Carmen DeArdo

[Opening BMW video/music.]

So thank you for being here today. I'm Carmen DeArdo, VSM strategist at Tasktop, and Rene's going to talk a little bit about some of the great IT that's in that car that's performing like that. So Rene, take it away.

Rene Te-Strote

Yeah, I hope you can forgive me, the weakness of my German accent and slightly my tech voice. All the interesting conversations here already are straining a bit my vocal cords.

So first of all, I wanted to show a bit about BMW Group at all, because BMW Group consists of the brands BMW, MINI, and Rolls-Royce. And BMW isn't only a car manufacturer, but also the leading motorcycle manufacturer in the world. We have a total of about 130,000 employees worldwide, and in 2017 we delivered a total of more than 2.5 million cars and more than 180,000 motorcycles.

The group's production network totaled 31 locations in 14 countries, and these comprise 19 BMW Group plants, five joint ventures, four partner plants, and three contract production plants. The same quality, safety, and sustainability standards apply for all these plants throughout the BMW Group production network worldwide. The 19 BMW Group plants comprise 13 automobile and engine plants, two plants for BMW motorcycles, three sites for producing components, press parts, and tools, and one supply center. In 2017, BMW was already producing electrified vehicles at a total of nine locations. So we always are talking about a worldwide network, a lot of people, and a tremendous massive amount of data.

First of all, I want to show that the BMW Group is not only a manufacturer of cars and motorcycles. BMW has a lot of services for individual premium mobility. Our premium car-sharing service already has over a million customers in 13 cities. A growing number of older people, families, business travelers are also using DriveNow. ReachNow service is available in the U.S. and since late 2017 in China, collaborating with a local partner.

The BMW Group is one of Europe's foremost leasing and full-service providers. The financial services segment's fleet management business, under the brand of Alphabet, offers commercial customers financing agreements as well as specific services. At the end of 2017, BMW had a total of 608,000 fleet contracts. And a little funny thing to Alphabet: this is a brand name of BMW, and BMW owns all the internet domains for alphabet.com, the E, whatever. You know what I'm...? Right. Yeah? Okay.

So finding a parking space is a problem for many drivers. ParkNow is the BMW Group's digital parking service. It enables ticket-free, cashless paying via app, both at roadside and in multi-story car parks. In January 2018, BMW acquired ParkMobile LLC, the largest provider of mobile parking services in North America. ParkMobile Group Europe, which also owns ParkNow amongst its brands, has been wholly owned by the BMW Group since April 2016. In Europe and North America, ParkMobile reaches over 20 million customers and offers digital parking solutions in more than 1,000 towns and cities.

In 2015, BMW, Daimler, and Audi jointly acquired Nokia's maps and local services business. HERE's digital maps provide the foundation for the next generation of mobility and location-based services. They form the basis for the new assistance systems right through to fully automated driving.

In the future, in the core business, the three areas of driver assistance systems, autonomous driving, as well as ConnectedDrive and all the services, are precisely the fields that concern us the most. I would like to give you some more insights.

Ten million BMW Group vehicles are already connected via ConnectedDrive. It takes about 25 million test kilometers to get an autonomous vehicle ready for series production. In 2018, our fleet of autonomous BMW 7 Series models will continue to collect data on the roads, mainly in Germany, Israel, and California.

In order to provide you as a customer with services such as online entertainment, each BMW vehicle has its own SIM card to guarantee network independently of your smartphone. All other services you have seen and will see will depend on ConnectedDrive.

Through ChargeNow, one more service, the BMW Group provides easy access to a constantly growing network of public charging stations, which, with more than 130,000 charging points in 29 countries, ChargeNow provides access to the world's largest charging network. Customers can locate the charging stations directly via the ConnectedDrive app, which is integrated into your navigation system in the car, via the ChargeNow app, or the internet page. And of course, the whole thing is not only integrated in the vehicle and the smartphone, but also in the wearable.

All of which means that our vehicles contain a lot of software, which are basically moving computers. The i8, which you can see outside, is controlled with all its functions by more than 100 million lines of code. How we can handle that, more than later, and then I hand over to Carmen.

Carmen DeArdo

Thanks, Rene. So that was the cool part of the talk, and now I get to talk.

So as Rene has laid out, there's a lot of software, there's a lot of need to connect that software. And so, in my dealings with BMW, I'm amazed at how many products, the product model they've already gone to. I think it's over 300 or some product models. So they are very advanced in what Mick talked about yesterday in terms of going from project to product. But for the rest of us, it's easy to look at that car and say, "Well, yeah, that's a product." If you're in an insurance industry, a financial industry, some other industry, what is a product? That's hard.

