 Welcome to everyone! I think it's time to start. As far as I know, there's not much space available for afterwards, so we should just get up on speed and I know that people will come in a little bit more, but I think it's a way to already settle in. Okay, so today I would try to explain to you the idea that we have and actually what is also going on. It's a lot about not what is Bosch or Bosch Software Innovations is doing. I'm an employee of Bosch Software Innovations, so that is a subsidiary of Bosch, but it's not a big Bosch. We always have to emphasize it in this context with a C. Okay, so let's just take one step back. What is ahead of us? In us, I mean primarily the industrial manufacturers. So software is becoming more and more important, especially now also for all the industrial players. And for example, if you're looking at cars, 90% of the automotive innovations is considered to be part or taking place in the software today. So that is changing a lot of things. It's not saying that like all the other stuff is no longer important. That's still really important. You can see it, everyone who is trying to manufacture things with Six Sigma, meaning you're trying to produce one million pieces and you only have a failure rate of with one digit. And like putting a sticker the wrong way on the piece is already a failure. So you can imagine how much that takes to do something like that, and Bosch is doing that a lot. But still, software is becoming a really, really crucial factor in everything, and that is like what we are facing in the anti-industry. And so it's no wonder that actually also General Electric CEO said every industrial company will become a software company, and that is a big change. So the interesting thing is, and that is what Mike Milenko is saying that, that every software company is an open source company. Why that? Because actually when you're looking at it, how many of the components, the stack that is in the software, more and more, especially in the internet stuff, the percentage of open source software in there is increasing more and more. Meaning at the end of the day, there are so many products, so many things that are out there that are used one way or the other, where you have 90%, 99% open source software in it. So that's not the case, for example, with every embedded stuff. That's not the case in every domain or industry at the moment. But you can see looking at the numbers that actually it's going this way. So there are various reasons for that, and it's always different. But for example, if you're just looking at this investigation of the price of software, from 1980 to today, it is more or less falling down to something like 0.7% of the price at that time. So if you're looking at what actually is going on, it's no longer the software that is sold that much, it's more the functionality and services around it. More and more, everything is becoming something like a service. Websites, apps, cloud services. So it's moving away from the actual one single component in an ecosystem or something that you are selling, and then you leave it to the customer to do all the integration and everything and keeping it up to date and whatever, even like trying to sell the next version of it a month later or a year later. It's more or less moving towards that you are selling the functionality. Or for example, in our case, a lot of the hardware that's coming with it. So another important factor is that things are moving from value chains to ecosystems, meaning that you have a huge complex setup. It doesn't work, for example, if you're trying to sell an app in an app store that is not working, that has not a significant market share, your app could be the best one. You cannot sell it. So that's a difference like, for example, to value chains. So you need to be in the right ecosystem, and the right ecosystem needs to have a big market share, otherwise the entire ecosystem does not work. So the dependencies are much more complex, and you have to make sure that everything is working together and fine in a way that your business is prospering. There's one problem with these ecosystems. You can watch that in IT again and again over and over. Typically, it's the big fish already eating the small fish. And the small fish have a big problem because at the end of the day, you just have, for example, one, two, three, or four solutions, ecosystems, something around a certain technology. So you don't have 100 different app stores for smart phones. You just have a few ones. And these view are more or less the thing that you have to live with and you have to live into this ecosystem. And so what can small fish do to make sure that actually for them it's working, that they are having good business? They can swarm. And they can build a swarm and they can go for, like, trying to do the same thing as the big fish. And that's actually what Simramji was talking about was a positive sum game. Just gather together in a community and just offer that. So have this technology in particular and try to make that prosper. Instead of trying to make your own technology individually prosper, what is exactly then a small fish? And the small fishes are eaten in the ecosystem business. They're eaten by the big fish. So that is also true, for example, for IoT. And it's interesting how many developers are there to be considered in the year 2020. In 2020, we see the forecast is that there will be 4.5 million developers in the IoT business. So when you want to have something like a platform or like a prospering ecosystem around IoT, you need to convince these 4.5 million developers that this is the right technology or at least a big portion of that. If you're not able to convince at least something, let's say, like 25% of them, it's very likely that your technology will not survive the consolidation of the entire IoT ecosystem fight. So it is essential to build up an ecosystem around your technology. And this ecosystem around your technology, that is what you have in the IoT industry with software. You have software, the software is building an ecosystem. The ecosystem is that what is like then afterwards making sure or like it's the crucial factor if you're actually making good business or not. And that's the reason why Open Source is a really interesting and good way to introduce the technology into the market with partners. As Cloud Foundry, it's the same idea. Cloud Foundry is providing a technology. This technology is widely used, building an ecosystem, and so you can stand up against vendors that are actually providing one proper IoT technology so where you have something like a vendor looking. So the community is actually the capacity. The more people, the more companies you can gather behind your technology, the better it is for your technology, for your ecosystem and then also for your