System and method for interconnection and intercommunication of trusted data spaces

Abstract

Disclosed in the present invention are a system and method for interconnection and intercommunication of trusted data spaces. The system comprises a plurality of interconnection and intercommunication components deployed in different fields. Each interconnection and intercommunication component comprises three parts, i.e., a key storage component, an encrypted data storage component, and an isolated compute pod. By authorizing sharing of encrypted data sets and matching keys, isolated compute pod resource information, and interconnection and intercommunication rule information with interconnection and intercommunication components in trusted data spaces, secure and lawful interconnection and intercommunication of trusted data spaces of different structures and different types is supported, so that a large-scale trusted data space network can be constructed, which can adapt to data circulation and sharing among intra-enterprise, industry ecological, and cross-industry trusted data spaces.

Description

A system and method for trusted data space interconnection Technical Field

[0001] This invention relates to distributed critical data technology, and in particular to a system and method for trusted data space interconnection. Background Technology

[0002] Data, as a new type of production factor, is the foundation of digitalization, networking, and intelligence. It has been rapidly integrated into all aspects of production, distribution, circulation, consumption, and social service management, profoundly changing production methods, lifestyles, and social governance.

[0003] With the rise of industries such as the Internet and IoT, many enterprises possess massive amounts of unique industry-specific data but lack powerful computing capabilities and the ability to design solutions for data value mining and enhancement. Meanwhile, some product and service providers deeply rooted in these industries, while possessing strong computing power and solutions, lack real-world data to validate their algorithms and solutions. This inability to establish a robust data exchange mechanism stems from the lack of effective data security and circulation solutions for data owners, and the absence of credible and trustworthy circulation platforms for data processors.

[0004] The 2022 research report defined a Trusted Industrial Data Space as a distributed critical data infrastructure built on existing information networks for data aggregation, sharing, circulation, and application. Through a systematic technical arrangement, it ensures the confirmation, implementation, and maintenance of data circulation protocols, resolving security and trust issues among data providers, users, and service providers, thereby achieving data-driven digital transformation. From a business perspective, the Trusted Industrial Data Space is divided into a distributed model and a middleware service model. In the distributed model, the middleware service provider only offers basic data compliance services and auditing services in data transactions. In the middleware service model, the middleware service provider offers data compliance services and data usage.

[0005] According to a report jointly released by the Industrial Internet Industry Alliance and the China Academy of Information and Communications Technology (CAICT), the Trusted Industrial Data Space System Architecture – From a technical perspective, the realization of the functions of the Trusted Industrial Data Space mainly requires the following seven categories of technologies: security technology, privacy computing technology, evidence storage and traceability technology, data control technology, management technology, computing processing technology, and OT technology. Security technology, privacy computing technology, evidence storage and traceability technology, and data control technology are the core technologies of the Trusted Industrial Data Space. Among them, privacy computing technology mainly focuses on the data processing stage, and mainstream privacy computing technologies are mainly divided into three directions: the first is cryptography-based privacy computing technology represented by multi-party secure computation; the second is technology derived from the integration of artificial intelligence and privacy protection technologies, represented by federated learning; the third is privacy computing technology based on trusted hardware, represented by trusted execution environments; and in recent years, the concept of multi-party mediated computation has also been proposed, which can actually be classified as an extension of the third category of trusted execution environments.

[0006] The first type, Secure Multi-party Computation (MPC), was created by Turing Award winner Andrew Yao in 1982 through posing and solving the Millionaire Problem. It refers to the collaborative computation of a target function by multiple parties in the absence of a trusted third party, ensuring that each party only obtains its own computation result and cannot infer the input data of any other party from the interaction data during the computation process (unless the function itself can infer the input of other parties from its own input and the obtained output).

[0007] The second category is Federated Learning (FL), also known as Federated Machine Learning, Joint Learning, or Consortium Learning. Federated learning enables multi-party collaborative machine learning training without requiring the original data to leave the local database. This is achieved through the circulation and processing of intermediate encrypted data. Participants in federated learning typically include data providers, algorithm providers, coordinators, computation providers, result providers, and task initiators. Depending on the distribution of the data among the data providers, it can be categorized into horizontal federated learning, vertical federated learning, and federated transfer learning.

