 Okay, hello everyone. Thank you for coming to this talk. So I would like to talk about Eclipse IoT as well. So we heard already a little bit about it. I want to show you all the things in there from a different angle. I'm from Bosch and so we have like... Yeah, IoT is a big problem for us or a big opportunity. But it means that you have to... Because we are producing a lot of devices. So for us it's a natural thing that we care about it. In particular, we are really interested in the automotive area. And so it was like a... A real car was a little bit too heavy to bring it here. So we also reduce it so like something could be like done also by... Yeah, like everyone. So there is a rover that is done from Eclipse. This little rover is just like this size. Actually it was also too lazy to bring it here. Or in the plane it was a little bit of a problem. And with this rover you can actually drive forward, backward and things like that. And connecting this thing to a cloud is like the BSC idea of Eclipse Cooksa. And you have like three parts. That's the cloud platform that you need. So Benjamin has already talked about things that are available here. Most of these are Eclipse IT components. And then you need something in the car. Typically Java is not that liked in the car. So that's the reason why we use here automotive great dealings. Which is running without Java. And then of course you need some development environment to do some stuff. And for that we used Eclipse J, was Eclipse. And so what we will see in the demo. I've just put together a small video. You will see first the rover. There's a camera drive around it. Then you will see how to build something in Eclipse J. And there it's about just a cruise control. Meaning that like a car is just making sure that the distance to the front car is hopefully always the same. So when there's something that is faster than the car can move. It will try to make it like again the same distance. So that will be deployed on the car. And then we will see okay that the adjusting cruise control does work. Afterwards we will just see okay. And then you can also steer it over the cloud. So you can do for it backwards and things like that. What you have to imagine is actually that is like a prototype to play with. But it's much more. It's about at the end what like the company's need is that they are able to connect a million cars. A million devices or at least a few thousand. And like the entire stack here for company like ours. It's about okay how to manage thousands and millions of devices. And so everything must scale. On the other hand it must be easy to learn easy to get familiar with. So it must be good for like everyone. So easy to start with but still capable of doing this huge things. And it's also you can imagine that like a lot of things like the number of messages coming into something like that is really big. So when you have like thousands and millions of devices connected. It's really like a lot of a lot of traffic going in. And then you also need to have maybe a lot of different applications using the different things from the background. So you have one device and you have maybe 50 applications behind it using it. Or you have like 10 different devices and one application wants to have like data and control over these 10 devices. So you have a lot of problems that you don't have in an easy setup. And in the future it looks like that I will talk a little bit more about it later on. That we will have more and more cross-vendor, cross-domain use cases in IoT. Which makes it all a little bit more complicated. And cross-domain, cross-vendor also means to get that work everyone needs to collaborate, right? And that means where it's like a good model to do that. Of course open source. So that's the reason why here Eclipse IoT is the thing. So I will show the small demo just to illustrate what you can do. And then we will see a little bit more about it. Hopefully you can see, yeah, that's good. So that is Eclipse J. Oh, sorry, that was in the middle of it. So that's the rover. You can see cook size on it. And there are cameras in the front and in the back. And there are some more sensors. Actually we are thinking about even adding some more stuff. And that's Eclipse J. So the development in like getting the stuff then on the rover and to write it is there. So we will see that we find the adjusted cruise control variable in there. There it is. Now we can do the configuration. Of course you can do now lots of different things with all that. So it's just one example setup. But it illustrates the idea. You can deploy and run that. Of course I will talk a little bit more about it. That is when you have one car. So that's the adjusted cruise control. Like the rover always tries to get the distance again. And now you see like, okay, now it's the feed from the telemetry. You can see like here you can also steer it. Like here's the steering controls. Here's the telemetry data. And here's like the event console. And here you can see what the rover is doing actually. So now you can also see the camera feed. What is important to understand is, for example, you can also use this now for trying to implement something with automated driving. That is one of the areas that would be really nice. If the vehicle is moving really slow, you can actually do everything in the cloud. So you just connect the data to the cloud. But not for regular cars. So they are already too fast. But, you know, if you have, for example, you can imagine being in the station. And there's like a helping robot that is just driving around telling the people, hey, do you want me to tell you where to go and things like that, or maybe like carrying the luggage or something like that. So these things, they could be done just feeding everything to the cloud and they can do the calculations and everything there. But if it's becoming more and more complicated, so automated cars still need it. Let's go back to the presentation. Automated cars still need it. They need a lot of data from the cloud. They need to have exact cards and also like events happening, for example, a crash and things like that. They want to know as much as they can in advance what is like in front of them. Typically, even there, the typical automated car is connected. So that is like the setup. So IT platforms, as I said, doing that on your own, like saying, OK, it's just us. Normally that's not a good idea because cross domain and cross render are the things. And the interesting things here coming up are actually the service-based products that are in the IT area now possible, especially for all the industrial manufacturers. Because for them, it's typically it was just selling hardware elements. But now it means the things are becoming connected, so there are other things that are like natural businesses coming up. And so they want to have something, they want to go into that, they want to use all the knowledge that they have in the hardware area. In the software area, you can see that it's more and more not about the typical value chain, what they're used to in the hardware. But you have more and more business ecosystems. And with business