business. So what does that actually mean for IoT? In IoT we have the situation that you have a lot of different use cases. For example, this is like one of them. It's 600 euro machine for like on building construction sets lying around and what happens often they're just getting lost. So this is the situation you want to know where are your tools and that's like track my tool, that's like a typical IoT use case for Bosch. Or you have like something totally different. You have like a garage and you want to know where are the empty parking, where are the available parking spaces. So that's like another thing, you can drive around everywhere or you can just have like a device that is just saying this parking spot is empty. And I could just go on like that for on and on because Bosch is actually producing one million devices per day. So we have thousands of these use cases. I'm not saying that we have IoT services for all these devices at the moment but the customer demand is going more and more towards this direction. So this is a crucial thing for us. We have to think about what devices need to be connected in what way and what services will the customers want to have in the future. And of course that means we also have to think about how to build this. And so the current situation today is typically that you have one stack for one solution. Meaning you have one device, you're building one stack for one solution. And so currently Bosch is much more in like all the industrial manufacturers. It's not only Bosch in the business of producing these devices but the future is going also into these services. So between the services and the devices there's something what we need. We don't care really about that much but we just need it and that's the middleware. And it's really hard to always like for a million devices per day for hundreds of different use cases. It's really hard to build it always from the stretch. So the idea is of course to have something like a generic platform in the middle that just speeds up the development of every kind of solution. The other thing is it of course would be so great if we could just combine these different devices and the solutions so we can have different category of devices in one IoT service. So that's the other idea behind it. So and at the end of the day it would be really great if people would just say oh we put everything out of our house that is not coming from one single vendor, like Bosch for example but typically what happens is they have a lot of different vendors so it also needs to be cross vendor. So you need to have in the future cross render and cross domain platforms for IoT. That is where everyone wants to go actually. It's not only the idea of us it's like just out there in the air. Everyone is thinking about this generic IoT platform that everyone can use every device can be connected and everyone can write applications on top of that IoT solution. That is an obvious solution for this problem. So everyone wants to get in there like IoT framework vendors, cloud providers, device manufacturers, connectivity providers, internet platform providers, IoT startups, system integrators, IoT solution providers and I'm sure I forgot other categories of companies that would like to go there. So there are actually two possible ways to do that. One is a cloud third-party IoT platform that is more or less than this of course vendor of this platform is as every company trying to maximize their profits which would be in a way not that good for the people that are trying to sell devices and they're trying to offer services on top of that because they would be in a bad vendor lock-in because that would mean that this platform will just rule them and if they want to be in that ecosystem business they have to accept whatever the conditions are from this vendor. So that's for industrial manufacturers not a good idea, that's not what they want because they are actually then harmed on both ways because they would like to sell their devices and they would like to sell the services on top of that and this platform in the middle would harm their business a lot and they also would have just little control and influence. So that's actually the idea just do it the other way around. Make sure there is an open technology and open source technology if possible that is just providing this middleware and then there will be providers of that technology that operate such a platform and you can choose between the different vendors or like service providers of this platform and then like the different companies in their industrial manufacturers or service providers they can actually have influence because they can influence this platform technology and they can also maximize their profit. What would mean is a lot of competition between all layers, between everyone that's also good for the young customers. So more or less that's just summing up exactly this thing if you go through it and think through it you will actually come out that what is needed is an open platform with open source and so it's really great that there is cloud foundry out there and we will come to the next part later. So the idea is more or less something like that you will have a pass layer that's an open IT cloud platform and our vision at least that's what we came up is cloud foundry is like the cloud part and then you need all the base cloud services and on top of that the IoT, the generic IoT stuff that's Eclipse IoT because that's what we see at the moment the most promising approach. It's a really nice idea and there are a lot of like 26 projects so I will tell a little bit more about Eclipse IoT. So there is one thing that is really crucial with all these devices is you need to connect them and connecting them is a huge problem because there are so many different protocols with all the different devices. For example you have low power devices you have gateways, you have wind turbines you have so many different kinds of devices industrial devices and all these things like companies that produce a lot of different kinds of devices they have to deal with all these different protocols. So we were actually looking okay how many technologies do you need to master communication with all these things connecting it to IoT and the answer we came up is it would be real gate great if this would just be one and that's Eclipse Honol. That's actually one of the newest Eclipse IoT projects that started and the idea behind this is actually have one technology where you have a lot of plugins for all the different protocols but everything that is like shared between them just working together. So you have for example the telemetric data that is going one direction you also have the command and control flow the other direction but you also have situations where it's really necessary that you get one