Huawei iMaster NCE One Map and One Master: Accelerating Towards AN L4 for FBB

Host: Hello and welcome to Huawei Tech Talks at Huawei Analyst Summit 2024. I'm here with Shanthi Ravindran, Principal Analyst in the Communication Software Domain for Appledore, and Stephen Shao, Vice President of the General Development Department at Huawei. Welcome.

Stephen Shao: Hello, Gary, Ms. Shanthi, and everyone. Welcome to the Huawei Analyst Summit 2024. In the ADN forum yesterday, Huawei released an Autonomous Networks Level 4 solution. The solution is an integration of "Digital Base and Telecom Foundation Model". Today, I'd like to share more about this solution and our innovation practice with you.

Shanthi Ravindran: Thank you so much. It's been a great session so far and it's always very happy to be here. Thank you.

Host: Okay, so let's take a crack with the first question. What challenges do operators face in the FBB field during the process of digital transformation towards Autonomous Networks? So Shanthi, if I could ask you that first.

Shanthi Ravindran: Thank you. So operators, as operators and vendors, we know that we have to make this move to make our networks more agile and autonomous. But operators have broad-field network. They have so much equipment that they have in the network that have been evolving over the many years. So definitely the first challenge that they face is how to make the basic changes in their architecture. That will set them up for this Level 4 change.

Stephen Shao: Now we are talking a lot about LLM in telecom domain. We all agree that LLM is a key technology for Autonomous Networks Level 4. Definitely, the first challenge is LLM application in telecom domain. But, Huawei believes that LLM can be more effective only if it’s associated with a strong digital base. The digital base offers accurate network service modeling, and all-scenario analysis and management capabilities. Then, LLM can achieve decision-making and a closed-loop. So the second challenge is building a solid digital base, which is prior to the LLM. In short, no digital base, no real autonomy.

I think two points should be considered when we evaluate the capabilities of a digital base.

The first one is full stack visibility of the network, such as data accuracy and comprehensiveness. Precisely, we need to evaluate whether object behavior, such as the traffic direction, can be monitored in real time across the entire network in the long term.

The second one is to evaluate whether the automation capabilities can solve operators' O&M issues, improve O&M efficiency, and guarantee service experience.

Host: OK. So how can operators leverage FBB Autonomous Networks to accelerate digital transformation, Shanthi?

Shanthi Ravindran: Stephen, that was a very good description of how to set up the foundation. We've been doing a lot of research in this base over the past few years, and we have a report which talks about the OSS change that is needed. So far, software was always in the operation support systems, but now there is more and more software in the network, including the LLMs. So the question now is how can we use that to transform our networks overall? And the key point here is, I would use the word "intent". We know intent, and we have been programming for this over many years. And many systems, including Huawei systems, have fairly robust intent modeling and intent operation, which means that inside a particular or autonomic domain, the management system can by itself control and has full independence of autonomy over it. So once that is there, no other system is telling it to do something. So it has full autonomy, it has visibility of the resources, etc. Then the LLM, which is available now, can come into giving some kind of language interfaces to the higher layers. And people who are managing these layers in the service domain, in the operators, they now have the flexibility to be able to talk to the network and get what they need. So that can lead to a higher level transformation. As long as we follow through into the service domain and towards the business outcomes, where we can describe what is the business outcome that can make this whole transformation occur. And that helps for operators to get their investments to make these changes. That's a key point.

Stephen Shao: As I mentioned before, in order to leverage Autonomous Networks, we should build the most advanced digital base first and then activate the most professional LLM in telecom domain .

And I have to say that AI agent is a very important key technology that combines LLM and the digital base. We should utilize the LLM chain of thoughts, we call it COT, to plan the optimal solution and network process, and then perform optimization and closed-loop based on data and tools of the digital base.

The digital base and LLM can be respectively regarded as a map application and the decision-making master. This collaboration can be named "One Map and One Master".

Huawei's "One Map" refers to the Network Digital Map of iMaster NCE, which implements the full stack visibility for FBB networks.

The full stack visibility means all the essential data from the network level to the application level from our network , network service status awareness, and real-time data collection.

By implementing full-stack visibility and unified management from active to passive networks, it can facilitate the digital transformation of operators and provide insights and a decision-making basis for LLM.

Based on One Map, iMaster NCE offers "One Master", which is an intelligent application based on FBB telecom foundation model.

As we know, corpus determines model performance.

Huawei is a leading E2E telecom solution provider, serving global carriers for more than 30 years. It has accumulated corpus, about tens of billions telecom specific tokens, from product documents, customer cases, expert experience, solution designs, network optimization experience, and so on.

The high-quality corpus will deliver the core value of the large language model. So, Huawei's "One Master" builds a unified FBB O&M LLM , which makes the most professional and optimal network decision.

The collaboration of "One Map" and "One Master" enables intelligent network decision-making by itself, just like a self-driving car in the road , for automatic service configuration, early risk detection, real-time traffic optimization, and fast service recovery, helping operators quickly evolve toward Autonomous Networks L4.

Host: What are the best practices implemented based on Huawei's "One Map and One Master"

Stephen Shao: The Network Digital Map offered by Huawei's NCE has been widely used in more than 50 operators worldwide, such as China Mobile, and MTN in South Africa.

China Mobile Guangdong is a typical digital transformation case. It is the most complex network as we know in the world . It has over 140 million subscribers and about 500 thousand 5G base stations. With such a large scale, it has a strong demand for network autonomy.

We've been collaborating on the network digital map with China Mobile Guangdong since 2019.

Technologies such as In-situ Flow Information Telemetry and intelligent incidents are used to implement precise monitoring and diagnosis of 5G traffic, achieving full-time and full-path monitoring and root cause analysis.

The cooperation between Huawei and China Mobile Guangdong helps China Mobile Guangdong keep its leading position in terms of network quality and user satisfaction in China.

In 2023, China Mobile Guangdong and Huawei jointly implemented the industry's first Telecom Foundation Model Application – Huawei NCE Net Master. Net Master delivers unique value for customers, including comprehensive telecom knowledge, and service intent understanding and fault self-diagnosis capabilities.

During fault diagnosis, it significantly reduces the complexity of network configuration and troubleshooting. It reduces the Mean Time To Repair from hours to minutes by implementing intelligent conversational interaction based on the AI agent.

In the service recovery scenario, intent-based APIs from the network digital map help our field engineers to use mobile apps to obtain network resources and fault information in real-time. Therefore, the entire troubleshooting process can be visualized and manageable, significantly reducing the site visit cost and the average fault response time by 83%.

It has been fully implemented in 13 cities across Guangdong province and has achieved remarkable results.

Host: Okay, and thank you both for joining us today and thank you for watching. Goodbye.

Shanthi Ravindran：Thank you.

Stephen Shao: Thank you.