
different interfaces. But at the end, if you want to build something out of that, you have to know where is it, how it looks like, and who can compare or combine it um with that. So therefore, it's a really helpful also for new comm community members to get familiar with all that stuff. what is in there in the big box. So therefore a kind of a structure
is needed and of course a transparent structure and that shows also how diverse the landscape within the STV working group or within the STV projects are because complexity without structure slows down the innovation. You all know the cloud native computing foundation. So maybe you know also the picture about it really big a huge amount of projects in there. But what is the the positive think about that?
So it helps also to the new users of that to understand what's in there, how the ecosystem looks like and it shows also how the community and the ecosystem is growing. It's CI CD Traven and it's a real real strategic instrument with the visibility to show do we have gaps do we have some dependencies on other stuff we have to overcome and what do you think do
we need something similar for STV at Eclipse I guess you know the answer Right here it is. So here's the first step I will say into the landscape project based on the structure of the CNCF. So it's also open source. So therefore you can take over the sources adapted and this is the current result at the moment. It's live. I will show it later on. So here
you are. you have the QR code step into that have a look on that so it's really really cool but I came a little bit later on to that and who is working on that stuff for example TLO myself and of course the technical advisory committee of the SDV working group why because you have to categorize it and to the matching categories. It's necessary that you have
here the devices in there. How does it works? So, it's also visualized like the CNCF on an interactive website. It's also integrated now in the Eclipse SDV website. I came back to that later on. uh the components are classified dependent on their STV functionality and projects are added automatically which is really really helpful so that you do not have to add everything by yourself it's just working
the project data are sourced directly from your Eclipse project page so this is really really an advantage so it works on the CNCF case. It enhanced the workflow of your data and the current landscape shows a real big diversity of the and yes they are categorized to their specific softwaredefined vehicle um functionality. So it's a kind of a workflow and you have to feed it with the
correct data. So therefore your input, your commitment to the projects is really really essential and it helps to build that. So now I jump to the website. So it's really really great. The best STV website I ever saw. So you have here. So for me it's also really new. So it's a real real um very good restructured uh website right now. So here within the projects you
have here the integration projects. If you go there scroll a little bit down then here So it's really really live. So I tried to make it a little bit bigger. So for example here you have is it big enough? No I guess not. So I'll try to make it a bit bigger. that you can also read it. No, that was too much. So on the left side
you have more or less the bigger categories like tooling for DevOps and you have your development tools inside. Just to make an example, if you hover over the project name and the project logo, you see a little bit more of information in there. So it's under Eclipse SDV of course and it's currently in in the incubating um state. You have here the link to the GitHub repo
and also to the documentation website. And if you want to have more then just click on it for example here on the website and you jump to the project website. So click directly into the open vehicle API. So now my scrolling won't work. So sorry for that. And you get here a bit more of um information about that stuff. For example, here the link to the GitHub
and so on and the current state. So you have things for testing, validation, um design, deployment, process and frameworks like the trusted software framework. Also we heard about uh thread X um the ACL certified open-source uh operating system. Then some example applications, middleware and frameworks and so on. So all that stuff is in there. And if you want to get a bit more familiar or you want
to check if you have it understand correctly, there's also cool link here. It's called the game. So you can check here if you know everything about that. So really cool thing, really new and yeah, surprising. Um you you can test here if you are familiar enough with your projects or to to learn a little bit more about that. So you can test yourself later on just before
the beer. So going back but we want to expand all that stuff. So to complete project information to have more projects in there like um HL and other uh opensource software initiatives. So we have to categorize the new areas of the software defined vehicle technology actively. So now it's also up to you what is your STV project? Bring your ideas. Think about bring it. But that's not
all. We have also um in the same organization array. It's called the automotive collection. It started as a hobby project by myself triggered by a former colleague of mine um feeding me with um all the open-source automotive projects I never heard about and I lost the overview. So therefore I started to collect all the links and bringing in to The hobby project as mentioned is now transferred
also to the Eclipse Foundation and um you can find it as well. So the main topics we have currently in there is OSS management, automotive software of course development and tools something about cloud so it's now more relevant for me. Um, open hardware. Open hardware builds also some GPUs for um, cloud stuff for example. Then other cool stuff maybe related to automotive other things. So to make
some examples what is in there? So about the operating system we have in there the Eclipse of Red X of course. than other stuffs like Sapphire, L4RE, some base software, middleware and communication protocols like um FAM IP for example, Eclipse ecal protocol interfaces API kvisa for example very important for standardization of uh APIs data management like uh open MDM uh the Eclipse data space container software and
