OpenTeams - The Future of Software
E18 podcast thumbnail


E18: Carl-Eric Mols, Former Head of Open Source at Sony Mobile on OSS in the automotive industry

In this episode, I chatted with Carl-Eric Mols, the former head of open source at Sony Mobile. Carl-Eric is currently working for Debricked which offers a new way to manage your open source dependencies.

Carl-Eric published a lot of research about open source in traditional industries such as the automotive industry which forms the bulk of our conversation today.

We also cover topics such as:

  • How attitudes evolved towards open source within Sony Mobile
  • The difference between open source in Europe and the US
  • How traditional industries are getting increasingly connected using open source
  • Why smart cities and the automotive industry will drive the open source movement in Europe in the coming decade
  • And finally, Carl-Eric explains the contribution strategy for automotive companies today

Note: Open Source For Business is produced for the ear and designed to be heard, not read. We strongly encourage you to listen to the audio, which includes emotion and emphasis that’s not on the page. Transcripts are generated using a combination of speech recognition software and human transcribers and may contain errors. Please check the corresponding audio before quoting in print.


Henry: Carl-Eric, thank you so much for joining me on the podcast today.

Carl-Eric: My pleasure. Thank you for having me.

Henry: So you’re currently a business development manager at Debricked, and before that, you were the head of open source at Sony Mobile for 10 years. So, can you walk us through your journey to give us an idea of how you actually got it?

Carl-Eric: Yeah, absolutely. So it all started where I was part of Sony Mobile, at that time, it was called Sony Ericsson, so it was a joint venture between Sony and Ericsson telecom company. I was part of the software strategy team of the CTO office. This goes back to 2007. And we knew already that Linux will make it to the handsets in one way or another.

And there were actually four contenders. We knew about Android before it was released. We were pretty sure that Android will make a hit in the marketplace. It then ended up into my lap to prepare for describing our strategy and how to handle open source because Android was based on Linux and that is open source.

So, I come out with a strategy in the beginning of 2008, at that time we had about four different mobile operating systems. We had our own mobile handsets or feature handset as they were called. We were part of the Symbian operating system together with Nokia, Motorola, and Samsung. And then we also had some windows phones, Microsoft windows phones. And in addition, we have a fourth segment and that was the so-called entry level, the budget phones, affordable phones. We had a specific entry level operating system for the easiest GSM handsets.

My prediction was that open source will make a dent into the mobile handset with Linux. And I predict that there will be about three to four years before we have to really to take into account of open source. Now, I was completely wrong, completely wrong. It took only 14 months from that strategy report until we close down on terminated our own mobile operating system for the mobile and turn our focus into Android. And you asked about three years later, we were working exclusively with Android phones.

So, that was quite a journey. In that process, I get an assignment to be the open source SAR, similar to in US, so I kind of get the similar kind of position. So my assignment was to prepare the company for working with open source in all aspects from both organization aspects, processes, but also to getting legal prepared, and also the business aspects of open source.

And so, I entered the position as head of open source in 2009, and then already then we had started the Android development, and our first handset was released in 2010. And that was the start of the journey. To be honest, it took a couple of years before we really get revved up in open source. I mean, what I come to understand open source is foremost, a culture change.

And this particular traditional industrial company was very protected about it’s intellectual properties. It was kind of a process of relearning and refocused the company in the culture of collaboration. And it took me actually until 2012 before the cultural shift was good enough to make it work more fluently from then.

Henry: Before the culture shift, what were the perspectives of internally towards open source software?

Carl-Eric: Well, curiously enough, there are different kinds of opinions about open source. First, there will be a great opposition from the top management – the executive levels. I knew that the development in general we were positive to open source, obviously because it makes the developers to be empowered and they have more fun and more engagements in a quite different way of doing what they’ve been told to as the old days.

But the topic executives were not a problem, the biggest issue what I found out is the middle management. And if you are to look at it in the perspective that if you’re being in a middle manager, having a teams in a department with maybe 60, 70 people, and you have a responsibility for certain components in an operating system that only Sony Ericsson had, and then suddenly overnight your competency is not needed anymore.