message through typically with telemetric data it's not that important you know if you lose one temperature information out of one million in an hour who cares but if you lose the emergency call of a crashing car that's bad. So that's one of the reasons why you cannot just have one silver bullet that is fitting in all with the protocols and with the situations so this technology needs to actually take care of a lot of different things. It's also this asymmetric information flow the telemetric information is a lot the command and control flow is just a little so it's not a simple technology that needs to take care of all these things and that's just one example for all the problems that are coming with IoT and if you're trying to build it up in a bigger platform so the goal actually is to have ready to deploy Microsoft's based IoT cloud platform and just to give you an example this is not like the final architecture or anything like it everything is just moving but this is just an idea of how this could be so this is just like covering the three use cases and with the different you already see like you can have there a mobile phone or you can have like a gateway or you can connect the device directly and then you need the different protocols for example you have MQTT as a protocol or lesion which is light with M2M or the REST interface all of that is possible and then afterwards you have these plugins into Hono and then Hono is in the back end there's like the application the actual application or Eclipse Hockbit which is a software updating service that's also another Eclipse project or TrackMyTool application so there are a lot of as I said there are like I think at the moment 26 Eclipse IoT projects out there this is just like a list of them like a short list of them so we are actually we started several of them and we are participating in some others and there's also like something new coming up that's Eclipse Kapoor the interesting thing about that is that's coming from your tech they open source the entire IoT cloud back end it's not based on microservices as far as I know so the code base has just been released it's not microservices based at the moment and it is also not like it has not integrated all the other parts of the Eclipse IoT ecosystem but they are working on it and so it's really interesting because that is the first back end complete back end that you can just use for having your IoT solution and it will be really really interesting exciting what is coming out of the all less so at the moment there's a lot going on at Eclipse IoT about that stuff and it's really great that we can be part of it so this is just a list of the members at the moment but it's continuously growing and so it's for example yeah I'm not sure if they are already in here General Electric I think has now also joined here I will come back to that later so now the second part of the solution I just talked now about the picture up there the Eclipse IoT layer that is the generic IoT layer but of course this needs to be today in the cloud infrastructure and for that I have my expert with me Wolf Weber also been working with us and he can tell you a little bit more about that because actually my knowledge is not that deep into that and I think for you guys here that are also into the technology you need someone who knows what he's talking about thanks for the introduction the very high one so my name is Wolf I work as a lead architect in Bosch for our cloud platform sorry and I would just like to take you through some of the challenges that we are facing on this platform which was mainly actually based on cloud foundry and to basically give you an idea of the challenges that we are facing within the IoT scope so don't get me wrong on this side I don't want to do any sort of blaming in the direction of the cloud foundry foundation I actually want to show you that they are the foundation is working on solving these challenges and it's actually more of a shout out to us to get involved into this to solve these challenges together so that we can all profit from these solutions together basically that's what I want to bring across here so the first thing that we are challenged with is as Stefan was saying is the protocol side so we've sort of got to, if you put it into a picture a massive wave of protocols coming in our direction of the platform with all these devices wanting to talk to us and we kind of sitting there and say whoa, where is this going and that's what we are seeing within the IoT use case is that it's not specific protocols that are coming that are, so we're talking really protocols on TCP based and UDP based side so we're not even talking about specifics like MQTT or co-app as examples so there's a whole bunch of proprietary stuff also coming in our direction so we need answers for these sort of challenges and so if we look at that one of the answers that the foundation is giving to us in this challenge is the TCP router which has come out and which is a very nice basic and generic implementation allowing TCP based traffic into the elastic run time so that we can start looking at TCP traffic and getting this terminated and have applications take up that traffic but on the other side we're still missing the UDP support which is also a very important support for us that we need to challenge and take up this challenge I was having a chat to a guy from the company GrapeUp who will be giving a talk I think this afternoon sometime that will be talking about an implementation that they've done for TCP routing as well and UDP routing so I'm very interested in that talk so also as information for everybody of you to maybe join up in that talk as well and see what they're doing to solve this UDP problem so if we move on to the next challenge that we're facing which we actually see in combination with the protocol support is the container networking so if we're looking into Cloud Foundry we have a container level which doesn't allow us to actually build up clusters within run times that we use for our applications so one of the trending run times that we are seeing is VertX for example which is being used a lot in the IoT use cases that are running on our platform and in this sense so I know that when we're talking about elastic run time it's more going into the fact of world 12 factor apps no state get it all out of there but if we now that's what I'm saying it's in combination with the first challenge look at long living TCP connections for example where a lot of our IoT use cases are trying to get into this bi-directional communication with devices which are a lot based on long living connections we have for example in the use case of a VertX multiple instances and nodes of a cluster holding these connections that means we're already in that sense of these connections have state where we