service orchestration anios symfony and also some open-source solution for in vehicle infotainment uh like uh the pilot drive. So this is not an Eclipse project right now but maybe in the future and a lot of more of course the same for the architecture and design um for IDE testing validation simulation and what triggered me in the morning during the keynote hey we have also something for formula
one in there in the collection not in Eclipse right now but maybe maybe should do you know the software heritage package if not it's also worth to have a look in that so they have or they do a kind of um I would say archive all the software open source software which is out there for example you find also an open source version of MS DOS for
the who knows MS DOS till now offer the double USL1 from Microsoft um the Leno emulator or the Linux um yeah now I lost my words. Um yeah um the Windows subsystem for Linux the version one is also in there for example and a lot of other stuff. So if you look for your own project, have a look in there. Maybe it's in there. If not, try
that it come in. Also about AI, AI is currently the hot topic. I will say everyone is talking about everyone needs it. Um there's a lot of things ongoing also in Europe within some public funded project. So really worth to have a look in there. For example, here we have the open- source AI definition. Then Eclipse Almos Aentic AI if I have it right in mind and
a little bit off topic I will say it's awesome robotic tooling but if you are dealing with um autonomous driving a lot of companies use the Ross framework for for example as a prototyping variant before they bring it to other stuffs so therefore it's also really worth to have a look in there if there is something you can use and yeah. So what is your project to
bring inside the STV landscape? Do you have one right or do you have some ideas what you want to bring in or what is missing or No, no idea. I know you are waiting for the beer and the wine. Yes, the microphone. Thank you. Thank you. Um, for going back to that quiz game that you showed, >> you want the game now. >> Uh, it's not about
the game per se. Um, have you given any thoughts instead of, you know, asking about what a project is? >> U, maybe it's some kind of quiz that whoever's answering, maybe it's a new contributor where they input their skill maybe. Sorry, didn't get it. Sorry. >> So, imagine a quiz where you know you are new to Eclipse SDV like I don't know anything about it and then
there are a few questions there that asks you which programming language you know or which kind of technology domain you are interested and then this questionnaire suggests you which approach you should start engage with as a contributor for example. Yeah, the the game is currently to to get familiar with with the stuff what is in there. >> Yeah. >> So it's related to that >> or I
misunderstand you. >> Yeah. About to figure out what the user knows what whoever is answer the questioner knows and how to match that knowledge which is needed by whatever project is in Eclipse SDV. >> Does that make sense? So, so, so, so currently it's it's just >> Yeah, you you you get an a question about what is in one of these >> No, for example, um you
get in there, you get random questions. Let's say do you like rust programming or do you like >> do you enjoy I don't know image building for operating systems do you like messaging frameworks >> and based on chose the answers of chose >> okay >> the assistant shows them you can engage with chose three different projects which are a match for your skills >> now now I
get it sorry >> eclipse yeah kind kind of um >> a a chat tool on the interface right now. But >> my I had another question is about um you have all this landscape information. Um it sounds interesting to eventually have some kind of a blueprint design builder where you can kind of uh from a you know from a just system design perspective you are able to
kind of pick whatever project is available so you can have a a visual representation of what you what your blueprint would look like. Yes, I will say um the current visual representation is here and then there's a project called STV blueprint where we try to to combine the first projects to an working >> yeah thing um I'll give an example um I have people in my company
that are there always some interest in contributing or trying to create a blueprint for Eclipse SDV but they don't know where to get started and they don't have the the whole picture of all the projects. Um so if they have something like this page is a good start. Um >> but have something that you are able I uh some kind of I know I don't know drag
and drop or just a form where you pick pros per category and we kind of build up some kind of a minimal uh blueprint design and that then they can use as a basis to actually propose a blueprint. >> Okay. Yeah. >> Suggesting but but it's a a great idea. Yeah. So what we are also talked in the the past a little bit about the the projects
is um yeah let's have a kind of a scanner who crawls all the projects and bring in the the information also bring in a kind of a pull request to put it automatically in into that stuff. And um yeah so what I also understand you need a kind of a chat tool with all the ask your questions. Um when I need something for communication then the answer
should be this this this and this this project or this project or the combination of that and that or uh spoken in in the next step use this blueprint or this one or yeah at the end maybe a kind of a I know it's the the wrong word a kind of a reference architecture. That's all I had. Okay, where were the questions? I don't know who raised
these answers, so I'm going to go closer. >> Um, so my question is, um, is there some API to define the SDV and then to do some co- simulation with your hardware code with HDL something? Is this planned? Um, currently I guess there's not a real API. Um, but the the collaboration between the projects is about to start if I have it right in mind or it's
in not in full swing but but it's as I have it right in mind. Uh, it's it's about to start but Sarah please correct me. If there's an API for the SDV to uh for example your HTL code to do hardware code simulation I mean some projects run on some >> but there is no standardized API there is no hardware project >> okay uh maybe second question
so the STV it's software defined vehicle It's currently only automotive or does it also include airplanes or ships? It's could be used the domain. >> So the the starting point is of course not of course it's the car >> but you never know for which you can use >> Thanks. >> So I guess communication protocols is not only needed in vehicles. I guess can more or less.
