Supercharge your software-driven transformation

Good afternoon, everybody. My name is Klaus. I'm Account Executive at BMW. And during the next 20 minutes, I'm going to tell you what keeps me at night during the nights. And it's not, it's not just here that we are in Las Vegas or any jet lag. But what really drives me is the business driven value for the end customer by the cloud. So business driven value for the end customer by the cloud. How about you? What keeps you going?

What keeps me going? Of course, our cars from BMW, they are very nice and I will show you some of them just in a few minutes.

Hello. Also from my side, my name is Your Graves. I'm with BMW and I'm a Cloud Solution Architect. I'm really excited to be here at the Rein event in Las Vegas in front of this great audience. Thank you for joining us today in our session. I can imagine that a lot of people here in the room dream of having an automated driving car and of course, this will be a BMW. So let us give you a glimpse today on how we work on making your dream come true today at BMW, we already have great premium cars like the Five Series which you can see here on the left side, your left side, the Five Series actually has already advanced level two plus driver assisted functions like the hands of option up to 100 and 30 kp. You can use this on the highway, for example. And we also have the Seven Series which you can see on the right side. And the Seven Series is also here in our expo. You can just walk over there and then you can also have a look at the Seven Series is the latest version. And we just announced this month that we will also have level three functions for our Seven Series available at the beginning of next year. So level three really meaning that not the driver has the responsibility, but the car takes over the responsibility from the driver fully.

But the BMW, we are actually never satisfied with what we have achieved so far. And therefore BMW's product range which has grown successfully over decades now will be realigned on the basis of the so called classic. And let me present to you the no classic. Unfortunately, we couldn't bring a car up here on stage, but of course, I have it on a slide with the BMW Nola BMW basically reinvents itself and it's a perfect play on words that we are now here at the AWS re invent to show you the significance of this initiative. I want to quote our ceo Oliver sip there who said the lawyer class means nothing less than the future of the brand BMW. This is our totally new generation of cars which will start in 2025. And on the left side here of the screen, you already see the BMW vision no glass as it was presented to the public at the IAA in Munich in September. This year, the BMW no classier combines and will be characterized by three key aspects which would dominate the mobility of tomorrow. Those are electrification, digitalization and circularity. It will have a completely redefined it and software architecture.

So some of you might say, ok, what is why on the slide in here, how do they fit into the picture and actually they fit in perfectly and are very essential part because to develop this royal class bmw has partnered with qualcomm, a top of the class tech company they bring in the system on the chip. This is the snap dragon, right? In our e as well as deep knowledge in the field of computer vision and computer vision software. And together with qualcomm, we jointly develop the safety functions as well as level two plus and level three features. And this common, this common development, why can even take and bring to the market as a product again for the validation and verification of all of those safety functions and the whole system we need to collect a huge amount of real world data. We need to store this process, the data and simulate it for that. We actually build up our platform and we will give you some insights about this in just a few minutes and maybe some of you have already spotted there on the on the bottom left. There's also cap gemini. Gemini is also one of our partners. They are actually operating, maintaining and further developing the platform we have built up so far.

So coming now to simulation repossessing, those are very two very essential parts in our whole two chain because we need to prove for all our functions that they are more reliable than a human rival. And as you can see here, a humid driver or is there's an accident with a fatal accident with a humid driver every 780 billion kilometers. So our systems need to prove this and we cannot do this on a classical way just driving over and over again and trying out the functions on the road. So we need to do this with a clever combination of simulation and repossessing where we use the data we have collected and then feed it into the system.

Now, I want to show you a little video about how we do this process. End to end at the beginning, you will see a test drive where we collect the data with a car and then this data gets ingested via copy stations into the cloud. Then the data gets visualized and the function developers can search for specific scenarios they are interested in and then they validate the scenarios with the repossessing and later on also with the with the simulation. But let's get started the car.

Uh can you understand me? Yes. Yeah, get a card. The guys, they are collecting the data. This is as i said, the the copy station where they just and they collect the data and it gets uploaded to the cloud. As you can see in here, this is one tool we have where the function developer then can really search for the scene he or she is interested in and then validate this. And now our repossessing comes into the play for the repossessing. You can see there are different object being detected and actually on the left side, you will see a little arrow which gets detected here, it was shortly seen and then we need to adapt the functions based on this. You can also see that we do some labeling of the and object segmentation of data we collect via cameras and and here in the last part of the video, we don't do our simulation. This is actually our automatic emergency brake and this gets simulated also in real life. As you can see they are from different angles. The car behind is driving to the other one and the car needs to break automatically.

