I have written before about the inflation of claims in wireless technologies, and so I have to ask what these groups are talking about. There is nobody in the world today that is delivering wireless that comes close to meeting the 4G specification. That spec calls for the ability to deliver 100 Mbps to a moving vehicle and 1 Gbps to a stationary customer. What is being sold as 4G everywhere is significantly slower than those speeds.
For example, OpenSignal studies wireless speeds all over the world. In February 2014 they reported that the average speed of US 4G networks was only at 6.5 Mbps in the second half of 2013, down from 9.65 Mbps the year before. The US speeds have rebounded some in 2014, but even the fastest 4G networks, in Australia, average only 17 – 25 Mbps. That is a long way from 1 Gbps.
Moreover, there aren’t yet any specifications or standards for 5G, so these announcements mean nothing in since there is no 5G specification to shoot for. The process to create a worldwide 5G standard hasn’t even begun and the expectation is that a standard might be in place by 2020.
I am not even sure how much demand there is for faster wireless networks. It’s not coming from cellular data for smartphones. That business in the US has been growing about 20% per year, or doubling every five years and it’s expected to stay on that pace. New demand might come from the Internet of Things, from devices that want to use bandwidth from the cellular network. IoT usage of cellular networks is new and, for example, there are utilities now using cellular bandwidth to read meters. And while industry experts expect a huge increase in this machine-to-machine traffic by 2020 I’m not sure that it needs greater speeds.
The other thing we have to always remember with cellular traffic is that it handles only a tiny fraction of the total data used in the country today. Reports from Sandvine have shown that cellular traffic only carries about 1% of the total volume of data delivered to end users in the US today, and landline data usage is still growing faster than cellular data. This is probably due to the expensive data plans that cellular companies sell and which have taught customers to be frugal with smartphone data. But it’s also a function of the much slower speeds on 4G compared to many landline connections.
Another limiting factor on 4G, or 5G or any G getting faster is the way we allocate spectrum. In the US we dole out spectrum in tiny channels that were not designed to handle large data connections. Additionally, any given cell site is limited in the number of data connections that can be made at once.
So I am completely skeptical about these announcements of upcoming 5G networks. I am still waiting for a cellular company to actually meet the 4G standard – what we are calling 4G today is really a souped of version of 3G technology. It’s very hard to foresee any breakthroughs by 2020 that will let cell sites routinely deliver the 1 Gbps that is promised by 4G. My guess is by the time that somebody does deliver 1 Gbps to a cellphone that the breakthrough is going to be marketed as 10G.
I don’t think that any of the groups that are promising 5G by 2020 are anticipating any major breakthroughs in cellphone technology. Instead the industry is constantly making tweaks and adjustments that boost cell speeds a little more each time. All of these technology boosts are significant and we all benefit as the cellular network gets faster. But the constant little tweaks are playing hell with handset makers and with cellular companies trying to keep the fastest technology at all of their cell sites.
We are not really going to get a handle on this until we have fully implemented software defined networking. That is going to happen when the large cell companies migrate all of the brains in their networks to a few hub cell sites that will service all of the cellular transmitters in their network. This means putting the brains of the cellphone network into the cloud so that making an update to the hub will update all of the cell sites in the network. AT&T and Verizon are both moving in that direction, but it might be a decade until we see a fully cloud-based cellular network.