[0008] The third type is the Trusted Execution Environment (TEE), which uses hardware and software methods to build a secure area within the central processing unit, ensuring the confidentiality and integrity of programs and data loaded within it. A TEE is an isolated execution environment that provides trusted applications running on the device with a higher level of security than a regular operating system (RichOS) and more functionality than a Secure Element (SE).

[0009] The fourth category, Multi-party Intermediary Computation (MPIC), is a novel privacy-preserving computation method proposed by Tan Li and Kong Jun. It refers to a data processing method where multiple parties' data, independent of the data providers and users, undergo analysis and computation through a secure and reliable mechanism, resulting in anonymized output. It is essentially a computation management system. In MPIC, the original data from each data provider is de-identified before being input into the intermediary computing environment or platform for computation. After computation, the data is immediately deleted, and the anonymized result data is output through a designated path after review. Under the specific environment and rules of MPIC, the identity of the information data is encrypted and identified. Due to the irreversible nature of the algorithm, it cannot be restored to its original state, thus fulfilling one requirement of anonymization: irreversibility. Simultaneously, because this de-identified information data is confined within a specific regulated environment or platform, another requirement of anonymization is objectively met: the inability to identify a specific natural person. Therefore, the processed data can be considered essentially anonymized, no longer belonging to personal information, and can enter the intermediary computing environment or platform for computation without the individual's consent.

[0010] Different privacy computing technologies are often used in combination to complete the computation and analysis tasks of the data while ensuring the security of the original data.

[0011] Security technologies typically borrow from existing cybersecurity and data security technologies. Evidence storage and traceability technologies are centered on blockchain and security auditing. Data control technologies aim to solve some of the security and trust issues in the transmission, storage, use and destruction of data, focusing on achieving controllability and auditability of data and participants.

[0012] Current data circulation and usage schemes that primarily rely on privacy-preserving computation have the following drawbacks:

[0013] 1. Limited applicable business scenarios: Various privacy computing technologies are only applicable to their specific business scenarios, lacking universality and powerless to address a large number of other business scenarios.

[0014] 2. High computational cost: Several privacy computing technologies either rely heavily on cryptographic algorithms, making computation several times or even tens of times slower than plaintext computation (multi-party secure computation); or there is a bottleneck effect among the participants, with the weakest participant affecting the overall computation progress (federated computation); or they require dedicated hardware investment, resulting in high prices per unit of computing power (TEE).

[0015] 3. It cannot solve the problems of authorization and ownership change. Privacy computing is essentially about solving the problem of secure sharing of data among multiple parties. It does not have the function of ownership change and authorization itself. It needs to be combined with other technologies such as blockchain to solve these problems.

[0016] 4. Interconnection and interoperability between trusted data spaces is costly and inefficient. Data interoperability requires adaptation to different business operations and privacy computing technologies, resulting in high costs and low efficiency.

[0017] Currently, in addition to data circulation and usage solutions based on privacy computing, there are also some solutions provided by traditional vendors. These vendors are either from big data product backgrounds, security product backgrounds, blockchain product backgrounds, or network infrastructure product backgrounds. Most of their proposed solutions are derived from their existing products. Some are biased towards data aggregation and circulation, some towards secure authorized use, and some towards auditing and rights confirmation, but they all lack an overall consideration for the cross-domain circulation and use of data. Summary of the Invention

[0018] This invention proposes a system and method for interconnecting trusted data spaces, which supports secure and legitimate interconnection of trusted data spaces with different structures and types, and can be adapted to data circulation and sharing between trusted data spaces within enterprises, industry ecosystems, and across industries.

[0019] The technical solution of the present invention is as follows:

[0020] A trusted data space interconnection system includes multiple interconnection components deployed in different trusted data space systems. Each interconnection component comprises a key storage component, an encrypted data storage component, and a closed computing module. Through authorization, it shares data encryption algorithm information (so that the closed computing module in another space can successfully decrypt the data after obtaining the authorized key), closed computing module resource information, and interconnection rule information with the trusted data space network (so as to select the best interconnection path and obtain the interconnection computing results at the lowest cost).

[0021] The key storage component is used to store the key corresponding to the encrypted data, and replaces the original key with a new key after the key is obtained, so as to realize the one-time key function.

[0022] The usage rules are as follows: the key storage component is controlled by the data owner or their authorized administrator, and the key owner has the authority to authorize other accounts; without authorization, only the key owner can obtain the key (including the new key after one-time one-key conversion); with authorization, only the key owner and the authorized account can obtain the key (including the new key after one-time one-key conversion), and the authorized account does not have the authority to authorize again.

