Huawei fleshes out “4.5G” ahead of commercialisation in 2016

LTE-M, MIMO and cloud edge to bring capacities up x3, cut latency by 80% and increase connections per cell x100.

It’s the end of his speech on 4.5G, and Huawei’s VP of Marketing, Wireless Network Product Line, Bob Cai, is wrapping up. He is happy that he’s made the case that 4.5G represents a series of achievable upgrades to deliver a) Gbps capacities b) 5-10ms end-to-end latency and c) the ability to connect thousands of devices per cell. “4.5G is not about marketing promotion” he says. Which just makes everyone in the room think, “Uh-huh?”

So what is 4.5G, as Huawei sees it, that takes it beyond being a marketing promotion? (Seeing as Huawei is promising commercialisation in 2016 it must have a pretty clear idea.)

Here’s what Cai says. 4.5G will involve a series of incremental steps, many of them being laid out within 3GPP R13 and R14, and the odd bit from R12, with the goal of going further than LTE-A. Where LTE-A has 300Mbps with Carrier Aggregation, 4.5G will provide Gbps capacities. Where LTE-A has a few tens of ms latency, 4.5G will have 5-10ms. Where LTE-A can handle 1,000 users per cell, 4.5G will have 100,000.

The idea is to enable immersive experiences. Where video is the key app for LTE, remote control automation and Virtual Reality will be the driver for 4.5G. Factories will be staffed by connected robots, humans (the few left with jobs, perhaps) will put on face masks and do imaginary jobs, drones will fly overhead providing connectivity

So how will this be achieved?

Essentially, according to Cai, 4.5G is mostly seen in 3GPP’s R13 and R14 schedules, “with a little bit of R12”. So we are looking at higher orders of MIMO (x4, x8 etc) and at higher modulation (256 QAM), at increased Carrier Aggregation, and at LAA (previously known as LTE-U). There will be a Cloud Edge – the distribution of intelligence through the network to bring down latencies.

These features will start to come on stream in a commercial manner from 2016, Cai says.

Importantly, however, 4.5G will be compatible with operators’ current LTE investments. (5G is too far off, operators cannot afford to strand their current investments.) However, Cai says that delivering these capabilities will also require a new interface.

LTE-M and Cellular IoT

Part of the new air interface is the support for LTE-M, the low power Machine Type Communications version of LTE that can theoretically connect tens of thousands of devices with very long battery life. Huawei reveals that there is chipset support now for LTE-M from its new acquisition Neul, and it shows a module based on that chipset integrated in smart meters and health monitoring wrist-bands.

Another part of the new LTE-M Cellular IoT air interface will be the ability to operate in a much narrower spectrum band.

Cellular IoT as defined in 3GPP R12 and 13 envisages a new class of device (Category 0) and a reduction of the possible receive bandwidth from 20MHz in R8 to 1.4Mbps in R13. This gives an uplink peak rate of just 200kbps. There are those who think commercialisation of R13 LTE-M specs will not come until at least 2017, so although Huawei is on the same path, it appears to be moving down it quicker.

Going even further, Cai tells TMN that Huawei will also enable narrowband LTE M2M deployment with a system bandwidth right down to 200kHz. This narrow bandwidth makes it easier to find available frequencies in the sub-1GHz spectrum that is better suited to wider coverage. One potential option could be using a guard-banded 200kHz GSM band. But moving Cellular IoT to such narrow bandwidths requires new synchronisation signals on the downlink and new random access preamble signals for the uplink. In other words, if 4.5G will give us narrowband, low power LTE-M, going beyond R13, then Huawei is actually pushing out quite far. There do appear to be others thinking similar thoughts, however. Nokia, for example, will have an LTE-M demo at MWC, with Korean operator KT.

As for the 4T4R/8T8R MIMO, you will also need to “add something” to the baseband unit, Cai says. Here again, Huawei is far from alone. Eight pipe devices will be demonstrated by other vendors, also at MWC.

4.5G – a term for the long term?

Questioned about the use of the 4.5G term, Cai says that perhaps the term won’t last: something else will come along in the industry to give all this a name. He adds that much of what 4.5G does can be seen not as a stop-gap before 5G but as a progression towards it.

Note also that some of what Huawei is promising is on plenty of other road maps as well. 4TR or 8TR MIMO, with 256 QAM is exactly what Nokia announced it will have on demo at MWC – but Nokia is not calling it 4.5G. An air interface to support low power IoT is something Alcatel-Lucent is very keen on. (Just to pick two examples to add to Nokia’s support for LTE-M as mentioned above.)

So if these features already “exist” in R13 and R14, and other vendors have or will have them anyway, then what is the need for the 4.5G term, if not for marketing? What Huawei means, perhaps, is that the technologies are real enough. This isn’t just vapourware. But the term itself is, indeed, marketing. And that’s OK, because it’s possible  the industry is going to need a name for something which is more than plain old LTE, but not 5G. EE, for instance, is already calling its Carrier Aggregation 4G+, to differentiate it from plain old LTE.

Huawei’s 4.5G Paradox

4.5G will be backwards compatible, but it will also require a new air interface.

4.5G will be backwards compatible, but it will also be a progression towards 5G (which will not be backwards compatible – at least at the air interface).

4.5G will be commercially available in 2016, yet it will be a progression to 5G, which isn’t close to being standardised.

4.5G is not marketing, but it probably won’t last as a term, something else will replace it..

Other vendors plan the same technology that Huawei has labelled 4.5G – but they don’t call it 4.5G.