Q&A: Gernot Heiser, Entrepreneur of the Year 2014, Engineers Australia

Pacemakers, cars, ATMs are all hackable. Researchers at the National Information Communications Technology Research Centre (NICTA) are working towards fixing this issue - they are researching and building software systems that eliminate these types of exploits.

Named Entrepreneur of the Year in the 2014 Sydney Engineering Excellence Awards by Engineers Australia, Sydney Division, Gernot Heiser leads 80 researchers, engineers and students at NICTA. Having had an illustrious career leading and developing operating systems, and with his technology running the security processor in all the recent Apple iPhones, Gernot Heiser is a wealth of information on technology, the future and Australia’s place in it. In the Q&A below, Heiser talks about the past, present and the future of software systems.

Q: As Research Group Leader of Software Systems Research Group (SSRG) at NICTA (National ICT Australia), what does your role involve and what are your group’s recent achievements?

A: I head a group of about 80 researchers, engineers and students working in the general area of software systems, with the aim of improving the reliability of software systems. The research I am personally involved in, which covers about half the group, is called Trustworthy Systems: we build systems that are provably trustworthy (ie, safe, secure and dependable). The aim is to make guarantees that critical systems cannot fail/be hacked etc. Present systems are far away from this ideal; pacemakers can be hacked, cars get hacked, ATMs get hacked. Our work is about designing software such that these exploits can be completely ruled out.

Q: You have collaborated with Apple, Boeing, Department of Defence Australia, Google, IBM and Intel to name a few; how is Australian research perceived overseas?

A: Australia is a bit of a backwater. Which is hardly surprising, given our small population, and the lack of the kind of high-tech industry that exists in the US, Europe and East Asia. Also, public-sector investment in research, in per-capita or percentage of GDP terms, is pretty low in Australia compared to other OECD countries.

For example, the US President recently said that economic prosperity in the 21st century will depend on cybersecurity. If he is right, and I think there is a lot of truth in what he is saying, then there is real reason to worry about our future prosperity.

Having said that, Australians have a track record of batting above our weight in research, and there are real highlights in a number of fields. Certainly, our work on L4 is widely known internationally, and I am aware of a number of plans for serious investment to utilise our technology.

I think generally the output/input ratio of Australian public sector research is much better than in most places. However, with inputs so small, there is a serious limit to what we can achieve. Certainly, NICTA has been phenomenally successful, performing much better than many international organisations that are much more generously funded.

Q: Having had an illustrious career leading the research and development of new operating systems and having just been recognised as Entrepreneur of the Year at the Sydney Engineering Excellence Awards, if you take a look back to where it all started, what motivated you to undertake your studies?

A: I got hooked on computers early on in my undergraduate studies, mostly because of the incredible wealth of things you can do with them. Computers are like the ideal construction material: you can build just about anything with them.

In fact, the software systems we build these days are way more complex than anything else ever built by mankind. Consequently, you need to take an engineering approach to build those things, or they will never work. The discipline is still young, and we are still working on establishing fundamental design principles.

Q: The US has Silicon Valley, a hub for high-tech innovation and development; is Australia doing all it can to foster this sort of innovative culture here, and if not, what could it do?

A: Short answer to the first question: No.

Silicon Valley is first the product of two great universities: Stanford and Berkeley - one private, one public. They had a critical mass of excellence. By now it’s self-sustaining, and the benefits flow back to those universities, and society as a whole. But the universities keep contributing a lot to the Silicon Valley environment. And there are cultural issues, and there’s a suitable infrastructure, including a large number of cashed-up venture capitalists, with excellent networks, but also excellent access to technical know-how (typically from the same two universities) so they can deeply analyse technology before investing. 

We can’t hope to produce something of the same scale (and other countries haven’t succeeded either). But we can try to build excellence in narrower areas. Key is to have critical mass of something. NICTA started to be successful once we stopped behaving like an oversized university department and instead decided to focus on a small number of areas where we could be the best in the world, and put significant resources behind those focus areas. That’s key.

Q: You led the team in the development of the L4 microkernel. Can you tell us what it does, what the need was/is and why it’s so important?

A: The L4 microkernel is the core building block in the design of trustworthy systems, the foundation on which we build everything else. It allows us to build more complex systems out of simpler building blocks, where, importantly, failures can be isolated and prevented from spreading though the system. This is in contrast to mainstream operating systems such as Windows, Linux, iOS, which are monolithic systems: a huge amount of complexity is piled together, far more than anyone can completely understand. The results are all those failures and security exploits which we have, unfortunately, become used to.

I’m in an age where I must expect that sometime in the not-too-distant future, my life may become dependent on a medical implant, such as a pacemaker. At the moment I wouldn’t feel too comfortable with that, knowing how those devices function and how easily they can be hacked. So by the time I need one, I want them to run our seL4 microkernel, to be protected.

Q: Tell us about your most significant research development and the impacts it had?

A: After years of trying to change the way mainstream (desktop and server) operating systems are designed, I realised that this was not going to go anywhere, there’s just too much inertia. The key insight was that the embedded space was about to undergo a major transition: classical operating-system technology then used in embedded systems (so-called real-time executives) was bound to reach its use-by date, because it would not be able to support the incredible explosion of functionality that was about to happen (phones are an excellent example). There would be a move to new technology, and I thought we had the beginnings of a suitable technology.

I was still surprised when it happened, and it did so for exactly the reasons I anticipated. But what enabled us to jump on the moving train was not just the research papers, but the fact that we had open-sourced our software. By now, our operating system has shipped in billions of devices, there’s probably half a billion people in the world that own a phone where our operating system runs somewhere. And all the recent Apple mobile devices (iPhones, iPads, iPods) run our L4 on the security processor. It’s a researcher’s wildest dream come true.

Q: You have been listed as one of Australia’s most influential engineers; what is your vision for the future of Australian tech engineering?

A: I’d like us to develop some real world-class capabilities that keep us competitive in the long term. We obviously cannot be best in everything, or even most things; we have to focus on a number of niches where we can excel. We need to build on what we’re good at.

Countries with a much smaller population, like Switzerland and Israel, are doing this very successfully. They have the advantage of not having a lot of natural resources, which forces them to be innovative.

One of our greatest assets is our lifestyle; natural environment, liveable cities. Smart people want to live here, and they want to work on cool stuff. And, judging by the many bright students I work with, there are plenty of smart folks who want to succeed, who want to show the world how good we can be. We need to support them, and the engineering and science disciplines are the foundations to success.

The possibly scariest medium-term development impeding this vision is the neglect of the STEM disciplines by governments and opinion makers. Computer science and engineering is maybe the crassest case. After the dot-com crash of the turn of the century, student numbers crashed from ridiculously high levels to ridiculously low ones. That was a worldwide phenomenon. But around the world, numbers have recovered to very healthy levels, except in Australia. We’re a total exception, and this is really scary. We’re in real danger of missing the boat in the technology that will continue to transform society.