Summary
computing power network is the key infrastructure for the construction of digital economy . We define the computing power network as an infrastructure that integrates cloud-net-end and integrates scheduling computing power resources. With the development of the digital economy, the rapid growth of data has led to the expansion of computing power scale. The computing power network can schedule unbalanced distribution of computing power resources between regions and directly provide computing power services through the network. We believe that the construction of 5G network has laid the underlying technical foundation for the integration of cloud and network for computing power networks. The construction of computing power networks can provide more flexible and low-cost computing power services for the development of metaverse and vertical industries. We believe that computing power network is an important infrastructure in the early stage of the development of the metaverse and an important driving force for the development of the digital economy.
moves from collaboration to integration, and East and West computing promotes the development of computing networks. computing power network is divided into infrastructure, orchestration management and operation service layers from the architecture. The orchestration management layer conducts computing power management, controls network scheduling, and provides computing network scheduling capability interfaces upward, which is the key "brain" in computing power network architecture. At present, the construction of computing power network is still in the initial stage of ubiquitous collaboration in infrastructure to the development stage of unified computing and network resource integration and unification. The three major operators are orderly promoting the construction of computing power network. We believe that computing power network is an integral part of new infrastructure, and the implementation of the Eastern Digital and Western computing policy will help the nationwide integrated computing power resource scheduling and infrastructure construction, laying a solid foundation for the construction of computing power network. The operator led by
and participated in the joint creation of an industrial ecosystem. computing power network is an organic whole composed of multiple digital infrastructure industries. We believe that operators are the leading participants in the computing power network industry chain, and have natural advantages in building the computing power network ecosystem: 1) Operators rely on network resources to seize the terminal connection entrance; 2) The high bandwidth, large connection and low latency of 5G network are the prerequisites for the data transmission of computing power network ; 3) The operator's own IDCh and cloud computing layout provide infrastructure resources, and the western IDC layout effectively reduces computing power costs; 4) The operator's massive individual users and government and enterprise users are the key user entrances to promote computing power services; 5) The industrial chain has rich resources, and coordinates with multiple parties to promote the construction of computing power network ecosystem. Cloud computing and IDC manufacturers are also key players in the computing power network industry chain.
empowers the application of the meta universe and vertical industries, and computing power promotes the digital transformation of social . The integrated service capabilities and technical advantages of computing power network can solve the pain points of computing power demand in metaverse and vertical industry application scenarios. As computing power network moves to the computing network integration stage, we believe that computing power network will be further widely used in many industries such as industry, energy, transportation, and construction, promoting the vigorous development of my country's digital economy industry.
risk
key technology research and development is lower than expected; downstream applications development is lower than expected.
Text
Report Framework
Chart 1: Framework logic—Computing Power Network: New Information Infrastructure of the Digital Economy
Source: CICC Research Department
Computing Power Network is a key infrastructure for the construction of the digital economy
Data has become a new production factor, and computing power network has emerged
Data is a production factor in the new era. Computing Power is the rate at which equipment processes data. The rapid expansion of data volume has driven a significant increase in computing power demand. With the development of the digital economy, we are in an era of exponential growth in data. Statista’s data statistics show that the global data volume is expected to be 47ZB in 2020 and is expected to reach 2,142ZB by 2035. In the article "Computing Process", the 2018 Nobel Prize in Economics, William D. Nordhaus defines computing power as follows: "Computing power is the amount of information data that can be processed per second according to changes in internal state of the device." Computing power is a positive indicator for measuring the data processing capabilities of the device. It usually uses the number of floating point number operations (Flops) per second as the unit of measurement. According to the forecast in the Ministry of Industry and Information Technology's "Three-Year Action Plan for the Development of New Data Centers (2021-2023)", by the end of 2023, the country's total computing power will exceed 200 EFLOPS. China Academy of Information and Communications Technology predicts that the global computing power scale will reach 3,300 EFLOPS in 2025, and the computing power scale will grow with a CAGR of more than 50% from 2020 to 2025.
Chart 2: Global data volume exponentially grow
Source: Institute of Information and Communications Technology, Statista, CICC Research Department Note: The mark is estimated/predicted value
Chart 3: Global computing power scale, 2016-2025E
Source: IDC, Gartner, China Institute of Information and Communications Technology, TOP500, CICC Research Department Note: This calculation comes from the China Institute of Information and Communications Technology, and is summarized based on the basic computing power scale, intelligent computing power scale and supercomputing computing power scale.
computing power network refers to the infrastructure that integrates cloud-net-end and integrated scheduling computing power resources. traditional network facilities only undertake the basic functions of information transmission and are the "data channel" connecting users and computing resources; while computing power networks can build integrated services for data, computing resources, and networks. For example, customers need to analyze and process a batch of data during their operations, and need to use diversified computing resources such as CPU, GPU, and storage. Compared to purchasing different software and hardware resources from different suppliers and then purchasing the operator network to schedule data, the computing power network can directly call various computing power resources accessed through the network to achieve one-stop computing power services.
computing power network builds the network foundation for the development of the metaverse. computing power network combines cloud computing, data centers, 5G networks and other infrastructures to provide basic computing resources for metacosmic applications and vertical industries' computing power needs. With the continuous integration and improvement of cloud and network technologies, computing power networks can solve the problem of expansion of network and computing resource demand brought about by the development of the metaverse, and provide more flexible and low-cost computing power services. We believe that the construction of computing power network is an important infrastructure that needs to be completed in the early stages of the development of the metaverse.
collaborative scheduling to reduce costs, computing power network helps digital construction
Currently the Internet has entered a transition period from Web2.0 to Web3.0 (refer to the report "Web3.0: New Paradigm Opens a New Stage of the Internet"). The development of the metaverse is accompanied by the birth of new applications such as digital twins, smart cities, and virtual people. The development of the virtual world is accompanied by higher requirements for network transmission and computing.
