Posted by: Chris | 02/05/2009

The Internet of Things: Reality or Hype?

Last week I had the privilege to attend SAP Research’s International Research Forum (IRF) in Dresden. The IRF, which took place for the fourth time, is an annual  global brain-storming event where SAP bring together industrial and academic partners from all over the world to discuss and debate (in a friendly way) topics of interest to SAP (Research) and their clients (which pretty much are all businesses big enough to care about automating their back- and

This years event was dedicated to the Internet of Things (IoT). We were asked to define what this might mean, when and if it would happen, who needs to do what to make it happen and how we knew it had happened :-).

We tried to identify potential “killer applications” – I hate that word – that would justify the investments and make somebody really rich. I don’t believe we came up with a single answer here, however we agreed that the energy business (and environment) would be first be impacted by IoT as government and private enterprise are most likely to invest there.

We learned from Prof. Dr. Janet Wesson, Nelson Mandela Metropolitan University. that the utilities in South Africa have started to equip homes with smart meters to turn off geysers (boilers for those of us who speak US English) when the load in the grid is getting too high. She explained that the utilities for more than ten years had not invested in maintenance of power plants and construction of new power plants, therefore this idea was born out of sheer necessity to deal quickly with the energy crisis.

A lot of technology for IoT already exists, however when it comes to wireless sensors and actuators scalability, reliability and energy consumption are still largely unsolved problems. When you think of an sensor solution with thousands of sensors, some of these are bound to be faulty and will either return wrong readings or no information. Do we understand how to deal with such faults at the application level?

Prof. Dr. Hubert Österle, Institute of Information Management, University of St. Gallen  gave a simple example of his home alarm system that had had numerous mal-functions because of faulty sensors. It became clear that IoT solutions needed to work and provide a good user experience otherwise they might never become adopted or used. Here in my opinion lies a huge challenge. For example, low power reliable wireless communication is still not solved. Just think of the difficulty of pairing two bluetooth devices. My car has one of those bluetooth enabled systems to allow hands-free phone calls. In 20% of the cases it seems to forget that it had paired with my phone and I need to redo that step once more.  Scalable Zig-Bee communication also has not been solved, I hear of cases where with systems of more than 60 sensors, the communication just breaks down. We have a research project on that topic, Mote Runner – I’ll be curious to see how far we can push the technology.

Another observation was that IoT solutions are mostly very interdisciplinary and more often than not have complex business cases as a lot more stake holders are involved.

A simple example is to optimize production at a chemical plant while also minimizing its energy consumption. There is an obvious business case that has an immediate ROI. However as the electricity is being paid for by a different organization in the enterprise from that running the plant, such a project would need to be sold as a strategic change project to the CEO rather than the production or real-estate manager. IBM observes the same when we talk about energy aware computing, the level at which this needs to be presented is significantly different from where we would sell a new piece of hardware.

Claudia Funke,  global ICT services practice leader at McKinsey & Company, challenged our industry (SAP and IBM) to adopt a more consultative sales approach and collaborate with strategy consultants (like those at McKinsey – I suppose :-)) to develop specific IoT solutions with our clients.

As every year, the IRF was a most interesting event – also the side program, which included a visit to Schloss Wackerbarth the Saxonian winery and to Volkswagen’s Gläserne Manufaktur where they produce their version of a luxury car, the Phaeton.



  1. […] The Internet of Things: Reality or Hype? « A day in the life as … […]

  2. […] More info here. […]

  3. Reliability, power consumption, and scalability are all serious challenges for wireless sensor networks. But these challenges *have* been solved, just not by Zigbee.
    Using the same underlying 802.15.4 radio, industrial automation giants Emerson, GE and others are all now shipping WSN products with better than 99.9% reliability and battery life of five to ten years. Systems of hundreds to thousands of sensors are routine. So it’s not just possible, it’s deployed in dozens of countries on six continents.
    The underlying technology for these deployments is in two industrial automation standards, and is now working it’s way through the IEEE and the IETF.

  4. TSMP is an interesting technology given what can be read from the freely available documents (a free download link to its specification or standard would be highly appreciated). In my understanding, though, the points made by Matthias are more related to wireless sensor networks of more powerful motes which can be easily programmed in object-oriented high-level languages, which can be setup and managed by the rest of us, and whose application set can be dynamically changed even after the WSN has been deployed. This way, more and more functionality actually moves into the WSN which, in turn, becomes more autonomous yet also has to deal with application-level problems such as wrong sensor readings by itself. Of course, TSMP could be an interesting building block for these types of WSNs, too, but there’s much more to it.

    • It sounds like we share the same vision of how WSN will be programmed and used in the long run. It’s been a long road from my original smart dust proposal in 1997 to large-scale deployments of reliable networks, and as you point out we’re just now getting started with the fun stuff!
      I do feel strongly that high-level applications are much easier to build on top of communication stacks that are intrinsically reliable, secure, and low power, hence my focus on standardizing TSMP.
      Regarding pointers: here’s a presentation that I gave at the IEEE 802.15.4e working group meeting a few months ago:
      We’re proposing Time Synchronized Channel Hopping, which is the DLL/MAC part of TSMP, to the 15.4e working group for standardization. The ISA100 DLL is based on this.

      The Wireless HART standard is fully described in the HART 7 spec ($995):

      Here’s a recent paper on TSMP:

  5. Great post, just compiled a couple of articles on the Internet of Things HERE. I’m amazed at the explosion in intelligent devices and the growth of the hacker communities, lots of innovation there.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: