Huawei’s annual analyst summit, hosted in Shenzhen last week, was dominated by the vendor’s own campaign to expand on the well-known triangle of 5G use cases (eMBB, mMTC and URLLC), and create three more that redress the balance in favor of the uplink and the enterprise.
These come under the banner of 5.5G (naturally), and consist of:
- Uplink-centric broadband communication (UCBC)
- Real time broadband communication (RTBC)
- Harmonized communication and sensing (HCS)
Huawei has been touting these three cases since its 2020 Global Mobile Broadband Forum (MBBF), where its executive director, David Wang, described them as cornerstones of the company’s vision for the next phase of 5G development. As Wang put it then: “5.5G is our vision for the industry. It is an enhancement and extension of the three standard 5G scenarios defined by the ITU – eMBB, mMTC, and URLLC.”
The unilateral nature of Huawei’s definition of these use cases means they are unlikely to be adopted, at least in this exact form, by the whole industry. They have even been downplayed by China’s own mobile operators. But they are valuable in highlighting the direction of travel for 5G and identifying changes or iterations that must be made for interactive robust and low latency communications in support of critical processes, some of which will require high data capacity in both directions.
Part of the confusion lies in the phrase ‘use case’. The existing triumvirate are really three pillars or focal points of wireless capability and performance with very broad applicability.
eMBB is about capacity or volume of data, that is the amount of traffic that can be carried in a very short time, the equivalent of having lots of lanes in a road transportation network. Transit speed is not increased, yet delivery times are shortened because the data pipe becomes less of a bottleneck and allows more simultaneous communication or parallelism.
mMTC then addresses volume of devices and sessions, so is about being able to support large numbers of small transactions in a very short period of time. Finally, URLCC is about data transit time, or latency, with the focus on reducing avoidable sources of delay, and then ultimately shortening the communications path to minimize signal transmission time, which is where edge compute comes in.
In reality, many real applications will intersect with more than one of these pillars, and some, such as extended reality in gaming at high definition, possibly with all three, since low delay, high bandwidth and support for large numbers of devices may all be required.
Historically, broadband communications has been biased in favor of the downlink over the uplink, because most consumer applications such as video and music streaming generate more data in that direction. That has been reflected in development of standards not just for wireless but also fixed line communications, as in the Docsis family of cable TV specifications.
However, even consumers have been generating far higher volumes of upstream traffic recently with the growth in multimedia messaging apps such as WhatsApp, as well as proliferation in user-generated content uploaded to YouTube and other sharing sites. For enterprises, demand for closer equality between the upstream and downstream is even stronger, driven by conferencing, telepresence and increasingly now in the wireless sphere by process control, robotics, transportation and UAVs (unattended automated vehicles). Such demand coincides in these industrial use cases with need for ultra-low latency.
It is in this context that Huawei considers 5G needs fleshing out and re-orientating around six pillars, which it depicts as hexagonal 5G. The company’s argument, expressed with increasing vigor over the past few months, is that the three existing pillars do not fully meet the demands of the emerging fully connected and intelligent world.
This is how William Xu, Huawei director and president of the company’s Institute of Strategic Research, put it during an opening keynote address at last week’s summit. “Currently 5G networks are far from being able to offer such fully immersive experiences even though now the 5G networks have been deployed at large scale,” he said, referring particularly to the situation in China.
Xu then described UCBC as a hybrid between eMBB and mMTC that would provide the more uplink-centric communication for those enterprise applications needing support for massive connection density at the same time as high upload bandwidth.
He went on to propose RTCB as an amalgam of eMBB and URLLC for real time interactive or control applications that again require support for large numbers of devices but this time with ultra-low latency, which could include private networks serving large industrial sites with many wireless-connected connected components.
Xu then finally described HCS, which is the most novel of Huawei’s three proposed use cases, integrating sensing and other forms of communication with 5G, driven particularly by the requirements of transportation and autonomous driving. It is not quite clear how this relates to work already ongoing around 5G in connected vehicles under the banner of C-V2X.
HCS is focused mostly on the V2I (Vehicle to Infrastructure) aspect of C-V2X, dealing with integration between vehicles and advanced lane markings, pavement markings, roadside sensors and other wireless sources of input that can help improve navigation and safety, even without full autonomy.
One of the main challenges lies in establishing redundancy such that vehicles are not dependent on any one source of data and can perform at full safely with just a subset of the total available. For example, GPS signals may be blocked by buildings and at the same time pavement sensors may be covered by snow, in which case a vehicle would rely on LiDAR sensors and cameras to determine local position with equal accuracy.
Lidar (light detection and ranging) exploits differences in return times from emitted lasers and also variations in wavelengths to map the surrounding area and track the vehicle’s ever changing position within that.
There is also a subtext of Huawei’s interest in expanding 5G with these three new pillars. China is at the head of the curve not just in roll-out of 5G infrastructure and recruitment of subscribers, but also evaluation of use cases. Huawei has admitted to having found that 5G, as currently envisaged, may not be able to support some of the more demanding applications, with the deficit largely lying in the uplink.
This raises the question of whether, and if so how, uplink capacity can be improved without a corresponding reduction in the downlink. One answer that Huawei will be considering lies in the application of non-orthogonal multiple access (NOMA) technologies, which have already been recognized firstly in academia and then among technology vendors as a promising direction for boosting uplink performance, with little if any sacrifice in the downlink.
As we know, OFMDA has long been deployed in broadband communications to improve bandwidth by allowing multiple closely spaced modulated carriers without causing interference. This is achieved by matching the symbols – the property of the waveform onto which data is modulated – to the spacing between the carriers such that the unwanted interference is cancelled out. This is a matter of tuning the system and leads to significant increase in data transmission efficiency over given spectrum, having been adopted both for WiFi and cellular communications under LTE.
NOMA avoids the need for orthogonality, with signals now allowed to be received concurrently without any traditional interference cancellation techniques. Until now this would simply have resulted in a collision, requiring the system to back off and try again, as is the underlying principle of the CSMA/CD (carrier sense multiple access with collision setection), as employed by Ethernet for example.
Instead, NOMA allows two signals to be received at the same time by exploiting the difference in strength between them. Using a process known as SIC (successive interference cancellation), the stronger signal is decoded, then subtracted from the combined signal to extract the weaker one from the residue remaining. In principle this could be extrapolated to multiple simultaneous signals.
Until recently SIC could not be implemented in practice because it required too much computational power at the receiving end, but advances in process density have brought it into reach. It will be one technique under evaluation for these 5G uplink enhancements.
It is worth pointing out that Huawei’s rivals agree that 5G will incorporate emerging technologies such as NOMA as it unfolds. But Ericsson for one prefers to see this described under the banner of ‘5G Advanced’, as a roadmap populated by milestones under the banner of successive 5G releases. “We do not see the need for a “shift in stance at present,” an Ericsson spokesperson told us.