ecosystems, it's typically you need to know, OK, the ecosystem must work and it must like all the people there are actually prospering and then you can sell your products. But if the ecosystem is not doing well, no matter how good your product is, you cannot sell it. And in the IT area, that's what's predicted in 2020 that there will be 4.5 million developers. So it's about that these developers think that's the best technology. It's not about that one manager somewhere up is saying, yeah, that's a good technology. It's about that these developers feel comfortable with this stuff. And they say, yeah, that's good and it's easy to use and it's like that's what we need. We also have this challenge and it's not only us, it's all the industrial manufacturers that have a variety of products that we need to have a lot of different use cases. You can see some here, like even oyster farms or like lawn mowers or like tools that are connected. And like for us, it's really hard to do something like that. Each IT solution one stack and we implement it again and again. And it's very obvious in a few years like nearly every bigger device will be connected somewhere, at least have the feature to connect it. And for a company like ours, we have the challenge that we produce one million device models per day. So it's not the pieces. So it's just the models, the different types. It's not likely that all of them will connect because it's also something like for small devices for the car, but many of them will be connected. And connecting them, the only way to do that properly is that you have something like a generic IoT cloud in the middle. And for that reason, as I explained earlier, if you want to have it cross domain, cross vendor, you need to have an open platform. That's the reason why we were looking around for, okay, what is the most promising open source IoT platform and we came to the conclusion that's Eclipse IoT. If everyone of you thinks like there's something else, there's more promising, please let us know and we would like to discuss that. So in there, there are several components. Benjamin already explained some of them and I will go through them in a little bit more detail. Some of them are in the setup that we've just seen with the demo. So the goal is to have a de facto standard for IoT cloud platforms ready to deploy microservice based. So a typical setup would be something like that. So you have different devices. You connect them somehow through a gateway like a smart phone or directly with a line with wired connection or with another gateway, Kura. And then you need, because you're using different protocols, there might be because with all the different use cases you might have hundreds of different protocols. So you don't want to connect them all differently but you want to have something like protocol adapters and then you have something like that is handling all the traffic. And then there's something in like Hockpit that is doing software updates. We'll talk a little bit more about that and there's Ditto that's a digital twin. So if you have an easy way for interacting with devices, so if they were just like an object somewhere. Okay, Ditto, digital twin, the idea is very simple. You have a physical device and you want to have just a very simple API that is like very similar to all devices no matter what device it is. And so you take something that describes actually how the device interfaces are and that's Eclipse Vorto. And then you have like a very structured way to interact with no matter which device you have and you can then interact with it and also the access management and the authorization, all these things can be done in a general way as well. So you have the description of all the different interfaces and then you can work on them. The other thing what I just told about is Hono another big problem, getting all the connection in. I think, yeah, that's fine. If you need more information about it we can maybe talk about it in the Q&A. Hockpit software update. The problem again is when you need to update like hundreds, thousands, millions of devices it's a totally different thing than just deploying it one-on-one thing. So what we've just seen in the demo is not like the case that is good for one car for one like the car but it's not a good idea to do that with the whole fleet. Typically you want to know, okay, the entire process worked so you're doing just a little proportion of it like 1% and then you see, okay, it worked out. It's good. Then you do the next. They call that campaign management. The other thing is maybe you have the devices all around the globe. You want to do it through a time of low usage so you do that during the night so you have to define time winners for that. And so like there are tons of problems in that and that's actually what Hockpit is about. It's just about the point to transport it on the device what is then like the update process on the device is something else. A big challenge when you're doing IoT solutions is that there's a lot of technology connected to it. And one area is that when you're like going deeper into the specific domains is that you need more and more technologies that need to be somehow aligned with all these generic technologies so one example is here Eclipse Unite. That is for the industry. It's like the idea how to improve the way to interact with the machines producing the stuff. I don't want to go into much detail but you can have a look at it. It's really interesting. So that's for that domain. So we already talked a little bit about Eclipse Cooksa which is the idea of a connected car platform. So that is also going into the domain and you have a lot of domain specific problems there. For example here it's as I said you don't use the Java stack, you use something else then you also have things like how do you do it for example propagating the changes to the different control units in the car and typically it's also you have somehow like an interaction with the infotainment system and then you also have the entire stuff about automated driving which is a really interesting and hot topic at the moment in the automotive industry. And so that is like when you're moving into the other contexts that are maybe not even has nothing to do with the IoT stuff you want to make sure that they are well integrated. So one thing is that is Eclipse Open ADX it's just like the tool chain for automated driving in there one part is already here connectivity based validation it's just that the car is sending back like the situations that they were not like thinking okay something was strange here maybe you should have a look at that. So that's one typical thing but it's the entire tool chain for automated driving is a huge thing there's a lot of artificial intelligence intelligence in there it's there's tons of data going on so one car like has to process in the automated driving area more than a petabyte per month and of course not all of that is somewhere stored or sent but it's like at some point it has to be filtered and then you have to decide what you want to store and then just a really small portion can be sent real time up to a server but as soon