4. Processing capacity in the devices
Addressing spectrum efficiency and availability in the previous posts I mentioned over and over the need for processing power to both exploit the spectrum (efficiency) and taking advantage of available one. For the former the transmitting/receiving device needs to use multiple antennas and combine/process the signals received (or the ones that need to be sent), for the latter the available spectrum lies in higher frequencies (the “good ones” are already taken) and you need more processing speed to tackle them. Additionally, there is more and more the need of managing multiple bands in parallel and this multiplies the need of processing capacity.
There are, obviously, two parties involved in the communications: the base station and the device (like your smartphone…). The base stations are controlled by the big guys, the Operators, that in principle can afford to invest money for advanced processing capabilities (although they also love to spend less !). Additionally, the base station and related antenna are industrial grade equipment not constrained by size and that can use bulky heat dissipation systems. The user device is quite a different story: it has to be affordable to the average user, it has to be small in size and it cannot dissipate too much (also because you don’t want to recharge the battery every half hour!). Because of the higher constraints of the device, this is the one we have to look at to understand the (possible) evolution.
This crucial role of the “terminal” is not something we discover now. It has been like that since the very beginning.
Today’s 4G phones would have required a truck for carrying them and their power supply (the first 3G “phone” that I saw was actually a pile of racks carried by a van) if they were to be built using the electronics of 30 years ago.
It is the evolution of electronics (Moore’s law) both in terms of performance and in ever lower power consumption that enables current wireless systems.
Power consumption and performance do not go hand in hand, even though we have seen the former decrease and the latter increase almost in “synch”. It is a matter of compromise. Apple decided to deliver the iPhone 11 without the 5G both because of the lack of 5G coverage and because the 5G chips that were available last year would have drained the battery much faster. In just one year the situation has improved and you can get a 5G chip whose power requirement compares (still a bit more power hungry) to a 4G chip.
Managing the lower spectrum bands (up to 3.8 GHz) and the higher -mm waves- spectrum bands (25-75 GHz) is not possible right now in terms of cost and power consumption. The 5G smartphones that you see on the market operate in the same spectrum bands of the 4G (hence they cannot deliver the hyped Gbps speed – from an engineering point of you, they can from the marketing point of view!).
If we cannot make today a fully compliant 5G phone, just imaging making a 6G phone with 10 times higher spectrum bands and massive MIMO, plus supporting all the node functionalities in addition to the terminal functionalities. However, in ten years time we can expect that electronics will be able to support (at least part of these) requirements (for a full support of what researchers imagine about 6G it will take some 15 years more).
The reason why we can be confident is the evolution record of terminals. One of today’s smartphone could control the NASA Apollo program, imagine controlling the huge Saturn rocket, the LEM with your phone! It is faster than the Cray-2 supercomputer of the 1980ies, and it is even faster than the computer used by NASA for the Orion spaceship for the Mars mission. Amazingly, all this performance rests in the palm of your hand and costs a few hundred bucks.
6G is likely to require a hundred time the processing capacity of today’s smartphone and even more critical it will require much better batteries since it will be using them for much more time (to communicate with other devices even when you are not using it, in order to establish a local processing fabric).