ADN Paves the Way Towards High-level Network Autonomy
Ongoing network evolution with high-level autonomy is a key driver of the intelligent world and autonomous driving networks are paving the way ahead.
Technological innovation is happening rapidly in multiple domains, with new technologies like 5G, cloud computing, and AI driving massive changes to economies and society, and positioning the digital economy as a core engine of global economic growth. The growth of the digital economy has significantly outpaced annual GDP growth over the last three years in 47 countries, and the combined value of the local digital economy jumped from US$30.2 trillion to US$38.1 trillion.
The digital economy has become an integral part of life and work, helping new digital applications go mainstream. The public is quickly adopting and adapting to applications like remote work, education, and healthcare; VR-based social networking and entertainment; scenario-based smart home experiences; and smart manufacturing and logistics. Even the 2022 FIFA world cup benefited from this new wave of digitalization, implementing a video assistant referee (VAR) system for the first time. The emergence of new applications is in turn facilitating rapid growth within the digital economy.
In 2022, 460 million people in China worked remotely and more than 350 million people took advantage of online education. Hybrid models of work, education, entertainment, and social networking that utilize both online and in-person interactions have become normal since the beginning of COVID-19.
Communications networks are the tool that connects individuals, homes, and organizations, creating new critical infrastructure essential to everyday life. The role of home networks now extends well beyond entertainment and information access. In today's world, it is essential for work, education, and even production, with enterprise networks acting as core production systems. The demand for clear and smooth live online classes, secure and reliable remote access, and ultra-low-latency interconnection in industry scenarios is setting a higher bar for networks.
And high-quality network connectivity has emerged as a core facet of the digital economy.
To ensure optimal service experience, networks with guaranteed SLAs need to be ready anytime, anywhere, and planning, construction, maintenance, and optimization must become automated and intelligent. To meet these needs, network acquisition, delivery, and management are shifting towards a Network as a Service (NaaS) model. In fact, most carriers have launched some kind of digital strategy to accelerate network evolution on the road to becoming NaaS and digital service providers. Research has found that more than 90% of global carriers have included network automation in their core business strategies for 2023.
Technological innovation and industrial transformation within the communications industry will also transform the landscape for industrial productivity. Twenty years ago, IP technology reshaped the forwarding architecture of communication networks. Ten years ago, cloud technologies profoundly impacted network management and control architecture. And now, AI technology is set to be embedded into every layer of network architecture to drive high-level network autonomy over the next 10 years. Upgrading the telecom industry with automation will unleash the full potential of carrier networks and empower industries with digital capabilities. AI will make automated networks intelligent, replace machine-assisted manual labor with human-assisted machine labor, and enable real-time experience awareness and optimization of on-demand services.
This wave of network automation and digitalization has resulted in a wide array of industry players coming together to define an industry-specific definition of "autonomous networks". TM Forum, standards organizations, such as 3GPP, the GSMA, ETSI, and CCSA; carriers, network equipment suppliers; and OSS software developers have all contributed to this definition, which includes a unified vision and grading criteria to ensure network upgrades are both effective and regulated. By the end of 2022, more than a dozen leading operators around the world announced their intention to achieve level-4 network autonomy by 2025, making it clear that they believe autonomous networks will be key to both digital and intelligent transformation as well as rapid growth.
Autonomous Driving Network (ADN) is Huawei's solution for the autonomous network industry. For three years, from architecture to application, Huawei has explored autonomous networks and evolved ADN to deliver level-3 autonomous networks.
ADN leverages innovations in high-precision digital twins, compressive sensing, parallel and incremental simulations, and self-closed loop of knowledge to achieve this. Huawei has also established joint innovation working groups with leading carriers around the world, such as China Mobile, China Telecom, AIS Thailand, MTN, and Vodafone to help carriers use ADN to simultaneously improve network quality and revenue, O&M efficiency, and energy savings.
These carriers can now offer upgraded home broadband services using new capabilities like precisely perceiving user experience and accurately locating QoE issues, reducing poor-QoE rates by 83%. Potential customer identification has also helped improve FTTR and gigabit service marketing success rates from 3% to 10%. These achievements have enabled proactive O&M based on user experience and improved carriers' brand reputations for home broadband services.
Intelligent fault analysis enables feature extraction, cleaning, and the aggregation of alarm and associated data. It also uses machine learning training and inference to identify correlations and derivative relationships between alarms to identify and locate fault root causes, increase fault identification accuracy to over 90%, reduce fault tickets by over 30%, and shorten average recovery time for each fault by at least 30 minutes. These improvements represent a massive leap in O&M efficiency.
As part of ADN, our base station energy-saving solution creates models for base station energy consumption, frequency bands, coverage, and performance, and dynamically generates power-saving policies. This enables site- and time-specific precise energy saving, and ensures optimal network experience and energy efficiency, reducing average energy consumption per site by more than 10%.
The ADN solution features AI-powered network elements (NEs), networks, and services that can help carriers achieve level-3 network autonomy by 2023 and level-4 network autonomy by 2025. ADN is designed to provide end users with a new digital network service experience characterized by zero wait times, zero touch interactions, and zero fault experiences. It also provides network O&M personnel with self-configuring, self-healing, and self-optimizing intelligent networks to increase the efficiency of their work. Looking to the future, Huawei will continue investing heavily in this ADN solution to achieve three key targets: native AI, single-domain autonomy, and cross-domain collaboration.
At the NE layer, ADN uses smart hardware and computing-network convergence to build hyper-converged perception capabilities. Innovative technologies like optical iris and intelligent optical modules, as well as real-time sensing devices, are deployed on NEs to enable active, millisecond-level sensing that replaces passive, minute-level sensing. Integrated technologies like low-power computing and stream computing pre-analyze and compress NE data, improving the local inference and decision-making capabilities of NEs.
At the network management layer, high-precision network maps provide real-time simulation services. To enable intent-based networking, ADN adopts predictive maintenance, multi-objective adaptive optimization, key technologies such as polymorphic configuration, high-precision simulations, on-duty digital employees, telecom network foundation models, and the dynamic identification of multiple fault models. These functions realize real-time online simulations, predictive global optimization, and adaptive closed-loop control on large-scale complex networks.
Cross-domain collaboration capabilities are improved at the service collaboration layer through platform evolution, service process optimization, and personnel transformation. Platforms are evolved from siloed support systems to unified systems that integrate all-domain data and expert experience, transforming service processes from network-oriented and driven by work orders to experience-oriented processes that require zero touch. The low-code development system makes it easier for CT personnel to go digital and intelligent, evolving O&M team members into versatile DICT talent.
Over the past three years, the industry has reached a consensus on autonomous networks in terms of vision, architecture, grading criteria, and core ideas. We are now well on our way to turn this strategy into reality. At the Autonomous Networks Summit held in October 2022, TM Forum, together with 54 industry partners, including CCSA, China Mobile, and Huawei, jointly released the AN4.0 white paper Autonomous Networks: Empowering Digital Transformation – from Strategy to Implementation. In this paper, the partners lay out critical technological challenges that must still be overcome to achieve high-level network autonomy. Overcoming these challenges will require collaboration between industry partners on interface standards, evaluation systems, and business models. They will specifically require the industry to:
The telecom industry has long viewed autonomous networks as a goal that is just out of reach. However, the potential to revolutionize the industry and create an ICT foundation for the future intelligent world where all things are connected with intelligence is undeniable. Autonomous networks will soon be a key production tool for the digital economy. And so, with its ADN solution, Huawei plans to ride this wave of network autonomy and collaborate with industry partners to realize this once remote ambition that will unfold a true intelligent world.