as you have it somewhere stored then you use it for artificial intelligence and for checking for simulations and for a lot of all the other things and then you have like again something because then you for example have a car that is driving through virtual reality and you're checking that if the car is behaving the way it should and using there again like connectivity things because of course the car the software then in this virtual car should be as close as possible to that then you have things like ROS robotic operating system the same thing that is also used in the automated driving stuff and from the virtualization to the cloud native stuff that is also something really important that you have to understand because last year I don't know who has seen like my talk last year we were really much interested in cloud foundry as the basis for all these things and it turned out that actually today Kubernetes is the thing that you use so if you're trying to do that all on your own you have to go this evolution by yourself and you have to implement all these stuff and that is a lot of work so it's much better to do it together with the big group and just to give you like an impression you know that is like the cloud native components here's Kubernetes that's just a small piece of it Keeklog is as well something that is really important because the access management we've heard some problems about security earlier here in this deaf room like admin, admin and things like that not a good idea so same thing of course is here with the cloud you have to make sure that everyone can just access the devices and the data and things like that that he is allowed to and for that you need a general solution and Keeklog seems to be a really promising one that is quite popular now in the Eclipse IT Working Group yeah if you of course that is a lot of talking a lot of theory if you're interested in doing that more like in real and playing around with this rover that we've just seen and like working on all that stuff there is a hackathon going on and in February like 21st and 22nd in Berlin it's like there are several ones and it's based on Eclipse IT to a large extent and Open80X as well so the one at least is here for the automated driving it's exactly with the cook-style thing and all these things together so a lot of things to play with and there will also be the people from Formula Student Driver List from the team from T.U. Darmstadt they will also like with their car they have a lot of sensors in there and they will also be part of this hackathon and so for students and for startups the participation is for free for everyone else who is really interested I can probably get your free ticket so then just send me an email and yeah then it's in total there like as you've seen the numbers 700 developers in total and like 30 hours of hacking like 80 things to hack with and in total it's 5 hack challenges so the one with the automated driving things that's just one of them yeah so that is like what I wanted to tell you about so if you have any questions good time up there so I didn't get this second question repeated afterwards first I go with the first question so the first question was why we moved from Cloud Foundry to Cloud Native which is Kubernetes so in a way we could say we didn't really move it just turned out that if you if you try to do Cloud Cloud Foundry if you want to use it as a base for your services for your IIT services like Hono and things like that it's not a good idea because of all the networking stuff in there it is good to use it as a service for Cloud Foundry but then you want to have it on something else that is doing all the scaling so that you have like a lot of workers and you can scale out and things like that and doing that inside Cloud Foundry was not a good idea that's what it turned out but again like if you want to have then a straight forward way to implement something you can use Cloud Foundry but then you provide the service like Hono for the applications in Cloud Foundry so in the second part so the whole architecture is based on putting your services into some type of managed you know, Cloud Foundry like systems also and Microsoft architecture you're going for so you can spin off more and more containers with your particular services is that the kind of general idea of how the system works? Yeah, the idea is that you have it on something like Kubernetes and that you can then scale by like putting up more and more containers exactly and so let's say five years ago people would have thought like different like using something totally different but that's the way to do it today right? More questions? Yeah Yeah How much is virtual computing to open source because you presented the whole ecosystem but I didn't quite get which part you were contributing to so actually like most of the Oh yeah, sorry like the question was how much Bosch is contributing to open source so actually the major project that you've seen here we are like supporting by in total I think 20 developers and we also like the project lead of several of them and we try to now because there are like several companies in the Eclipse IoT Working Group and the challenge for this year is getting all the different pieces all the different projects together in a consistent platform and that is what we are also working on together with all the other companies like Red Hat and EuroTech Yes? What component for storing of yes, sensor data for instance for analytics later is this in these frameworks So we see it getting all the data but what about analytics? Yeah, so the question was what database is used there or how is the data stored in this setup and how is it about analytics so when we started it as a first thing we had like okay it should be we had like one database and that was like part of that and we thought like every customer would store it then and we realized that's short-sighted because there are so many different setups so that's the reason why like Hono is now exactly the idea to use it there's just a published subscribe system so whenever data is coming in you can set up one database that is just like subscribing to it and then you can also set another database to it so everyone can choose from the variety of the corresponding cloud platform to do it so we for example there's like a commercial offering for that that's called the Bosch IoT Suite where like many of these components or services are in and there we have like a databases that we choose for the corresponding customer and say hey what do you need okay maybe your last question yeah have you considered IOTA what part do you think it will play in the ecosystem sorry I didn't get the IOTA IOTA have you considered it do you have any sense of how what role it will play in the ecosystem no not yet but I guess we can talk with Benjamin about it maybe like we can figure it out okay so if you have more questions or if you're interested like in anything just come up to talk about it also my contact details are here on the slides over there like published would be yeah fantastic if you contact us we're always interested to hear what you want to know and what you think thanks