AI-RAN Alliance is established, 6G and AI, no Chinese manufacturers participating?

The Radio Access Network (RAN) is an invisible thread that weaves the entire Internet world, just like the neurons in the human brain that are tightly connected to form a large network, except that RAN connects people and devices in the Internet world.

In the past, we regarded this network as a hub for transmitting information and wisdom. Now, after 5 generations of technological evolution, some manufacturers are already considering how to make this network more intelligent, especially now that we are getting closer and closer to 6G.

AI-RAN Alliance is established, Chinese manufacturers are absent

At this MWC, while major operators and communication equipment manufacturers are showing their talents, NVIDIA, AWS, Arm, DeepSig, Ericsson, Microsoft, Nokia, Samsung, Northeastern University, T-Mobile, and SoftBank jointly established A new alliance called the AI-RAN Alliance.

AI-RAN

AI-RAN

To summarize the slogan of the official website of the AI RAN Alliance, AI is used to enhance RAN performance, build an infrastructure for AI and RAN to collaboratively share information, and at the same time create new AI applications that run on RAN.

Judging from the current members, the AI RAN Alliance includes chip manufacturers, cloud service providers, IP providers, telecommunications equipment manufacturers, and operators. However, judging from the current list of alliance members, although the weight of the participants is not low, the number of participants is small, especially the lack of telecom operators.

The AI-RAN Alliance points out that member operators will take the lead in testing and implementing these advanced technologies through collective research work by member companies and academic institutions. However, when the only participating operators are T-Mobile and SoftBank, will they still be able to attract more attention in the future? It is still unknown whether more operators will join.

What deserves attention is the participation of the three major RAN equipment manufacturers Ericsson, Nokia, Samsung, Arm, and NVIDIA. This lineup seems to be an attempt to weaken the dominance of Intel and x86 architecture in the operator data center and vRAN market.

Another point worth noting is that among the limited participants in the AI-RAN Alliance, there are no Chinese manufacturers. Chinese operators, cloud service vendors, and equipment manufacturers have not participated in this alliance.

For this newly established alliance, it seems a bit arbitrary to claim that 6G technology is blocked. After all, China’s 6G RAN does not necessarily need to follow the same AI technology route.

Is Open RAN abandoned by these vendors?

Now that some manufacturers have chosen to follow AI-RAN, has Open RAN, which was once considered popular, been abandoned? This is not the case.

The current development of Open RAN mainly focuses on 5G and 5G-A networks. The development of related 6G open RAN networks is still under planning, while AI-RAN is mainly geared towards the development of future 6G wireless access networks.

Considering that the AI RAN Alliance has not been established for a long time, its planning and standard route are still unclear, and it is difficult to say whether it will eventually join the open route of Open RAN.

Open Radio Access Network (Open RAN) is a project group launched by the Telecom Infrastructure Project (TIP) in 2017 to build RAN solutions based on common, vendor-neutral hardware and software technologies.

To increase the promotion of Open RAN, China Mobile, AT&T of the United States, Deutsche Telekom, NTT DOCOMO of Japan, and Orange of France established the O-RAN Alliance in 2018, and has attracted a large number of manufacturers to join in the past few years.

From the essence of Open RAN and the O-RAN alliance, operators promote cross-vendor interoperability of RU and DU to improve the interoperability of white-box hardware from different suppliers, thereby reducing deployment costs.

Various manufacturers participating in the AI-RAN alliance have more or less contributed to the construction of Open RAN.

For example, with the help of NVIDIA’s GPU and Aerial SDK, it built the first GPU-accelerated 5G Open RAN network for NTT DOCOMO, Japan’s largest telecom operator. Samsung also cooperated with British operator Vodafone to build the first 5G OpenRAN site in the UK and launched a large-scale deployment plan last year.

Not to mention equipment manufacturers such as Nokia and Ericsson, have spared no effort to promote the implementation of Open RAN and are vying to cooperate with operators.

Currently, many manufacturers are still insisting on building Open RAN networks, but progress is still relatively slow. The CTO of Dell’s telecom systems business also said in an interview at this year’s MWC that the popularity of Open RAN is slower than expected, but the current explosion of generative AI applications is likely to bring a turning point for Open RAN to accelerate its implementation.

As for the domestic market, although domestic operators have joined the O-RAN Alliance, they are still in a wait-and-see state. From the perspective of equipment performance, the white-box solution of Open RAN is currently significantly inferior to the black-box solution of domestic equipment manufacturers in terms of efficiency and cost.

Therefore, even if the vision of openness and interconnection is good, analysis of the current situation of several foreign operators that have adopted Open RAN shows that not only the problem of integration of different manufacturers is always difficult to solve, but also the network performance cannot meet the standards, and the operation situation is also relatively bleak.

On the contrary, some Middle Eastern countries have built high-performance 5G networks with the advantage of low cost and rapid deployment after adopting black box solutions provided by Chinese equipment manufacturers.

It can be seen from this that not all alliances need to be mixed to be successful. If the goals of the alliance cannot be achieved, the market will naturally answer.

5G network

“Traditional” telecom operators build next-generation AI RAN networks

So will AI-RAN be a new blank check? It can’t be seen that way. AI-RAN is more like a new gimmick thrown out to promote the strong binding of 6G networks and AI. AI technology has long been listed as a key technology for 6G wireless access networks by domestic operators.

Taking the 6G Simple Wireless Access Network proposed by China Mobile as an example, it mentions the introduction of AI technology for intelligent resource control and network self-optimization. , self-control.

China Unicom has also long planned the research direction of closely integrating AI and RAN networks. In the “China Unicom 6G Synaesthetic, Intelligent, and Computing Integrated Wireless Network White Paper”, it is mentioned that the future wireless network will move towards the direction of integrated synaesthesia and intelligent computing. development, the computing power on the base station side will move from stacked chip computing power to heterogeneous converged computing power, and then to cloud computing power, providing flexible computing power services to the outside world; base station intelligence will shift from supporting intelligent reasoning to supporting small The direction of large-scale intelligent training will evolve, and eventually endogenous intelligence will be realized; synaesthesia will evolve from 5G-A to 6G, and relevant standardization and trial commercialization will be promoted.

From 2023 to 2025, it will be the first stage of China Unicom’s promotion of integrated wireless networks with sensory intelligence and computing. It will increase computing power for RAN by using ASIC-customized smart chip overlay, and deploy low-computing power intelligent reasoning models on RAN.

Its strategy is consistent with the research direction of the AI-RAN Alliance, which also includes application cases for improving RAN performance such as intelligent energy saving and intelligent orchestration, as well as scenarios where AI and RAN collaborate such as synesthesia fusion.

write at the end.

It can be seen that even if no Chinese manufacturers are participating in the AI-RAN Alliance, it will not hurt the future development of domestic operators. After all, in the process of evolving from 5G-A to 6G, we are already building the so-called AI RAN network.

There are endless so-called “minor alliances” in the communications market, but ultimately it is market feedback that determines their success or failure. Since Chinese manufacturers have already gained market advantages in the 5G era, how can they abandon these advantages and choose this kind of “small circle”?

Update cookies preferences

Get a Quick Quote!

x

    Get a Quick Quote!

    x