Earlier this week, we sat down with John Wall, the COO of Products and Services at QNX to ask him about the QNX Everywhere initiative that was recently launched at CES.
Third Law:
If we were to summarize QNX Everywhere, we’d say that you recognized the industry needs a standardized, hardened platform – something that is reliable, safe, secure, and certified. And so, you’re making significant changes to your product availability, ecosystem, and business model to fill that gap. Is that right?
John Wall:
Yes, I’d say so but let me give you a little bit of background. The notion of “openness” – in other words, adopting some aspects of the open-source model – is something that we at QNX have felt necessary for quite some time. Up until now, senior management at BlackBerry wasn’t comfortable with it. But over the last two years, we’ve had a lot of discussions with a lot of customers, and the number one piece of feedback we’ve received is that the industry is pushing for Linux. Not because they don’t like QNX but because there’s not a big enough QNX ecosystem.
One company told me: “We’re hiring a lot of students out of university, and they come with their ‘pre-configured’ environments – the way they like to work. And that’s not QNX.” Developers know Linux environments and Linux tools. And they’re passionate about it … something that I understand. As a developer back in the day I remember we used to have religious wars around editors like vi versus Emacs.
QNX has been increasingly gaining traction in the mission-critical, highly secure, and highly safe world, but we’re not really attracting enough developers to build a big, competitive ecosystem.
We believe there’s room in this world for more than just Linux. QNX fills a niche that is going to grow substantially over time as things become more and more software defined, more connected, and more interactive. This will demand safety certifications, mission-criticality, and real time. People forget about the importance of real time. When you’re talking about robotics for example, it’s very important that the system behaves deterministically and that there’s no jitter.
We concluded that we needed to grow our ecosystem, and so far, it’s been very well received.
We’re not excluding Linux at all. But we do think that current in-market open-source solutions don’t address all needs.
Third Law:
That seems like a very intelligent way to go about solving the problem. It makes me think about an interview we did with Till Adam and Matthias Dalheimer from KDAB. They are concerned about how the industry is building all these complex systems with really serious safety and reliability demands on a foundation of low quality software and “poorly designed internal architectures”. It seems like QNX Everywhere is a part of the solution to this problem.
John Wall:
Exactly. We’re not excluding Linux at all. We have a virtualization solution that’s an extension of the operating system, so we embrace Virtio, and the fact that this allows us to interface with Linux very simply. It allows us to interface to Android in a very simple way as well. It creates a needed abstraction from hardware to software, which also allows us to have a cloud-first approach. So, we don’t believe it’s an “either or” situation. But we do think that current in-market open-source solutions don’t address all needs. Again, it’s not just about safety and security requirements; real time and determinism are very important for a lot of today’s applications as well.
We’re also taking QNX Everywhere as the overarching philosophy for a new highly integrated vehicle platform that we’re working on with Vector and TT Tech. So, it’s available for free for non-commercial use. Eclipse SDV, Score, SOAFEE – all these different initiatives that are being driven by either Eclipse, or the OEMs, or the European Union – can all take this platform and build on it. Although there will be a cost associated with commercial use.
Third Law:
Can we dig into this a little more?
John Wall:
Sure. There needs to be a commercial model behind software that companies rely on, especially when it’s safety critical. Purely open-source efforts are so difficult to use in production automotive, because you absolutely need liabilities, warranties, and indemnification. You need a commercial entity standing behind a software product. I think one of the mistakes the industry is making is trying to create all these free solutions to solve truly hard problems. There’s also the concept of value-added resellers. QNX will be the maintainer of this core platform, but distributors can take it and bring their own value-add to it. That’s similar to how the Linux Foundation provides a core Linux platform and then Red Hat and others take it and create derivatives.
Third Law:
I think that’s right. You’re pointing out the inability for free software to supply aspects beyond just the source code. Who’s the “throat to choke” that the OEM can chase down when there’s some kind of a bug? You’re not going to go sue some guy who coded on this open-source project in his spare time ten years ago. I mean, it just doesn’t make any sense.
John Wall:
This is a pet peeve of mine; the entire model is broken. OEMs want to create a free software platform. And organizations like Eclipse SDV or SOAFEE want commercial companies to contribute their software into an open-source platform that’s free. They want those companies to change their business models to be more service oriented. But then they use examples like Kubernetes as a classic example of a great open-source project.
