Abstract: In the context of China’s “Country in the East and Compute in the West” strategy and the integrated development of computing and networks, computing power is a common demand. In order to solve the complex distribution pattern and distribution area of computing power nodes in the computing power network, as well as computing network services, perception, and Due to various factors such as connections and scheduling capabilities, problems such as computing power network collaboration and complex transactions have arisen. Baidu Blockchain and China Unicom Research Institute have jointly developed a computing power service matching and trading platform to promote the development of computing network integration services for domestic companies. . This project uses distributed BaaS as the basis of chain management to build a distributed computing power service matching and trading platform. Through the matching algorithm, the production capacity and demand of the computing power service matching party, the demand side and the supply side are matched, and through artificial intelligence application scenarios , showing the application effect and efficiency of matching services.
This case has been included in the "China Blockchain Development Report" (2023) published by Social Science Literature Press
✦
Research Background
✦
(1) Computing power is the infrastructure of the digital age
Computing power is the infrastructure of the digital age. With the rapid development of information technology, computing power is as important as basic resources such as water, electricity, and gas. The "2021-2022 Global Computing Power Index Assessment Report" points out that as the global digital economy continues to grow steadily, the proportion of the digital economy is expected to reach 41.5% by 2025. The trend of computing power and GDP shows a significant positive correlation. For every 1 point increase in the computing power index of 15 key countries, the country's digital economy and GDP will increase by 3.5‰ and 1.8‰ respectively. Computing power has become the core support and driving force for mining the value of data elements and promoting the development of the digital economy. In 2022, from the strategic level, energy policy and technical perspective, the country will comprehensively launch the national project of "Calculate in the East and Compute in the West", coordinate the layout and build a national hub node of the national integrated computing power network, and promote the infrastructure of computing power. In the process of computing power infrastructure, a key computing power coordination system needs to be formed, and a blockchain-based computing power matching and trading platform emerged at the historic moment.
(2) Emerging industries require super computing power, and computing power is unevenly distributed
Computing power consists of a large number of computing devices. Due to constraints such as chip technology, the room for growth in computing power is limited. It is very important to efficiently and rationally utilize computing power distributed in the cloud, terminal and network edge, and to make full use of computing power resources distributed across the country. Emerging industries such as AI large models, the Metaverse, and the Internet of Things rely on large amounts of data and super computing power. Tens of thousands of companies that are deeply involved in such application scenarios are the demand side of computing power and lack computing power infrastructure. Communications operators such as China Unicom have a large number of edge computer rooms and edge computing power advantages after network cloudification, and are the owners and exporters of computing power. After the establishment of the computing power matching and trading platform, the flow of computing power between the demand side and the output side is freely allocated by the "invisible hand" of the market, becoming one of the necessary solutions to solve the scarcity of computing power and efficiently allocate computing power resources. .
(3) Introducing blockchain technology to make the computing power trading platform safer and more credible
The current centralized computing power trading platform is difficult to meet the requirements of high-frequency and trusted transactions. After the introduction of blockchain ledger and trusted computing and other technologies, the opening and service-oriented computing resources have changed from traditional centralized and static methods to decentralized and dynamic ones. , safe and trustworthy new way of development. Operators open up the computing power of edge computing, consider combining it with the blockchain in the decentralized transaction model, and introduce the concept of "token economy"; technically, they adopt more efficient alliance chain technology to better meet the needs of supervision and auditing. . In actual deployment, blockchain platforms or applications can be installed and deployed on edge computing servers to provide blockchain technology and capability support for different application scenarios.
✦
Preliminary exploration of direction
✦
Blockchain is promoting the transformation of the "Internet of Information" into the "Internet of Value". For the computing power network, the blockchain transforms computing power into a quantifiable and transferable "commodity" on the network, helping computing power to exert key value as a production factor. Specifically, computing power services and transactions in the computing power network rely on the decentralized, low-cost, privacy-protecting and trusted computing power trading platform of the blockchain. There are three roles in this computing power trading platform.
1) Computing power supplier: Contribute idle computing power, such as mobile terminals, computers, game consoles, small data centers when enterprises are idle, etc. The supplier will notify the computing power trading platform of the scattered computing power through registration or update;
2) Computing power demand side: Put forward computing power requirements to the computing power platform, obtain the container nodes allocated by the computing power platform, and make payments, perform tasks, etc.;
3) Computing power trading platform: The computing power trading platform serves as an intermediary role in dealing with sellers and buyers of computing power, maintaining, managing, and dispatching computing power resources, and providing economical, efficient, safe, and reliable computing power services.
In previous transaction models, information between buyers and sellers was not transparent. In future ubiquitous computing scenarios, the network can provide computing power to users as a transparent and open service capability. In the process of computing power trading, the contributors of computing power (computing power sellers) are separated from the users of computing power (computing power buyers). Through scalable blockchain technology and containerized orchestration technology, the computing power contributors’ resources are integrated. Fragmented computing power provides economical, efficient, decentralized, real-time and convenient computing power services for computing power users and other participants in computing power services.
