6. Network Architecture
The telecommunication network, seen from a mile above, has kept evolving towards a flatter and flatter architecture. It was very hierarchical in the beginning, a call from Siena to Philadelphia would have to be routed via Florence, Rome, London, New York to reach its destination in Philadelphia. The reason was that all routing was pre-cabled and mechanical routers (see the photo and watch, hear the noise, the clip) with no intelligence whatsoever had to route the call through subsequent steps.
This hierarchy started to flatten with the advent of computers, the so called Intelligent Network. Now a computer could look at what resources (wires) could be available and will instruct the network to allocate a path to that specific call. At that point a call from Siena to Philadelphia could follow a routing and the next one follow a completely different one.
Wireless networks (or better, networks supporting wireless terminations) appeared at a later stage when computers were already widely used and therefore were born with much flatter and flexible architectures.
Actually, subsequent “G” saw an ever increasing adoption of software and computers (the 2G, also known -in Europe- as GSM -for Group Special Mobile- was dubbed by technicians as “Great Software Monster” to emphasise the big role payed by software, in perspective a small one if we compare 4/5 G use of software to the one in the GSM!).
The Internet and IP (Internet Protocol) was co-opted in 3G and of course it is an integral part of 4/5G and, no question about it, will be part of 6G.
It was not just the network to become computerised and softwarised, devices, and in particular cellphones (now smartphones) have become computers running software. What is most interesting is that smartphones have been progressively taken the upper hand in supporting services. As smartphones got more and more involved in the support (and delivery) of services these have progressively left the (Operators’) network migrating into the Cloud. This is providing flexibility and is decoupling the network from services. More than that: it is also shifting the ownership of services (and related revenues) from telecom Operators to third parties.
This evolution is particularly evident in the wireless network (systems) to the point that we now talk about a service network and an infrastructure network (the one comprising wires, base stations, antennas…).
Actually, the “service network” is completely separated from the infrastructure network and it is completely flat: services are created independently of the infrastructure and can work on any infrastructure, any “G” can support any service, provided a sufficient performance is available.
This is great for the users, this is a nightmare for the Telcos!
- The users can use services wherever they are, independently of the network available in that particular location and no longer perceive the network, only the service. As long as the performances are sufficient a user will be happy.
- The service creators, and providers, design the services in such a way that they can work on the broadest range of networks and can operate in a wide range of performances, since they are making money from the use of the services and the more people can use them the more revenues.
- A new G (like 5G) clearly can deliver better performances but service creators are unlikely to create services that need these higher performances because that would restrict the number of users that could use the service. Much better for them to create services that can run on Gs and smartphones that are widely available.
- Telcos are facing the problem of deploying a new G for which there are basically no services that would make a difference and that might sustain a premium price. Take 5G: what sense does it make telling a user that he would be able to download a movie in a few seconds whilst on the move whereas with 4G it would take several minutes when streaming services are designed to work perfectly well within 4G performances?
The separation of the service network from the physical (classical) infrastructure has reached a point of no return, also because the service network is actually enabled by an intermediate layer, the service enabler layer (cloud, servers and much more, I’ll address this in a later post).
Whilst in the past an Operator could design services to leverage its new network features and thus provide an incentive to customers to buy onto their new network even if it was in the earlier stage of deployment, now the service providers fuelling the service network do not have any interest in investing to promote a new network (a new “G”) until that becomes widespread and can sustain a large market.
In spite of all this progress we are still far from a really flat network infrastructure and from a completely flexible one, so there is still plenty of space for further evolution.
I’ll address that in the next post.