The current computing power supply and demand distribution shows decentralized and unbalanced characteristics, and the computing power network is required to schedule. With the development of 5G, AI and other technologies, the Internet of Things has become possible. Various terminals such as smart homes, smart cars, and smart factories may become output devices for computing power, and the distribution of access devices shows a trend of decentralization and sinking. In addition, my country's computing power resource deployment is characterized by unbalanced characteristics. The computing power supply in developed eastern regions is in short supply, while the IDC computer room listing rate in rich resource areas in western regions is insufficient. The development of the digital economy requires a network that can connect various devices in the cloud, edge and end to coordinate computing power planning.
Figure 4: From cloud computing to ubiquitous computing, the evolution of computing network architecture
Source: China Unicom "White Paper on Computing Power Network Architecture and Technology System" (October 2020), CICC Research Department
Metauniverse Era, the application of new technologies such as VR and AR is closely related to large bandwidth and low latency networks. When we make video calls and click on web pages, the delay response of 100-200ms is still acceptable, but in 3A games [1], the same degree of delay will cause players to feel stuck, resulting in a significant decline in experience. According to Subspace statistics, for every 10ms increase in game delay time, the user's game duration will be reduced by 8%. The application of VR and AR technology has also further expanded the bandwidth requirements. The application of computing power network can directly dispatch computing power resources for the network for access, provide cloud rendering, and achieve low-definition high-definition video enjoyment. In addition, the computing power network can deploy large-scale computing resources on western computing nodes, match them with the demand in the eastern region through direct connection channels, reduce business costs with low latency requirements, and promote the promotion and popularization of commercial applications such as AI and high-definition video that require a large amount of computing.
Figure 5: Requirements for delay and network speed of VR and AR (single device)
Source: CICC Research Department
Figure 6: Real-time data of 5G cloud testing platform speed in the first quarter of 2022
Source: China Academy of Information and Communications Technology, CICC Research Department
Computing power Network investment drives economic output and has become an important driving force for the development of the digital economy. According to the information disclosure of the Ministry of Industry and Information Technology at the 2022 China Computing Power Conference, in the past five years, the scale of my country's core computing power industry has increased by more than 30% per year on average. By the end of 2021, the computing power scale ranked second in the world. Among them, in 2020, the scale of my country's computing power industry reached 2 trillion yuan, driving direct economic output of servers, computer room equipment, etc. by 1.7 trillion yuan, and bringing productivity efficiency improvement through computing power resources, indirectly driving economic output of 6.3 trillion yuan. According to the calculations of Information and Communications Institute , every 1 yuan invested in the computing power network industry will drive an average of 3 to 4 yuan of GDP economic growth. In May 2022, the "Package Measures to Solidly Stabilize the Economy" issued by the State Council once again mentioned promoting the construction of new infrastructure and the development of the digital economy. We believe that the construction of computing power network will constitute an important part of the transformation and development of the digital economy, and computing power network is a key foundation for new infrastructure and digital economy construction.
crosses the wind of 5G, the computing power network pre-supply
reviews the development history of 1-5G, and the construction of basic network facilities has led to subsequent changes in communication media and productivity. From to 4G era, the evolution of communication technology and the improvement of data transmission speed have changed people's lifestyles and promoted innovation in corporate business models. Networks and communication infrastructure determine the rate of information interaction, and on this basis, directly drives the improvement of production efficiency.
Figure 7: Major changes in the 1G-5G era
Source: Ministry of Industry and Information Technology, C114 Communication Network, CICC Research Department
From the perspective of the transformation process of 1G-5G, the development of communication technology is always preceded by the transformation of productivity. 5G technology further increases the speed limit of wireless data transmission to more than 1Gbps, which makes it possible to connect smart devices to each other. Not just mobile phones, but also all kinds of wearable devices, furniture, cars and even various machines and equipment in factories can become part of the communication network. eMBB, uRLLC, and mMTC constitute the three major application scenarios of 5G, corresponding to enhanced mobile broadband, low latency and high reliability communication, and massive Internet of Things communication. Among them, eMBB corresponds to mobile mobile phone communication, while uRLLC and mMTC further expand the B-side industry Internet application scenarios, such as vertical industry scenarios such as industrial Internet and Internet of Vehicles, providing a technical basis for data transmission for future intelligent manufacturing. Therefore, to achieve substantial changes in future production relations, it is necessary to first complete the construction of computing power network as the basis.
