By Serge-Paul Carrasco on August 2, 2015

Deciphering the Internet of Things With a Developer Eye

(T) When I worked at Sun Microsystems, we used to say “The Network is the Computer”…now maybe that has changed and… “The Sensor is the Computer”. It took thirty years to downsize a Mainframe computer to a Personal Computer. It took another thirty years from client/server networks to the recent mobile Web to connect all types of new personal devices over both wireline and wireless networks. Will it take another thirty years to make the emerging Internet of Things (IoT) a reality? I will let you think about that question.

I have attended the last two years, a few IoT developer meet-ups and hackathons and I have come up with a few and practical observations:

Segments

I can see three different segments based on different technologies and architectures:

IIoT (Industrial IoT or what GE calls the Industrial Internet) in that case the IoT devices are part of the manufacturing and supply chain of an enterprise
CIoT (Consumer IoT) in that case the IoT devices are for the consumers and just have RFID, GPS, WiFi and an IP protocol stack
Note that either the IIoT or the CIoT can be WoT (Web of Things) in that case the IoT devices leverage the present Web protocols, interfaces, and browsers.

Devices

I think that some of the predictions made by analysts about the number of connected devices are wrong. I think that not all devices will be connected for sure.

Do not expect your stereo system (except if it is made by Sonos), your piano, your coffee machine, your fridge, your washer or dryer to be connected in the soon future. Most likely the devices that will be connected have already some computing and/or electronics parts and short life cycles such as:

Cars
A few appliances
Smart TVs
New wearables
Smartphones and tablets

But yes, in the meantime, someone is manufacturing an IoT toothbrush, an IoT tennis racquet, a pair of IoT shoes…

Services

I think that what fundamentally is going to spark the IoT market is services. If you think about it, we are using the Internet for many services: music from SoundCloud, movies from Netflix, search from Google…

The same will be true for the IoT. Examples are for instance:

Disney MyMagic that uses a combination of a website, smartphone app and bracelets called MagicBands to track guest’s movements at Disney World. The MagicBands store tickets, hotel keys, debit and credit card information and can be used anywhere just by tapping a sensor
NFL pad trackers based on RFID that provides details analyses of the player moves on the field
Seniors’ health wearables that are connected to a call center to monitor their needs and any emergency situation
Sen.se devices for helping with monitoring daily tasks in the house

All of those devices provide a real “service” to their users. The IoT device is just a component of the total service.

Platforms

There are now many development platforms available for developers:

Raspberry Pi, Arduino, Intel Galileo and Edison
IoT extensions for Eclipse, Java…
A few protocols to choose: MQTT, CoAP, OMA LWM2M

I have run as well into two open source IoT platforms: RIOT and Contiki. And, Google at its I/O conference in June this year announced its Project Brillo, an IoT O/S based on a Linux kernel with the basic Android system layer with a communication stack Weave. Brillo will be open sourced at the end of this year.

We shall see if either RIOT, Contiki or Brillo will take any market adoption.

Players

Two very active public companies are probably GE and Intel. GE for the IIoT and Intel for its IoT developer platforms (Intel somewhat missed the mobile market but does not seem to be willing to miss the IoT market?).

Note: The picture above is from Wikipedia – Wilgengebroed on Flick.

Categories: Internet of Things