Partner content: 5G deployment continues, but work has started on 6G which will be AI native, and require the openness and collaboration of open source
Some fundamental questions about 6G still need to be clarified. For instance, whether the air interface should largely correspond to that of 5G or be redesigned, and there are similar discussions about the core network.
The debate is because for some, 5G has not fully lived up to expectations. Its adoption in the industrial sector is progressing slowly, so mobile network providers still rely on consumers’ use of mobile broadband as 5G’s main revenue generator. Other use cases, such as Ultra Reliable Low Latency Communications and Machine-type Communication, have not played as big a role as some expected.
This has driven an increased focus on cost efficiency for 6G, and the need to build on existing hardware so that the switch from 5G to 6G can be realised via a software upgrade – reducing investment.
The foundations for this have been laid with 5G Standalone (SA) being cloud-native, decoupling hardware and software to offer a high degree of reliability, scalability and flexible infrastructure management. By the time 6G is ready – its commercial introduction is expected in 2030 – the core networks of mobile network providers are likely to have been converted to 5G SA. In other words, a software upgrade to 6G would be possible in principle, if the industry decides to go down this route.
More immersive experiences
Whether all 6G’s planned features can be implemented in this way, or whether hardware replacements will be necessary in some cases, remains to be seen. After all, the next generation of mobile communications will use additional frequency bands, particularly in the upper range of the frequency spectrum, to improve data throughput, capacity and latency.
Among other things, this will enable higher video resolutions for applications like AR, VR and XR that will be more immersive experiences. The projection of holograms would take interaction with people and objects to a new level, while video glasses would allow gamers to immerse themselves even more deeply in game worlds than is possible on mobile phones, gaming PCs or TV game consoles.
6G will offer special functions to acquire and utilise sensor data in conjunction with communication services to improve localisation and positioning accuracy in space. This capability will help with autonomous vehicles and other unmanned systems, as well as contribute to the design of smart cities in which countless IoT devices are connected to each other.
The lower latency and expected higher energy efficiency of 6G also come into play here, enabling the reliable and economic provision of mobile services at the edge. The industry will likely initially focus on those new functions and 6G services that it can monetise quickly.
Integrated AI
AI will be an integral part of 6G – both in the form of intelligent services and applications, as well as in the management and optimisation of the network itself. The groundwork for this is already being laid with 5G Advanced, which serves as a stepping stone to 6G and an AI-native mobile network. Among other things, 5G Advanced uses AI to organise the complex arrays of small mobile phone antennas that help to target mobile phone users. In this way, more power can be extracted from the radio access network and its energy requirements can be reduced.
However, the efforts of mobile network providers in some regions to skip the intermediate step via 5G Advanced and instead switch directly from 5G to 6G could pose a challenge. The companies and organisations involved in the development of the new mobile communications standard still need to come up with a way to reconcile the desire for as few changes to the network as possible in order to keep investments low, and, on the other hand, the implementation of many new features to create new opportunities for service revenue generation.
Controlling resources
Ultimately, however, native AI support is essential to improve the control of resource usage and utilisation in the radio access and core networks. The goal is an autonomous network that makes intelligent decisions and dynamically adapts and optimises itself without manual intervention. The fact that AI is currently still very energy-hungry does not have to be an obstacle – the technology is developing rapidly, and it has recently been shown that there is a lot of room for improvement.
For example, small language models (SLMs) are better suited to numerous use cases than large language models (LLMs), as they have more specialist knowledge in one area and consume less computing resources and energy. Agentic AI and vLLMs are also interesting developments for the telecommunications industry. A vLLM, for example, can use the GPU memory on an inference server very efficiently to perform calculations quickly and save energy – and thus contribute to the sustainable use of AI in mobile networks.
More open source
With the cloud-native nature of 5G SA, open source has already become an important component of many mobile networks and with 6G and the stronger focus on AI, the importance of open source solutions will continue to grow. This is because open source drives innovation, especially in the AI sector, where models based on the contributions of a large community offer many advantages. They include lower costs, flexibility and adaptability for security and transparency.
However, the community concept can also help with the design of innovative 6G services. If companies and organisations from different sectors join forces, they can work together to ensure that the next mobile communications standard meets a wide range of requirements and creates a large ecosystem of services. There is already a trend towards more openness and collaboration, including from telecommunications companies, and the industry needs to continue on this path to ensure that 6G will become a powerful, sustainable technology that puts people’s needs first.
The author is Rimma Iontel, Chief Architect, Telecommunications at Red Hat
