Subscribe Find out first about new and important news
  • Share via email
  • Subscribe to blog alert

What is different in the Internet of Things?

what is different Source: offenblende

Many different kinds of organizations are embracing the Internet of Things (IoT) e.g.: start-ups, global internet companies, established software vendors, telecommunication operators, or established “things” companies. Over the past years I have been experiencing, that the IoT is not only a big change in business model, culture, and technology for established manufacturers like Bosch, but also for all other players in the future IoT arena. So, I asked myself:

  • What is different in the Internet of Things for “things” companies?
  • What is different in the Internet of Things for “internet” companies?

That’s when I realized that when we think of the Internet of Things, we see it from two perspectives: internet companies and things companies.

The problem with the term “Internet of Things” is that we tend to think it’s simply about connecting objects. In reality, there’s not much value behind that. For example, connecting a fan to the internet doesn’t have value; but if it produces something that provides a service to us, then it has a benefit for society.

When I started working in the IoT in May 2009, in the left corner of my whiteboard I taped up a chart of the 15 global challenges for humanity. The chart reminded me that I was working towards a purpose, and kept me inspired if I ever felt lost or tired. I found myself referencing the number one challenge: sustainable development and climate change, and the global convergence of IT and energy.

In the IoT, we find value and profit in services, therefore turning connectivity into a commodity. Of course, there were a lot of failures when we started our first IoT projects, but we also found patterns in our learnings.

The key takeaway: collecting data based on size and time isn’t one-size-fits-all

One of the first things we learned in the IoT was to consider two dimensions: time and data. Are you collecting data in milliseconds, months or years? How much of that data are you really interested in? A few bytes, or megabytes, and petabytes of data? Solving the requirements of collecting data based on size and time doesn’t have a one-size-fits-all technical solution, but instead requires different approaches. Consider collecting field data from cars versus video stream processing. With cars, you won’t know what specific data to collect at the beginning to understand what can be adjusted, so you collect everything. For video streaming, we know that we should filter locally first in order to process it quickly and avoid extra costs of transmitting “empty” video frames. Therefore, both cases require different communication protocols, databases or enterprise service buses.

System of coordinates illustrating the amount of time on the x-axis and the amount of data on the y-axis. Source: Stefan Ferber
System of coordinates illustrating the amount of time on the x-axis and the amount of time on the y-axis. Source: Stefan Ferber

Different applications require different patterns of connectivity and deployment

Starting in the seventies, a very early pattern of connectivity (often called M2M), included users establishing a VPN connection to an enterprise IT system. The problem with that was that it’s not really ready for tenants, since it’s basically meant to run in one company. Applications can then only scale so much, because enterprise IT is not really made for scaling. You must be granted access from IT, who then must set up new systems, and so on and so on.

Infographic illustrating two different patterns of connectivity and deployment. Source: Stefan Ferber

Thus began another pattern of connectivity from consumer electronics: smart remote controls. For example, if you buy a smart camera, you most likely use a local connection such as Wi-Fi or Bluetooth to interact with the device. Your iOS or Android device then serves as a smart remote control. The difficulty with this is the lack of flexibility in one-to-one mapping. There’s one user, one tenant and one application per device. Trying to share data with your family at once from the camera isn’t going to be that easy.

Smart home technology is a prime example of the third pattern of connectivity. A central gateway gathers local protocols, connects to the cloud and then scales as needed. It works with multiple users, tenants, applications and has the ability to connect to thousands of IoT devices. IoT gateways are now entering also other domains enabling for instance Industry 4.0. And there are some good reasons for connecting devices in a more indirect way to the cloud via gateways: getting more independent of internet connectivity, reduce the data volume to be transferred to the cloud (which also ends in minimized costs), and – last but not least – ensuring privacy demands since master data can be stored and processed locally. However, in todays world users suffer from having one gateway for smart home, one for the car, one for the dishwasher, and so on.

Infographic illustrating patterns of connectivity and deployment for an IoT gateway and an IoT platform. Source: Stefan Ferber

IoT as a platform: the pattern to solve all problems with “pattern 42”

The way to make the Internet of Things a really purposeful system is to connect all devices with semantic interfaces to the same platform. Then, invite thousands of developers to build the applications to serve the market. Otherwise, we end up with vertical solutions of nice tiny applications without the benefit of an interconnected infrastructure. The way to push this technology forward is through an open system community, with open-source solutions and standards.

In order to build pattern 42, the corporate, government, NGU, and open source communities embrace a lot of change.

What is different in the Internet of Things for…

So back to our original question: what are the differences for internet and things companies in the IoT?

  • To start, the internet wasn’t really built with reliability and security in mind. Instead, it was built for interoperability, and openness. But with that comes concerns for security, privacy and cost. The trick now is to make the internet a trustful system for “real world” applications, and allow connectivity to be affordable. That means a major culture change for software and internet companies, who will have to embrace liability and 20+ years of service delivery for this new class of applications.
  • As for the “things” companies, the difference lies in culture and continuing the cycle of innovation. It’s more difficult for things companies to adapt to the “always on” nature of internet customers. Therefore, a shift in business model is necessary to continue making progress: if you are always connected to a customer, consider a recurring revenue, subscription or premium model and continuous customer feedback on service delivery and product development. In addition to culture, it’s difficult for things companies to put the developer at the forefront. By 2020, we’ll need about five million developers to create applications for IoT. There’s no one company on planet Earth that can hire so many, which is why it’s important to emphasize the role of the developer.

Get started with your IoT project

Thomas Alber provides five things you should consider beforehand.