What’s After 5G

This year’s IEEE Symposia on VLSI Technology and Circuits (VLSI 2020) included a presentation by NTT Docomo that looked far into the future of cellular communications, setting the stage for a broad industry shift in communication. This is far from…

This year’s IEEE Symposia on VLSI Technology and Circuits (VLSI 2020) included a presentation by NTT Docomo that looked far into the future of cellular communications, setting the stage for a broad industry shift in communication.

This is far from trivial. 5G only just recently entered the commercial world, and — especially with the higher millimeter-wave (mmWave) frequencies — it has a long way to go before it reaches its full potential. Even so, some researchers and industry members are looking beyond 5G to figure out what comes next.

NTT Docomo’s Takehiro Nakamura took a stab at both how 5G would be extended through an “evolution” phase and what would happen with 6G beyond that. He sees both 5G evolution (5GE) and 6G continuing the trend towards increased bandwidth, increased coverage, and increased volumes of data. [Note: In the US, some companies are calling extensions of 4G prior to 5G “5GE,” causing possible confusion. NTT Docomo’s references to 5GE follow the more traditional meaning of coming after 5G.]

Nakamura started by reminding us that 5G went live just this last March and that it will be the dominant technology through the 2020s. He said that 6G is a technology for the 2030s, although dates and definitions will likely vary, depending upon the vendor.

Fig. 1: 5G pre-service launched in Japan with the Rugby World Cup last year. Commercial service officially began on March 25, 2020. Source: IEEE/NTT Docomo

Part of that rollout includes the use of higher frequencies, which can carry more data, but don’t travel as far and don’t bend around corners as well. Circuit implementations also will be more challenging. Exactly when we move to the mmWave regime will vary by country, because each country auctions off its own spectrum allocations.

Being a Japanese company, NTT was able to speak to its plans within Japan — which, at present, don’t come with a specific deadline. “New frequency bands, 3.7 GHz, 4.5 GHz, and 28 GHz, allocated in April 2019 in Japan, will be deployed for 5G [on a] case-by-case basis,” said Nakamura, “taking into account frequency bandwidth, propagation characteristics, co-existence conditions, etc.”

Fig. 2: 5G will use frequencies higher than what are used for prior generations. In the case of mmWave frequencies (28 GHz and up), they will be far higher, offering more bandwidth but bringing new implementation challenges. Source: IEEE/NTT Docomo

Part of what will drive the next generations will stem from what we learn with 5G. Nakamura identified three specific challenges: mmWave coverage and mobility, uplink performance, and demands from industry use cases. In particular, he envisioned a cycle that largely reflects the model that the Internet of Things (IoT) brought about, only with more data and high performance. He described a cycle where data is derived from real-world entities — analogous to IoT sensors, then communicated to the Cloud, where decisions are made. The results of those decisions are sent back to ground for action – or, literally, actuation.

Fig. 3: Data acquired in the physical world will be transported to the cyber-world, where decisions are made and — where possible — the future is predicted. That prediction can then be used to launch an action in the physical world. Source: IEEE/NTT Docomo

6G then will focus on solving social issues and a closer fusing of the physical and the cyber-worlds, enabled by an expanded set of higher-bandwidth communications options and by more sophisticated fusion between the physical and cyber realms.

Technological implications
Nakamura noted that, while early wireless generations were marked by specific technology differences, later ones haven’t been so clearly delineated. Instead, existing approaches have been refined and improved. 5G and its evolution will be characterized by:

Extending the frequency range;
Moving from MIMO to massive MIMO (mMIMO);
Moving to polar codes for control channels and to low-density parity codes (LDPC) for data channels, and
The appearance of ultra-reliable low-latency communication (URLLC) and massive machine-type communications (mMTC).
6G might even extend the frequency range into terahertz territory, and it could come with a new waveform to complement or replace orthogonal frequency-domain multiplexing (OFDM). A new network topology would need to provide extreme coverage, enhancing the 5G new technologies and making AI available everywhere.

Fig. 4: Early generations of cellular technology replaced prior technologies outright. Later generations are building on earlier ones, selectively adding or enhancing new technologies to address specific challenges. Source: IEEE/NTT Docomo

The design of the networks themselves is likely to change. Existing approaches to placing cells strike a compromise that maximizes coverage while minimizing overlap between adjacent cells. He sees this changing in the future, with overlapping cells leveraging multiple propagation paths to improve bandwidth and reliability.

The excursion into mmWave territory presents opportunities courtesy of the challenges that such high frequencies present. Because of increased line-of-sight requirements, it will be necessary to reflect beams to ensure coverage in “darker” corners. But materials don’t always act as expected at higher frequencies, making it difficult to create mmWave reflectors.