[0023] The encrypted data storage component is used to store the encrypted dataset and, in conjunction with the new key generated in the key storage component, updates the original encrypted dataset to a new encrypted dataset to achieve a one-time key function.

[0024] The usage rules are as follows: The encrypted data storage component is responsible for storing ciphertext data generated by encryption using the key controlled by the data owner or their authorized manager. The ciphertext data can only be decrypted and used within the closed computing module.

[0025] The encrypted data storage component can store data including algorithms, models, tools, and image files.

[0026] The closed computing chamber is used to provide a closed computing environment and performs the following actions: opening the data transmission channel, acquiring the encrypted datasets that need to be interconnected, closing the data transmission channel, decrypting each dataset, executing computing tasks, destroying plaintext data, ciphertext data and keys, opening the data transmission channel, and exporting the computing results. The closed computing chamber is composed of a group of containers, virtual machines, physical machines or other computing devices, individually or in combination. Its data transmission channel is immediately closed after acquiring the ciphertext data and the corresponding key authorization, thus maintaining a closed state during computing.

[0027] The interconnection components are implemented in the form of a trusted space appliance, a trusted space platform, or a business-specific appliance.

[0028] The Trusted Space All-in-One Machine is an interconnected component implemented in the form of an all-in-one machine. Its usage includes: a single machine for data circulation within a single space; multiple machines interconnected to enable data circulation between different regions within the same industry; and multiple machines interconnected to enable data circulation between different industries within the same region.

[0029] The aforementioned trusted space platform is an interconnected component implemented in the form of a medium-to-large-scale cluster, typically serving as an industry-root trusted space platform or a supercomputing center trusted space platform. Usage methods include: interconnecting the industry-root trusted space platform with trusted space all-in-one machines from various units within the industry, acting as an industry data center to aggregate and store massive amounts of industry data, and providing more powerful computing resources than all-in-one machines; interconnecting the industry-root trusted space platform with other industry-root trusted space platforms, providing a cross-industry, multi-source data usage environment; and interconnecting the supercomputing center trusted space platform with multiple industry-root trusted space platforms, providing ultra-high computing power to address complex scenarios with large data volumes and high computing demands, solving problems that other industry-specific trusted space platforms cannot handle due to performance limitations.

[0030] The dedicated business appliance is configured with customized models, algorithms or other software. When in use, the model, algorithm or software, along with encrypted data and keys, is imported / accessed into the closed computing compartment of the dedicated business appliance for business processing to complete tasks in specific business scenarios.

[0031] A method for trusted data space interconnection and interoperability, comprising the following steps to enable cross-domain data circulation and use:

[0032] Step 1: For the interoperability requirements of different trusted data spaces, the encrypted dataset and key of the data to be interoperable are stored separately in the encrypted data storage component and key storage component of the interoperability component in the corresponding application area of ​​the requirement scenario. Based on the closed computing cabin resource data and interoperability rule data shared by each trusted data space, a network transmission channel is established.

[0033] Step 2: The data user submits a usage application to one or more target domains for the dataset. The usage application includes the encrypted dataset to be used and the specified data usage domain. The data usage domain defines which demand scenario the dataset to be used for and in which closed computing chamber the data will be processed.

[0034] Step 3: After obtaining authorization for use, the data user can use the interoperability component of the data use domain specified in the application to pull the authorized encrypted dataset and its corresponding key stored in the interoperability component of the target domain to the designated closed computing module.

[0035] Step 4: After the encrypted dataset and key are obtained, the key and encrypted data that have been authorized in the interconnection component of the target domain are updated to the new key and new encrypted data. After the designated closed computing module has retrieved the required encrypted dataset and key, the closed computing module performs the following actions in sequence: close the data transmission network channel, decrypt each data, execute the computing task, destroy plaintext data, ciphertext data and key, open the data transmission channel, and export the computing results. At the same time, the trusted data space updates the shared closed computing module resource data in a timely manner, and the key logs of data key authorization and closed computing module execution are uploaded to the blockchain.

[0036] Step 5: Data users obtain the calculation results exported from the closed computing module, which means that the trusted data space has completed one interconnection.

