Shocking headlines always get attention, so already we see stories of the "forget about 5G, 6G is coming" sort. It's rubbish, of course. The material basis for building 6G do not even exist yet, so it makes no practical sense to ignore the tools we have, the tools that are on their way, in favor of tools that do not exist.
All we can say is that 6G will grow out of 5G as 5G itself has grown out of 4G.
None of that is going to stop researchers, who do not yet have much of a clue about what the final 6G platform might really look like, from starting work. They have to do so, as international consensus has to be built, enabling technologies perfected and capabilities defined, before 6G can launch, in perhaps 10 years.
Mobile networks in the digital era--starting with 2G--have been launched about every decade, though the typical useful life of any generation is perhaps 20 to 25 years. As 5G now is starting to be commercialized, research attention will begin to turn towards 6G, in seminal form.
Up to this point, it has been fairly straightforward to characterize the distinguishing features of each platform, in terms of speed and latency. Each digital generation--starting with 2G--has boosted data speeds by at least two orders of magnitude, typically in two-step fashion. Latency performance likewise has been reduced in each generation.
It is also possible to characterize each generation by key application and use case innovations.
This graph shows 2013 typical latencies on real-world networks. The basic point is that latency performance improves with each digital generation. As real-world performance always lags the design specs, 5G will not generally feature one-millisecond latency. But it would not be unreasonable to expect that 6G architects will aim for functionally zero latency, while seeking to boost effective speeds by one or two orders of magnitude.
If 5G is any indicator, the 6G standards will be a mix of specifications relating to the core network itself, but also the way apps and sessions interwork with other key networks, as the 5G spec includes interworking with Wi-Fi, and ways to handle core transport using satellites, for example.
It is also possible that 6G is created with general classes of use cases in mind, as 5G was created as the first mobile network deliberately designed to support humans using phones and machines talking to other machines.
It seems safe to predict that as small cell architectures have become necessary for 5G to support new millimeter spectrum and higher speeds, so 6G will incorporate and extend such approaches.
As 5G was developed with new core network transport, virtualization, edge computing and internet of things in mind, so it is possible that 6G will be created with other frameworks in mind.
So far, it is way too early to speculate on what such frameworks might include. But it seems hard to believe that advances in artificial or augmented intelligence will fail to play a new role, in terms of the design of the physical, network and transport layers.
Some might wonder whether internet protocol still will be the choice for transport and network layers. Others might wonder whether “devices” as such will be necessary, or whether peer-to-peer networking between embedded devices might reduce the need for cell site antennas.
About the only safe bets are that speed will increase by two orders of magnitude, that we will rely even more on millimeter wave spectrum and that latency performance will decline close to zero.
It is harder to describe how 6G might change the realm of possibility, but Nokia Bell Labs researchers believe 6G will bring some new synthesis of the physical, digital and biological realms, in terms of applications and use cases.
No comments:
Post a Comment