And we've kind of gone through this journey where we said, "Hey, we've done cool things in IT. We've done Agile, we've done DevOps. We have all these new technologies, Docker, Kubernetes." I gave this at an ITSM conference, so I'd put something about server certified... Oh, I shouldn't have said that. I love that.

But as Mick said, is the business really happy with this? Do they really care? Do they really care about these things like proxy metrics? And if you talk in this conference and you go back to 2014 or even before, what you'll find is Agile's great. Agile's a great place to start. The Nationwide journey yesterday talked about they started Agile 10, 12 years ago. And the code to cloud, you need code to cloud.

However, if you ignore that left-hand side and you can't connect back to the business, and you can't flow everything from the portfolio, and not just flowing it, but actually talking the language that the business understands, I think you end up where Jeffrey Stover talked about. If you're a cost center, well, I want to reduce the cost. If I'm a profit center, I want to increase the investment. And that's one of the powers, I think, of going to the project to product journey.

Because otherwise, this is from John Smart. Are we agile? Yeah, we're agile. There's 14 steps there. One out of 14, we're agile. I don't think so. He told this better than me because he got a lot of laughs in London. Okay, John, I don't know. Shouldn't have used your slide.

But Kevin Fisher talked yesterday about Nationwide and somewhere around 2.5% of their lead time of a story is spent in actual development. So I don't think this is uncommon in terms of having to approach this from a value stream perspective.

So what flows in a value stream? Well, there's five kinds of elements that flow. One is features. And I think this is where you get a stark contrast between projects and products. When you align around projects, it's typically about features. The incentive for project manager is about features. It's about cost.

However, there's defects. And certainly, if you think again about a car model, who wants a car that has a lot of defects in it? You got to address your defects. There's certainly risk. You heard Rene talk about all the autonomous capabilities. You don't think BMW's giving priority to risk? But we all have that risk. We all have the PII data. There's GDPR. We have to address risk, and it has to be a first-class citizen. It can't be a security officer or system owner trying to beg a project manager to upgrade a Struts library. That's not going to work.

It has to be looked at holistically, and then you have to invest in debt. And I asked Rene, I go, "Well, what would you consider debt?" And he talked about the cable harness, and the fact that at some point they realized with all that data that you saw, they had to re-engineer the cable harness. So those are product decisions that you have to look at and take holistically if you're really going to be able to succeed, as Mick said, through the turning point.

So what are those core flow metrics? They're how much value are we delivering, and they have to be in the customer's language. So how much value are you getting for your business? And that's flow velocity.

How much time are you waiting on work? I think what we find is most of the time that we're sitting around, teams are waiting for work. They're either waiting for work to come to them, they're waiting for another team to do something, they're waiting for an environment. That two and a half days from Nationwide shows that you may have lead times of 40, or flow times of 40, 50 days, and a lot of times that work is just waiting around to go to the next step because you're missing something.

How long does it take to get from request to actual release? I mean, that's your flow time. And it's not just code to commit. It includes that creative process from idea all the way through the entire value stream.

And then, this is Dominica's big talk, and she's talking after this. WIP. We know that WIP is a killer. We learned that lesson from The Phoenix Project, or we should have learned that lesson. Because I think I hear people say it, I hear executives say it, but yet, do they actually go to Gemba and see all your WIP and do something about it?

I think Mick had a very compelling story yesterday about what happened at one point at Microsoft when they said, "We're just going to pay down debt for a while, and then that's going to increase our velocity of delivering features." And the happiness of your teams is going to go up. And so are we actually using this information appropriately to actually make sure that we're able to deliver on that value and also deliver in a way that keeps our teams happy?

And then finally, the right capacity, and I think this really hits home from the last slide. Do you actually have a way to talk from a perspective of those four flow items and saying, "Okay, we're going to dedicate this much to risk or this much to defects," maybe at the expense of features for a while because that's what's most important right now. And I think in a product model, it's much easier to do that.

So again, here's the flow metrics. Here's an example of a dashboard which shows all these things to a team. And I really saw the power of this just in showing flow time was very powerful in changing the mindset of folks because it was data. It was data you could show executives. It meant more because this was actually data that we were collecting.

And then on the right side is you relating it to the business in languages that they understand. So if this is an i8, you're talking about the value, you're talking about the cost, you're talking about the quality, and then key, as we heard yesterday in some of the keynotes, is the happiness of the team. The happiness of the team is a good indicator of how well things are going. And if your team isn't happy, they're not going to be productive. If they're not productive, it's going to get you in a spiral.