[0037] In step 1: the closed computing module resource data shared by each trusted data space shall include at least one of the following: data encryption algorithm, available computing power, network status, storage status, and security level; the interconnection rule data shall include at least one of the following: preferred cooperative data space routing configuration, data space reputation routing configuration, data price priority configuration, data comprehensive score priority configuration, closed computing module price priority configuration, and security level priority configuration.

[0038] Through the above steps, a universal data space connectivity protocol can be established, thereby building a large-scale trusted data space network. In this trusted data space network, the shared and circulating closed computing module resources and data, as well as the key steps executed by the computing modules, are updated on the blockchain in a timely manner. The trusted data space can quickly select routes based on the data of its own blockchain nodes to obtain the most comprehensive and optimal interconnected data and computing power, achieving data sharing and circulation at a more cost-effective cost.

[0039] In building a large-scale trusted data space network, there are idle resources such as data center computing power and storage space. By installing trusted data space interconnection components, these idle resources can be quickly revitalized and generate value. For example, AI training that requires a lot of computing power can be directed to a trusted space with abundant AI computing power.

[0040] The beneficial effects of this invention are as follows:

[0041] 1. This invention has strong business adaptability, whether it is model training, algorithm development, data reprocessing or providing streaming / batch processing services, it is suitable for various businesses.

[0042] 2. Compared to privacy-preserving computation and encrypted computation technologies, this invention does not waste computing resources on various privacy-preserving algorithms, nor does it rely on expensive TEE hardware. For privacy-preserving computation and comparison results, this invention can reduce resource consumption by 10 times, or even more than 100 times, resulting in high overall computing power utilization and low cost. 3. Data circulation costs are low; users only need to purchase a Trusted Space all-in-one machine and connect it to the network to achieve data sharing and circulation with other nodes in the network.

[0043] 4. Resources are allocated reasonably. Businesses that can be handled by local resources (computing power, storage, data, etc.) can be processed locally; businesses that cannot be handled by local resources can be processed by borrowing resources from the upper-level node.

[0044] 5. Resource expansion is convenient. Each node can independently adjust its resource configuration according to business needs without being affected by other nodes.

[0045] 6. Data ownership is clear. Original data is always stored in its own domain in encrypted form and with a key. Authorized data is completely destroyed after use in the closed computing chamber, effectively preventing secondary dissemination and ensuring that the data remains firmly in the hands of the data owner. For data ownership changes in special scenarios such as data reprocessing, the ownership of the processed data can be negotiated by both parties when the original data is authorized, and the owner of the processed data will retain it.

[0046] 7. Data usage is controlled. Data usage takes place in a closed computing chamber within the authorized domain. Data owners can check data usage through container logs and other means to prevent violations such as overdue use, illegal tampering, and secondary dissemination, which could infringe on the economic and reputational interests of the data owner.

[0047] 8. Provide a unified data circulation method to reduce the cost of data circulation and connection between various fields, making data circulation and aggregation between fields simpler and more efficient. Attached Figure Description

[0048] Figure 1 is a schematic diagram of the architecture of the present invention for point-to-point interconnection between two trusted data spaces.

[0049] Figure 2 is a schematic diagram of the architecture of the present invention for interconnection and interoperability of multiple trusted spaces. Detailed Implementation

[0050] Example 1

[0051] The trusted data space interconnection system in this embodiment includes multiple interconnection components deployed in different trusted data space systems. Each interconnection component comprises a key storage component, an encrypted data storage component, and a closed computing module. It shares data encryption algorithm information (so that the closed computing module in another space can successfully decrypt the data after obtaining the authorized key), closed computing module resource information, and interconnection rule information with the trusted data space network (so that the optimal interconnection path can be selected and the interconnection calculation results obtained at the lowest cost). Figure 1 shows the architecture of point-to-point interconnection between two trusted data space systems, and Figure 2 shows the architecture of interconnection between three trusted data space systems. Of course, the architecture can be infinitely expanded to include interconnection between multiple trusted data space systems according to the pattern in Figure 2.

[0052] The key storage component is used to store the key corresponding to the encrypted data, and replaces the original key with a new key after the key is obtained, so as to realize the one-time key function.

[0053] The usage rules are as follows: The key storage component is controlled by the data owner or their authorized administrator (the account storing the key can be assumed to be the key owner). The key owner has the authority to authorize other accounts. Without authorization, only the key owner can obtain the key (including the new key after one-time one-key conversion). With authorization, only the key owner and the authorized account can obtain the key (including the new key after one-time one-key conversion). The authorized account does not have the authority to authorize again.

