One thing I find quite interesting from a historical point of view is, that the first 5G networks will be based on TDD instead of FDD, at least on the 5G side of the equation.
In most parts of the world, all previous and current mobile network generations are based on Frequency Division Duplex (FDD), i.e. there is one channel for the downlink and a separate channel for the uplink. However, there are huge spectrum allocations for Time Division Duplex (TDD), especially in the US and China. This is why a few years after LTE networks started mostly as FDD networks, we also saw LTE being deployed in a TDD variant in those geographies. For completeness sake, there are also a few TDD deployments in Europe, e.g. in Sweden, but that is so far the exception rather than the norm.
So this was the ‘order of things’ in previous network generations. Not so in 5G NR! Here, there first EN-DC deployments that will use 4G LTE as an anchor will use 5G in the TDD variant first. For those network operators deploying 5G in 3.x GHz, 3GPP band n78 will be used, which is TDD. And for those network operators going for mmWave, TDD is also the only option. Of course, there are also 5G bands defined for FDD operation, mainly those in which LTE is already used today, probably required for peaceful coexistence and regulatory reasons. But from what I can tell from press reports, those will not be the bands used at first. Everybody is focused on either 3.x GHz or mmWave to get that significant speed boost out of 5G.
How interesting, TDD, the ‘ugly duckling’ of cellular wireless will finally get its prime time with 5G globally!
5G is going to use technologies such as Massive MIMO and beamforming, and, as far as I know, those technologies are easier to implement with TDD rather than with FDD.