Got a call yesterday from a Dutch journalist who was just back from an Ericsson presentation on 5G and wanted to know about 6G. Looks like Ericsson managed to make him curious on what might the future be like…
We discussed a little bit the general evolution trends, since 6G does not exist not even in vaporware, the only sure thing is that 6G will come after 5G and the time frame, as for the previous Gs will be spaced by a 10-15 years windows.
I’ll try to summarise the thoughts I shared with him, soliciting your comments and … vision. Come on, it is a basically empty space you can roam freely with your fantasy!
I see basically two (engineering) goals in the evolution from a G to the next one:
- better use of the spectrum (increase spectrum efficiency)
- use more spectrum
Better use of the spectrum
Since Marconi’s first experiment we have continuously managed to become more and more efficient in using the spectrum, in recent decades as we moved from analogue to digital that meant squeezing more bits per Hz of available spectrum. With OFDMA adopted in 4G we have basically reached the Shannon limit, we cannot go any further in that direction. However, Shannon pointed out the limitation imposed by noise on a given channel. It did not say anything about the use of multiple channels and indeed, already starting with 4G (but this is a trick already being used in the 802. – WiFi- area) we have started to use MIMO, Multiple Input Multiple Output. Here there is basically no theoretical limit: it has been shown that if you can have n+1 receivers spaced at least a wavelength from one another they can sort out the signal generated by n transmitters (similarly spaced).
Here the possibility to increase MIMO (moving to massive MIMO and ultra massive MIMO) is a matter of increasing the processing power available (and possibly decreasing the wavelength to make it practical to use many antennas in a small form factor). As times goes on it seems that researchers find ways to increase the processing power (this is still true although the demise of the Moore’s law is making progress much more difficult that it has been in these last six decades) and if we move to higher frequencies we get smaller and smaller wavelengths. We should, anyhow, be realistic and understand that although it is likely that ultra massive MIMO will be able to increase the efficiency, practical limitation will remain.
Use more spectrum
This is by far what is happening with 5G, use higher frequencies in the 3.5, 27, 70GHz range allowing the apportioning of a bigger slice of spectrum (this has kept increasing as we made use of higher frequencies, from the single digit MHz spectrum in the early Gs to the 20MHz spectrum of 4G and up to 30MHz for the 3.5GHz band of 5G, much more for the higher frequencies, see chart).
If you allocate more spectrum, than you can transport more bits… as easy as that. Well, not actually. The problem is that the available spectrum is not a good one. The good one has already been taken by the previous Gs and the fading out of an old G (like GSM) or even the fading out of some parts of the spectrum used by analogue television is not enough so we are forced to use higher frequencies. In turns, these higher frequencies require faster processing (increasing processing capabilities) -that is manageable within limits- but they are not propagating as well as lower frequencies (it is not by chance that Marconi used very low frequencies in his first radio transmission -in the 100 of KHz range with antennas hundreds of meter long). Low propagation means that you need to have smaller cells, and that is good because with smaller cells you are going to have fewer users and therefore you can deliver more capacity to each of them. The only problem is that the more cells … the higher the cost (and the network gets more complex in managing handover).
Given these two basic evolution trends we can predict that 6G will increase the efficiency by circumventing Shannon (ultra massive MIMO) and by using higher frequencies. Indeed the research that is going on in Finland is exploring the use of sub-THz frequencies.
If return on investment on 5G is making telcos big bosses scratching their heads, having to face investment to deploy 10 times as many cells as they had to deploy for 4G, just imagine a scenario where the number of cells needs to be multiplied by 100-1,000 times…
More on next post