G network uses virtualization technology NFV to fully launch the core network to the cloud, laying the architectural foundation for cloud network integration. NFV technology applies standardized network functions to unified hardware devices, changing the strong binding relationship between software and hardware in the original network equipment. The core network function after decoupling of software and hardware can be flexibly deployed on public cloud, private cloud and hybrid cloud. This cloud-native architecture realizes the first phase of "cloud network integration" - that is, cloud network resource layout can be flexibly connected and managed in a unified manner. In addition, the 5G core network has many characteristics such as service architecture, control and bearer separation, slicing, fixed mobile fusion, etc. [2], which laid the foundation for the subsequent construction of computing power networks.
moves from collaboration to integration, and East Data and West Calculation promotes the development of computing network
computing power network architecture and key technologies: From infrastructure to computing power services
computing power network architecture can be divided into three layers, from bottom to top, the infrastructure, orchestration management and operation service layers.
► The infrastructure layer is an organic combination of "computing" and "network" resources. The infrastructure layer consists of a variety of distributed computing resources and underlying networks.Among the key technologies in the infrastructure layer, all-optical network (OTN) can provide data transmission with large bandwidth and high speeds; SRv6 can improve network flexibility and realize network programmability; in addition, it also includes edge computing , cloud native and other technologies.
► Orchestration management layer is the core management center of the entire computing power network. The functions that the orchestration management layer needs to implement mainly include: 1) managing the computing power resources of the infrastructure layer, metrics, perceptions, identifications, and outputs of computing power; 2) controlling network scheduling to realize multi-dimensional computing power and network resources connection, addressing, and allocation services; 3) providing computing network scheduling capability interfaces for the operation service layer. Among the core technologies of the orchestration management layer, OpenStack and K8s assist in unified management of computing network resources; container technology realizes flexible resource scheduling. In addition, deep learning, big data and other technologies can be orchestrated and scheduling.
► The operation service layer provides a one-stop supply of products and services. computing power operation service layer is the top layer of computing power network, realizing the encapsulation and operation of computing network resources, and combining blockchain technology to create a trading service platform for social computing resources. Customers only need to care about the service they receive in the end, and do not need to consider the location and status of the computing power deployment, so as to achieve intelligent services without any effect. Core technologies include blockchain, computing power trading, computing power packaging, etc.
In addition to the three-layer architecture of the computing power network main body, security technology is also needed to protect the security of the entire data, and green energy-saving technology reduces the energy consumption during the operation of the entire computing power network.
Figure 8: Computing power network architecture
Source: China Mobile's "Computing Power Network White Paper" (November 2021), CICC Research Department
Figure 9: Key technologies of computing power network
Source: China Mobile's "Computing Power Network White Paper" (November 2021), CICC Research Department
Computing power network development stage: From computing network collaboration to integration symbiosis
According to China Mobile's division in the book "Computing Power Network", the construction of computing power network can be divided into three stages: starting, developing, and leap. ► Initial stage (2021-2022): ubiquitous collaboration between computing networks. The core goal of this stage of is to optimize the layout of computing power network infrastructure and realize the docking and collaboration between cloud computing and network resources. At this stage, the computing network facilities at the orchestration management layer and the infrastructure layer are still independent of each other, but through resource collaboration, the integrated computing network output can be achieved to the customer layer.
► Development stage (expected from 2023 to 2027): Unification of computing and network integration. In this stage of , the computing network breaks the independence situation at the infrastructure level, realizes unified management, orchestration, scheduling, operation and maintenance at the orchestration management level, and continues to output integrated services at the service level. At this stage, the logical architecture of computing and network gradually converges, resource management and service scheduling are integrated to output the true "computing network products". This stage is a key stage in the development of computing power networks, and determines the final "integration and symbiosis".
► Crossing stage (expected from 2028 to 2030): The calculation network is integrated and symbiosis. is the ultimate goal of computing power network, to achieve complete symbiosis between computing and network in protocol and form, and to break the technological boundaries. From the user's perspective, storage, computing and the network are becoming more integrated, and users can directly use multi-level and unified computing resources.
Figure 10: The development stage of computing power network and the current construction status of various operators
Source: China Mobile's "Computing Power Network White Paper" (November 2021), "China Telecom Cloud Network Integration 2030 Technology White Paper" (November 2020), CICC Research Department
We believe that the current computing power network construction progress is still in the process of moving from the initial stage to the development stage. At present, the construction of computing power networks of the three major operators is progressing in an orderly manner. For example, China Telecom announced at Digital China Construction Summit that it has entered the cloud network integration 3.0[3], emphasizing factor aggregation and realizing a new form of cloud network integration at the infrastructure level. China Mobile released computing network service 1.0[4] at the 2022 China Computing Power Conference and actively assisted in the construction of a unified international standard system related to computing power networks.
New infrastructure and East Data and West Computing help the development of computing network
With the vigorous development of my country's digital economy, the society's demand for digital transmission and other industries has further expanded, and computing power networks will become the infrastructure for the construction of a new generation of digital economy. In April 2020, the National Development and Reform Commission explained the meaning of " new infrastructure " for the first time [5]. The computing power infrastructure represented by data centers and intelligent computing centers belongs to information infrastructure, and encourages local governments to actively make arrangements and accelerate construction progress.
