Of course, 5G is much more than the radio part and spectrum efficiency/use. It has a lot to do with network and service architecture. It is through a novel architecture that much lower latency can be achieved. It is the novel architecture that support the session management at the edges, inside the terminal. Managing the session at the terminal level means, basically, that the terminal becomes a network node, i.e. it is no longer a terminal. Add to this the possibility of network slicing, SDN and NFV potentially controlled by the terminal and you get a revolution in the way the network is managed (potentially – it remains to be seen how far current network Operators will agree to open up their network resources and hand over the control to the edges…).
As it is already happening today with the tethering function (you create a local wireless network with your smartphone to let your computer connect, via the smartphone, to the big network) we are going to see more and more wireless cells created by a device, a smartphone, a car, a watch… you name it. In the coming decades it is very probable that most appliances, devices, buildings, lamp posts … will create their own wireless cell and will open it up to other devices, effectively becoming access points. More than that. These wireless cells will overlap with one another, most of the time in a dynamic way, creating mesh networks at the edges. Depending on their location a device may both create a wireless cell and be an access gateway to the big pipes.
Radio towers are bound to multiply, as denser and denser coverage needs to be supported, and they will get smarter and smarter, aggregating data and processing power as part of their functionality (in addition to electrical power and optical fibre connectivity) effectively creating a cloud at the edge and offering edge computing (by the way, this seems to be the way to deliver one digit ms latency). This cloud will seamlessly overlap with the fog created by devices (shared processing and storage capability).
There is no question that this kind of scenario will require intelligence at the edges and intelligence in the devices. This is a challenge today given the amount of data and processing power required to infuse intelligence locally. Yet, this is the path industry is walking on. At a recent meeting with some key people in Huawei I heard them saying that intelligence in the smartphone is inevitable if you want to increase service capabilities. More than that. Having data processed locally on the device in your hand can make privacy and ownership issues easier to manage. AI is going to play a fundamental role in the future decades networks architecture. This will require improved hardware capabilities, not available today, both in terms of performance and in terms of lower power consumption.
6G, with its reliance on higher and higher frequencies, will require many more cells making a centralised investment unfeasible (from an economic standpoint but also from a practical standpoint: how can you negotiate the deployment of tens of millions antennas?). This necessarily shifts the burden to create a 6G coverage to the devices (clearly, there will be plenty of gaps in coverage if you rely only on devices and self assembled mesh networks but these may be flanked by a few cells deployed by Operators and -most likely- by public institutions, like municipalities). It does not seem so far fetched to think of wireless coverage as a basic infrastructure, so that all new buildings might be required to provide wireless coverage in their surrounding/connection to the big network pipes, as today they are required to provide parking places and to connect to waste and power infrastructures. Same may go for light poles, advertisement panels, shops and malls…
The Digital Transformation is most likely to affect our ideas of what a network is and who owns it, from something that is out there owned and managed by private parties to something pervasive that is resulting from the way artefacts and civil infrastructures are designed, deployed and managed. Communications infrastructures will shift, and morph, into communications fabrics with the contribution of many players and of users.