So just briefly, you've seen the Flow Framework. Nicole talked yesterday about product modeling, and here's an example of what we do with customers like BMW to model their artifacts to make it real. I found it was very important to hang this on the wall and show people this is what I mean when I say we want to visualize the flow of work in our value stream.

And then finally, again, the metrics. The power of these metrics, the power of actually seeing this is not what we hope things are, this is not what we think things are, this is how things are. And they're not just from an IT perspective, but how they relate to the business. So with that, Rene, I'm going to let you talk about how some of this applies to what you're doing from an IT perspective.

Rene Te-Strote

Yeah. Thank you. Based on what Carmen had said and what I told before, let's take a look at the BMW challenges that we have currently.

Today, BMW as a company faces a whole series of challenges. One of the biggest challenges and the basis for everything else is the agile transformation. And in the first place, we will no longer speak of projects, but of products. We no longer start with the old project methods and start a project, but everything is agile and a product in the future.

A very important basis for this is the agile tool chain. We call it ATC. This allows us on a large scale to bring new products to market quickly and adapt them to customer needs. To achieve this, it's not enough to have a few selected departments do something agile. It has to be agile, the whole IT, but also the whole business, and bring them closely together. And by that, I also mean such areas as finance and controlling, HR, purchasing, and so on. The whole company.

To achieve this, and to make it better, we are also renovating entire office complexes so that we can better facilitate collaboration and provide the latest communication technologies for our internal and external teams. The integration of our partners is an extremely important aspect for us.

Then all the activities for the development of autonomous driving have to be integrated into the entire company, because gradually everything has to be integrated into the vehicles of the future and built in large quantities. And in the end, all this must mean that in the future we can and will develop our products much faster and even better than before.

Let's take a quick look where we come from. In the past, we already had the approach to automate the software development process. Even then, Tasktop was the central part, or the central hub, for a tool chain that allowed us to create the most seamless integration possible, from check-in code, through the build process, to test automation. We succeeded very well, but it still was not real agile. It was not a real agile way of working.

We use this kind of tool chain to realize IT projects, but also to drive automation and virtual vehicle development. After all, such a tool chain does not care if it builds, tests, and deploys software or just simulation models. It doesn't matter for that tool chain. This is an output of such a tool chain as it is still today in use. Here is a simulation model for the calculation of vehicle properties, in this case, chassis and tires. It's generated by the already described tool chain, and the simulation film is created for visualization, and all this completely automatically.

Let's talk about the ATC. Without an SDLC, agile software development is impossible. All dev depends on the scalable, integrated, high-performance tool chain. With that kind of tool chain, we started with an Atlassian-based stack to find out what's best for us.

The ATC is constantly adapted and improved, and after we get to new insights, which let us search for new solutions. The ability to run software development on a grand scale with many teams and products over a long period of time is one of the biggest challenges. In the future, it must become clear whether the selected products fulfill all its expectations, whether products have to be added or something has to be eliminated. Currently, we are discussing the use of Octane from Micro Focus as an add-on or replacement for Jira because of the requirements in the large-scale enterprise environment.

A very big challenge is ECU development. ECU is electronic control unit. Everything that is with electricity inside a car or software. Yeah. So ECU development is a very big challenge. E/E for vehicles. Today we are at the point of having to replace the Micro Focus ALM.net instance, which is the largest in the world, in order to be prepared for the future tasks already mentioned. Therefore, we are in the process of replacing ALM with Micro Focus Octane in order to make the next steps and also to make this part of the vehicle development completely agile.

After the replacement of ALM by Octane, the next big tasks are directly for us. So we have to shortly integrate the emerging R&D center in China into the whole process and connect the system landscape and the processes. Here, Tasktop Integration Hub takes over the data supply. Then it's my job to realize and build this R&D center in China. I hope this will work.

Carmen DeArdo

We have confidence in you, Rene.

Rene Te-Strote

Yeah. In the E/E world of the future, we are talking about much larger amount of data than today. Today, we have about 50 million test cases and perhaps 500 terabyte of data in our ALM instance. In the future, we are talking about billions of test cases and 250 megabyte per second generating a single autonomous vehicle. Now you can calculate what thousands of vehicles produce in nonstop operations, 3,600 seconds per hour. Everything has to be managed. For that, again, we are in close contact with Tasktop so that everything works well, because integration has to transport all the data.

I would like to briefly discuss the scenario of integrating the R&D center in China. In China, we are building exactly the same landscape that exists in Germany, including the HPC cluster. HPC cluster is high-performance computer cluster, but a bit smaller than we have in Germany and in Iceland and in Sweden. However, the applications that are being built in China initially contain no data. I think it's clear.