Chart 11: Related national policies for computing power network
Source: Official websites of various departments, CICC Research Department
In view of the current situation of unbalanced computing power distribution, the eastern and western computing promotes the integrated resource linkage of computing networks. In May 2021, the National Development and Reform Commission and four other departments triggered the "High Implementation Plan for Computing Power Hub of National Integrated Big Data Center Collaborative Innovation System", and proposed to lay out the national computing power network national hub node structure in eight regions: Beijing-Tianjin-Hebei, Yangtze River Delta, Guangdong-Hong Kong-Macao Greater Bay Area, Chengdu-Chongqing, Guizhou, Inner Mongolia, Gansu and Ningxia. In addition, this "Plan" also mentions that "network integration" and "computing power integration" require the construction of a direct network for high-speed data centers. This direct-connected network is used to support large-scale computing power scheduling and promote resource linkage between multi-clouds, between clouds and data centers, and between clouds and networks. This marks that the "computing power network" has been included in the development scope of national new infrastructure. We believe that the implementation of the Eastern Digital and Western computing policy can effectively solve the problem of uneven regional distribution of computing power resources, and through the construction of a direct connection channel, the demand in the eastern region is matched with the computing power resources in the western region, so as to achieve coordinated scheduling of computing power and data. We believe that the East Data and West Computing project is a key link in the early stage of computing power network construction, and can realize the overall planning and reasonable scheduling of computing power, network, data and energy nationwide, laying a solid foundation for subsequent computing and network integration and industrial services. The
operator leads the company and participates in the joint creation of the industrial ecosystem
computing power network is an organic whole composed of multiple digital infrastructure industries. We believe that operators are the leading participants in the computing power network industry chain, and have joined hands with many participants such as upstream and downstream cloud manufacturers, IDC manufacturers, etc. to build a new underlying facility for the digital economy.
Chart 12: Computing power network industry chain
Source: Institute of Communications Industry, CICC Research Department
Operator: Leading the construction of computing power network ecosystem
We believe that operators have natural advantages in building computing power network ecosystem, which is mainly reflected in the following points: 1) In terms of connection , the operator's network has access to a large number of and rich types of terminal devices, such as mobile phones, computers, wearable devices, VR/AR devices, smart homes, Internet of Things terminal devices, etc. Operators rely on network resources to seize the terminal connection entrance. ) In terms of transmission, 5G network has the characteristics of high bandwidth, large connection and low latency, which can be used as a prerequisite for data transmission of computing power networks. ) In terms of computing power resources, the number of IDC available resource pools and cloud computing power scale of operators are ahead of other third-party data center manufacturers in China. At the same time, the western IDC resource layout is complete, which can provide large-scale computing power while effectively reducing costs for customers. ) In terms of market , operators have a large number of individual users and government and enterprise users, and a wide range of customer resources are the entrance for operators to promote computing power services. ) Ecological aspects: , operators have rich industrial chain resources, and can collaborate with the strength of various industrial chain participants such as communication equipment merchants, terminal equipment manufacturers, cloud service manufacturers, software service providers, etc. to create a computing power network ecosystem. Based on its existing unique advantages, we believe that the three major operators can build computing power networks from four dimensions: "network", "computing power", "commercial services" and "ecology".
"Network" - Technology improvement, Optimize the network basics
operators accelerate the upgrade of 5G network technology and optimize the computing power network basics. The ultimate goal of computing power network is to meet the cloud needs of thousands of industries and provide differentiated and high-quality services. This requires that its underlying network, namely 5G network, can flexibly adapt to diversified business needs with one network.At present, operators are accelerating the construction of a new generation of information infrastructure represented by 5G. According to data from the Ministry of Industry and Information Technology, as of the end of June 2022, a total of 61.854 million 5G base stations have been built nationwide, covering all cities, counties, urban areas and more than 90% of towns and towns in China; fixed broadband has been upgraded from 100 megabytes to gigabytes, and the proportion of fiber optic users has increased from less than 10% in 2012 to 95.2% at the end of June 2022.
How to use one network to meet the diversified requirements for network performance in different application scenarios? The best solution is 5G network slicing. network slice can transform the original single physical network into a multi-level virtual logical network independent of each other. Different logical networks correspond to different performance and application scenarios, which is equivalent to turning an ordinary knife into a multi-function Swiss army knife. And a subnet can be further divided again, for example, eMBB slices can also be divided into virtual reality subslices, smartphone subslices, etc. Operators have also developed a new business model, which can create network slicing functions and rent virtual networks with different performances to service providers engaged in different businesses. After use, they can also cancel slicing, which significantly improves resource utilization and network flexibility.
Chart 13: 5G key technology - network slicing
Source: Netmanias, CICC Research Department
G network is divided into access network, bearer network, and core network. The implementation of network slicing function needs to be completed by three parts. The access network is a neural terminal that plays a role in connecting all types of terminals. For example, a base station can be used for wireless access, fiber-optic access, etc.; the bearer network is an neural network , which mainly uses fiber-optic cables to transmit and schedule data between the access network and the core network; the core network is the brain, composed of network element devices with various functions, managing and controlling various services. In the 5G era, the various components of 5G network have undergone key technological changes to continuously meet innovative business needs. Chart 14: Basic composition of 5G mobile communication network
Note: SDN is a software defined network (Software defined network), SRv6 is the abbreviation of SR (Segment Routing) + IPv6 (Internet Protocol version 6, Internet Protocol version 6), UPF is the user plane function (User plane function), and MEC is edge computing (Mobile Edge Computing). Source: CSDN, CICC Research Department
"Computing Power Resources" - the three major operators combine infrastructure to layout the computing power network
The three major operators are not only builders of the underlying network, but also important suppliers of computing power resources. According to data from the Forward Industry Research Institute, in 2020, telecom operators accounted for nearly 50% of the IDC market market share . The operator's IDC resource layout is highly consistent with the national "East Data and West Calculation" hub node layout. It gives priority to seizing scarce computing resources in first-tier cities, builds super-large data centers in the western region, and has the advantages of leading layout and the increase in the listing rate of Western data center brought by the East Data and West Calculation policy. We believe that operators are expected to usher in a longer dividend harvest period compared with third-party data centers.
