At Heavy Reading -- and here on the pages of Light Reading -- we've been following the emergence of 5G closely because, quite simply, it's a technology that will change the world. It has the potential to revolutionize entire industrial sectors, it heralds the Tactile Internet, it promises a connectivity layer for vehicle communication, and so on… it's a great story and we’re "all in" with the vision.
And progress, so far, has been good. Earlier this year the Next Generation Mobile Networks (NGMN) Alliance's 5G White Paper established a framework for R&D and market development. And the International Telecommunication Union's (ITU) IMT-2020 process -- the ultimate arbiter of what 5G actually is -- to define the performance requirements is ongoing. However, much of this work has been, in essence, aspirational: What do we want? What could we conceivably develop? At some point aspiration must translate to the real world.
That time has now come. At a special 3GPP RAN workshop starting tomorrow in Phoenix, Arizona the industry will formally start work on an implementable 5G standard.
It also gives us some clues about what 5G might actually look like. Using the documentation posted by operators, vendors and R&D organizations to the 3GPP website in advance of the workshop, we get a first look at how important players view the development process and, by extension, what type of 5G the industry is likely to produce.
So, what can we glean from this document dump about the 5G development process?
(Keep in mind that the 3GPP won't use the term "5G" and that the intent is to develop a standard that meets the requirements for IMT-2020 technology, which is still to be formally defined by the ITU.)
The first thing that becomes apparent is that 5G development will span at least three 3GPP releases (R14, R15 and R16) and will be delivered in Phase I, which could be commercialized by 2020, and Phase II, from 2022 onwards.
Phase I is being driven by the operators' desire to demonstrate prototype 5G systems, with a new, high-speed air interface by 2018, and launch the first commercial services in 2020. Even though it won't meet the full IMT-2020 requirements, this technology will, inevitably, be marketed as "5G."
Phase II won't be specified until Release 16 (currently expected at the end of 2019) and probably not commercial until 2022. This is closer to the 5G story the industry has been selling (True 5G?) and by the time it exists will probably be known as 5G+, 5G Advanced, or similar. We all know how this works.
An indicative timeline for the 3GPP process is shown below:
What is Phase I 5G (pre-5G?) likely to look like?
Well, probably not the Tactile Internet, network slicing, remote surgery, massively scalable sensor networks, self-driving cars, factory floor automation and all the really transformational stuff -- that is all being pushed to Phase II. Much more likely is that Phase I will be about enhanced mobile broadband (known as eMBB) -- or, in other words, what we have today, but faster and more power efficient.
This phased approach makes sense -- the 5G goals are extremely ambitious and the industry obviously can’t do everything all at once -- but it does mean the original goal of developing major new commercial opportunities will have to wait. In this scenario, operators will be more or less restricted to essentially the same business model of selling gigabytes by the bucket until Phase II.
Quite few of the operator and vendor presentations posted ahead of the event express concern about this (and this is more acute in private in "corridor conversations") because they fear the opportunity to transform the service model will be unnecessarily delayed if the sole focus of Phase I is on eMBB.
In my view, therefore, two really crucial objectives of this week's kick-off meeting should be: (1) That Phase I is forward-compatible with the new capabilities earmarked for Phase II; and (2) That Phase II development work on more advanced capabilities starts right way and runs in parallel with Phase I.
So with that as context, here's my short summary of what Phase I might look like based on the 3GPP document submission:
- As already noted, it appears that Phase I 5G will be driven by enhanced eMBB with a focus on small cell access. (With some desire to also support macro cells with inter-site distances of up to 500 meters.)
- Also as expected, the focus on sub-6GHz spectrum, with 3-6 GHz considered the sweet spot. There are quite a few players that would like Phase I to extend to 30-40 GHz.
- Everyone expects to use wide carrier bandwidths with request ranging from 40MHz up to 100MHz, or even 200MHz.
- The expectation appears to be that 5G will not operate in a standalone mode in Phase I, but will be anchored by an evolved LTE network and connected to an evolved 4G core. This is called LTE Assisted Access and appears similar, in principle, to LTE-U/LTE-LAA.
- Hyper-scale IoT (known as massive Machine Type Communication or mMTC) capability looks likely to be pushed to Phase II. One prevalent view is that LTE-M (R14) can do the job in the interim, although some influential organizations would like to see mMTC in Phase I.
- Overall, there appears to be support for a single air interface that will eventually also support ultra-scalable mMTC and ultra-low latency. But clearly some players are open to a multiple air-interface strategy (e.g. for very high-band spectrum).
- There is not yet a consensus on the choice of waveform. OFDM with flexible numerology seems to be the most popular, but support is by no means universal.
- There are quite a few proposals for a new waveform, or various forms of filtered OFDM, but no real consensus among the proposers. Some also argue for an additional single-carrier waveform (i.e. non-OFDM) for the higher and very low bands in Phase II.
- There is agreement that the focus will be on dynamic TDD with a new frame structure; not many calls for FDD support.
- Broad agreement, in principle, on lightweight connection and mobility management, and on unscheduled transmissions. Very little detail on what this might mean in practice.
- Lean design (blank sub-carriers, blank sub-frames, etc.) seems to be the proposed solution for Phase I to Phase II migration. Again, there's relatively little detail on how this might work.
This is all, obviously, subject to change -- that's the point of the workshop after all -- but it does I think give a pretty good overview of the state of 5G in the real world as it stands today.
— Gabriel Brown, Senior Analyst, Heavy Reading