2020 Will be a Tipping Point for IoT

IoT has massive implications for businesses of all kinds, and for individuals at all organizational levels as well. Bart Schouw, IoT solutions director at Software AG, recently reflected on the changes taking place and explained why 2020 will be a critical year for IoT.

What is SoftwareAG’s role in the IoT market?

The IoT market is incredibly fragmented. There are literally hundreds of players in every area.  At Software AG, we don’t see ourselves as an end-to-end player. We will play to our strengths and from our history that is enterprise integration. And, going forward, we see a major role with our Streaming Analytics Platform capabilities within our Digital Business Platform. We will use those streaming analytics capabilities to help companies to operationalize their predictive analytics, machine learning, and even artificial intelligence.

What are the key verticals in which streaming analytics is important?

As a software provider, we support virtually every industry, but where I see a lot of uptake on a natural level is in manufacturing. Retail is gearing up quickly also; Amazon has come up with Amazon Go and retailers will be smart to get serious. We will see it in hospitality where customer experience is key. We will also see it in insurance where companies will start thinking about new business models based on sensor data and the behavior of the customer in relation to risk. In transportation and logistics, companies are using it to understand where packages and goods are shipped. I think every industry will be touched.

What is the level at which companies are leveraging IoT at this point?

It is the very large enterprises that have the capacity to invest. It will be a challenge for smaller and midsize companies to keep up with that digitizing of the products. I have been traveling around the globe talking to a lot of our customers and prospects and many of them have a strategy for 2020—and, for the last few years they have been exploring ideas for their 2020 vison. However, if they really want to have that strategy bearing fruit they have to start investing now.  Over the last year, I have seen people deploying pilots and PoCs to try to understand how to implement IoT.

How are views about IoT changing?

In the keynote presented by Peter Sondergaard of Gartner at its ITxpo, he said a digital platform is going to be key if you want to build an ecosystem on top of your environment. I think many companies are realizing that it is not about siloed solutions. It is really about getting a digital approach and that means building a digital platform.

There have also been many conflicting predictions about the number of connected devices and sensors by 2020, with some estimates going as high as 50 billion. Do you think of that year as being a tipping point?

It will absolutely be a tipping point. It will be very interesting to see how companies will execute after 2020 once all their strategies are in place. But companies will have trouble getting from the pilot stages to production if they don’t take the platform approach. People are also finding that getting those things out there in the real world is very difficult. It is still very expensive to have connectivity. Adoption will accelerate for sure but whether it will be 40 billion or 50 billion devices is not really relevant. It will be a lot.

Is security a big concern for companies?

Yes, but at the same time we see more and more standards being adopted. There is a difference between consumer IoT and enterprise IoT. Unfortunately, consumers are influenced by price and often go for the cheapest options available and that means that enterprises have to compromise—which means compromising on security, as well

However, enterprises that are building for enterprise usage are putting serious thought into how to build those architectures from a software perspective to make sure that the security is in the foundation. But IoT is made up of physical devices and those devices might have a very long lifespan. We need to think about whether the architectures that we deem to be very secure now will still be secure then.

How is AI being put to use in IoT architectures?

We have a very specific view on this because we are not building our own artificial intelligence capabilities, but we want to support companies in executing AI. A lot of companies have data scientists on board that are building those AI models to do deep learning, machine learning, and predictive analytics with very advanced tools on big data and that means that they can build, train, and test a model.

But that doesn’t mean it is easy to deploy it into production because after the data scientists are done, IT has to take over—and there is a completely different ball game. That is where we come in. Through our platform, IT can support those data scientists by taking ownership of it. Give us the model and we will deploy it in production and provide governance and we will make sure that it can execute in real time in the way the data scientists designed it.

What is changing?

Up until now for many IT departments, this meant completely rebuilding the data scientists’ models in C++ or .NET and building a very bespoke implementation which could take months of work and loads of testing before any kind of model originally built by a data scientist could be deployed.  The way that we have set this up, a data scientist can hand this over and we will generate an executable model out of it so that IT can take ownership and get those models into production in time frames of hours rather than days and do that regardless of the kind of cloud infrastructure where tests are run. That is the next step: How to get those models that data scientists are building as quickly and reliably as possible into production and do this kind of analysis at scale in real time? That is the next big thing for IT software companies.

When you look at the IoT field globally, is there anything that concerns you?

From a personal perspective, I really wonder about the future of humans in enterprises. I think that many decisions by humans in the future will be taken over by machines and in very unexpected ways. It will even influence the jobs at the top. We have already seen a company that has assigned an algorithm at the board-level to evaluate the decisions of the board.

Where do you see this going?

It is only the jobs that are done with your hands like the work of a gardener that will not be replaced by IoT but everything else that can be an administrative role and involves decision making will be industrialized. It is similar to the Industrial Revolution that began in the late 18th century. However, this is not a concern for this year or even for the next 3 years. But, for the next generation, these are questions that will have to be answered.

Do you have any advice for organizations starting on the IoT path?

Yes. When they start off, most companies say that their things don’t “talk,” so they want to start small and get an understanding.  And, the first thing they always want to do is visualize data; they want to see a graph showing the temperature rise and the curve that their machines go through. Then, as soon as they have that, they come to the realization that the visualization alone is not valuable to them, so they want alerts on those visualizations telling them that if the graph is going out of the curve in an unexpected way they want an alert. It is the alerting and the automation that is going to bring value but the issue with alerting automation is that you need a different kind of architecture for it. This means that you have to architect for the future solutions. You have to think along the maturity curve.

If you only try to solve your most immediate issue for IoT, you will fail in the end because it is not visibility that is the use case. IoT in the end will influence your business model and you have to architect for that. You have to keep that in mind. If you don’t do that from the very beginning and get business involved, and if the business sees it as an IT project only for doing something technical, they have already lost the game in IoT.



Subscribe to Big Data Quarterly E-Edition