Now you’ve been handled a piece of software that is open source that someone else has developed, and you’re supposed to work with that instead of your legacy code. So, that was actually what I found, surprisingly, that biggest culture resistance came from middle management.

Henry: So they were developing the software and almost like their babies, they didn’t want to let go, even though it was entirely in the interest of company to do so?

Carl-Eric: Yeah, absolutely so. But interesting enough after a while the cultural change, actually hit the dumb too, because then they’re starting to discover that their department, their team was performing great with open source software. And they’re starting to get the sign splattering on that or what their developers were doing.

So quite a few examples of managers that were there from the beginning, it was quite opposing but they cannot really oppose strategy to have to do with what was decided on the executive level. But I’ve seen them turning around much more positive views the last couple of years when they’re starting to see the results. Well, it’s kind of funny sometimes to experience that managers… and I know them being sometimes against open source. Now starting to talk about, yeah, it’s the greatest thing since sliced bread kind of talks.

Henry: And then your next step was then to join Debricked. Can you talk a bit about what led to that change and what you’ve been doing there?

Carl-Eric: That happened much later. I would so about two years ago. I was let go from Sony Mobile due to the company doesn’t perform very well in the marketplace. I had since a couple of years being engaged in research in open source, so I was engaged in the European research program. From that, I wrote a book it’s called Principles for Industrial Open Source, taking much from my experiences from Sony Mobile, of course, but also in that research project, we actually had about 30 companies that we did studies on.

So, we had a couple of case studies from other companies working with open source. So it took all these experience, my own and from the research, put that in the book. And I had the idea because I have encountered similar kind of companies over the years that I would start my own business on doing open source strategy, consultancy services. So the step actually to go to [Adlot], which is a smaller consultancy company which were part of the partner in the research program. I took part in [Adlot], became a consultant, and then soon after I got an assignment as business development manager at Debricked and I’ve been working there since May last year.

Henry: I know you’re currently based in Sweden you’ve been working and living there and you’ve worked a lot with us companies. I’m really curious to know when you look at open source in the US and when you look at it in Europe, what are the key differences that stand out to you?

Carl-Eric: You’ve come to my favorite to talk about. So, what I observed of course is that US development is of course ahead of the rest of the world. And that is very much of the fostering that is made from California and the tech giants and the advancement in IT technologies. My observation being working in Sony Mobile, which, although it was a telecom or mobile telecommunications technology company, it’s still resembles very much traditional industrial company.

We did produce handsets, so we have tangible products that we produced. So I started to realize there is a difference US at large compared to Europe. And that is, in US, you have the tech giants, and this is still the fan companies; Facebook, Amazon, apple, Netflix, Google and Microsoft, but their products very much based on softer products.

Yes, I know that some devices, but mainly there is a software companies, whereas in Europe, you have huge old industrial companies for instance, in the automotive industry. And that is maybe not so very recognized, not by themselves. They actually are heavily engaged in the digital transformation. So, what is the digital transformation? That is software.

And what is software today? It’s open source. So in fact, what is happening now today is that a lot of all these age old industrial European companies are starting to be moving to be more like a software company, but they are still working with some open source setting or have industrial products. And I mentioned the automotive industry, and in fact, I’m studying now the automotive industry where I started off working in Europe and doing questionnaires on standard to probe the automotive industry in how far open source is starting to make it into the automotive industry.

So, there is a difference that is in US is fair dominated by software companies, traditional tech giants. I called European old industrials Titans. Those Titans are realizing more and more that there are becoming the software companies foremost, and I can give you an example. I happen to gather information that today, Volvo, which is one of the famous Swedish car brands, Volvo alone has 8,000 software developers.

And I think that is in par with the rest of engineering in Volvo, if not larger. So Volvo actually has started to realize, oh, they are software company nowadays, at least there’s a lot of software going on our business. And they joined in the Swedish NGO for software, it’s called Swedsoft – it’s about two years, they don’t need it. So that is happening now in Sweden and the rest of Europe that the big industrial starting to realize, Hey, our business has become a software business, so we have to turn around and look into software in better ways. And as I said, today, software is really open source. You cannot do anything really modern software development without taking the account of open source.

