IT is the basic challenge

I dag indbygges eller indlejres software i næsten alt, i stadigt mindre systemer og med minimalt forbrug af energi. Mere it gør vores samfund smartere, men gør også vores systemer mere komplekse – med større risiko for fejl. It er en forbundet grundudfordring, et underlag for alt andet, vi vil løse på, mener Kim Guldstrand Larsen, som udover indlejret software også forsker i at sikre de stadig mere komplekse it-systemer med ’modelbaseret udvikling’.

 - It is actually hard to imagine a society in which embedded it - pervasive it - does not play a crucial role, Kim Guldstrand Larsen, director of Centre for Embedded Software Systems and director of the InfinIT innovation network, explains.

- That is why it is quite good to work with these challenges, he continues in reference to the main challenges, described in FORSK2020, that Danish research is to focus on in the future:  - But if we can see, that IT is an important element in all of them, then we also need to invest properly in IT. Not just as a given technology, but as something that is continually changing. We need to be good at developing it, and we need to be good at utilizing it - the way it is at a given time.  - So actually, it would be nice if we had a specific grand challenge called IT, he says.

Smarter systems, larger complexity - and risk

In other words, according to Kim Guldstrand Larsen, IT cuts across the challenges we as a society are facing: IT is a kind of basic challenge - something that is connected to all fields and topics.

   Among other things because IT by now can be found in everything: it is increasingly integrated in our everyday lives, we use it in an increasing number of contexts - and we are increasingly dependent on it. Not least on its working as it is supposed to - also when the complexity of the systems increases, and the risk that something goes wrong is heightended accordingly.

   Kim Guldstrand Larsen's own research in a way works on both sides: on the one hand it focuses on embedding IT in our everyday lives and make them smarter. And on the other it focuses on ensuring applicability and security in the increasingly complex systems.

   Things are connected. 

Embedded software in almost everything

The main field of Kim Guldstrand Larsen's research is called embedded software.

   Embedded software is software that is integrated in mechanical appliances where it helps controlling, surveilling and managing the device it is part of.

   Compared to mechanics or ordinary electronics, embedded software opens up for far more characteristics and applications of a device or system - things become more "intelligent" and reliable.

   The potential is extremely big and wide: by now, the embedded technology can be designed to take up next to no space. And to use a minimum or no power at all - that is, be self-sufficient. The technology can be applied in solutions within pretty much any field and in pretty much any challenge - obviously in e.g. green growth or health technology.

   In the power grid of the future, embedded software can e.g. contribute to controlling a series of functions in the intelligent home such as light, heating and laundry machine - and at what time of the day green and cheap power is drawn from a possible smart energy grid.

   Within health care, it can contribute to the remote surveillance of pacemaker patients - and maybe even no-battery pacemakers that use embedded software for harvesting energy from the movements of the heart. Or the software can be integrated into the personal health checkup, biometric measurements, where the results could be shown on an iPhone.

   All in all, there are a lot of possibilities in the technology, Kim Guldstrand Larsen says. And we have only just started working with them.

Model-based development - avoid errors, delays and 'deadlock’

But the more IT is propagated and integrated in our everyday lives - especially if we as a society focus on it - the more it makes sense that the technology needs to develop in a desirable way.

   In Kim Guldstrand Larsen's words, the IT systems over time become 'incredibly complex'.

  - Because the complexity of the systems increases, so does the probability that something at some point will go wrong, he says.

   That challenge - that we become better at ensuring that IT and the complex systems do in fact function as intended - is what he works on within the field called model-based development.

   Model-based development entails the use of design models for analysing and calculating how a system will perform. And preferably before it is put to use.

   As an example, Kim Guldstrand Larsen mentions how, if you want to build a Femern bridge, it would be sensible to create some design models beforehand in order to find out how much strain the bridge can take in any imaginable situation. The same can be done for IT systems: system models can be set up and then thoroughly analysed using the proper methodic tools: how much strain can the system take, hwo much energy will it use, will it be done in time?

IT as our nervous system

In the brake system of a car, e.g., it is pretty vital that the signal that turns on the ABS system works when you need it.

   Today, a car has several hundred small computers embedded in it, Kim Guldstrand Larsen says. All have 20-30 sub-functions, each of which is highly safety-critical and which all 'compete' for the tasks on the computers. For instance, if the car hits a tree, the air bag must be released within very few milliseconds, and not go into waiting position because another part of the car's system has gone online to download new maps for the GPS.

   Kim Guldstrand also mentions the IC4 trains as a spectacular example of how something can go wrong in the system design, with large consequences. Or much worse: the Chinese high-speed train in Wenzhou last year, which was hit from behind by a following train and thereupon derailed in a massive accident with 40 casualties, probably because of a software error.

   Our entire communication infrastructure is an IT infrastructure that for instance the global financial system is a part of. And in the entire system within which power is traded among countries today, the nervous system is IT.

Strategic advantages in being able to document IT security

Methodical tools within the model-based approach can analyse and calculate extreme amounts of data and catch safety-critical 'cornercases’ – that is, all and also extreme cases. This could e.g. be in connection with important testing of medical equipment for diabetes patiens or ventilators.

   The fact alone that a Danish ventilator company can document a high level of skill and security in precise testing of teh technology is a strategic advantage, not least on the export markets.

   The model-based approach can save large amounts of development time. And it can furthermore be used for planning when what needs to happen when in a system, so that the energy consumption of the finished product is minimised.

Sort out the ICT

In other words, there is a large potential simply in being able to say very precisely that the IT technology works and is safe. Something very fundamental, and something a network such as InfinIT needs to contribute to emphasising, says Kim Guldstrand Larsen:

 - One of the things I really want to achieve is to make it clear how important sorting out your ICT is. If we do not solve that challenge, we cannot solve all the other challenges. So it is a basic grand challenge, he says. 

Kim Guldstrand Larsen is a professor at the Department of Computer Science at Aalborg University. He is in the steering committee of InfinIT.