[0054] The encrypted data storage component is used to store the encrypted dataset and, in conjunction with the new key generated in the key storage component, updates the original encrypted dataset to a new encrypted dataset to achieve a one-time key function.

[0055] The usage rules are as follows: The encrypted data storage component is responsible for storing ciphertext data generated by encryption using the key controlled by the data owner or their authorized manager. The ciphertext data can only be decrypted and used within the closed computing module.

[0056] The encrypted data storage component can store data including algorithms, models, tools, and image files.

[0057] The closed computing cabin provides a closed computing environment. It comprises a group of containers, virtual machines, physical machines, or other computing devices, individually or in combination. The actions performed by the closed computing cabin include: opening a data transmission channel, acquiring the encrypted dataset to be interconnected, closing the data transmission channel, decrypting each dataset, executing computing tasks, destroying plaintext data, ciphertext data, and keys, opening the data transmission channel again, and exporting the computing results. The closed computing cabin, composed of a group of containers, virtual machines, physical machines, or other computing devices individually or in combination, maintains a closed state during computation by immediately closing its data transmission channel after acquiring the ciphertext data and corresponding key authorization.

[0058] A method for trusted data space interconnection and interoperability, comprising the following steps to enable cross-domain data circulation and use:

[0059] Step 1: For the interoperability requirements of different trusted data spaces, the encrypted dataset and key of the data to be interoperable are stored separately in the encrypted data storage component and key storage component of the interoperability component in the corresponding application area of ​​the requirement scenario. Based on the closed computing cabin resource data and interoperability rule data shared by each trusted data space, a network transmission channel is established.

[0060] Step 2: The data user submits a usage application to a specific dataset in one or more target domains. The usage application includes the encrypted dataset to be used and the specified data usage domain. The data usage domain defines which demand scenario the dataset to be used for and in which closed computing chamber the data will be processed.

[0061] Step 3: After obtaining authorization for use, the data user can use the interoperability component of the data use domain specified in the application to pull the authorized encrypted dataset and its corresponding key stored in the interoperability component of the target domain to the designated closed computing module.

[0062] Step 4: After the encrypted dataset and key are obtained, the key and encrypted data that have been authorized in the interconnection component of the target domain are updated to the new key and new encrypted data. After the designated closed computing module has retrieved the required encrypted dataset and key, the closed computing module performs the following actions in sequence: close the data transmission network channel, decrypt each data, execute the computing task, destroy plaintext data, ciphertext data and key, open the data transmission channel, and export the computing results. At the same time, the trusted data space updates the shared closed computing module resource data in a timely manner, and the key logs of data key authorization and closed computing module execution are uploaded to the blockchain.

[0063] Step 6: Data users obtain the calculation results exported from the closed computing module, which means that the trusted data space has completed one interconnection.

[0064] Example 2

[0065] The design of methods and systems for interconnecting trusted data spaces includes three specific implementation forms of interconnection components, as follows.

[0066] 1. Trusted Space All-in-One Machine: An interconnected component implemented in the form of an all-in-one machine. Its features are low price and plug-and-play, but it is limited by the performance of a single machine, and its computing and storage capabilities are limited.

[0067] Usage methods include: standalone use for data exchange between departments within an organization, such as data exchange between departments within a municipal emergency management bureau; interconnection of multiple machines to enable data exchange between departments within the same industry but across different regions, such as data exchange between the environmental protection bureaus of City A and City B; and interconnection of multiple machines to enable data exchange between departments within the same region but across different industries, such as data exchange between a municipal emergency management bureau and the land and resources bureau.

[0068] II. Trusted Space Platform: Interconnected components implemented in the form of medium to large-scale clusters, with strong computing power, storage capacity and system reliability, generally serving as industry root trusted space platforms or supercomputing center trusted space platforms.

[0069] Usage methods include: interconnecting the industry root trusted space platform with trusted space all-in-one machines of various units within the industry, serving as an industry data center to aggregate and store massive amounts of industry data, and providing more powerful computing resources than all-in-one machines; interconnecting the industry root trusted space platform with other industry root trusted space platforms to provide a cross-industry multi-source data usage environment; and interconnecting the supercomputing center trusted space platform with multi-industry root trusted space platforms to provide supercomputing power, addressing the performance limitations of various industry trusted space platforms in complex scenarios with large data volumes and high computing demands.