Henry: And upon first glance, looking at the automotive industry, people might be a bit taken back when you say that they could be the key drivers of open source in the future. Because historically cars haven’t used much software, but the cars today, the Teslas, the Volvo, like you explained, they have to be connected.

And I know one technology we talked about was 5G technology, and they’re really going to use that to connect the cars and connect the networks and this technology, the foundation of it is open source software. So, can you talk a bit about a pattern that you’ve seen within traditional industries, like the automotive industry, for instance, and explain how they’re getting connected and how they’re actually starting to use open source more and more today?

Carl-Eric: Let’s start off thinking about the car and you were immediately thinking about an engine and petrol [unclear14:44], but also things. But in fact, a car is really a smartphone on wheels. It’s a lot of computing power in a car today. And it’s worse than that because in the more modern handset, you have maybe two CPU’s, one for the general CPU and one maybe for the modem, and you maybe have some additional graphics CPU as well. In our normal car, you have about 90 to 120 different computers call it ECU, but there is computers – microcontrollers in many cases, so they maybe not that big capacity, but nevertheless. So, it’s everything you could expect with a communications network, a lot of computers working together someone is controlling the engine, others are controlling the brakes. So that is already happened.

What is happening now is that the automotive industry is going through a tremendous transformation. First you have the drive of electrical vehicles CV. Then you have to more or less restart with the whole platform, with the whole software platform. With petrol engine anymore, you really have to rewrite the whole software for controlling our now electric motor that is driving the car. And then it continues, all the software tools for developing software are based in open source. There are some examples of software platforms for electrical vehicles vehicle based on open source and open source solutions.

The Porsche, for instance, the electrical Porsche have released the platform as an open source and the Teslas has done similar things. And then it continuous; cars of tomorrow will be very well connected. They have to, for instance, for self-driving cars to enhance the path they’re driving, they’ll be in constant connection with the cloud, where they have some navigation software that guides the cars.

So you need a lot of connectivity, high speed connectivity, 5g, and boom, you’re into open source again, because everything thing about 5g is open source. And then it continues if you look into the machine learning algorithm that is driving the self-driving cars, that is also open source. When you starting to scratch the surface, you’re seeing a lot of open sources starting to getting into closer, close to the car infotainment systems, instrumentation systems, that is already now based on open source. Automotive grade Linux, for instance, it’s a project that’s run by Linux. So, I would say that it’s yes, has entered into the car, but everything around the car is already there.

You already have a lot of open source starting to getting into there. We had an interview with our last Swedish truck company with the head of open source. They haven’t had an open source in that truck company, and he is prediction was that within five years, he doesn’t think it’s going to be a huge change – within 10 years, everything will change.

So he himself thought that it’s going to be slow, but then it’s going to be an avalanche. Within five to 10 years, it’s going to be an avalanche with open source in the automotive sector. And if you thought that the entrance of Android was disruptive for the mobile industry, just wait till you see what open source is going to do to the automotive industry.

Henry: you wrote a research paper in the past titled Scaling Up Software, I believe, is that correct? And in that you predicted the IT in the automotive industry would be the fastest growing sector in open source software, so can you talk a bit about.

Carl-Eric: That is exactly the point I made that if you look into the European industry, the two more important ones that will take an uptake of open source, that is the automotive industry. And as I mentioned, that is already happened or is happening maybe. Just now, it’s maybe not seen as a very large, but it’s going to grow and grow exponentially, but it also has another sector where I think actually Europe has a lead on, and that is smart cities.

That is a lot of things happening in Europe. And that is the idea of smart cities and how to connect buildings and a collection of buildings and data, and then connect everything in the city like traffic management, water management, waste management, everything; you can imagine, those projects already taking part in Europe. Barcelona is well-known to be quite ahead on this.

Stockholm is making these moves in Sweden, and there’s a lot of initiatives taken in a lot of European countries as well, so that was our prediction. Automotive and smart cities, those are the two drivers in Europe.

Henry: It’s really interesting because I know you also did a lot of research in the last few years. Another research paper I was quite interested in that you wrote with the European chapter to-do group was “Why open source software matters to your enterprise?” And I know that’s where you covered a lot about the automotive industry and these traditional old age industries.

