Asia raises the 5G bar in the next digital revolution

Performance-enhancing technology

So, what specifically is 5G-Advanced? For a number of reasons – very much including its real introduction five years after standard 5G and probably five years before the introduction of 6G networks – it is tempting to think of 5G-Advanced as a mid-cycle, halfway house between 5G and 6G. But in its own right it brings clear performance upgrades and supports new use cases compared with the first iterations of 5G.

5G has become the fastest developing generation of mobile communications, but fundamentally 5G-Advanced is designed to bring continuous enhancements in mobile network capabilities and use case-based support to help mobile operators with 5G commercialisation. Moreover, it is the first mobile standard specifically built to take advantage of artificial intelligence (AI), machine learning (ML) and network energy savings for a fully automated network and a sustainable future.

Indeed, AI and ML will become essential for future networks given the predicted rapid growth in network usage and use case complexities that can’t be managed by legacy optimisation approaches with presumed models. System-level network energy saving is also a critical aspect, with operators needing to reduce the deployment cost but assure network performance for various use cases.

Compared with standard 5G, 5G-Advanced is designed to bring more capabilities to enable wider use cases and to enhance radio performance to a new level, driving enhanced network performance towards higher bandwidth, lower latency and greater reliability. It will offer improved support for applications such as extended reality (XR) and promises monetary opportunities to consumer markets, with applications in gaming, video streaming, remote working and virtual training.

5G-A and AI get to work

5G-Advanced will create its own ecosystem for vertical markets, and ongoing development in this area is designed to bring improvements in traffic throughput, network coverage, power saving and anomaly detection. And thanks to the actions of the Asian operators, as showed clearly at this year’s MWC Asia event in Shanghai, this ecosystem is up and running, with the operator community being joined by the comms tech and infrastructure giants such as Huawei and a burgeoning number of AI developers looking to cash in.

The first standards release for 5G-Advanced – 3GPP Release-18 – was officially frozen in Shanghai on 28 June 2024, a key milestone for the 5G industry making this the first year of commercial 5G-A deployment.

Assessing the prospects for 5G-Advanced on the eve of the show at a 5G-A and AI industry roundtable hosted by comms tech giant Huawei, the founder and executive chairman of research firm CCS Insight, Shaun Collins, noted that in the new comms world, artificial intelligence is going to transform the comms industry, in particular the operators that provide services with 5G-Advanced, boosted by AI, adding value across the whole market and offering “a fantastic and a very rich” experience.

“There are three major areas that we are considering at the moment, more broadly in the technology space, where AI and 5G-Advanced will make a significant difference both to our lives and to our customers’. Probably the single most important at the moment is that one right at the top of this triangle, which is AI in the realm of [the radio access network (RAN)]. We’re looking at improving the radio performances and optimising energy usage as we deliver some of the expectations that our societies have in terms of being able to do things on a better level in terms of sustainability.

“And, perhaps in the not-too-distant future, even seeing artificial intelligence become a significant part of non-RAN workloads at the edge of the development of the orchestration of those networks. We consider the AI in the core, both at the BSS and OSS level, as equally important in terms of how we are delivering the experience for customers and for networks.”

In terms of what was needed technologically to support AI in the RAN, Collins noted that networks need to hit new levels of performance for responsive data transfer with digital twin simulations improving the reach of networks. For AI in the core, Collins noted that to benefit from AI within 5G-Advanced, 5G standalone (SA) would be the linchpin.

We’ve probably got about a three-year window where we continue to talk about AI. From about 2028, it’ll just be what we do, just part of the network

“Increasingly, we’re seeing artificial intelligence for cloud and devices being a very neat endpoint to how AI is being delivered, both in terms of just the sheer scale of the optimisation of networks and the optimisation of applications, and also how we’re able to move that data around at scale with intelligence, making it agile, useful and valuable to an operator or as an end user. AI models will be in every part of business.

“We’ve probably got about a three-year window where we continue to talk about AI. From about 2028, it’ll just be what we do. It’ll just be part of the network. It’ll just be in everything that we do – for sure, it will make 5G and possibly 6G more efficient, more valuable, more agile and more flexible. And it could deliver some new network products going forward.”

Collins was adamant that the revolution was already underway in what he called the more advanced networks, particularly in the US, Middle East and China. He also stressed that massive compute loads were starting to be deployed on the silicon required to support RAN AI, with progress in this part of the ecosystem by the likes of Nvidia, Intel or Qualcomm. The bottom line, he said, is that AI is playing a massive part in the development of the comms industry beyond just moving into a new generation going forward.