Nakamura presented some work that creates reflectors out of metamaterials — that is, materials engineered by humans that don’t exist in nature. By creating these materials, engineers can design a specific index of refraction and determine both the direction and shape of the reflected signals. In fact, not only can the material have differing indices of refraction for different locations on the reflector, but those indices could even be dynamically controlled, making the reflector far more adaptable.

An extreme example of this would be the ability to route light around an object, effectively cloaking it — that is, making it invisible. And by controlling the ratio of reflected to transmitted signal, surfaces could be made alternately transparent or opaque.

Fig. 5: Engineered materials that play with the index of refraction can create effective cloaks of invisibility. Source: IEEE/NTT Docomo

The radio access technology (RAT) also will see some upgrades, first through the expansion of mMIMO technology. In addition to that evolutionary move, new technology will be deployed to enable faster-than-Nyquist (FTN) communication.

Communicating under Nyquist criteria means using orthogonal coding to avoid issues with inter-symbol interference (ISI). But we trade off bandwidth for that clean operation. The idea is that, by allowing some ISI, we can stuff yet more data down the channel, maintaining the same bit-error rate (BER) until the channel finally becomes too crowded. New algorithms already have been proposed to combat the resulting ISI, enabled by advanced silicon technologies that make possible faster and more complex circuits than have previously been feasible.

Yet another development for 5GE and 6G will be the availability of private industrial networks. Today, all cellular users — whether consumers or corporate entities — use the same network. So in theory, people watching videos can interfere with, or crowd out, industrial uplink of massive amounts of mission-critical data. By creating private networks, industrial companies can get better control of their performance, with guaranteed availability and response. Mobile network operators (MNOs) might play a role in those networks, as well.

Coverage will be enhanced with numerous additional technologies, some of which may not be cellular. In particular, airborne or space-based transceivers will play an increased role. While geostationary (GEO) satellites are currently in use, Nakamura presented an increased role for low-earth orbit (LEO) and high-altitude pseudo-satellites (HAPS) (also referred to as atmospheric satellites, or atmosats).

The idea here is to provide a large number of transceivers accessible from anywhere on land or sea to bring adequate coverage to areas that are currently under-served, or where there is no coverage at all. Different vehicles having different bandwidth characteristics can address a wide range of scenarios that are not well addressed today.

Fig. 6: With bandwidth running from the 10-Mbps range to over 10 Gbps, different vehicles will be able to serve a wide range of applications. Source: IEEE/NTT Docomo

Finally, AI will be pervasive in this view. The image below shows the range of areas where AI will add value. Most of those areas relate to the operation of the network itself, but general access to AI remains a part of the promise.

Fig. 7: AI will feature heavily, largely for the efficient operation of the networks, but also as a service for business or even individuals. Source: IEEE/NTT Docomo

Key takeaways
This view was shared by NTT Docomo, with specifics focusing on the Japanese market, naturally. But the work done so far, even if exploratory, will impact all companies involved and all geographies. We’ve distilled some summary points for agreement, disagreement, or additional depth as an opportunity for comment by a few other companies. Those summary points and the responses follow.

5G evolution will focus on better uplink and towards more delivery guarantees (as opposed to “best effort”), with a focus on ultra-reliable low-latency communication (URLLC).

Suresh Andani, senior director of product marketing for IP Cores at Rambus, agreed. “The initial 5G use cases will focus on enabling enhanced mobile broadband (eMBB) for UltraHD, AR/VR, 360-degree streaming video, etc,” he said. “Over the next three to five years, the 5G use cases will need to enable massive machine type communications (mMTC) and ultra-reliable low latency communications (uRLLC) to realize the 5G vision. The uRLLC use cases, such as intelligent transportation, industrial automation, and remote healthcare, will require guaranteed (vs. best effort) and the lowest-latency communication between the 5G machine and the network. As 4G and 5G infrastructure are key Rambus focus areas, we are enabling the networking chip and telecom equipment manufacturers with the lowest and deterministic latency interfaces required to efficiently move data across the network.”

David Vye, technical marketing director for AWR Software at Cadence, likewise expanded on the details. “A key goal of 5G has been to expand the reach and value of wireless technology beyond the individual mobile subscriber in support of mMTC and URLLC,” Vye said. “Expanding connectivity to include network-to-smart device communications, combined with artificial intelligence and the Internet of Things, will usher in a new industrial wave and offer greater business value for both industry and society.”