need to know which node is holding which connection to which device which brings us also into messages coming in from the back end which want to be delivered to the devices we have to find out where are these connections so that is state that we have to bring in there and this is also where the foundation once again is bringing an answer with the container networking feature which I think is going to be a very very interesting feature and I really really hope it's going to be coming in the new future and I think it's not just going to open up any sort of IoT use cases I think it's also going to open up a whole bunch of other use cases and ideas when you start realizing with the cloud family system when you start being able to build up clusters and container networks to start clustering systems within the elastic runtime so I think that's going to be a very interesting feature coming on that side and last but not least we also have the challenge of the services so this is where we're basically seeing two major points within the IoT segment is that the first is the quality of the service because what we are seeing increasingly with our IoT use cases there's not much time from pilot to production to evolve a service to be able to carry the load and reach that production grade that's required for all of these millions of devices that everybody's talking about and when such an IoT use cases started they kind of explode so there's the massive load simultaneously coming onto the system as this thing goes live and that's why these services actually when they are established and brought up and offered they have to have a grade A production readiness to be able to carry that sort of grade of production and a load that's coming and the other factor is the quantity or the variety of services so if we look at the elastic runtime the apps that are running there are very reliant on the backing services that we require and what we are seeing in the IoT use cases is that certain services are actually for certain use cases but for others they don't so we've gone down the road of misusing services it's not fun so we need a more variety also on the service offering for such use cases so the one says the one service fits for me better for my use case but for the other one says another service fits better so the variety is actually important so this is where also the foundation on that side is bringing in the answer of BOSH we call it BOSH because of our company name that brings in confusion a lot of the times and on this side we see a great platform for automating such rollouts and management of such services or products that bring these services and the actual uniqueness on this side is the abstraction to the infrastructure so you can independently create these so called BOSH releases that describe your rollout packages independently to what infrastructure you are running on so this is definitely an answer for us but we basically would love to see much more BOSH releases to give us the variety of the services coming in there and most of all the production grade that these really bring quality of grade A production systems and with that I would like to close with a note from Colin Humphries the production grade doesn't have to be able to survive through nuclear wars so we can always keep calm and just make sure that we bring the variety with that I'll give you back to Stefan Thank you so and the third part of the solution is actually no one can do it or IOT alone so as I already like said at several points BOSH has decided to participate largely into the open source ecosystem so the two we've chosen is Eclipse IOT and so you can see like we started our first own project for the company like watch this big step sometimes in 2014 that was Vorto Eclipse Vorto then we came up with the next one Eclipse Hawk Bit and because we realized this is an important ecosystem for us we became a strategic member of the Eclipse Foundation another project we started then is Eclipse Hono we started that together with Red Hat so it's also an interesting step for us to start collaborating on a large scale into these ecosystems so we started with Red Hat which is great and then we know there's another project we started just recently Eclipse Unite and this will not be the last one for sure and just this week we announced that we've become a member of the Cloud Foundry Foundation for the moment just a silver member but who knows in the future and it's also great that we were able to partner or like have a partnership with GE because it seems to be that we have a lot of overlap in the interest and we see it at least in some parts the same way and so it's really great that we were able to establish this partnership with GE in the Eclipse IOT ecosystem so they also brought their first Eclipse project there ACS which we would like to participate as well in and it's really interesting what is going on there and it's GE with Predix and Bosch with Bosch IOT Cloud we both have like a cloud installation that is based on Cloud Foundry and in the future hopefully more on both sides on Eclipse IOT but that is not all of it we hope of course as I said you cannot do it alone and like everyone who is interested in that we really hope that you will join these approaches because you are here I think you are already like inside the Cloud Foundry ecosystem so that's great but if you want to get into the IOT stuff in the long run I hope something like Cloud Foundry will establish and for me it looks like at the moment that this will be Eclipse IOT so I hope you will join there as well and so what can you actually do in detail Eclipse IOT become Eclipse member if you are not already most of the big companies are if you are not for the smaller companies it's not that expensive join the working group then and adopt and contribute to Eclipse IOT components and let's hope that we will be soon be able to establish a complete network and Cloud Foundry join the Cloud Foundry IOT special interest group I'm the co-chair there and it would be really great if you can participate just to making like this IOT use cases more aware for the Cloud Foundry foundation so we would get all the problems solved that are in the Cloud Foundry layer contribute to the Cloud Foundry core to result the IOT issues Rolf had just pointed out a few of them there will be more but I think these are the major showstoppers at the moment contribute to the CF integration the Cloud Foundry integration of the Eclipse IOT stuff so we would need to have a way to easily deploy that and also contribute to the Cloud Foundry integration of relevant base services that are needed for that so yeah let's do this together I hope we will be successful thanks for everyone for listening if you have any questions I think we are over time already but you can just come up and ask us thanks a lot