If the IoT isn’t just hype, then how is it different from M2M? Why are decoupled devices and permissionless innovation key for the IoT to succeed?
In a previous blog How Will the Next 50 Billion Devices Shape Our Future?, Krisztian Flautner shared his thoughts on the big Internet of Things trends. Now, it’s time for a deep dive into what is really happening in the IoT world. What technology, standards and tools must be in place for the IoT ecosystem to grow and flourish? To answer these questions, I’ve turned to Zach Shelby, Director of Technology at ARM. What follows is a portion of that interview.
Blyler: Is the Internet of Things real, from a software developer’s point-of-view?
In the late 90’s as Machine-to-Machine (M2M) technology evolved, developers realized that the embedded bus wasn't enough and a larger communication network was needed for embedded devices. [Editor’s Note: M2M, a precursor to the IoT, communicated with remote devices and applications via a network instead of a local bus.] M2M companies created highly customized networks with special radios, power lines, protocols and the like. Later, these companies began using the Internet but still maintained most of the customized hardware and software.
Today, most M2M vendors use the Internet but still maintain proprietary protocols, data formats, security, etc. It is really an M2M of silos.
Blyler: As long as you stayed with a given M2M vendor, then your embedded device worked great. What was the downside? Were you restricted to a particular vendor with special drivers and firmware?
Shelby: Yes – your deployment was done in a highly siloed way with tight coupling between the embedded device development and the specific back-end application and infrastructure. The entire M2M industry has grown around this siloed mode of operation.
By contrast, the IoT movement is not about using the Internet as a special proprietary network but by using all of the technologies, infrastructure, and services of the Internet to connect things - plus IoT is really about services. What do we want to do with things as human beings? We want things to participate in services, e.g., managing the energy efficiency and operation of our street lights. Also, we want those things to make our home safer and more comfortable, such as with the Nest thermostat. That is a particularly good example since the Nest is an intelligent device, local but still in the cloud. It is about things on the Internet participating in services. That is different from the older M2M siloed business model and technology. IoT is a new thing, regardless of all the hype.
Blyler: Will the value of the IoT be mainly in services? What will enable revenues and hence real growth in IoT?
Shelby: One very interesting trend is that 80% of the actual value of the IoT will come from the software application. That is the service that will be generating real value. The other 20% of value will be split between the devices, the network, and the middleware - but most of value will be in applications. [Gartner: “Half of this activity will be new start-ups and 80 percent will be in services rather than in products.” ]
The value of the application is also what we see today on the internet Infrastructure network vendors, people creating web servers and the like do generate value but its small compared to the e-commerce and everything else that runs on the Internet.
Blyler: Isn't that just another way of saying that the hardware is becoming commoditized?
Shelby: Hardware is becoming commoditized and yet it can be a differentiating factor. We are seeing a revival in cool hardware, products that look great and hook into services nicely - but hardware is not where most of the value comes from, even though it may help to have cool hardware that you give away to help sell the service.
The other big IoT trend is the importance of start-ups. Gartner predicted that, “by 2018, 50% of IoT solutions will be from start-ups less than three years old." In other words, start-ups that don’t exist today will represent 50% of the market in just three years! These start-ups will emerge from the ideas that creative people are having right now. That’s why it is said that the IoT is about new innovation just like the web was in the late 90's. It was about who could bring out the coolest, most interesting ideas.
The critical aspect of that trend is that we must enable "permissionless innovation" in which innovation can flourish with a minimum of authority. This approach was why the web and the Internet were so successful. Almost anyone could put up a web server, spit out some HTML and every browser in the world would work with it. The technology let you innovate in your own way - you didn't have to ask anyone for permission. Imagine the challenges if every developer who wanted to put a new feature on a web page had to get authority from Apache or even the W3C. Had that been the case, the Internet would never have grown. We have to achieve this same creative environment with IoT, where start-ups have free or inexpensive access to basic IoT technology and services.
Blyler: The two big IoT trends will be the financial importance of applications and the dominance of start-up companies. What needs to happen for the trends to be realized?
Shelby: Since a sizeable amount of the value of IoT will come from applications, we must make the development tools, communications and middleware technologies as generic as possible - and they need to be scalable otherwise the people creating the new applications won’t have a real business. There will be little growth in chips, in communications, in the networks themselves unless there is this massive amount of these devices.
The proliferation of start-up companies will come from a "permissionless innovation" approach. To make this happen, you need to have the technology ready and into the hands of the developers in the right way - and that leads into a talk about open standards. (Continued in Part II to be published next week)