So what will happen next?
I closed the (skimpy) overview of the present of telecommunications emphasising the growing role of data and AI. These are the key technologies that will transform communications in the next decades.
Talking of decades, we have seen a basic cycle of ten years in wireless “evolution”. It took ten years to move from GSM to 3G, ten more from 3 to 4G and ten more from 4 to 5G. We can expect this to continue which will place the full deployment and mass adoption of 6G around 2035.
Why am I focussing on wireless networks? Well, that is because wireless will dominate the landscape and not just from the perspective of the users that like to be untethered, but most importantly from the perspective of investment and pace of evolution. We are already seeing today that many people are dropping their fixed line subscription and even those that are holding on it are using it mostly as a data link (accessing via WiFi or Ethernet).
From the perspective of investment, wireless is characterised by a shared investment and scales much more gracefully (it is a shared investment since users are buying their own smartphone and this is actually taking over part of the functions that used to be provided by the network, hence supported by the Operator’s investment; it scales more gracefully because with an antenna and base station you cover a large area and you, as Operator, will need to invest in a new antenna to sustain a growing traffic but that means -in principle- that you are getting more revenues whilst on fixed network you need to deploy the last mile in a pervasive way to reach not actual customers but potential customers that will not generate any revenue until they become actual customers).
6G will transfer, likely , even more the investment to the edges, to the customers, it will be a network made up bottom up, actually a meshed network, whilst present networks are designed, deployed and “funded” in a top down way, by the big guys (Telecom Operators and Telecom Manufacturers with a few “big ones” steering the evolution). For 6G we are likely to see the burden of network building shifting to the edges, to big department stores, to municipalities, to industries building a network for their manufacturing plants (we are seeing the first signs today: Siemens is starting to set up private 5G networks in manufacturing plants), to automotive industry delivering cars that are network nodes and capable of creating dynamic mesh networks. On a microscale, smartphones and IoT will become nodes and gateways creating networks at the edges.
This is going to change the rules of the game and the relations among the various players mudding the water and fading away the boundaries among providers and users. Networks of the future, at least the ones being perceived by the end users, will be created bottom up, from the edges, will be much more flexible than today and network resources allocation will be determined mainly through autonomous decision taken at service level. Hence quite a different landscape compared with what we have been used for hundred and fifty years of telecommunications.
Notice that envisaging a network built and controlled from the edges does not mean that the core network will cease to exist. Quite the contrary. The core will keep growing to offer cheap and huge pipes supporting the ever increasing demand of traffic but it will be, basically, a network connecting data centres, huge ones and smaller ones, billions of them. From the network viewpoint voice will no longer be a service to be provided. Today’s network are rooted in a design catering for voice, tomorrow they will be designed for data.
Just as I was closing this post Gabriele Elia, a researcher and telecom innovator I had a pleasure to know for many years, pointed me to an article with the latest data on the growth of ICP (Internet Content Provider). As shown in the graphics (source Ovum) their spending is now exceeding the one of the major Telecom Operators meaning that evolution in the optical network equipment will be steered by them. The telecom network is indeed becoming a data network interconnecting data centres.