Vyte pointed to gaps in the current technology that need to be plugged. “5G deployment and related trials have shown there is room and need for improvements to the coverage and uplink performance of non-line-of-sight (NLOS) environments and heavy traffic use cases. To fully achieve the promise of this next wave of communications, continued enhancements are needed to guarantee the high reliability and low latency necessary to close the gap between the cyber and physical worlds.”

Achieving URLLC performance with sub-1ms latency and up to 99.9999999% reliability for factory operational technology and networks based on edge cloud computing will require a massive number of small cell radio access points and distributed antennas with AI and ML-controlled beam steering.

“Implicit in the requirements for this vast deployment of non-orthogonal networks will be the need to drive down the cost of the beamforming antenna arrays and complex receivers — especially if faster-than-Nyquist signaling techniques are applied — while greatly improving beyond the current best-effort uplink technology,” he added. “This new network topology, combined with 6G’s move to even higher mmWave spectrum (94GHz to 3THz), will lead to a wide range of design and integration challenges.”

5G evolution will see the implementation of a loop between the physical world and the cyber world, where volumes of data on everything are sent to the cloud to be processed and used for “predicting the future.” Those predictions will result in actions (or actuation) that’s send back down to the physical world.

Simon Rance, head of marketing at Cliosoft, agreed. “Big data analytics in 5G will coincide with this loop between the physical world and the cyber world to aid with key technology and business drivers such as predictive maintenance and cognitive analytics, for starters,” he said. “By leveraging AI, 5G, cloud computing, and big data analytics, predictive maintenance will help to predict failures before they occur and will play a crucial role in communication, applications, and many other areas. Similarly, by leveraging machine learning, cognitive analytics in 5G can go way beyond traditional descriptive analytics and learn in real time from the context and predict what will happen next — or adapt let’s say, from analyzing past behavioral patterns. Cognitive analytics in conjunction with the loop between the physical world and the cyber world will help to make optimal decisions in real time and will have a range of applications from autonomous driving to future intelligent applications.”

Andani said multi-access edge data centers, such as telco data centers and central offices, will play an important role in this loop between the physical world and the cyber world. “Not everything from the physical world should — or needs to be — sent to the cloud,” he said. “Artificial Intelligence (i.e., predicting the future and taking actions) will be split between the edge and the cloud, particularly for uRLLC use cases where keeping the latency of the required action to a minimum is extremely critical. AI at the cloud/edge requires high-speed SerDes and memory interfaces such as 400G Ethernet, PCI Express Gen4/5, HBM2/2E, GDDR6 – all part of Rambus’ 4G and 5G portfolio.”

6G will roll out in the 2030s.
Some countries have started early research work on 6G, said Andani. “The 6G commercial deployments are expected to start from 2030. However, there is a heavy dependency on how 5G and 5G Evolution roll out in this decade.”

Today we have fixed networks with cells designed to minimize overlap where possible. In the future, we’ll have overlapping cells that can be reconfigured dynamically.

Vye amplified on this. “Current activity in mmWave front-end design, including antenna-in-package (AiP) phased arrays, large-scale beam-forming RF integrated circuits (RFICs), multi-technology integration, and system-level electromagnetic (EM) analysis all will contribute to realizing new radio access technology that can be cost-effective and easy to install, supporting the small cell networks that will achieve URLLC performance.”

He noted that previous cellular communications were based on networks of hexagonal cells spaced far enough apart to avoid signal interference with neighboring cells. “6G may employ a spatially non-orthogonal, overlapped and dynamic topology to increase path selection. Beam control through AI/ML will help to reduce inter-cell interference at a cost of complexity. This architecture will also require new antenna design (conformal as well as phased arrays). The move to higher frequency bands in 6G will help to reduce the size of these antennas, making efforts to shrink component footprints easier.”

However, the antennas, feed networks and package interconnects all will be more susceptible to parasitics and unintended coupling. That will require rigorous EM analysis and design verification. “Strategic design partitioning, leveraging of optimal semiconductor processes, and multi-fabric assemblies will undoubtedly be utilized, calling for a range of simulation technologies, design and manufacturing flows and tool interoperability,” Vye said.

Andani agreed this is a must-have. “To meet the service model of 5G and beyond, the infrastructure cannot afford to be fixed. Software-defined infrastructure that is dynamically composable (reconfigurable) will be necessary to meet the service-level agreements for all 5G-enabled use cases,” he said. “For example, if a default cell is over-subscribed, then the wireless infrastructure must be able to service the end user with another cell without compromising on the quality/experience of service.”

High-altitude pseudo-satellites/atmosats and low-earth-orbit satellites will be deployed to extend coverage into the air, over the sea, and even into space.

“Various commercial companies (e.g., Viasat, SpaceX, Amazon, etc.) have started taking steps in this direction,” said Andani.