Well, Kubernetes was successful because it was Google that wanted it. Google staffed it and Google built it for their cloud business. It was a means to an end. It wasn’t because they wanted Kubernetes per se; they wanted to displace AWS and support their cloud product.
If OEMs really want this software platform to be created for “free”, they are going to have to build it. They can’t expect others to build it; that’s a fundamental flaw in their approach. I don’t know if it’s because they don’t have the capability or if they have the blind spot that “open source is going to solve all my problems”, but the examples we have today show that this has not worked.
Third Law:
That’s a good point. I think people are very attracted to the benefits of open source, but they often only look at the positives.
Let’s talk a little bit about the free part because you say it’s freely available for some purposes. So, if you’re a hobbyist, or you’re basically tinkering, or you’re not trying to create a product based on it, then you can download and use QNX for free? But you need a commercial license if you’re going to be building products and selling those products. Is that the dividing line that you’ve made?
If OEMs really want this software platform to be created for “free”, they are going to have to build it. They can’t expect others to build it; that’s a fundamental flaw in their approach.
John Wall:
Correct. Even if you’re prototyping. Let’s say you’re an OEM or you’re a tier one, and you’re prototyping something that you think you might sell in the future, you can still use QNX for free. I mean, today what they do is they grab Linux, they do their prototype, then they have to change it to QNX. Now they’ve got source full of “Linuxisms” – like Unix domain sockets – that might not necessarily translate well into the microkernel world. QNX Everywhere is all about more people using it, more people prototyping with it, more hobbyists, more universities, more students. You really have to have the intent to go commercial before you have to worry about purchasing QNX.
Third Law:
That’s simple. Other companies have a similar approach, like Qt. If you’re just going to use the product to create your own personal projects, then knock yourself out. It makes sense.
John Wall:
I think it’s important to note that it’s not just the free aspect that’s important, it’s also the openness. We’ll be opening the source code. The vehicle platform will be built in an open fashion. However, instead of calling it open source, we’re calling it “transparent development”, because we don’t want to inherit some of the other assumptions that are associated with open source – like that it’s free, or that it isn’t backed by a company.
The vehicle platform will have a formal governance with key members that contribute to it. There’ll be adjacent members too that can use it and contribute to it (although we may have to do some of our own work here). We’re working through all the details – it’s a multi-year journey. We don’t expect everything to happen in one year, but this is our new philosophy going forward.
Third Law:
When you talk about “openness”, how open are you going? Does it mean that you’re also planning on opening up the kernel or some of the other core components? Or is that yet to be determined?
John Wall:
The kernel is really the only part that’s not been determined yet. We’ll be opening up the networking stack and all the driver frameworks. We’ll be opening anything to the extent that we can. Obviously, we have drivers that are proprietary that come from silicon partners that are not freely available today, but there’s going to be very little off the table.
Third Law:
That makes for another interesting differentiator. Because of the way that GPL licensing tends to work, it does inhibit people like Nvidia or other cutting-edge silicon vendors who might have proprietary stuff that they don’t really want to release to the world. Those vendors can more easily work with QNX because you’re saying that not absolutely everything must fall under an open-source license.
John Wall:
Right. There’s nothing that forces you to open source your code just because you’ve linked against certain parts of the operating system. We want to be transparent, and part of the reason is that we want to show our customers that we’re not trying to lock them in. We don’t want a situation where they can’t see what’s inside the software that their products rely on. I think that’s created some concerns over the years, especially considering there isn’t a lot of competition for what we’re doing, and those hidden aspects can create discomfort for customers.
Third Law:
Let’s move on to talk about markets outside of the automotive space. Obviously, automotive is an area where you have had a lot of success. But you mentioned robotics earlier in the conversation, and there are medical systems and many other places where system critical and/or mission-critical software could be used. How do those markets fit into QNX Everywhere?
John Wall:
The number one driver for QNX Everywhere is growing the ecosystem. We want our existing customers to have more access to more developers, to grow the developer community and the ecosystem of QNX-compatible products. So, one of the things that we’re doing is not just migrating many projects to QNX, we’re porting source with all the test vectors, and we’re pushing it back out to the community. The goal is to incorporate QNX variants into popular open-source projects, rather than having us maintaining all these forked QNX-specific versions. Hopefully we’ll see the ecosystem continue to support that and grow those efforts.
