Internet of Things

The most pervasive applications of the Internet – so far – have been humans communicating with each other, mainly via email, the Web and now social media. But increasingly machines are being connected to the internet in their own right. As in the early human internet, the early connections are “1:1”, put in place for a particular purpose, and often “M2H”, having a machine at one end (collecting sensor data for example) and a human at the other (reading that data on a web page).

However, the number of computers “embedded” in our daily lives is increasing, and so is their connectivity. The information they can offer increasingly has value not only to humans, but also to other machines. And as more machines connect, so the value to other machines of connecting rises too – just as with the human internet. We could see the modern trends of Open Data and the Semantic Web as a part of this trend too – making the vast corpus of live and historical web data accessible to, and understandable by, machines as well as humans.


Here is a post and talk about how Semantic interoperability could help in bringing-about the Internet of Things.


Click the picture below to watch my 2011 talk (about 20 mins), or browse the bullet-points below.

Convergence in the Internet of Things

“Information Wants To Be Free. Information also wants to be expensive. Information wants to be free because it has become so cheap to distribute, copy, and recombine—too cheap to meter. It wants to be expensive because it can be immeasurably valuable to the recipient.”

Stewart Brand, 1984

Embedded Data and Control plus Ubiquitous Connectivity = Kaboom!

  • Data … escapes its silos. Becomes Live.
  • Devices … can become “Smart”

But thinking top-down…

  • How does the recipient get value?

–      Convergence: e.g. Mashups

Value Pyramid:

  • Data at the bottom: Of many types, collected live and sitting in a data warehouse.
  • Information can be inferred from the data
  • But we need to get all the way to Value – revenue or utility (usefulness).

For example, a set of data showing the temperature in your house is of little use by itself. But you can infer information from it, for example how fast your house cools at night. And then you can drive value from that information, for example by installing insulation.

How does this map onto possible business models?

  1. In the Internet of Things we are often tempted to think bottom-up, of hardware products. These offer revenue but quickly become commoditised (especially given the huge volumes that most IoT applications have).
  2. We can sell software to turn that hardware-gathered data into information. Software has the great advantage of scalability – you can “write once, sell many”. But once everyone has a copy of your app, you’ve saturated the market.
  3. Which leads us on to the third model, which is one of service provision. I think this is probably the model which has most lasting value in the Internet of Things.

So far the early players in the space, including AlertMe, have been forced to build the entire infrastructure themselves, from end to end. Building the products to collect the data, storing it, turning it into information, presenting it on user interfaces. So a complete horizontal delivery platform. Which is fine, but it involves a lot of heavy lifting. In the future, vertical players will deliver small slices of this. A great early example is Pachube, which is “just” a cloud storage service for the Internet of Things. They leave the hardware and data-collection to other people.

Embedded processors are core to the Internet of Things, and the winner in this space is ARM. Observe their share price compare to Intel over the past five years.

Using AlertMe as an early example of the Internet of Things, what are the qualities that a consumer might expect to see of an IoT application?

  1. Live 2-way
  2. Intelligence spread seamlessly
  3. Easy to install and care for
  4. Ignore the hardware and “live” the benefits
  5. Subliminal User Experience (not modal)

Get all this right and it “just works” as if by magic.

  • From Horizontal to Vertical
    • Early players having to do some “heavy lifting”
      • Build full horizontal services
      • But future growth will be more emergent
        • Vertical specialists
      • Putting devices online is necessary … but not sufficient for IoT
        • Transition:
          • From 1:1 to 1:many
          • From M2H to M2M
      • Problems:
        • Re-inventing the wheel
        • Burden of multiple platforms & gateways
        • 1:1 thinking
  • Need Management, to make IoT seem simple
    • Despite limited capabilities of end-devices
    • Despite unreliable/high-latency connectivity

Barriers to the Internet of Things?

  • Doesn’t require any more rocket-science
  • Does require practical implementations

–      Cost, comms standards… => usability

–      E.g. chipsets for ZigBee, RFID, NFC etc.

  • Is Emergent

–      The more we use it, the more useful it becomes

  • The Key challenge? Usability.

–      It must “just work”

Define the Internet of Things?

  • “Real-world agents, autonomously expressing their capabilities and needs, and getting those needs met, to the benefit of us all


Leave a Reply