TechCasts

Read the Transcript

Ben: Embedded software has transformed dramatically over the past decades, evolving from single-core simplicity to highly complex, multi-core systems. As industries demand more power, efficiency, and security, companies face growing challenges in balancing innovation with certification requirements. In today’s episode, we’ll explore the shifting embedded market together with the role of open-source software. How do we navigate these changes while ensuring reliability and safety in a more and more connected world?

Hello everybody. I am Ben from the Marketing and Communications team and today I'm joined with Mark Tootell. Mark is our regional sales manager at SYSGO and based in the UK, taking care of a wide range of projects and customers all around the world. So, welcome Mark.

Mark: Good morning.

Ben: Could you please introduce yourself: What is your work here at SYSGO and how did you actually come to us?

Mark: So, I am the regional sales manager for UK, US, Canada, Australia and New Zealand. As you said, customers all around the world. I joined SYSGO because an old friend of mine was working for SYSGO as an FAE and he reached out to me and said, interesting work, great company. Do you want to come and join us? We're looking for a sales guy. And strangely enough, as I'd known him for more than 20 years, said, okay, I'm interested and ended up here at SYSGO, which happened about six and a half years ago now, where I have no idea where six and a half years have gone. But I've been in the embedded and more laterally the IoT space for nine or 30 years now. So, it's been a lot of changes in the market space.

Ben: That's quite impressive. The embedded market is definitely a unique and complex space, presenting plenty of challenges. Over the years, we’ve seen significant shifts in technology and development approaches. From your perspective, what do you see as the biggest challenges companies face today when working with embedded systems?

Mark: That's a great question. One of the things I've seen in the market space is going from single-core and external chips and almost having a simple system. I'm not saying they were simple, and I'm not saying they were easy to integrate, but a single-core with an external IoT chip was kind of the norm when I started. We now have significantly more complex systems, a multi-core massive amount of IoT, but we're abstracted away from that complexity by operating systems. The Linux community does an amazing job building drivers and building support for very complex systems. And as a consequence, I think the knowledge that we gained when we had single-core systems and external IoT chips has almost gone. 

We've almost forgotten, as an industry, what happens under the hood. We're very good at configuring operating systems and building applications, but the fine tuning and the low level stuff has got harder, but at the same time we have less people doing it. I've seen less and less people coming into the true embedded market over the last 20 years, which is an interesting challenge. It's a big gap to fill, and I don't know where that skill set and that knowledge is going to come from for future generations.

Ben: Embedded systems are becoming more powerful, especially with advancements in multi-core processing and graphics applications. At the same time, there’s a constant need for energy efficiency, with size, weight, and power as key considerations. How do you see this evolving in the future? How can we balance the demand for more processing power with the need for greater efficiency from a technology standpoint?

Mark: Yeah, again, great question. I see the silicon vendors making ever smaller, ever more efficient system on chips. I've seen some vendors, the system on module vendors, taking those and developing derivatives of those SoCs or COTS platforms, but the SoC vendors are also adding more functionality, which as the size of the silicon goes down, but the complexity goes up, we're almost balancing out the gains we get from better technology with greater levels of integration.

I think in the world that SYSGO lives in, one of the challenges is how do we remove the things we don't need on a chip that's already built. How do we make sure that we're not putting power into things that we're not using? Actually, how do we simplify the complex? How do we remove the technology we don't need? The reason I say the technology we don't need is because finding silicon that is purpose-built as in it was built for one purpose is becoming more and more rare. We're seeing ever higher levels of integration where we're adding more technology into a single system, but we don't necessarily need all of that technology for the use case that we have. It's an intriguing problem and it's a growing problem. If you have the luxury of not worrying about how much power you're using, then it's not a problem you have. If you're looking to reduce power, reduce technology, and in our world, the SYSGO's world certify, it's definitely a challenge that we're faced with on a daily basis.

Ben: You mentioned open-source software earlier, so how important is open source in today’s embedded systems landscape? Do you see it as a key driver of innovation, or does it present more challenges than benefits?

Mark: Ironically, I think it's incredibly relevant. I spent a large proportion of my career where open source was seen as the enemy. I'll say that because I spent a lot of my time working around Windows embedded and Windows embedded business and partners. So selling a closed source operating system with partners that are providing hardware that sat on. One of the benefits of that was Windows is one of those operating systems that supports pretty much every technology on x86 platforms, for instance. 

However, I've seen the Linux is the enemy or Android is the enemy message change to actually we need to embrace open source and use it and benefit from it. It's a different commercial model and it does take some thought compared to a close source operating system. But the right answer is now a mix of both in that open source provides a huge amount of resource to keep up with changes to provide support for new technologies to drive new technologies into places where perhaps they wouldn't have been if it was down to a single company to support.

And I think the answer now, especially since I've joined SYSGO, is mixed criticality or using Linux and open source and Android for the right reasons, having deeply embedded real time alongside open source, but just making sure that the two are used for the right reasons, which if you'd asked my 20 year ago, 30 year old self, who was saying open source is the enemy, I probably would be surprised at me saying that.