The other driving factor behind QNX Everywhere is awareness. We’re very well known in automotive. There’s no problem picking up the phone and talking to OEMs or getting a meeting. The other markets are much more fragmented as it’s much more diverse and we’re not as well known. A big part of QNX Everywhere is to get more QNX awareness into these other markets. That’s its secondary purpose. If you have universities talking about it, professors teaching courses around it, and students coming out of school knowing it, it’s a big advantage. I think we’re a hidden secret in some of these other markets.
Third Law:
People who use QNX become very passionate about it, so getting people to use it early in their careers will have a big payoff.
John Wall:
That’s exactly right. I mean, GE still calls us after 25 years of using QNX saying how much they love it.
Third Law:
So, presumably are you putting resources against this educational initiative?
John Wall:
Absolutely. This is not exclusively an engineering job. It’s about having evangelists, doing hackathons. It’s not about knocking on one university door at a time; it’s about going out very broadly. So yeah, we have quite a number of resources set aside for this. The good news is we have the investment and we have the backing from senior BlackBerry executives to do it.
Third Law:
That leads us to the next question, “Why now?”
John Wall:
Like they say, timing is everything. If you look back to the years before the pandemic, late 2015, early 2016, QNX was losing the market for infotainment, which was our bread and butter in automotive. Hypervisors weren’t really that big yet and domain controllers were just starting.
In 2001 and 2002, we had gained a lot of traction. Our roots in infotainment later worked to our advantage since that needed a higher performance background: we ran on the big CPUs in the car, and we had the graphics frameworks. We built a better architecture that worked well on SMP and on four cores, whereas other solutions did not. I think those things helped propel us to where we are now. We came back because we focused on the right things: safety and security.
When we started the development of SDP 8, we knew we had to perform well on next generation CPUs. That was going to be our competitive edge with Linux. If we weren’t able to do it, people would have simply continued to work on Safe Linux.
And then too, we had a change at BlackBerry at the executive level. I suddenly felt a bit emboldened to go out and tell people, “We’re taking on an open approach”. I tend not to ask for permission, I ask for forgiveness. So, I announced this publicly because I thought it was important.
Without this change, I think it would have been very difficult to make the next steps we needed to make. The non-automotive market is much bigger than the automotive market. Without that market, you’re at the mercy of the ups and downs of automotive. Plus, a strong position in non-automotive markets helps you balance out; if one market takes a bit of a hit, you’re not affected as a business. We can’t forget that we’re a publicly traded company.
Third Law:
You mentioned earlier in this discussion that QNX Everywhere isn’t an instant turnaround. How long will it take to see the adoption you’re looking for?
John Wall:
That’s hard to tell. I believe we’ll start to see some payback in two to three years from now with some of the initiatives and investments that we’re making. Our focus is on universities to start with. We want to reach students and professors to get more mindshare.
We expect to see the vehicle platform take shape in the next year. We’ve already announced some of our partners, like Vector and TT Tech, and we also have OEMs that are participating – we’ll announce those shortly. All that is going to act as a flywheel to start generating more and more awareness. In the non-automotive space, indirect sales are more important than direct sales. And the way we’ll get companies interested in reselling our product is if they can make money. If they can’t make money, they won’t resell our product.
If you have a commercial company that is doing open-source software, and you want to create a QNX variant, or anything that contributes to the ecosystem, you can use the product for free and get our support.
Third Law:
It sounds like you’re going back to your roots. QNX once had a lot of general-embedded value-added resellers in the late 90s.
John Wall:
Oh, we did, absolutely. We not only need to create an incentive for resellers to want to sell our product but we also need to give them the chance to create added value. And that’s why we’re creating a distribution model with the automotive vehicle platform. It’s important that a tier one or an integrator can take that platform, add value to it, and resell it.
Third Law:
QNX has been around a long time, and even if it doesn’t have much public awareness, it does have a lot of fans. If you’ve got fans that want to help contribute to the growing QNX ecosystem, how should they get involved?
John Wall:
I’m glad you brought that up, because we talked about non-commercial use – QNX Everywhere is also for commercial companies that want to create QNX ports. So, if you have a commercial company that is doing open-source software, and you want to create a QNX variant, or anything that contributes to the ecosystem, you can use the product for free and get our support. We really want to encourage people to contribute and be able to grow the ecosystem around QNX, and we’ll have different incentives that we will create over time for people to do that.
I think that this is the most exciting thing I’ve seen since I’ve been at QNX. And as you know, I came with the building, so I’ve been here a long time. But it’s really exciting, and it’s really fun to see that there’s still a lot of energy and enthusiasm to make this happen.
About the author