[0070] 3. Business-specific all-in-one machine: A business-specific all-in-one machine is equipped with specific functional models, algorithms or other software. When in use, the model, algorithm or software is imported / accessed together with encrypted data (including keys) into the closed computing compartment of the business-specific all-in-one machine for business processing to complete the work of specific business scenarios.

[0071] The three implementation methods mentioned above can be combined and can include many different forms.

[0072] Example 3

[0073] The use case for interconnectivity components can be data flow between departments within an organization, specifically using: a trusted space appliance + a business-specific appliance.

[0074] For example, the Emergency Management Bureau of a certain city has departments such as the Safety Production Supervision Department, the Materials Management Department, and the IT Management Department. Among them, the Safety Production Supervision Department is responsible for the safety production supervision of key enterprises in the jurisdiction; the Materials Management Department is responsible for the management, distribution, and dispatch of emergency materials in the jurisdiction; and the IT Management Department is responsible for the research and development and management of the unit's IT facilities. All departmental data is stored in the unit's trusted space all-in-one machine in the form of encrypted datasets and keys and is bound to the department's account. Without authorization, each department can only obtain its own encrypted data and keys through its own account.

[0075] The IT Management Department needs to obtain key enterprise data from the Safety Production Supervision Department and material management data from the Material Management Department. Based on information such as key enterprise type, enterprise location, emergency material storage locations, and the types and quantities of emergency materials stored at each location, and combined with the IT Management Department's own AI algorithms, an AI model will be trained to assist the Material Management Department in replanning the emergency material allocation scheme when new key enterprises are added. This includes whether to add new emergency material storage locations and how to adjust the types and quantities of emergency materials stored at existing locations. Main implementation steps:

[0076] 1a. The IT Management Department selects the data to be used on the Trusted Space Appliance, including key enterprise data, material management data, and AI algorithm data.

[0077] The Trusted Space Appliance recommends closed computing bays within the current Trusted Data Space network that meet usage requirements based on selected data and AI algorithms. This includes the preferred Trusted Space Appliance for the organization and alternative industry-leading Trusted Space platforms. The recommendations are based on factors including, but not limited to, computing resources (idle CPU / GPU configuration, unit computing power price, etc.), storage resources (idle storage capacity, storage reliability, etc.), security levels (security compliance level, data security management certification, etc.), network conditions (network bandwidth, geographical location, etc.), and other custom factors (opening hours, preferred partners, reputation, etc.).

[0078] 1b. Based on system recommendations, the IT Management Department decided to complete the computing tasks within the closed computing bay of the Trusted Space Appliance within the organization. It then applied to the Safety Production Supervision Department and the Materials Management Department for key enterprise data and materials management data, respectively, specifying that this data would be used within the closed computing bay of the Trusted Space Appliance. Authorization was granted after both departments agreed.

[0079] 1c. The IT Management Department uses the Trusted Data Space Appliance to import key enterprise data, material management data, AI algorithm data, and their corresponding keys into the closed computing bay of the unit's Trusted Data Space Appliance. After decrypting the data in the closed computing bay, model training is conducted. The AI ​​algorithm belongs to the IT Management Department and does not require authorization. During the process, the Trusted Data Space Appliance updates the shared closed computing bay resource data in a timely manner, and the key logs of data key authorization and closed computing bay execution are uploaded to the blockchain.

[0080] 1d. The IT management department exports the trained model and stores it in the form of encrypted data and keys in the encrypted data storage component and key storage component of the business-specific all-in-one machine.

[0081] 1e. When new key enterprises are added, the Materials Management Department applies for model usage authorization and new key enterprise data usage authorization from the IT Management Department and the Safety Production Supervision Department respectively through the Business Dedicated All-in-One Machine. The new key enterprise data is stored in the Trusted Space All-in-One Machine, and the application is made by forwarding messages from the Business Dedicated All-in-One Machine to the Trusted Space All-in-One Machine.

[0082] 1f. The Materials Management Department's dedicated business terminal pulls the authorized model and new key enterprise data into the terminal's closed computing module to obtain new emergency material allocation plans.

[0083] Example 4

[0084] The use case for interconnectivity components can be cross-regional data flow within the industry, specifically using: Trusted Space Appliance + Trusted Space Appliance.

