For half a decade, the predictions for growth in the Internet of Things and Machine-to-Machine markets have been staggering:
- 2010, IBM: A world of 1 trillion connected devices by 2015.
- 2011, Ericssons CEO, Hans Vestberg: 50 billion connected devices by 2020.
- 2013, Cisco Systems: 50 billion things will be connected to the Internet by 2020.
- 2013, ABI Research report: 30 billion by 2020.
- 2013, Morgan Stanley report: 75 billion devices connected to the IoT by 2020.
- 2014, an Intel infographic: 31 billion devices connected to Internet by 2020.
- 2014, ABI Research updated report: 41 billion active wireless connected devices by 2020.
- 2015, Gartner Research: 9 billion connected things in use in 2015 and will reach 20.8 billion by 2020.
Although the specific predictions and the numbers differ, what is remarkable is that the numbers predicted for 2020 have been consistently extremely large over the years. The IoT markets are experiencing explosive growth around the world, and the numbers are still performing at what Gartner calls the peak of inflated expectations in its well-known Hype Cycle diagrams. The Gartner Hype Cycle showed the Internet of Things peaking on this curve in 2014, so were moving beyond mere hype into reality.
But how realistic are these massive numbers? Even the most conservative prediction Gartners 20.8 billion connected things by 2020 is predicated on a steady 30% annual growth. Ciscos oft-reported 50 billion connected things is dependent on linking up tires, roads, cars, supermarket shelves, and yes, evencattle by 2020, according to the companys blog. No one can know if either of these things will happen.
Why the Big Numbers Are Hype
First, adoption rates for new technology tend to spike over an initial introduction period and then trail off. Consider the smartphone market in 2014, worldwide sales passed one billion units, and yet in the fourth quarter of 2015, worldwide sales growth was the slowest since 2008, at 9.7%. Its difficult for any technology to maintain spectacular growth over the long term, as markets reach saturation, standards change and evolve, and applications are enhanced to take advantage of new advances in processors, sensors, and communications chipsets.
Second, while the Internet of Things is a broad category, suggesting that we could slap an IP address and sensor on every possible thing on planet Earth is honestly stretching the concept. It is true that IoT connectivity could enhance most any industry, but just because it can, doesnt mean it will and not necessarily by 2020. Some business and people wont want to connect any time soon, and, more importantly, some cant connect their things yet. According to the United Nations, as of 2014, only 40% of the worlds population has Internet access. There is not an Internet of Things without the Internet, and counting only the things is putting the cart before the horse.
But What If the Numbers Are Real?
Lets say those huge numbers are realistic. If an estimated 6.4 billion things are connected today, as Gartner estimates, weve got four years to connect another 14.4 billion to 43.6 billion. How will this work? What do businesses, government, and standards organizations need to do to prepare for this growth? Turns out, things dont just connect themselves, and the impact of that many connected devices is not subtle.
As we mentioned, connectivity is not yet ubiquitous around the world. To get more things online, we need to expand cellular or perhaps satellite connectivity across the vast, mostly rural spaces. The U.N. reports that over 90% of those not yet online are in the developing world, and getting these people (and their things) connected will be a challenge for businesses and governments.
Certainly, the benefits of IoT technology are already being proven for these distant areas. The organization Sweet Sense has teamed with government and NGOs to put IoT sensors on water pumps in rural Africa. This enables the NGOs that install the pumps to track the pumps functionality. In a Rwanda study, only 56% of the water pumps were working consistently. After adding the Sweet Sense technology to track the pumps function via cellular IoT systems and analytics, the water pumps were able to be repaired more quickly, and 91% of the pumps could be kept working on a regular basis.
With projects like this, IoT connectivity can help provide clean water more days out of the year for more people however, these systems can be counted in the hundreds of devices right now. Theyve yet to scale up to the enormous numbers in Gartners or Ciscos predictions.
Can We Scale Up to 50 Billion Devices?
Scalability is at the heart of this problem. Even if all these people and places can be connected with all these things, that will generate massive amounts of data. Take the example of a relatively simple cellular IoT/M2M device that generates a mere 4MB per month (far less than most cellphone data monthly plans). If the world does reach 50 billion devices by 2020, then the number of cellular IoT/M2M devices would be about 2.5 billion (given that these are likely to be about 5% of all connected devices). The data this small subset of all connected things would generate is a whopping 10,000,000 terabytes per month or 10 exabytes of data. And again, thats about 5% of all the data that 50 billion connected things could possibly generate.
How will all this data be transported? How will it be stored? How will it be analyzed? How do you search through it? How will it be kept secure and private? Each of these issues must be addressed. Finding timely, actionable information within these vast data stores could be difficult and expensive. The costs of metered transport, storage, analytics, etc., have to be accounted for by industry, and distributed processing of information is vital.
Consider data security. When large numbers of things are creating and disseminating increasing amounts data, security measures have to be built into the process from the start. How do you do that? Today, the news regularly reports on breaches of financial information or hacked devices from baby monitors to connected vehicles. If 50 billion devices are connected by 2020, undoubtedly some of the data they generate will need to be carefully secured, whether its financial, health, or other personal statistics.
Further, what if a problem is discovered in a system that was considered secure, and you have billions of these connected things out in the field. Just picture any hack or security bug requiring such a fix, and imagine it scaled up into the billions with a variety of these devices, small and large. The time and expense required to keep these billions of connected things secure is staggering.
We dont really know if the world will reach 20 or 50 billion connected things by 2020, and frankly, we are not prepared for it. Industry needs to scale up, piece by piece, and build the Internet of Things as a whole. The connectivity must be available globally, the data storage must be distributed carefully, and security cannot be an afterthought it must be designed in from the beginning!