Why 6G needs to build on 4G
In “The 5G Myth” (2016) I said that 5G as proposed back then, with its tens of thousands of new cells, sub 5ms latency, mmWave spectrum and more, did not make economic sense and that the MNOs would not deploy anything like that, instead they would deploy something closer to 4.5G and call it 5G.
I would argue that this is broadly what has happened. Firstly most – over 75% - have not implemented a 5G core, but instead are using NSA which runs on the existing 4G core. That percentage has hardly changed for over two years and there are no signs of a widespread move to SA.
Recommended by LinkedIn
Secondly, I’d argue that the 5G radio interface is just an enhancement of 4G – more 4.5G than 5G. This is because it uses the same basic technology – OFDM – and similar framing structures. 5G does enable broader bandwidth and shorter timeslots, but these are evolutions rather than revolutions and do not deliver material improvements in spectrum efficiency.
Thirdly, the network has not materially changed. Most MNOs have deployed 5G RAN on existing cells rather than building a much denser network.
Finally, 5G has not changed the services offered by the network, or any material aspects of the MNO business, except in those countries, primarily the US, where FWA has been a success.
Broadly, 5G has delivered enhanced capacity, which was much needed, and a higher headline data rate which has been used in a zero-sum PR game. If we had called 5G “4.5G” and provided the extra spectrum then we could have achieved pretty much the same result.
Does any of this matter? It does in so much as those working on 6G should consider that the starting point is 4.5G. It may be that not all MNOs decide to adopt SA so 6G should be able to work with NSA. The base station density will not be higher, so any new frequency bands (if needed) must work with the existing grid structure. Few of the 5G objectives, such as 1ms latency, have been met, so 6G should not aim to improve 5G in these areas but understand why metrics were not met and whether they are still important.
I have set out my own views of what 6G should be in “The 6G Manifesto” which work well with a mix of 4G, 4.5G and 5G networks and a mix of NSA and SA architectures. And the 6G Reset initiative is fostering a community who can think about 6G not as 5G-on-steriods but as a sensible evolution from where we are now to something that is economically viable and delivers what consumers want.
Wireless & 5G Tech Expert (vRAN, MEC/Edge, IoT, Sync etc). Independent consultant. M2M/IoT CSD Innovator and Co-founder, Ex-Ericsson & Ex-Huawei.
4moAgree. I guess #5G and #5GSA will fully mature when Enterprise solutions becomes successful and profitable to the industry (ie cost efficient). It's not just to swap to a new phone, as with the current consumer in #4G.
Telecom Expert (10+ Years) | Tech Lead & Telco Software Engineer (Python/Golang) | Solution Architect (4G/5G, eSIM & IMS) | Driving Innovation in Private & FWA Networks
4moGreat article, William Webb. I’ve been working on deploying and coding 4G and 5G SA networks for a while, and I’m always struck by how much of the procedures, protocols, messages, parameters, and logic are essentially the same across generations—just rephrased. While I truly enjoy working with mobile networks, it’s frustrating to spend so much time redeveloping what often feels like the same thing under a new name or with barely justified incremental changes (slicing, anyone?). I wish we, as an industry, could lead a truly customer-centric evolution. It feels like we only go backwards, though—it could just be a song.
I think you are looking at this with a heavy incumbent MNO and Consumer centric view. 5G was defined to be so much more, and have flexibility to cope with private wireless solutions, and the not-traditional big macro CSP's. It would be an error to define 6G on 4G technology. I can spin up a 5G SA core in 4 hours, suspend and resume it; add and remove slice-types. I don't want to go back to old EPC tech for sure. Only legacy MNO's have them. More simplified and lower cost chipset's for 5G SA is needed to drive growth.
I really don’t agree that it makes sense to build 6G on a 4G core. The 5G core is based on a modern service based architecture and it would be straightforward to add functions. Meanwhile forcing an even longer lifetime for 4G spectrum arrangements means lower efficiency.
Move fast or get left behind. Model - Measure - Act - Repeat. Then automate.
4mo"If we had called 5G “4.5G” and provided the extra spectrum then we could have achieved pretty much the same result." This sounds familiar, and AT&T did something similar - light up 4G and call it 5Ge.