Chart 15: Operator IDC resource layout is highly consistent with the layout of the East Digital and West Calculation hub nodes
Note: The map comes from the official website of National Ministry of Natural Resources , and the data is as of the end of 2021.
Source: National Development and Reform Commission, company official website, CICC Research Department
► China Mobile : According to the "China Mobile Computing Power Network White Paper" released by the company in 2021, the company will focus on computing power and the network as the foundation. The goal is to build a computing power network that integrates multiple factors such as network, cloud, digital, intelligence, security, edge, end, and chain (ABCDNETS), and ultimately realize "computing power ubiquitous, computing and network symbiosis, intelligent orchestration, and integrated services."
In terms of computing resource layout, according to the company's announcement data, the company has planned and deployed the "4+3+X" data center, built a "N+31+X" cloud resource pool system, and reserved a number of site resources for edge access.As of the end of 2021, China Mobile has 407,000 IDC racks available to the outside world, and has put into production of more than 480,000 cloud servers and more than 480,000 new cloud servers; in 2022, China Mobile's planned capital expenditure will reach 48 billion yuan, and the company expects to add about 43,000 new external racks available to the outside world in 2022 and more than 180,000 new cloud servers.
Chart 16: China Telecom Cloud Network Integration Target Technical Architecture
Source: "China Telecom Cloud Network Integration 2030 Technology White Paper" (November 2020), CICC Research Department
Chart 17: China Unicom Computing Power Network Architecture
Source: China Unicom "White Paper on Computing Power Network Architecture and Technology System" (October 2020), CICC Research Department
► China Telecom : The company continues to deepen its cloud network integration strategy. Its goal is to form an integrated integrated technology architecture through virtualization, cloudization and service, and ultimately achieve a simple, agile, open, integrated, secure and intelligent cloud network integration resource supply.
In terms of computing power resource layout, according to the company's announcement data, China Telecom has built an integrated cloud and IDC resource layout of "2+4+31+X+O". As of the end of 2021, the company has more than 700 data centers, and the scale of external service racks has reached 470,000, and the computing power scale has reached 2.1 EFLOPS. China Telecom plans to invest 6.5 billion yuan in IDC in 2022, and the company expects to add about 45,000 new externally available racks; cloud resources plan to invest 14 billion yuan, and 160,000 new cloud servers are expected to add 160,000 new cloud servers, and the computing power scale increases by about 81% year-on-year to 3.8 EFLOPS by the end of 2022.
► China Unicom : The company has built a CUBE-Net3.0 network innovation system. With this as its target architecture, China Unicom has set up a new "5+4+31+X" data center to create an integrated computing and network service of "connection + perception + computing + intelligence". In terms of computing resource layout, according to the company's announcement data, as of February 2022, China Unicom has more than 880 data centers, with a total scale of more than 300,000 and carrying more than one million servers. The data centers are mainly distributed in the Beijing-Tianjin-Hebei region, the Yangtze River Delta, the Pearl River Delta, Shandong-Yu, Sichuan-Shaanxi-Chongqing, Mongolian-Guizhou and other regions.
"Commercial Product" - Enriching Industry Solution
Computing Power Services launched by operators have become the first step for computing power networks to develop commercially. At present, China Telecom and China Mobile have released commercial computing power service products. Taking the product of Telecom as an example, on July 12, Shanghai Telecom released its first public computing power service product for operators. This product is based on the advantageous resources of China Telecom Tianyi Cloud , providing a variety of services for general computing power, intelligent computing power, and supercomputing computing power, supporting real-time and flexible scheduling capabilities of cross-regional computing power networks, and realizing integrated operation services for computing power, network, and storage resources cloud networks. We believe that the launch of operators' commercial computing power service products is an important progress in the construction of computing power networks. By directly providing customers with efficient and flexible computing power support, we promote industry solutions innovation, realize the integration of its own cloud, network, computing and other resources, and expand the scale of digital services.
Chart 18: Introduction to Shanghai Telecom Public Computing Power Service Products
Source: Official official account of China Telecom Shanghai Branch, CICC Research Department
As the scale of computing power resources expands and computing power services gradually becomes rich, we believe that a trading platform integrating computing power resources will become a new computing power network business model. Considering that computing power resources are more complex than hydropower, we believe that technologies such as blockchain and AI will become key technologies for trading platforms. The computing power trading platform can gather distributed scattered computing power in various places, promote the flow of idle computing power between various regions and service providers, realize the allocation of and fair pricing of computing power on demand, and at the same time reduce the threshold for enterprises to use computing power services.
"Ecology" - Take the lead in co-construction, mutual benefit and win-win
operators have become the most critical link in the computing power network industry chain due to their cloud network resource advantages. They should actively explore opportunities for industrial chain cooperation and take the lead in building a win-win ecosystem.