“The sheer scale of some of the things being offered in terms of applications or use cases around the world are generating such immense amounts of data that it would be physically impossible on a standard computer system to be able to make the most of that information and, in some cases, to make sure that critical application, if it isn’t deemed a critical application, can be delivered, given everything else that’s going on in the network, he added.

“AI allows us to deliver this in a way that is effective, valuable and can deliver what’s required in the midst of immense amounts of data and delivery. We see AI not only as an important development for networks, and an important development for customers, but also transformational for how network operators are able to deliver that experience in a better, more efficient, more reliable and more valuable way. Those three elements of AI in the RAN, AI [in the] core, and AI for end users and devices mean that, together, we’ll be able to deliver a much more interesting, engaging, private, possibly even safe environment going forward.”

China leads the way

This AI-enabled revolution is already being delivered in China. Research from global mobile industry trade association the GSMA calculates China’s 5G market is set to add almost $260bn to the Chinese economy in 2030, with connections set to top one billion this year. It added that Chinese operators have demonstrated a clear interest in using generative AI (GenAI) for network operations and management.

At the pre-MWC Asia roundtable, Deng Wei, duty director of the wireless and terminal technology department at the China Mobile Research Institute, said now was the time for AI and connectivity integration, with the former being a driver for new use cases and plotting out how the telco would be aiming to take advantage of the new 5G standard.

He stressed that technical gains on a platform perspective could be generated from general development, including parameter optimisation and dynamic policies. In base stations, he saw the new technologies as combining for load-based prediction in an energy-saving capacity, with optimisation-driven, site-specific, energy-saving policy generation.

With 5G-Advanced, Deng Wei predicted a world of intent-driven communications in networks that, compared with previous mobile generations, are more adaptable, working in a multi-modal way to conduct computer-intensive scenarios. Moreover, networks that are more intelligent will guarantee premium experiences, meeting the more intensive requirements of future services.

Yang Jianjian, director of network and digitisation planning research centre China Unicom Research Institute, said 5G-Advanced was the logical next step with its advantages in terms of latency, in particular a response to meet rising needs in industrial scenarios for mobile communications.

He drew an analogy with housebuilding, whereby with “standard 5G” it was a case of building a house with “standard” bricks. In future mobile projects, he suggested users would need “different bricks for 5G projects to offer different connectivity”, using higher-quality networks. He remarked that as an enhanced version of 5G, 5G-Advanced has shown high maturity in technological innovation, industry ecosystem and commercial scenarios, and offers a “promising prospect”, bringing new opportunities for industries to “go digital and intelligent”.

Reinforcing the need for a robust ecosystem, Yang Jianjian said China Unicom was committed to collaborating with all parties in the industry and working with the wider ecosystem. He identified specific key areas of development – accelerated open gateway implementation with rich application programming interfaces (APIs), advanced AI for application innovation, and industry-wide collaboration.

At the heart of this, he said, would be initiatives like the GSMA Open Gateway – designed to accelerate the growth of digital services and apps, unlocking the full potential of 5G networks and commercialising network APIs, boasting 53 mobile operator groups, representing 67% of global connections – doing more research that would be more available to the industry in general.

He concurred with the general consensus that integrating AI further into the comms industry was the right future direction. As well as innovation in AI chipsets and on-board intelligence, Yang Jianjian noted the need to focus more on AI models for the various industries in which the technology would see use.

Consensus was also very much on the mind of Eric Zhao, vice-president and chief marketing officer of Huawei Wireless Solution, who unveiled a plan to bring AI to networks in what he called the 5G age. In the Yang Jianjian analogy, Huawei is one of the companies manufacturing the different bricks for future 5G projects, and as he addressed the roundtable, Zhao stressed that Huawei would be part of the mission to bring AI to the network, boosting intelligence in the RAN and increasing network capabilities.

He noted that 5G-Advanced had a clear set of advantages, displaying “quite impressive” performance in terms of network, business and device developments. Specifically, he noted that 5G-Advanced networks would support downlink speeds of up to 10Gbps and uplink in the 1Gbps region and support up to 100 billion internet of things (IoT) connections.

Yet he also cautioned that 5G-Advanced presents its own set of challenges, with mobile networks facing more complicated operations and maintenance, differentiated network characteristics and diverse experience-driven operations. AI, he noted, would enable 5G-Advanced to defuse growing complexities “at ease”, offering a platform for autonomous operations and maintenance, energy savings, and optimal and deterministic mobile experiences. In short, making state-of-the-art networks more intelligent.

• Read part two: Continuing our look at 5G-Advanced, we find out what products and services the new network will support, and which operators are already planning ahead for a new era in communications.