On June 3rd, 2024, an ETSI NFV cloud-native roundtable was held as an activity following the ETSI NFV#46 F2F meeting hosted by ZTE Corporation in ZTE Hotel, Xi’an, China. This roundtable focused on sharing views and challenges for the current and future Telco Cloud and its relationship with AI technologies. Representatives from Chinese organizations like China Mobile, China telecom, and China Unicom, China Communications Standards Association (CCSA) SP1, China Academy of Information and Communications Technology (CAICT), as well as delegates from vendors discussed  hot topics and technological prospects in the industry of Telco Cloud native & AI, so as to promote the development of relative standardization activities.

Wang Weibin, Chief Scientist of ZTE Product Planning, mentioned: "The NFV technology was born more than 10 years ago. With the modernization of networks and the large-scale commercial use of 5G networks, NFV has achieved supporting large-scale deployments in the first ten years. Technologies such as virtualization, containerization, cloud native technologies, SDN/NFV convergence, edge computing, and hardware acceleration are emerging, which have become the new subordination model of telecom networks. At present, the emergence of GenAI and LLM has given new impetus to the development of NFV, and the network is facing the challenges of increasingly complicated structures, diversified scenarios, and higher requirements for reliability and automation. Through this NFV round table, we hope to bring together the strength of the industry to jointly promote and accelerate the development of NFV technology in cloud native and intelligent standards, laying a foundation for the development of network technologies and the communications industry in the next ten years”.

Lin Yang, Chief Planning Engineer of ZTE Computing and CN Products, delivered a speech entitled "Cloud Native and Intelligent Evolution Enables the Digital Transformation of Operators", describing ZTE's thinking in the direction of cloud native and AI standardization of telecom networks. He pointed out that due to the lack of standardization, different operators and equipment vendors in the industry follow their own trends in deploying cloud native networks. To speed up the standardization and commercial implementation of cloud native and AI in NFVs, Lin Yang made multiple suggestions and said that ZTE is willing to work with global partners to help operators transform and build digital infrastructure networks that meet the needs of the intelligent society in the future.

Zhang Hao, Director of the Institute of Network and IT Technology Institute of China Mobile Research Institute, said: "With the development of 5G and Generated Artificial Intelligence (AI) technologies in recent years, operators' services are rapidly shifting from traditional voice and data connection services to diversified and multi-industry application services. Service innovation is more frequent and requires diverse computility, higher network bandwidth and lower latency, as well as intelligent capabilities. Meanwhile, the current NFV network with simple virtualization is unable to support these new services, need to evolve in the direction of hyper-distributed architecture, deepening of cloud native, in-depth computility and network convergence, and intelligence, to satisfy the demanding requirements of services for the network".

Wang Xuliang, Deputy Director of the Cloud and Network Operation Technology Research of China Telecom Research Institute, introduced how China Telecom achieves in-depth integration of telecom service platforms and network functions through the cloud and network integration strategy. This strategy not only promotes service cloudification, but also provides customers with one-stop integrated services that integrate cloud and network concepts through cloud infrastructure. Since 2011, China Telecom has been exploring service cloudification and evolving its service model. By 2020, China Telecom had clearly proposed the cloud-network integration strategy to accelerate the cloud-based transformation of network services. In his speech, he emphasized the importance of technological innovation, proposed the goal of enhancing infrastructure security and disaster recovery capabilities, and looked forward to the future vision of building a unified service and technology architecture, covering the development of intelligent computing, enhanced PaaS, high-reliability architecture, software and hardware integrated computing, and distributed hybrid multi-cloud architecture.

Doctor Chang Cao, Director of the Research Department of Next Generation Internet of China Unicom Research Institute, pointed out: "With the development of network business demands and network architecture, from the perspective of technological evolution, the development of communication’s underlying infrastructure will be directed towards nativity and intelligence. Communication’s underlying infrastructure is currently undergoing evolution through three distinct phases: phase 1 is "Hardware-software Integration/NFV plus AI for Net", phase 2 shifts towards "NFV/Cloud Native plus Net for AI", and phase 3 will enter the realm of "Cloud Native plus AI Native". This trajectory illustrates an overarching trend of enhancing functionalities while simplifying the architectural design".

Haining Wang, Vice Chair of CCSA SP1 from Intel, said: "CCSA SP1 (NFV) is the first ad hoc cross-TC standards project team established by CCSA. Since its establishment, CCSA SP1 has formulated more than 20 industry standards related to the management and orchestration of NFVs. Based on the ETSI NFV specifications, CCSA SP1 has added specific requirements for NFVs deployment of Chinese operators, providing good support for the development and deployment of China's NFVs industry, and meeting the urgent needs of the industry in recent years. At the beginning of the second decade of NFV development, CCSA SP1 plans to continuously promote the evolution of the open NFV architecture towards agility, reliability, and intelligence in accordance with the future telecom cloud requirements, enhance cooperation and coordination with ETSI ISG NFV, and jointly promote the development of the global NFV industry".

Zhiruo Liu, a senior engineer of China Academy of Information and Communications Technology, said: "With the cloud-based transformation of network NFV, network flexibility and openness are improved through software and hardware decoupling, and large-scale and multi-vendor product integration and interoperability are difficult, which has become a huge challenge that operators must break through to promote network cloudification. This report will sort out the development trend of intelligent network in terms of standard, industry, and deployment, analyze the key to empowering intelligent network O&M, and share the considerations on model evaluation and standardization".

Lian Yang, an architect of Telco Cloud from Huawei, made a speech, which mainly described the challenges and opportunities in the evolution from Telco cloud native to AI native, including three parts. The first part described the telecom cloud native evolution roadmap from Huawei's perspective; the second part mainly involved the challenges and opportunities faced by the telecom cloud in the AI era, which are briefly described from two directions: Cloud4AI and AI4Cloud; the third part attempted to define the reference logical architecture of AI Native in 5G-A and provides some suggestions for future development.

Shuling Wang, Chief Scientist of Computing Network from AsiaInfo, said: "Facing the business needs for a large amount of edge computing power, AsiaInfo propose the computing native network solution, which means without the need for additional hardware devices, by decoupling the idle computing power of a large number of 5G BBUs from communication services, it provides computing applications with computing force support, achieving flexible and low-cost computing power supply. AsiaInfo's computing native network is realized "communication and computing in one" by enhancing the functions of 5G BBUs and 5G MEC without changing the basic functions and architecture of the network".

Two panel discussions were held around "Telco Cloud Native" and "AI".

In Telco Cloud Native panel discussion, Yoshihiro, ETSI ISG NFV Chair from DOCOMO was the moderator and guests discussed in detail how to reconstruct the telco cloud system architecture to evolve to a cloud-native architecture, how to provide a foundation for telco cloud infrastructure to provide cloud-native network functions, and finally discussed in-depth the technical challenges in telco cloud evolution to cloud-native coordination with orchestration systems.

In the AI panel discussion, Lingli Deng, ETSI ISG NFV SOL Vice Chair from China mobile was the moderator and guests discussed the future trends, opportunities, and challenges of introducing AI into the telecom field, typical application scenarios, and key elements of AI in NFV and cloud-native networks. They fully confirmed the necessity of introducing AI into telecom networks, especially in large models, which will bring enhanced customer service experience, simplify network maintenance, create new market opportunities, ease risk management, and facilitate new service innovation.

In this regard, the ETSI NFV will continue building consensus and promote prosperity in the telecom industry, smoothly incorporating new innovative technologies and concepts by pushing global cooperation and standardization.