So i hand it over to klaus now. But you want me to do this first. Ok, then. Yeah, our requirements when building up this platform was actually, as i said, we have a very extensive need for this validation of verification because we cannot drive all the scenarios with all the functions all the time. So we need to do the repossessing and the simulation. Um this is really the the basis of uh our, our um function approval and we have chosen to build this platform with aws. But also, as i said before, cap gemini is a big partner for us in further developing the platform together with us and we are collecting terabytes of data during one drive and could even our could even collect data up to one petabyte totally. When we have the whole fleet running, we gather thousands of kilometers and we need to even simulate those millions of kilometers during the repossessing and simulation shops. Ok. Thank you.

So what do we do to build a solution out of those requirements? So at the end of the day, we assembled all experts within capgemini experts from ada from data analytics cloud platforms, from engineering from de de ops and also from infrastructure with the aim to really build a large scale data driven cloud platform. And by doing this, we, we did our big data and hpcs technologies and bmw asked of course to stay agile and staying agile means we have a full de ops approach to develop the platform and to develop also the applications um to get this p bytes platform running, we have one certain no one special issue that means the worldwide data collection. So you can't go on test drive and collect all those data, you just said. So so hundreds of terabytes or p byte and just plug in the hotel at the evening and ingest the data. So this doesn't work. So what needs to be done is really you need a worldwide data logistics where you physically get the data to an access point and ingesting the data to the cloud. And considering also all legal constraints you have by data export control. And additionally, of course, you need to to build on, on on really network physical network to get this done. And what is important for gemini is always the innovation and sustainability part. So a future proof solution.

Now on the right side, you see the data logistics coming in. So here all the shipment, the fleet tracking. So you know where your fleet needs to be and where where you get the new cartridges in your car handling and so on and put this data then into the copy stations to the cloud. What you can see here in the middle is that we use the aws infrastructure. And on top of this, we build up the space platform together with aws covering different aspects like the cloud room provisioning network and security, the fs part. And even on top of this, we have a whole data management layer which orchestrates the data and the processing of the data throughout all of the different as we call it realm. So those are different applications but even sit then on top of this. And yes, you can see here, we have to ingest again, we do also some signal extraction of course and feed and into the simulation and repossessing process is also what i showed you before in the video in the end. Now there's also this data portal. This is where the um the developer then can really search for the scene he or she is interested in and then feed this back into the whole processor and really supporting our ad function developer and the platform developers.

We have here 22 de ops areas. So the platform team and the application teams and currently we are running seven parallel de ops teams worldwide, distributed worldwide. And to keep the platform updated, to keep the applications updated and to serve this. And of course, we are very closely collaborating with our partners from aws and one thing i haven't mentioned so far and this is this is hidden over here, the hardware in the loop test benches. So you need hardware in the loop and the hardware in the loop uses power and uses space. So we thought there must be something better. We need to have a better idea how to cope with that and what we came up with a solution to say, why don't we put hardware in wind turbines at the moment, at the very beginning, when we suggested this, everyone asked us, what is wrong with you? What, what, what, what are you doing with the wind turbines? But the more we discussed and find out we had a partner or we have a partner where we can do this. And um just even if you, if you consider there's no no wind and no sun, we are still doing 95% better than if you use the classical data centers with the current electrical power mix we have in germany and one additional point which is important, you don't use any space in frankfurt anymore. You use space somewhere really on the outskirts and the space is there any home? So if we are going two steps back and think, ok, what is the solution? Why is, why is this such important thing also for bmw but also for cup gemini? And this is what we talk, this is intelligent industry really at its core and intelligent industry for cup. Here is the remaining formatted transcript:

gemini means the following. You have the engineering part engineering where you say, ok, you need to collect all the data that means driving around the complete world, getting the data ready, ship the data as we just said and operate the heart and the loop and do the verification validation. You have the complete infrastructure part here with the it part, infrastructure de se of cloud security and the data at scale. And what i said, what keeps me awake during the night is the business value for the end customers. And we don't do this on our own. So that's not just a cup gemini game. We do this, of course with bmw, we do this with qualcomm and we do this with aws really in a partner ecosystem. And last but not least what is important for us in a very sustainable way.

So, thank you very much and I'm open for questions and York and myself will be here for the next couple of minutes to discuss. Thank you.