Yes. But can is also used in other domains I would say. So therefore feel free. So thank you very much. Uh I'm coming back to your analogy at the beginning of the Lego bricks. I want to extend it a little bit. >> Okay. Um so um we have seen now a little um amount of uh bricks. So you have stones, wood and we have some tools like
saws and everything else. But uh what I'm missing completely is uh the big picture, the architecture. Why do we need such a big uh f uh window? Why do we need a small door, a red door, a green door, whatever? Everybody's throwing something in but we do not know if it's really necessary. It's it's just a singular use case or is it really something everybody can uh
profit from. Um so this is uh what I would wish. So I don't expect an answer but everybody should think about it. Where is really the use case for more than just one single um project or yeah initiative. I I guess it's also a little bit the the tradeoff of um hindering companies from open source or yeah enable them. So if you give them also the I
would say the platform to bring their stuff in then the the first hurdle is overcome. The next step is then which project fits together which with the project or how to to build and the next uh thing. So for um to to to use your um pictures at the beginning, you never know if you need small or big windows and therefore it's good to have them all
and then let's see what is really needed to build the house. So >> to answer your question >> design by design. So that was a decision at the very beginning of the working group that in our principle this we are not maker it means if someone brings to us 10 orchestration projects as you saw this morning uh in the excellent example by Nachi of having a blueprint
that highlights you know the plus and the pro and the cons of the single projects. So because the thing is since there is no planning there is no one side fits all. So depending on your needs you can pick and select and this is the beauty of open source also and if there is one project which is much better than the other you don't know this at
the beginning but it might evolve in something and then depending on who wants to chip in who wants to collaborate. So and um and as it comes to examples of use a good thing to do to look at are the blueprints. Yes, they show you how to interface one project with the other and then there is of course escort which is uh that one starts from an
architecture design and kind of brings the building block to it. So we kind of try all the approaches so to allow as many use cases to be covered as possible but this is by design. It might not be the most efficient way to go but it is the only one to go in open source when you don't have a product to sell. Andi, I liked your overview
on that uh automotive open source collection and thanks for your private initiative to put that in. Uh just what is in there? Is it a list of those artifacts of those tool or did you do as well a comparison of those things amongst them with strengths weaknesses? >> Currently there's no comparison in there. It's just a list of URLs to the GitHub repos. I started also to
bring in the the current license and a short description what is it and it's also a little bit of a clustering um for example yeah in vehicle stuff um then there's also something for for ABS for uh engine control and so on I try to do that but I know it's not complete it will never be complete I guess but it's a starting point and yeah I
would say it could also be a backlog for for the STV projects. If there's a gap, you can have a look in in there and see is this something Yeah. you can you need you can use further questions or something you want to bring in? No, no. Um, so then I will say, "No, this was the wrong one." Still valid. We need you. Help us. Bring in
your ideas. So, for example, like yours. I like that. And thank you so much for being here. And um, yeah, I guess now it's beer time. Thank you uh Andy. That was awesome. Um that concludes our OCA track today. But uh don't miss out on tomorrow, right? We'll start at 11 for about um open source basics for automotive, right? How to be a good open source citizen
in the automotive world. >> Yeah. Then you see me again. Yeah. So >> we can stay here. Stay here. >> Very good. So good. Uh have a nice evening. Um there is a lot of things happening out there and see you tomorrow again. So you
