As more and more Communications Service Providers (CSPs) launch 5G, they find themselves grappling with the challenge of network complexity. The CSPs are introducing Network Function Virtualization (NFV), cloud-native computing and Software-Defined Networking (SDN) technologies in the 5G networks even as they continue to operate the legacy networks. This combination of legacy and new-age networks is creating unprecedented network complexity. The traditional network, built for siloed networks, are not capable of managing this complexity. 

Powered by Artificial Intelligence (AI), autonomous networks will be crucial in managing the new-age networks. The CSPS and vendors have already started to explore intelligent autonomous networks to enhance network efficiency and mitigate this complexity. The CSPs use AI to ensure superior network performance, provide exceptional customer experience, improve sustainability, and maximize the returns from network investment. 

Recently held AI Summit Conference – Intelligent Autonomous Networks at Mobile World Congress Shanghai highlighted how the telcos are using the power of AI to cost-effectively enhance network efficiency and improve customer experience. It also explored how AI is enhancing human creativity and paving the way for a better future. Several prominent members from the communications industry participated in the AI Summit and discussed the opportunities and challenges of using AI in network automation. 

Industry consensus has been formed 

Global mainstream industry organizations, including TM Forum, ETSI, 3GPP, ITU, IEEE, IETF, NGMN and CCSA, are collaborating to support the telcos in their journey towards Autonomous Networks. 

In 2020, with more than 25 CSPs and vendors participating, the industry organization already published several research reports about Autonomous Networks. In October 2020, TMF and six standard organizations, including 3GPP, ETSI and GSMA, among others, collaborated to set up an Autonomous Networks standard promotion working group. The multi-SDO group will release the Autonomous Networks standardizing progress every six months. All this indicates the industry has already reached the consensus that the future of the telcos is Autonomous Networks.

Further, TM Forum released the Autonomous Networks whitepaper version 2.0, Autonomous Networks empowering digital transformation for smart societies and industries. In this whitepaper, the industry promotes how to migrate to Autonomous Networks, building the three-layer and four closed-loop technical architecture to build three operation capabilities of self-serving, self-fulfilling, and self-assurance which will provide zero-wait, zero-touch zero-trouble experience for enterprises and end-users. 

Network automation picks up pace 
CSPs have already started moving from legacy networks to autonomous networks to be able to provide superior quality to their end-users. The latest market survey by TM Forum indicates that 58% of the interviewed service providers had a well-defined Level 3 autonomous networks strategy for the next two to four years.
China has taken the lead in leveraging AI technology to enable the service providers to leverage network automation to improve network quality and bring down operating costs. China Mobile is a case in point. It runs the world’s largest mobile network with 59,000 O&M personnel and invests about 140 billion annually on the Operations and Management (O&M) network. The service provider, which has already launched 5G services, uses automation to enhance this complex network’s efficiency. 
"Currently, we are in the Level 1 to Level 2 phases of Autonomous Networks. We will continue to increase investment and accumulation in network O&M automation and intelligence in the next five years, and strive to reach Level 4 by 2025 and have advanced network O&M support capabilities," says Li Huidi, Vice President, China Mobile.
"We must further accelerate the intelligent transformation of network operations and management data, fully utilise automation and intelligence technologies to implement end-to-end O&M support capabilities oriented to customer perception, and achieve agile service provisioning and efficient network O&M," Huidi adds. China Mobile won TM Forum’s `Outstanding Contribution’ award, with the collaboration with Huawei to evolve its network to autonomy. 
Huawei is playing a crucial role in empowering the telcos to transform the network operations for new-age use cases, like Industry 4.0. Huawei’s Chief Digital Officer, Dr Haiping Che proposed the concept of `autonomous domain’, which "accelerates the transformation of traditional networks from execution functions to execution intents and implement intelligent closed-loop with cognitive capabilities." Leveraging autonomous domains as the basic management unit, they can develop a flattened architecture featuring single-domain autonomy and cross-domain collaboration. 
In his address at Mobile World Congress Shanghai, Dr Che also spoke about the growing need to build an industry O&M knowledge ecosystem, which can be used to train the AI model. “The industry should jointly accelerate the maturity of data O&M (DataOps), model O&M (ModelOps), and development O&M (DevOps) engineering technologies to improve the performance, scalability, explainability, and reliability of algorithm models.” 
“Huawei ADN uses public cloud services to provide data asset management, model training, and update services for the entire industry. It uses the cloud-ground synergy technology architecture to continuously inject algorithm models and gradually improve algorithm models, bridging the gap between machine learning applications and software continuous integration and development,” he added.
Journey From L1 to L5 Network automation 
Network automation is at the center of network transformation for Telefonica, a prominent service provider in Europe. "We are creating the infrastructure and we are creating the culture to put automation at the core of our network and IT strategy. We are not using any platform which is not virtualized. This new model is an extraordinary challenge for us, but it is the biggest opportunity for us to change the traditional way of things in IT and network …we are facing the future bravely," says Enrique Blanco, Chief Technology Innovation Officer (CTIO), Telefonica.  
China Unicom is another Chinese telco that also started leveraging AI to add intelligence to its networks. "Network intelligence is an inevitable trend. We would be deploying AI along three main lines: applying AI to all scenarios of planning, construction, maintenance, optimization and operation; using AI for holographic insight of service objects and for intelligent matching among terminals, network and business," elaborated Cheng Xinzhou, Director of Network Intelligent Operation Research Center, China Unicom Technology Research Institute at the MWC Shanghai event held recently.  
The telcos journey towards autonomous networks is not devoid of challenges. "It is important to keep systemic perspective in mind while designing AI-centric systems because communications networks are very complex and applying different algorithms in different parts of the network may not be very efficient," says Elena Fersman, Research Director, Ericsson. 
With several global CSPs starting to leverage Autonomous Networks, it is clear that there is a consensus in the industry to adopt technologies that promote Autonomous Networks. The joint efforts of industry parties have helped the company reach an agreement on the concept definition, technical architecture and standard collaboration of the Autonomous Networks. Further cooperation among the stakeholders will help in the refinement of grading standards and evaluation methods for different capability levels, clarify specific objectives of each development phase and jointly define the industry pace based on innovation practices, and accelerate network interconnection to achieve the vision of L5 Autonomous Networks.