✦
Application landing
✦
Baidu Blockchain and China Unicom Research Institute work together to independently develop a computing power service matching trading platform to promote the development of computing and network integration services for domestic companies. Build a trustworthy computing power matching trading platform of Suanwang, including:
Build a decentralized storage framework for the trusted computing power matching and trading platform of the computing network to enable data matching to participate in the trusted collaboration of multi-party services;
Build a computing network blockchain network to support the development of relevant smart contracts, support data compliance storage and use, manage the storage network, and meet the needs of efficiency, credibility, and privacy protection;
Build an agent computing network to support the running of machine learning algorithms in the TEE environment, realize computing power control and orchestration, obtain demand data and match the data required for calculations, and output computing power suitable for customer needs to the computing power control and orchestration platform to achieve data scheduling.
Based on the China Unicom Chain framework, assist in building the core framework and technical components of the computing power matching and trading platform of the computing network chain and demonstrate and verify them to enhance the collaborative service capabilities of multiple computing networks.
Figure 1 Design framework of China Unicom’s trusted computing power matching and trading platform
This project uses distributed BaaS as the basis of chain management to build a distributed computing power service matching and trading platform. Through the matching algorithm, the production capacity and demand of the computing power service matching party, the demand side and the supply side are matched, helping the computing power supply side to Unused computing power is sold to companies in need, creating new sources of income for the supply side.
China Unicom Network's trusted computing power matching and trading platform is based on China Unicom Chain, making full use of China Unicom Chain's existing capability components, including multi-party secure computing services, digital identity services, cross-chain services, etc., combined with CUBE-Net3.0 Computing Network The characteristics of collaborative services, build the core framework of the trading system. The system has several unique features:
1. Institutions can join the computing network’s trusted computing power matching and trading platform as a computing power provider. Each computing power provider in the computing network can provide a number of blockchain nodes to form a computing network service chain.
2. Each institution can also provide storage services and join the decentralized storage network to become a data storage node. Storage networks store metrics for different computing resources and tasks.
3. For a single computing power provider, the blockchain is built into the provider node and the service orchestration and management node (or closely integrated) to provide chain-native trusted service capabilities. These built-in logical functions interact with the computing network service chain to serve the computing power provider's service collaboration and data storage needs.
4. The computing network service chain interacts with the computing network brain, and uses the computing network service chain to record and store information related to computing network orchestration, scheduling, management and service.
5. Shuangwang Brain has trusted computing capabilities and solves the data island problem caused by data privacy and security among multiple parties through a secure computing model based on TEE. At the same time, TEE hardware is used to provide an enterprise version of a secure hardware server and a corresponding basic operating environment for multi-party data calculations.
6. Computing Network Brain models the measurement indicators of different computing resources and tasks, determines scheduling goals, and provides relevant scheduling algorithm support. In addition, for distributed architecture, a fault-tolerant scheduling method for distributed multi-nodes is provided, and the final scheduling strategy is generated based on the scheduling target for the scheduling strategies on multiple nodes.
7. The computing network brains of different operators form a computing network brain collaborative agent computing network. Through the computing network brain collaborative agent computing network, the computing network brains of each operator interact trustworthyly to improve the relationship between computing networks. cooperation. At the same time, different computing network service chains collaborate with each other through cross-chain and other technologies. With the support of the computing network brain collaboration agent computing network, computing networks of different operators can cooperate with each other in a trustworthy manner.
✦
Outlook
✦
With the network as the core for information exchange and computing power as the core for information data processing, the close integration of computing power and network has become an industry consensus. Communications operators represented by China Unicom are actively exploring new business forms and application scenarios in the era of computing power networks to adapt to future artificial intelligence, cloud games, live broadcasts, autonomous driving, intelligent security and industrial machine vision and other computing power and powerful technologies. Interactive business requirements.
Communication operators use blockchain technology to build trusted computing power infrastructure, which can achieve more effective resource allocation and management, improve network efficiency and create new sources of revenue. The application and development of blockchain in computing power networks will promote artificial intelligence Innovations in the fields of intelligence (AI), distributed applications (DApp), and the Internet of Things (IoT) have consolidated the important role of computing network infrastructure in building a future digital society and helped the digital economy flourish.
references
[1] One of a series of interpretations of the "Eastern Digital and Western Calculation" project | The "Eastern Digital and Western Calculation" project helps my country comprehensively promote computing infrastructure, 2022/03/07, China Government Network
https://www.ndrc.gov.cn/xxgk/jd/jd/202203/t20220317_1319467.html
[2] Tang Xiongyan, Cao Chang, Zhang Shuai, He Tao, Liu Ying, Du Wei, Zheng Ruobin, Tang Xiaoqin: "China Unicom Computing Power Network White Paper", 2019
[3] Zhang Shuai, Liu Ying, Pang Ran, and Cao Chang of China United Network Communications Co., Ltd. Research Institute: "China Unicom Computing Power Network Architecture and Technical System White Paper", 2021
[4] International Data Corporation IDC, Inspur Information, Tsinghua University Global Industry Research Institute: "2021-2022 Global Computing Power Index Assessment Report", 2022
[5] China United Network Communications Co., Ltd. Research Institute, China United Network Communications Co., Ltd. Guangdong Branch, Huawei Technologies Co., Ltd.: "White Paper on the Evolution of Cloud Network Convergence to Computing Network Integration Technology", 2021