Looks like there is quite a difference between the two numbers, to get a better idea would you like 20 euros or 1,000? Hence, it is difficult to understand how a company like Ericsson is forecasting 50 billion “things” connected to the Internet by 20120 whilst a company like HP puts the number in the trillions.
I was yesterday at the ETIS annual gathering where I gave a talk on IoT and before me there was a guy from Ericsson who made the 50 billion forecast. Nothing new, actually. It seems to be the magic number Ericsson is pushing on all tables.
Both Ericsson and HP are well known and respected companies, so why such a big discrepancy?
I think I spotted the problem, and it is an old one, going back at least to the 80s. It is rooted in the different perspectives of telecom vs computer industry.
On the one end the world seen by the telecom companies is the one at the termination points of their network, a network that is fully controlled by them and that can be shaped to their liking.
On the other hand the computer industry sees as its world whatever embeds a microchip, be it connected or not. And of course, these days those chips are more and ore connected, even though they may be connected through networks that are not telecommunications network. Well, these latter simply are invisible to telco’s guy, they do not exist. The problem is, of course, that by taking this approach you, as a Telco, are limiting the scope of your market and in a world where volume makes a difference you are undercutting yourself.
So let’s go back to numbers: is it 50 billion or 1 trillion?
Let’s do some simple math on the back of a handkerchief. Today there are 7 billion people on the Earth. By 2020 there will be a bit more. Already today each of us have several sensors in his context.The computer I am typing on has an accelerometer, a light detector, a multitouch, a temperature sensor, my cellphone has accelerometer, a compass, a touch screen, a temperature sensor, same goes for my iPod and iPad. All of them have virtual sensors, that is it is possible to know the location they are, how much memory, how much power is used, what type of interactions are taking place and so on. Just looking at me, then, I count over 20 sensors. Multiply it by 7 billion and you get 140 billion “things”. Clearly other people may have 0 sensors (although there is close to 1 cell phone per person in the world and that alone would qualify for 35 billion sensors!). Other people, on the contrary, may have a few more.
And of course I have sensors at home (several for the anti-intrusion system, temperature sensors, weight sensors, light sensors, I counted over 50 at home and it is likely I missed several more…). Then I have sensors in the car, over a hundred of them… And there are sensors in the building I live, on the roads, in the city….
50 billion means 7 sensors for each person on the Earth. Just with the quick calculation above it is easy to see that today we are already well over the 50 billion mark.
One trillion would mean less than 140 sensors pro capita. Not that much if you think about it! And I have not mentioned all the other things that can be connected (or just logically connected) to the Internet, like the ones having a tag. Any suits, t-shirt, towel in a shop is likely to have a tag that is intercepted and identified by the shop security system.
What will happen when all bar codes will become electronics? How many trillions of objects will be related to the Internet?
The crucial point is that you can look at IoT has things that have a SIM card (the Telco approach) or as things that have a mirroring element in the Internet (a number of bits attached to them). The latter is the computer world vision and it is in this latter vision that most of the biz opportunities lie.
When a video camera is connected to the Internet via a WiFi link Telcos add 1 to their IoT, but the computer world adds 1,000 since that camera brings in images of thousands of objects captured by its lenses and identified by image recognition applications.
Eventually both Ericsson and HP are wrong in their forecast, but HP has an head start placing it 20 times ahead of Telcos… And so do SAP, Oracle….