The ITU’s report “IMT Future Technology Trends Towards 2030 and Beyond” is expected to be released in June 2022. This document will be the input for the ITU recommendation “Vision of IMT beyond 2030”, which will help shape the vision for what 6G will do and be. So far, a wide variety of different ideas and opinions have been expressed on the topic, from “6G will deliver what 5G promised” to some truly revolutionary concepts that will have ramifications for the shape of the telecoms industry. Organisations such as India’s TSDSI are busy submitting their ideas and requirements.

The question of why 6G might be necessary at all – and therefore what it would need to deliver – is a vexed one. A number of governments have already put stakes in the ground by sharing their own visions, such as China and the EU. There is one repeated call which is likely to make many telecoms engineers struggle, however, and that is in the pursuit of more sustainable future networks.

Unlike traditional Key Performance Indicators that are related to technical outputs, societal and sustainability outcomes – Key Value Indicators or KVIs – are a unique discussion point in 6G and create a source of internal conflict.

There is good reason for both telcos and society to want improvements in this area. A recent white paper by the ATIS Next G Alliance spells out why.

“Telecommunications specifically consumes 2-3% of the global electricity supply,” the paper notes – which forms a considerable operating cost. “The broader ICT industry currently consumes 5-9%, but with the rapid growth in digitization this may rise up to 20% by 2030.”

Meanwhile, the materials involved in building and running our technology infrastructure create huge strains on the planet’s resources, including in water resources used intensively in mineral extraction. Without something changing drastically, the addition of billions of connected devices and considerable network densification would demand considerable additions to energy infrastructure as well.

The challenges this poses to the development of 6G are laid out eloquently in this article in the ITU Journal on Future and Evolving Technologies by Riccardo Bassoli, Frank H.P. Fitzek and Emilio Calvanese Strinati.

The three academics explore some of the use cases, KPIs, and KVIs that have been suggested for 6G, along with their practicability. Reducing energy consumption is one of the main challenges they highlight, as almost every other KPI or KVI tends in the opposite direction.

“The very high data rate, the massive application of virtualisation, native AI with continuous data mining and processing (so‐called connected intelligence) will increase in‐network computing to an unprecedented level,” they point out.

“Next, an ’always on’ 6G network (i.e. experiencing no failures) will require the presence of backup virtual network functions, always ready for a replacement. Moreover, a large density of User Equipment, together with several of them needing Tbit-per-second data rate, will greatly augment the computing required per user at the Baseband Unit (BBU). Especially, this will be a virtual BBU according to the C‐RAN paradigm. Thus, this huge stress on edge and cloud data centres will significantly increase the end‐to‐end energy usage.”

At the same time, there are already complaints that 5G drains phone batteries too fast; reducing power demands in the phone will likely demand more power from the network. With some understatement, the authors note that “all these observations about energy underline some inconsistencies… that will be extremely hard to solve by just evolving the current ’classic’ technologies.”

Answers From Beyond The Norm

The exciting thing about this is that, to solve these problems, it demands creativity outside “classic” technologies and opens the door to new approaches.

One of them is Semantic Communications. While traditional signalling focuses on ensuring the correct transfer of all data packets, including re-sending any data packets which are lost or corrupted in transit, Semantic Communications aims to understand the intent or purpose of a transmission – the meaning that is being conveyed overall, or the goal to be achieved. So long as this is effectively managed, re-sending or duplicating packets is unnecessary.

Humans instinctively understand semantically, so that we can tolerate a certain amount of disruption or interference in speech and still understand another person; for example, people speaking through masks or in a noisy background. This is especially true where the message is a common one. Being asked for a cup of coffee is much more usual, and therefore more intelligible, than being asked for a cup of chopped bee. We do this by understanding the patterns of speech, language and common habits of what is going to be requested – the context.

Understanding this context will involve networks supporting machine learning algorithms and knowledge representation tools at the edge. As well as improving the efficiency of the communications it would also improve security by flagging up what doesn’t appear correct in context.

Other approaches targeting the RAN – an area where the telecoms industry can make the biggest impact – are spelled out in a recent NGMN white paper on green future networks. One significant area of opportunity involves developing a much more granular understanding of the equipment behaviour and usage in the radio and the base station, then turning off specific elements when they are not required.

As the white paper points out, 20% of cell sites account for 80% of usage, so there are considerable efficiencies to be made by running enough of the elements within any given site to cope with demand, but no more. A more modular approach to site management and maintenance – with designs that allow for simple repair or replacement of specific parts, not whole units – would also reduce e-waste and improve the sustainability of the overall network. The challenge will lie in persuading vendors to adapt their designs and manufacturing approach to accommodate.

NTT’s IOWN project also aims to make radical energy savings in telecoms networks by moving away from electronics towards greater use of photonics.

They note that heat generated in electronic wiring inside of chipsets limits the performance of those chips. As an aside, it also marks a reduction in electrical energy and makes power demands for cooling systems.  Photonics systems, by contrast, do not generate heat. Again, turning such fundamental innovations into commercial realities will take time and considerable research – NTT itself projects a 2030 timeline for this project.

These approaches have a common theme. Even where fundamental research is not required, such as in the case of making equipment more modular, there are considerable business implications. Overcoming the inertia of entrenched habits and interests will be a huge challenge. It will demand a willingness for taking bold steps in business relations and commercials to apply the radical innovations in technology.

However, if the telecoms industry can avoid increasing their energy bills, or even lower them, the financial as well as ecological outcomes will be highly rewarding.