And for those listening, I’ll actually leave a link to the research paper in the description. But in that paper, you talked about specifically entertainment systems within cars and how the software they’re using is gravitating towards more open source solutions, so can you touch on that.

Carl-Eric: We gave the example of Polestar 2, and Polestar 2 is the brand of Volvo for the electrical vehicles. So it’s the electrical vehicle of Volvo is Polestar that was the latest make they released. I think it was last fall. I know a little bit about the reasoning, but they’ve decided to include Android car or as they say comes with Google, and then they have inherit of course, open source because it’s is based on open source. And one of the reasons for them making that new is that they’re starting to realize that it’s not only Android that is important, but it’s all the surrounding technologies that are provided by Google, for instance, in voice recognition. And they did their own voice recognition software, but they just understood the years behind what Google already done. And that show us a little bit about the power of both open source development, but also how far you can come to when you starting to collaborate with others.

And obviously, Google has taken the lead here and voice recognition. So then they make the decision that is no chance that we’re going to catch up what Google has already done in voice recognition. And that is making a strong case for us to moving into Android in the cars. Then you start to how to handle open source because Android is open source. There are people that are arguing that Android is maybe not a real open source project and that could be debated, but nevertheless, the source code all released as open source and you have to handle it nevertheless. And then you’re starting to get dragged into these universe of open source, whether you like it or not; that is the way forward.

And so, I think actually Volvo is in a path for discovery. And that was one of the reasons why the join into the Swedish NGO for software industry is because they’re starting to realize this is software, their businesses become more dependent on software. It’s going to be open source. The future model of developing software is more and more collaborative, which is open source. So you see what it is, it is leading parts and bits of the current is going to be untangled and then open source centers that.

Another very important part is you mentioned about the infotainment systems on the instrumentation, but it’s equally important to look into what is happening with the communications. And that comes from the legacy. If you want to make your device, your product, to be connected to the cloud for a lot of reasons. For instance, if you want to have surveillance of where your product of the maintenance needs, then you want to have some kind of census.

You can have a census on the ball bearings picking up the vibrations, and then you starting to realize when the ball-bearing doesn’t work any longer, the car needs to be getting an appointment to a workshop to get those replaced that is connected to the cloud.

And I assume you’re starting to talk about connectivity that is that traditional in the telecom mobile industry business is today totally dominated by open source. I think that people doesn’t really realize how far open source has entered now and what’s happening in the telecoms industry. It’s gone to the extent today that today you don’t build telecommunications network in a country you do not virtualized network and based all nodes in this virtualized network is based on standard components and standard open source software.

Henry: In a past podcast that I did with Patrick McFadden, who’s the VP of developer relations at Data Stacks. He talked about how automotive companies actively contribute to open source projects when a certain software component isn’t necessarily considered to be a competitive advantage of the company.

The example that he gave was the GPS system, and he explained how competitors are happy to collaborate on the GPS system, because at the end of the day, no one goes and buys a car or doesn’t buy a car because of that GPS system. Whereas on the other hand, you have say the dashboard user interface, which may actually impact a purchase decision. So, is this something that you’ve seen in the automotive industry?

Carl-Eric: Not in automotive, but it’s very familiar from my days in Sony Mobile; that was the reason all the time should we keep the social media contributed and released it open. That was one of the things that I preached within the company – when to release and when not to release. And basically you can divide it into two. If you have something that is commodity, something that you call it the qualifier – GSM software, for instance, that will be a typical software that’s a qualifier. Important piece of software, but it’s not uniquely new property. And then you have what is the differentiators? What uniquely you have as a business proposition and that, of course, it should release as open source. You will lose, you advantage then.

I recognized that kind of thinking from constantly that that is what my work was actually very much in the Sony Mobile as the head of open sources is arguing about the business prospects of releasing something against the needs for a company, maybe to protect some IP intellectual property. And that led to a paper that is what are called contribution strategy, using actually is some sourcing methodologies on about external assets.

Henry: What is the contribution strategy currently for these automotive industries? What are they contributing? What aren’t they and why?

