As you know, GreenSI is a blog on which we follow subjects over the medium term. We therefore regularly return to subjects identified a few years before, to assess the speed of development and the curvatures, even the changes in trajectories.
In 2019, Amazon announced Sidewalka network for connected objects, when the 5G was promised great development. Where are we today and who are the candidates for the connected object networks of tomorrow?
Without a network, an object cannot be connected, and the growth in the deployment of connected objects has not slowed since 2019. The need for IoT networks is therefore there, growing stronger, and new networks are emerging.
In 2010, we reached 1 billion connected objects, and in 2017, Cisco predicted 30 billion in 2022.
In 2023, Strategy Analytics estimated in a study their number at 22 billion worldwide (smartphones, speakers, televisions, watches, tablets, laptops/desktops, etc.) which is less than expected, but which nevertheless confirms double-digit growth which does not not bend. The professional and connected home sectors have the strongest growth, the latter being Amazon’s territory. Today, we are announcing 38 billion connected objects in 2025 and 50 billion in 2030.
Of course, there won’t be a single network to connect them all, even if this Hollywood title makes more than one dream. Let us therefore try to understand the levers which are influencing these future winners, and thus better understand the factors of influence in the deployment of new technologies, beyond the IoT in fact.
Satellite, Sidewalk, NB-IoT, LTE-M… new physical topologies (generally schematic description of the network layout, including its nodes and connection lines) are arriving for IoT networks.
When we talk about connectivity between an object and an application, it is for example to display its state, its measurements, and to activate processing algorithms, more or less intelligent. How can I monitor the temperature in my room from my smartphone, where a connected object that measures it live is placed? And conversely, how can you act on the radiator to lower the temperature?
In the house, WiFi reigns supreme and the internet operator’s box, or a hub connected to it, allows objects, applications and terminals to be connected to the Internet. But when you move away from home, and when you reach the limit of repeaters or energy, you need a network to take over. This is where we imagine a network of antennas to listen to these objects, like those of our operators to bring us 4G or 5G, by investing in the deployment of such antennas.
But another possibility is to look higher for these antennas, with new low-orbit satellites which are the basis of the network Starlinkbut also other networks like Eutelsat OneWeb. Objects communicate with satellites, which communicate with earth stations, which are networked. But launching these satellites and managing them to operate the service requires significant investments.
Amazon’s initial idea with Sidewalk is completely different!
Instead of looking for a higher point, Amazon looks to see if there isn’t already local connectivity, a sort of “low point” to use the same image. This idea is to reuse the already existing infrastructure and bandwidth to operate the equipment Alexa (speakers connected) and Ring (doorbells, cameras) sold to its customers. Each piece of equipment sold by Amazon thus becomes an element of a mesh network. and can give you connectivity to access objects in your home, even when you are not there. It’s a bit like Apple’s AirTags that geolocate your luggage using the iPhones and bandwidth of other Apple customers. To avoid criticism of data security, the protocol for exchanges on this network is extremely secure and separate, like a VPN, from the Amazon network used for the Alexa or Ring service. Amazon customers can exit the program and turn off the Sidewalk feature on their equipment, but they are enrolled by default.
Today, this Sidewalk network is a reality in the United States and developers are invited to submit applications. It covers 90% of the American population, almost by “design”, since 90% of the population is visibly close to an Alexa device at home or at one of their neighbors if they do not have one. And this without investment in infrastructure…
For comparison, this high investment in infrastructure (and therefore the debt that had to be repaid) was fatal to the network Sigfoxdismantled after going bankrupt while it was operational and had customers.
Now, this strength is also the Achilles heel of this Sidewalk network. Indeed, for it to cover an entire territory, and not just where people live, it would be necessary to deploy millions of Alexas (and power them) in the forests or in the desert. It’s impossible. The uses of this network will therefore be more easily linked to home automation, than to monitoring a herd in the mountain pastures, or rather bison in Wyoming. This latter use will be better covered by a low-orbit satellite. This Sidewalk network is also low-speed, and 5G will be more suitable for uses requiring more speed or less connection latency.
High point, very high or very low, with an operator or live without an operator, we see that the technical topology of networks is today a differentiating factor, and that each type will be more suited to certain use cases. There will thus be several networks in the future to cover the different use cases which will finance them. The use cases bringing the most value, such as those that can generate savings for financing, will be able to afford access to networks requiring more investment. On the other hand, the low value of the data to be collected means that it will only be exploited if a network has little infrastructure to amortize, or even not at all as with Sidewalk.
But topology is not the only factor that will influence the emergence of these networks, geopolitics and the environment too.
Concerning the environment, the environmental impact of IT is distributed between data centers, the network and terminals, including connected objects. Efforts on data centers and the proliferation of connected objects mean that the latter have become the bad students. However, their number must still be at least doubled. There is therefore certainly a good chance that regulations will limit this energy consumption. This efficiency will be an element of choice for networks depending on uses.
Low-consumption networks already exist (LPWAN – Low Power Wide area network) and are deployed for objects which are not connected to an electrical network and for which we do not want to change the battery for at least 10 years. They could therefore be used more in the future, for objects that would like to be less talkative and more energy efficient. This suggests the development of those of these networks not yet widely deployed, such as NB-IoT (Narrow Band IoT) and LTE-M (Long-Term Evolution for machines), a variation of 4G specialized for IoT.
Regarding geopolitics, GreenSI wrote in 2021 about the breakup of the Internet and the fact that we now have to speak of the Internets.
Since then, Russia has adopted its strategy of disconnecting from the Internet, which China and North Korea had already done. The internet has become a militarized space where attacks are produced to disrupt public services and support other attacks in the real world. The Internet of Things will also be a militarized space.
China is massively engaged in NB-IoT. It drives global volumes (in millions in the graph opposite), which boosts the market for the manufacturing of these modules in China and the adaptation of connected objects to operate with this standard.
Even if the United States is in the standardization bodies for these networks (3GPP), the massive arrival of equipment from China to implement these networks is not to please those who put cybersecurity before technology. Chinese patents also take advantage of NB-IoT and create another dependency. The dominance of NB-IoT is therefore strongly centered on Asia and much less on the Americas.
LTE-M, coming from the evolution of operator networks, is considered less risky in terms of sovereignty, and we have seen that Amazon had not chosen NB-IoT either. Europe, under American technological influence via Big Techs, has agreements with China of which it is an important trading partner. Europe will have to reconcile the goat and the cabbage, and certainly once again end up with a patchwork of solutions, depending on the Member States…
In the end, what do we remember about the IoT network of tomorrow to support its digital strategy?
Firstly, there will be no single network, nor interoperability between several networks in the medium term. Then the dominant network will depend on where we are. Finally, the network topology, and the value chain to implement it, will guide the use cases and make connectivity costs very variable.
But another lesson is that Amazon has taken a big lead in the United States to create synergy between its Cloud services and a low-speed collection network in the Edge.