Klaus-Dieter Walter, Member of the Management Board at SSV Software Systems: "We have to make our engineers fit for dealing with IoT data and algorithms."

© Computers&AUTOMATION

Mr. Walter, is the topic of IoT already gaining momentum in the industry?

Klaus-Dieter Walter: Absolutely - until recently, we were in a phase in which connectivity played the most important role. However, we have now reached the data phase: Analogous to the development in the classic Internet that began around 20 years ago, we will now be dealing less with interfaces and protocols, but much more intensively with data and the information it contains.

So everything is well on its way?

Klaus-Dieter Walter: Unfortunately not: the generation of large amounts of data - keyword big data - already works quite well with the help of suitable sensors in the Internet of Things. But very few people can really do anything with this data. This can be seen in the countless sensor-2-cloud solutions that are emerging: Sensor data is simply streamed into the cloud because the operators of these platforms promise us that sophisticated tools exist in the cloud to extract valuable information and knowledge from this data with the help of suitable data analysis. In principle, this is true, but unfortunately this data analysis process is highly complex and requires a great deal of specialist knowledge, which only well-trained data analysts have at the moment - unfortunately, there is still far too little focus on equipping engineers for data science tasks.

So companies need to look around for data scientists?

Klaus-Dieter Walter: Not necessarily! We had a similar situation in the seventies: Only experts could program computers - with the tools available today, any high school student can do it! We will experience something similar with data analysis. That's why we shouldn't wait for data specialists to familiarize themselves with our specialist world. Instead, we should rely on our engineers and make them fit for these tasks now.

What should engineers be taught?

Klaus-Dieter Walter: Primarily two disciplines:

1. Dealing with machine learning and artificial intelligence: this involves algorithms, some of which we have known for decades from the world of statistical learning and neural networks. They have been used in the IT world for a very long time, for example to monitor credit card transactions in real time or for automated high-frequency trading in shares and other assets.
2. Digital business models: We technicians usually focus primarily on the functionality and technical features of an assembly or system. The economic aspects that arise from the user's perspective are often ignored. In the Internet of Things, this is no longer enough to remain competitive against the USA and China. In the B2B world in particular, you have to remember that customers don't want to buy products, they want to achieve quantifiable results. As a provider, you should start an IoT project with this basic attitude.