Only by coupling a data hub, which we realize with Tasktop Integration Hub, are the systems in China supplied with the data according to what data is really needed. We call this the need-to-know principle. This goes hand-in-hand with a common software lifecycle of all these applications, which is realized by a common build environment in Munich, placed in Munich. Especially this means in Munich, there is a build environment around Jenkins and several slaves and some more tools, which controls and executes all deployments in China and in Germany, and this is done via a DMZ to China. And now I would love to hand over to Carmen for further insights.

Carmen DeArdo

So thanks, Rene. I want to make sure we get to the video, so I'm kind of looking at the time. So maybe you want to talk about how you think this can be applied in the car lifecycle, and then tell us more about the work that we're doing in the China BBA.

Rene Te-Strote

Okay. So, when we talk about the agile transformation today, or about something else that we want to implement or change in the business, we start by considering what output we expect. What should be the goal? What benefit should the activity have? What should the product be? Then we try to find out which bottlenecks we have and how to handle them. Only then we think about what the solution should look like. And we do not look for existing solutions and see if we have a suitable problem for that.

This search and consideration for the best solution for a given task can be found in the Flow Framework perfectly, because I see very quickly whether something works or not. In the many discussions between people from BMW and Tasktop, we found out very fast. Yeah, in the discussions between the people of both companies, but also about Plant Leipzig, how it was built and how it works today to build agile and extremely efficient vehicles, it has become clear that what we call Flow Framework today, that's exactly that, what contributes a perfect companion to good product development for us at the moment.

That's why BMW and Tasktop launched a POC to find out how we can best measure the health of our products. And this requires that I have a monitoring where I can indicate, depending on the needs, which flow distribution I have related on my features, defects, whatever, depending on what I'm currently interested in.

And regarding the car lifecycle, in the future our partners will have to be integrated into our vehicle development cycle even faster in order to simply become much faster in total, and with the ever-increasing complexity of our products, and at least a constant quality. And to achieve this, we need to be able to integrate different systems, heterogeneous systems, and to quickly see if everything is going as intended.

Then regarding China, I have now a video for you. Okay, first of all, let's take a look at the video.

Video

Virtual reality is a technology which enables immersive experience for users, so you enter a new virtual world and can experience different scenarios. And right now, the focus is more on 3D geometry, and in future we will also enter the field of functionality, car functionality, like numerical simulation inside virtual reality.

It's a paradox that we view and analyze our 3D simulation models on 2D screens. And we had this idea to look at full vehicle crash in virtual reality because it's quite tricky to understand the kinematics and the damage to the vehicle at the first glance. Normally, to analyze a 3D model, you will need to perform an analysis in your pre-processor, post-processor, and carry out a series of cross-sectional cuts. And subsequently, you will create a PowerPoint presentation in which you'll pass your findings to your team. This takes a lot of time. In virtual reality, we can perform exactly the same actions faster and simpler.

For the virtual reality crash simulation demonstration, we use BMW 5 Series crash data, off-the-shelf pre- and post-processor, VR Sim software, and typical VR hardware headset. To improve the quality of the demonstration, we rendered the materials such as plastic, glass, and steel.

In the end, by leveraging virtual reality technology, we are able to produce a fully immersive experience of a crashed vehicle. You can literally tour the car inside and outside and see the crash for yourself.

Also, a collaboration aspect of virtual reality plays a big role in distributed teams, especially now when BMW becomes an international and global company. One of the biggest advantages of virtual reality is the independence from your physical environment. Especially in the case of BMW, where we have different global locations, one design team in LA, engineers in Beijing, or other main engineers inside BMW at Munich, they all are able to enter one virtual room and discuss, interact with a car model in real time together with the other colleagues. And one advantage is also that you can point to some car parts and the other colleagues will see it in the virtual world, which is not possible with the 2D techniques right now.

Rene Te-Strote

So, by the way, you are the first audience outside of BMW that had ever seen that video. So, in this video, you can see what the future collaboration looks like. We are currently rehearsing what you have seen for use in China in order to optimize worldwide collaboration with suppliers in our deployment center in Munich. Also, VR is a key technology. The whole thing only works with the right data supply. Without the right data, it's nothing.

And we always talk about deployment or development in a large-scale enterprise environment with a huge amount of data. So, I think we have to hurry up a bit.

Carmen DeArdo

Yeah, I apologize.

So, Nicole talked about this tomorrow. Stop by our booth to actually figure out how to connect, model your products. Certainly, we could use help. I think the community could in how you're going on this journey, defining products. How are we doing this so that as a community we can move forward together? And we'll leave you with this little ending video, and thank you.

Closing

Thank you for your attention.

Thank you.