► On the underlying resources of the computing power network, building a computing power network for the whole society also requires introducing partners from all walks of life, such as cloud manufacturers, IDC manufacturers for infrastructure, and third-party service providers that provide vertical industry services. Operators provide cloud and network resources, connect all links of the industrial chains of all parties through the network, play a key role of "scheduler" in the computing power network, identify the computing power needs of downstream customers, and schedule computing power resources in different locations in real time as needed to realize automated service deployment, optimal routing and load balancing.
► In computing power network application services, operators can join hands with industry partners to continue to deepen cooperation around computing power network technology transformation and vertical solutions, jointly develop products, share technical experience, and explore more application innovations of computing power networks in various industries such as manufacturing, transportation, medical care, education, and ports.
In addition, operators actively promote the construction of a unified international standard system related to computing power networks, laying the foundation for future sharing of multi-platform and multi-vendor capabilities. , for example, China Mobile has joined hands with 26 units from industry, academia and research to work together to achieve technological breakthroughs in multiple directions such as computing power routing and computing power native, and promote the construction of international standard systems such as ITU/IETF. We believe that operators promote the construction of a unified computing power network standard system to help build a unified and generalized computing power ecosystem and enhance the influence of enterprises.
cloud vendor: an important participant with cloud resources and technical advantages
cloud computing vendor is an important participant in computing power network. With the help of existing cloud computing resources and technical advantages, it can participate in computing power network construction from multiple levels of infrastructure, cloud network services, and security management.
cloud computing vendors are the main providers of public cloud computing resources and cloud technology, and are important suppliers of computing power infrastructure. According to the IaaS+PaaS market data of China's public cloud in the second half of 2021 released by IDC, Alibaba Cloud, Huawei Cloud and Tencent Cloud successively occupy the top three market share, with a total share of more than 60%. In addition, Internet manufacturers have made early efforts in the field of cloud computing, combining the advantages of cloud computing technology with the transmission and scheduling advantages of computing power networks, which can provide customers with a better service experience.
Chart 19: China's public cloud IaaS+PaaS market share, 2021H2
Source: IDC China, CICC Research Department
Chart 20: Tencent Cloud Integrated Computing Power Infrastructure
Source: Tencent official website, CICC Research Department
and operators work together to launch cloud network products to serve corporate customers. cloud manufacturers can use SD-WAN, dedicated line and other technologies to provide customers with cloud dedicated line, cloud VPN, 5G cloud dedicated line and other products by renting bandwidth from operators. Taking Tencent Cloud's 5G cloud dedicated line as an example, we cooperate with operators to use Tencent TSAC (edge access and acceleration platform) to connect with the UPF gateway of the operator's core network in the core network, directly dispatch operator's core network resources, ensure the reliability of cloud dedicated line services, and achieve a high integration of cloud computing and 5G network technology. We believe that as the development and maturity of computing power networks in the future, cloud manufacturers can participate in the construction of computing power network systems as enablers of underlying cloud computing basic resources and technologies.
IDC manufacturer: Dongshuxixi computing drives computing power layout optimization
IDC manufacturer is an important participant in the infrastructure layer of the computing power network, providing rental services for various distributed computing power resources. In the process of computing power network construction, IDC mainly lays out development directions such as east-west computing, green and low-carbon, and computing and network integration. The computing resource layout and performance revenue of
IDC manufacturers are showing a good growth trend. my country's IDC manufacturers continue to increase their computing resources layout. According to data from the Ministry of Industry and Information Technology, the scale of data center racks in my country has steadily grown. According to the 2.5kW standard rack statistics, as of the end of 2021, the scale of data center racks in use in my country has reached 5.2 million, with an average annual compound growth rate of 33.03% in the past five years. Driven by the huge demand for computing power in the digital economy era, my country's data center performance revenue continues to grow rapidly.According to data from the China Academy of Information and Communications Technology, in 2021, the market revenue of my country's data center industry reached 150.02 billion yuan, with an average annual compound growth rate of 30.69% in the past three years. We believe that the rapid growth of IDC manufacturers' scale confirms the expansion of the market's demand for computing resources.
Chart 21: my country's data center rack scale
Source: Information and Communication Development Department of the Ministry of Industry and Information Technology, CICC Research Department
Chart 22: my country's data center market size
Source: China Information and Communication Research Institute, CICC Research Department
is a game of calculation in the east, and the western region undertakes some eastern computing power demand. The demand stratification carried by domestic data centers will be clearer. The eastern data center is close to the city and is responsible for businesses that require low latency such as industrial Internet, AI reasoning, edge computing, etc.; while the utilization efficiency of computing power resources and green energy in the western region will be significantly improved under the concentration of demand, serving large computing power scenarios such as algorithm analysis, video rendering, supercomputing, storage backup, etc. We believe that the layout of East and West computing fully utilizes the unique advantages of the eastern and western data centers, and is highly consistent with the requirements of computing power network for IDC optimization layout.
Green and low-carbon is the only way for IDC to develop. East Digital West Calculation Policy requires the PUE (electricity utilization efficiency) requirements of the four western regions (Guizhou, Inner Mongolia, Gansu, Ningxia) to be less than 1.2, and the PUE requirement of the four eastern regions (Beijing-Tianjin-Hebei, Yangtze River Delta, Guangdong-Hong Kong-Macao Greater Bay Area, Chengdu-Chongqing) to be less than 1.25. According to ODCC data, the current PUE level in each region is higher than the requirements of the eastern regions to be more energy-saving space. In addition, IDC manufacturers are also actively solving their high-energy consumption problems, optimizing energy consumption through technologies such as liquid cooling and AI energy efficiency management to improve efficiency. We believe that policy support and technological innovation will effectively promote the evolution of data centers toward green and energy-saving, and help the green construction of computing power networks.