Carl-Eric: Well, I will guess though, I’m not sure at this point. I mean, there will probably active in Linux and that is what I’m seeing, and everyone will agree that Linux is from our automotive companies viewpoint. Linux is a very good operating system, but still a commodity. You need everything well working with Linux, maybe adopted for some special needs in a car, it should be maybe more suitable to safety, critical systems, but nevertheless, Linux is a commodity from the car company viewpoint.

So I will expect that are active in what they are seeing as commodities, like Linux, maybe in tools. I mentioned that you have self-driving cars; they’re using a lot of machine learning algorithms and software. Maybe they are participating in developing machine learning algorithm and software, but then that surely doesn’t contribute the data sets and all sorts of things; that is where you have their differentiation and their assets, so to speak.

I’m seeing the same kind of process we had in the mobile industry that in the beginning you tend to be very protective about your IP, your intellectual property, until you starting to realize, hey, this is just a mobile operating system. That is not making the magic of selling things. People doesn’t care about it.

They just want to have a fancy phone, a car with a lot of features, of course, and those features, that is what you’re selling, and that is what you have your business aspects. So in fact, that is actually a fact of open source. Open source in reality is used for things that are considered to be a commodity. There is one of you really have the benefits of open source, because if it is commodity, it could be collaborative developed with your competitors and the making the whole thing much better for the whole industry. And that gives you a competitive advantage when you can work together with the competitors for something that is seen as a commodity, because then you add all your development resources to the level that no one can match it. Microsoft tried it with windows phones, but yet they I gave up in 2017.

So, that is something done, talking a lot today is about the business impact of open source. And I can give you an Android as an example, Google launch Android in 2008, within just two years, it was on 40 plus mobile operating systems, commercially available mobile operating systems. Like the one that I mentioned about, we source it in from a French company for interference, that market just died within two years.

And the last remaining commercially available mobile operating system was Windows phones – cease to exist in 2017. So it took Google about two years to kill the majority of the market than the entire market for commercial available mobile operating systems in just nine years.

Henry: Okay. I think that’s definitely a trend that we’ve seen in a lot of the space when it comes to open source and them coming in and disrupting and completely removing the need to have proprietary technology. I’d like to shift gears and just ask a final question that I think is really quite interesting.

What is the difference between industrial open source and enterprise open source?

Carl-Eric: Not very much. If there is something that differs, I would say that in industrial open source, we take more concentration on the fact that based on my experiences working in an industrial company, there is a bit larger need for an industrial company to protect the IP, or at least the how to put the question, what IP is worth to defend and what IP could get go.

Sometimes it’s actually beneficial to let some of your IP go as open source, because then you kind of destroyed the market for others to take the patent in that area. So Sony Mobile had a defensive patent portfolio strategy. And what that means is that, we often took patents just to be sure that not somebody else would take a patent in that area. And then the patent office in Sony Mobile discovered that, hey, if we have a patent in this area and we release it as open source, then we really have a protection from barring anyone else to take a patent in that market.

So if you have a defensive patent strategy, then open source is a good strategy. And so that will say that is maybe that flavoring that differs industrial open source from enterprise open source that are with my experiences working in industrial that you have to take more concentrations on IP. The second thing I’d do somewhat different from enterprise open source, by taking also the business aspects in account to a larger extent. If you look into my base there, I have a five stage model, which I call the Sony Mobile Open Source Maturity Model. They’re just one, two, three or four are equivalent to what it’s describing the enterprise open source. But my fifth states are obvious actually there, the Google stage, the Google level when you dominate the market.

Yes, showing what you can do if you are. You can be very disruptive in the market and change the whole market logic. And that is not really described in an enterprise open source. So I have a, maybe a little bit more business aspects, business model aspects in industrial open source.

Henry: Well, thank you so much for your time today. It’s been really great talking with you Carl and getting to know you, and I appreciate it.

Carl-Eric: It’s been my pleasure. Thank you very much.

Henry: And to everyone listening, thank you for listening. If you’re watching this on YouTube, then please leave a like and subscribe to see more content like this. And if you’re listening to the podcast on apple podcasts, then please leave a review and let us know what you think. It really does help us to grow and get feedback on the podcast. So, thanks very much everyone. Thank you, Carl. And until next time, see you all.