"4G changes lives, 5G changes society." But ADN is transforming 4G and 5G, and even the entire communications industry. 5G is an epoch-making technology, another dimension of "autonomous driving network", the significance is obviously not only epoch-making. It is often not perceived by the public, often confused with automotive self-driving technology, but it is the ultimate dream of the communications industry.
The "ultimate goal" of network development
The Autonomous Driving Network (ADN) has been, the most ambitious goal ever. It has been described as the "final state" of network development. MWC2022 includes this technology in the annual "Best Network Software Breakthrough Award" short list, which is also intended to express the communications industry's pursuit of the "ultimate goal".
Or it may be more vivid as this metaphor. In 1886, Beijing had its first railway; in 1971, the first subway in Beijing was officially opened, in 1992, the "Second Ring Road" was opened to traffic; in 1993, the Capital Airport Expressway was completed. After that, just like Yue Yunpeng's cross-talk: "Ah, ah, ah, five rings, you are one ring less than the six rings." ”
Stripped of the ridiculous elements, how similar are the development of urban transportation and communication networks. Beijing's railways, subways, ring lines, high-speed, ring lines are like network communications, transmission networks, mobile networks, data center networks and other different domains. The domains are crossed and crossed, and each domain itself is "expanding" at a very fast rate.
This is where the problem arises. Over the past decade, data has grown rapidly, networks have expanded rapidly, operational management costs have remained high, and operating expenses in the communications industry have grown faster than revenues. Not only that, the O&M model of operators is also facing great challenges: long service provisioning cycles and difficulty in dynamically adjusting bandwidth; manual-based passive O&M is inefficient and urgently needs to be improved; and users need to provide differentiated SLA experiences.
It is in this context that the "autonomous driving network" is proposed, hoping to deeply integrate digital technologies such as "artificial intelligence, cloud, and big data" with network technologies such as SDN and VNF in the information and communication industry, so as to realize fully automatic deployment of services and adaptive and self-evolution of the network.
The thinking of the self-wisdom network is very representative
It can be seen that in the face of challenges such as operating costs, business models, and operational and maintenance pressures, "autonomous driving networks" have become the only way for operators to transform digital intelligence. In this direction, ORAN is also a kind of thinking, but the spirit of Don Quixote can not save the challenges faced by the communication industry, software quality, platform monopoly, fault location, etc. are problems, and "computing" is not easy to go to the edge, is it necessary to go back?
The "self-wisdom network" has obviously been supported by more organizations and enterprises, and the progress from standard formulation to application landing is quite fast. After 2019, international standardization organizations such as TMF, ETSI, ITU-T, and GSMA have launched the "Autonomous Networks" project to formulate industry standards from the levels of architecture and grading (L1 to L5).
Many operators around the world have also developed strategies based on this standard and proposed a clear timetable for evolution. According to the survey, 70% of operators plan to reach L3 self-awareness networks by 2024. Leading operators in the industry such as China Mobile, China Unicom, and MTN have released their strategic targets for L4 by 2025.
But there's no denying it. The challenges are certainly there, and they are not small. After 30 years of development, the communication network has evolved into a super complex body, carrying multi-service, multi-specialty, multi-field, multi-vendor, and the complexity of China's communication network is the first in the world. To move toward autonomous driving, it is essentially an evolution from a super complex to a super intelligent body.
In this regard, Huawei's thinking is very representative. The autonomous driving network is a super intelligent body, but its target state is to direct the super intelligent brain of the whole network, or a group of distributed cerebellums; is it only necessary to transform the OSS from the top down to achieve automatic driving, or does it need oss/control/network element three layers to achieve automatic driving; is it to achieve automatic driving in the old network and mature business first, or to achieve automatic driving on the new network and new business?
In 2019, Huawei officially released the "Autonomous Driving Network" solution and innovated with major operators. With in-depth practice, the above questions have been answered one by one: complex agents cannot be achieved overnight, and distributed intelligence is a more practical solution. Since it is distributed intelligence, it is necessary to inject intelligence into the network element layer and the network control layer business layer to achieve automatic driving together. At the same time, any transformation should put business value first, consider the complexity of the network, and give priority to the innovation of autonomous driving in new services and new networks.
L3 is not far away
It is with the refinement of this idea that Huawei has built five architectural capabilities of "environment awareness, situation analysis, adaptive awareness, online planning and adaptive decision-making", and built and launched a leading L3-level autonomous driving network solution based on the principles of endogenous intelligence, single-domain autonomy, and cross-domain collaboration.
In fact, the autonomous driving network is quite similar to the automatic driving of the car, and it is always inevitable to be compared. Both have set the evolution standard of L0~L5, and they are currently in the evolution process of L2~L3. For example, the L2 level of autopilot is human-based and car-supplemented; L3 level is car-based and human-assisted.
The same goes for autonomous driving networks. In 2021, China's network as a whole is in the L2 scale application stage, and advanced provinces have begun to make breakthroughs in L3's dot-like applications in terms of 5G base station energy saving, home WiFi configuration self-closing loop, and SPN service self-activation. Among them, the home business, as the field with the highest level of integrated automatic driving, is in the leading position and is expected to take the lead in achieving L3.
Huawei's autonomous driving solutions are the first to reach the L3 level, or even the L3.5 level, and for different application scenarios, it also reflects the value of different aspects. In the toH quality home wide scenario, 85% of the quality difference business self-optimization can be realized, taking Zhejiang Mobile as an example, QoE poor users have dropped by 43%; the marketing success rate has increased to 68%; in the toB private line e-commerce scenario, 60% of the private line business is self-issued, and the order issuance and order issuance time is shortened from month to minutes and days.
Taking Tianjin Mobile as an example, its market share increased by 4% within one year, and more than 8,000 tob dedicated lines were newly ordered; in the toC intelligent O&M scenario, 60% of fault self-diagnosis can be realized to improve O&M efficiency. Taking Henan Mobile as an example, the number of transportation problem sheets has been reduced by 48%, supporting the identification of 27 kinds of transportation failures.
The ultimate goal of autonomous driving networks
Returning to the Mobile World Congress (MWC2022), Huawei's ADN autonomous driving network and MTN Group have entered the short list of the "Best Network Software Breakthrough Award", which is a recognition of the above ideas and the solutions for the implementation of ideas. MTN Group, South Africa's largest telecommunications operator with operations in more than 20 countries in Africa and the Middle East, has developed the "Ambition 2025" strategy for the first platform and first connectivity in the "Pan African Region".
Huawei's ADN solution's innovative self-intelligence network architecture injects intelligence into MTN Group's fixed network, mobile network, O&M, and other fields, enabling the network to continue to move towards automatic, self-healing, and self-optimization. Huawei's ADN solution's efficient network management capabilities are also helping MTN Group achieve one-stop service access to the cloud, intelligent fault repair, and network traffic tuning.
The overall resource utilization rate of mtN bearer network is increased by 30%, and the optimization efficiency of congestion traffic is shortened from 3 hours to 5 minutes. MTN deployed a dormant cell in its South African subnet for detection and repair pilot, and used Huawei's ADN fault self-diagnosis and self-repair capabilities to successfully reduce the overall time of discovery, analysis, and recovery of the scenario from an hour to 5 minutes.
Of course, this is just the beginning. The autonomous driving network is blocked and long, but the road is coming. MTN Group is expected to achieve the level of L3 network automation in 2023 and L4 in 2025, gradually changing the network with wisdom and achieving its ultimate goal. The ultimate goal of Huawei's autonomous driving network must also be to continuously improve the network value index and provide users with an automatic, self-healing, self-optimized, and autonomous network experience.