Chart 23: Average PUE and calculation requirements for data centers in various regions across the country, 2021
Source: ODCC, CICC Research Department
calculation network integration. In the early days, the computing power network of mainly focused on computing network collaboration, and will eventually evolve to the computing network integration stage. In the future, computing power network can coordinate the scheduling of multi-dimensional resources in the data center, providing efficient, flexible and professional services for upper-level applications, and achieving the best computing resource utilization rate, optimal network efficiency and optimal user experience.
empowers the meta-universe and vertical industry applications, and computing power promotes the digital transformation of society
meta-universe
computing power network will serve as an important infrastructure to empower the meta-universe. has entered the meta-universe era. The popularity of VR and AR will make the transmission of high-definition videos a wide demand. The wide application of technologies such as virtual humans and real-time rendering will further increase the computing demand for heterogeneous computing power such as GPUs. The high-definition VR and AR videos of the Metauniverse have typical characteristics of big data and high speed. In the application scenarios of the Metauniverse video, the computing power network can realize the following functions: 1) Scheduling heterogeneous computing power resources to meet real-time cloud rendering and reduce the needs of C-end hardware equipment; 2) Large-bandwidth network access provides guarantees for the smooth experience of multiple users; 3) Using edge nodes to complete distribution and partial rendering needs, reduce operation delay, and improve user experience. At present, computing power network has been applied in the fields of live video broadcasts, high-definition videos, etc. We believe that relevant capabilities can continue to be further promoted in meta-universe scenarios.
Chart 24: Schematic diagram of the application of computing power network in the field of live video
Source: China Unicom "White Paper on Computing Power Network Architecture and Technology System" (October 2020), CICC Research Department
Intelligent Driving
The low latency, high bandwidth and cloud-side collaboration capabilities of computing power network provide key technical support for real-time data computing of intelligent driving. In the era of vehicle connections, smart driving, as one of the most typical applications in high bandwidth, low latency and high reliability scenarios, is promoting changes in the traditional automobile market. Autonomous driving has created a huge demand for data transmission. According to Google's test, the amount of data generated by a vehicle can reach 1GB per second. For driving safety considerations, autonomous driving puts forward strict requirements on the stability and delay of data transmission. The computing power network's ultra-low delay, high bandwidth and cloud-side coordination capabilities provide key guarantees for real-time control signals such as vehicle steering, acceleration and braking.
In intelligent driving, the computing power network can provide high bandwidth and low latency networks for the car itself and road infrastructure, as well as computing power services with edge computing power. The data generated by the vehicle is initially processed at nearby edge nodes and uploaded to the cloud before returning, and contact the traffic system data such as light poles and traffic lights of smart cities, and provide driving information for the vehicle after cloud analysis.
Chart 25: Cloud, edge and end collaboration scenarios
Source: China Unicom "White Paper on Computing Power Network Architecture and Technology System" (October 2020), CICC Research Department
Wireless Medical
Computing Power Network is expected to continue to empower the upgrade and extension of wireless medical scenarios. is based on key technologies such as 5G, artificial intelligence and cloud computing. Wireless medical care can make full use of limited medical manpower and equipment, break the bottlenecks of insufficient medical resources and uneven development in various regions, and provide informatized, mobile and remote medical services in disease diagnosis, medical monitoring and surgical treatment.
In wireless medical application scenarios, each task puts forward high requirements for data transmission and connection speed. For example, remote ultrasound based on AI and haptic feedback, theoretically requires 30Mbps data rate and 10ms maximum delay. The computing power network can carry the needs of low latency and high bandwidth:
► 1) For remote surgery, in terms of video and image transmission, the low-latency computing power network can ensure the quality of picture transmission and meet the real-time high-definition transmission requirements required for precision operations; in terms of surgical instructions, the low-latency computing power network improves the accuracy of one-end instruction issuance and the timeliness of the command response on the other end;
► 2) For remote diagnosis, the support of the computing power network can solve the problem of bandwidth and network delay in the video network during the diagnosis and treatment process, complete the integration of remote auxiliary diagnostic examination operations, remote consultation and other information resources between hospitals, and realize remote operation and real-time diagnosis and treatment guidance;
► 2) For remote diagnosis, the support of the computing power network can solve the problem of bandwidth and network delay in the video network during the diagnosis and treatment process, complete the integration of remote auxiliary diagnostic and examination operations, remote consultation and other information resources, and realize remote operation and real-time diagnosis and treatment guidance;
► 1► 2) For remote diagnosis, the support of the computing power network can solve the problem of bandwidth and network delay in the video network during the diagnosis and treatment process, complete the integration of remote auxiliary diagnostic and examination operations, remote consultation and other information resources between hospitals, and realize remote operation and real-time diagnosis and treatment guidance;
► 1► 3) For the coordination of first aid resources, the computing power network can timely process complicated first aid information and plan and allocate first aid resources in real time, support real-time remote dispatch of first aid vehicles and personnel, timely update of first aid patient information and status, etc., to gain valuable time for saving lives.