There's also outside of the technological challenges, there's also the licensing challenges, and that can be probably more difficult to navigate, whether it's an MIT license or a GPL or whatever flavor of open source license that the technology is released under, the management of the software better materials and the licensing of the technology, if you are taking care of that can also be a major challenge. So I think open source used in the right way and licensed in the right way and managed in the right way is a very, very valuable tool in the world we now live in.

Ben: That’s a really interesting perspective on open source and licensing. Managing software components while ensuring compliance with different licenses is clearly a major challenge. And speaking of compliance, that brings us to another critical topic — certification. This might be a good time to shift our focus to avionics. Given your experience and geeting into closer discussions with our customers and partners, what do you see as the unique challenges when developing embedded software for aviation?

Mark: Wow, and a great question. And actually, my history in industry is business critical rather than safety critical. So joining SYSGO, focusing more on the Avionics and defense space than I have historically in my career, actually I had some very interesting lessons. Avionics is special rather than business critical, which has one type of driver behind it. Safety critical is another league again. So it was a bit of an eye opener for me.

I think the challenges with Avionics will almost go back to the level of complexity that we can now see in the SoCs that are coming out from the big vendors. Historically, Avionics only wanted to support what it needed. If an SoC didn't have technology, that was actually a good thing because we didn't need it. We didn't need USB, we didn't need some of the IO. We as an industry could focus on silicon and technologies that we needed and only enable those technologies and only enable those features.

As the world is getting more complex and the SoCs that we can buy from the SoC vendors are adding more technology, multi-core, massive amounts of IO, incredibly fast Ethernet. The ability to refine the technology down to the bits that we need is becoming harder. We're almost pairing what we've got down to bare minimum. The choices of silicon out there are driving decisions. We're still trying to work out how to properly certify multi-core in the context of the silicon that's available today.

We are still trying to work out how to take something from perhaps the automotive market, which has high volumes and is attractive to the silicon vendors. Then take that and remove the technologies we don't need from the silicon or disable the technologies we don't need from the silicon and then apply that silicon to the Avionics world. Even though the Avionics world has huge amounts of money and huge amounts of requirements, the actual volumes compared to things like automotive are significantly lower. No silicon vendor would or very few silicon vendors, I should correct myself, but look at the Avionics market as a separate market and therefore that also forces the industry to behave in a way that perhaps it wouldn't if the volumes were similar to that in automotive.

The challenges are there because we need to make sure that we do things properly. We need to make sure that the aircraft and the technologies we're deploying are fit for purpose. The rigor that we go through definitely provides a whole bunch of challenges to the industry.

Ben: Yeah, it's really interesting. When we talk about Avionics, it's not just Avionics in the traditional sense anymore. We're seeing a shift with civil Avionics, defense applications, and the rise of UAVs and eVTOLs. The market that we've known for decades is evolving, and it feels like it's undergoing a significant transformation. At the same time, we're seeing a lot of new impulses coming in, especially with the increasing need for integrating software across various use cases.

Mark: Yes, and I think that's also driving the industry in a new way. Avionics or the aviation market as a whole suffered from people leaving and let's be honest, we're all getting older. So lots of the knowledge and a lot of the capability is no longer in the avionics industry. And now, as you say, there are new use cases and new technologies, eVTOL, UAV, AI, autonomy, high altitude, pseudo-sathlights. There's all these new use cases and new things that we can do with Avionics that we haven't done before, just as a lot of the capability and the knowledge is quite rightly going off to enjoy retirement in some instances. And I think it's a unique challenge. How do we apply the right processes and procedures that we've built up to make sure that the airplanes that we sit in that are taking us thousands of miles do so safely? How do we apply those technologies and techniques that we've known and have been used in a particular way because the world we live in as in the avionics world has slowly but surely changed, but we've managed
that change? How do we apply those to the new use cases, the new shiny, shiny technologies that are coming out like the UAV and UAM world, where the technology underlying is different, but the requirement is still the same?

The requirement is still to make sure that we apply the right level of rigor and the right level of testing and the right level of process to make sure that people are transported from A to B safely, or that the things that don't have people in them but are flying above people don't certainly stop and cause injury. How do we apply that level of rigor to technologies that aren't quite as familiar as the ones that we've worked with for the last 20 years with less resource, with less people as well, and with silicon that is coming out of silicon vendors that is far more complex. The center of the Venn diagram is very small with lots of challenges being put around it.

So yeah, it's a very interesting and for me every conversation I have is fascinating because people are pushing the envelope and coming up with solutions. And so every conversation you have in this modern avionics world is different and you learn something every day, it's always a great place to start your day knowing that you're going to find
out something new.

Ben: We’ve touched on some fascinating aspects of the evolving embedded systems market, and the complex challenges in avionics. As we continue to explore this world, it’s clear that the road ahead will require even more innovation, precision, and collaboration.

In our next episode, we’ll dive even deeper into the specific challenges of the embedded market and discuss how solutions like our PikeOS RTOS and Hypervisor can help navigate this evolving landscape. Don’t miss it — we’ll explore how PikeOS is addressing the need for both high performance and certification compliance in complex systems. Until then, stay tuned and keep your systems running smooth!