Chart 26: Wireless medical network logic architecture
Source: Huawei "Wireless Medical White Paper" (January 2018), CICC Research Department
Risk warning
key technology research and development is not as expected. Currently, the construction of computing power network is in the early stage of development. Technologies such as computing power orchestration and scheduling, computing power perception network are still in the early stage of R&D, and relevant companies need to invest more resources. The computing power network must achieve unified scheduling of existing computing power resources across regions, cross-blue-end, and cross-computer power types, and carry out load optimization, etc. There is still uncertainty in technology research and development. The research and development of key technologies is not as expected, which will affect the overall progress of computing power network construction.
downstream applications develop less than expected. is currently less commercialized in the metaverse, and the demand for large-scale computing power is limited. Considering that the development of downstream industries requires all-round development support such as chips, AI, software and hardware, we believe that the development of downstream applications may lead to the commercialization of current computing power networks being less than expected.
[1] 3A refers to a lot of time (A lot of resources), a lot of money (A lot of money). Simply put, it is a high-quality game with huge investment in resources.
[2] Source: Huawei official website https://www.huawei.com/cn/technology-insights/publications/huawei-tech/78/cloudcore-5g-evolution-cn
[3] Source: People's Posts and Telecommunications Chairman Ke Ruiwen: Give full play to the advantages of cloud and network integration to empower high-quality development of the digital economy https://mp.weixin.qq.com/s/HQF-Jrapq48PygRvs0ByDw
[4] Source: Mobile Cloud official account, create the future with computing network: China Mobile held the "Computing Power Network, Innovation and Development" forum https://mp.weixin.qq.com/s/2wujYHblsTI2BegfYVcyzA
[5] Source: CCTV Finance https://baijiahao.baidu.com/s?id=1664461956525441441wfr=spiderfor=pc
Article source
This article is excerpted from: "The Metaverse Series of Computing Power Network: New Information Infrastructure of the Digital Economy" released on August 9, 2022
Yu Mingyang SAC Practice Certificate No.: S0080121060043
Qian Kai SAC Practice Certificate Number: S0080513050004 SFC CE Ref: AZA933
Tong Siyi SAC Practice Certificate Number: S0080122070064
Legal Statement
Special Tips
This official account is not the publishing platform for the research report of China International Financial Co., Ltd. (hereinafter referred to as "CICC"). This official account is only forwarding some of the views of CICC's published research reports. If subscribers use the information contained in this official account, they may have ambiguity in understanding the key assumptions, ratings, target prices and other contents in the information due to lack of understanding of the complete report or lack of relevant interpretation. If subscribers use this information, they must seek guidance and interpretation from professional investment consultants.
The information and opinions contained in this official account do not constitute the bid or levy for the purchase and sale of securities or financial instruments, and the analysis and judgment of ratings, target prices, valuations, profit forecasts, etc. does not constitute investment advice on specific securities or financial instruments at specific prices, specific time points, and specific market performance. Such information and opinions do not constitute targeted and guided specific investments for anyone at any time. Subscribers shall evaluate the information and opinions in this official account, make investment decisions independently based on their own circumstances and bear investment risks.
CICC does not make any express or implied guarantees regarding the accuracy, reliability, timeliness and completeness of the information contained in this official account. CICC and/or its affiliated personnel shall not bear any responsibility for any consequences caused by the basis or use of the information contained in this official account.
This official account is only for CICC mainland Chinese customers. Any subscriber who does not meet the above conditions, please evaluate the appropriateness of receiving subscribed content before subscribing. Subscription to this official account does not constitute the basis of any contract or commitment, and CICC does not regard subscribers as CICC’s customers due to any simple subscriber to this official account.
General statement
This official account is only for forwarding some of the views published by CICC. The views contained in the profit forecast, target price, rating, valuation, etc. are based on a series of assumptions and prerequisites. Only by understanding all the information in the relevant reports can subscribers develop a relatively comprehensive understanding of the relevant views. For complete views, please refer to the complete report on CICC Research website (http://research.cicc.com).
This information is delayed forwarding compared to the report officially released by China Securities Corporation, and may no longer be accurate or invalid due to changes in the situation or other factors after the date of the report release. The opinions, assessments and forecasts contained in this information are only the views and judgments on the date of the report. These opinions, assessments and forecasts can be changed at any time without notice. The price or value trend of securities or financial instruments may be affected by various factors, and past performance should not be used as a premonition or guarantee for future performance. At different periods, CICC may issue research reports that are inconsistent with the opinions, assessments and forecasts contained in this information. CICC’s sales staff, trading staff and other professionals may publish market comments and/or trading opinions that are inconsistent with this information oral or written based on different assumptions and standards and using different analytical methods.
Under the circumstances permitted by law, CICC may mention in this information that the company is establishing or striving to establish a business or service relationship. Therefore, subscribers should consider that CICC and/or its related personnel may have potential conflicts of interest that affect the objectivity of this information view. For disclosure information related to this information, please visit http://research.cicc.com/disclosure_cn, and you can also refer to the specific research reports on related companies that have been released recently.
This subscription account is an official subscription account established and maintained by the Research Department of CICC. The copyright of all materials in this subscription account is owned by CICC. No organization or individual may forward, reproduce, copy, publish, publish, modify, copy or quote the content in this subscription account in any form without written permission.