[0085] For example, a large lake spans two prefecture-level cities. Environmental protection management of the lake is jointly handled by the environmental protection bureaus of the two cities, A and B. Each bureau is responsible for water quality monitoring, pollution control, and ecological restoration within its own jurisdiction. Historical data is stored on their respective storage servers. However, due to concerns about data leakage, it is difficult to integrate the data to form a unified work plan. To better facilitate work, promote data fusion, and improve environmental management efficiency, both bureaus have introduced a trusted space appliance. The water quality monitoring department stores historical monitoring point information and water sampling information in encrypted datasets and keys on its trusted space appliance. The pollution control department stores historical pollution source information and inspection information in encrypted datasets and keys on its trusted space appliance. The ecological restoration department stores historical restoration measures and results information in encrypted datasets and keys on its trusted space appliance. Without authorization, each department can only access its own encrypted datasets and keys through its own account. Both parties currently require the aggregation of their historical work data to train an AI model. This model will learn from the governance measures and water quality changes of the lake over the years. Once trained, the model can provide overall suggestions for the next year's lake environmental governance plan based on the current year's water quality, monitoring points, pollution source information, patrol information, and remediation measures. This includes adjustments to the monitoring point layout, patrol times and routes, and a list of key pollution sources. Each party will then complete the corresponding portion for their respective jurisdiction. The model training task is currently being undertaken by the R&D department of Environmental Protection Bureau A. The main implementation steps are as follows:

[0086] 2a. The R&D department of Bureau A selects the data to be used on the Trusted Space Integrated Machine of Bureau A, including monitoring point information and water quality sampling information datasets from Bureaus A and B, pollution source information and inspection information datasets, remediation measure information and remediation result information datasets, and AI algorithm data to be used. Among them, the data resource information in the Trusted Space Integrated Machine of Bureau B has been transferred to the Trusted Space Integrated Machine of Bureau A through the data catalog.

[0087] 2b. The A-site Trusted Space Appliance recommends closed computing modules within the current trusted data space network that meet the usage requirements based on the selected data and algorithms. This includes the preferred A-site Trusted Space Appliance and alternatives such as the B-site Trusted Space Appliance and the industry-root trusted space platform. The recommendations are based on factors including, but not limited to, computing resources (idle CPU / GPU configuration, unit computing power price, etc.), storage resources (idle storage capacity, storage reliability, etc.), security level (security compliance level, data security management certification, etc.), network conditions (network bandwidth, geographical location, etc.), and other custom factors (opening hours, preferred partners, reputation, etc.).

[0088] 2c. Based on system recommendations, the IT Management Department of Bureau A decided to complete the computing tasks within the closed computing chamber of Bureau A's Trusted Space Integrated Machine. Therefore, it applied to the water quality monitoring department, pollution control department, and ecological restoration department of Bureaus A and B respectively for datasets of monitoring point information and water quality sampling information, pollution source information and inspection information, and restoration measure information and restoration result information. The applications specified that this data would be used in the closed computing chamber of Bureau A's Trusted Space Integrated Machine. Authorization was granted after the relevant departments of Bureaus A and B agreed. Specifically, the application for data resources from Bureau B was made by forwarding messages from Bureau A's Trusted Space Integrated Machine to Bureau B's Trusted Space Integrated Machine.

[0089] 2d. The IT Management Department of Bureau A uses the Bureau A Trusted Space Integrated Machine to pull datasets, AI algorithms, and their corresponding keys—including monitoring point information, water quality sampling information, pollution source information, inspection information, remediation measure information, and remediation result information from both bureaus—into the closed computing compartment of the Bureau A Trusted Space Integrated Machine. The AI ​​algorithms are also stored in the Bureau A Trusted Space Integrated Machine in the form of encrypted data and keys, but belong to Bureau A's R&D department and can be directly imported without application.

[0090] 2e. After all datasets are aggregated into the closed computing module of the Trusted Data Space Appliance in Bureau A, the closed computing module decrypts the data, and Bureau A's R&D personnel begin AI model training. Upon completion, the trained AI model is exported, and the closed computing module automatically destroys all data and keys within the module. During the process, the Trusted Data Space Appliance promptly updates the shared closed computing module resource data, and data key authorization and critical logs of closed computing module execution are uploaded to the blockchain.

[0091] 2f. The R&D personnel of Bureau A will encrypt the exported model and place it in the form of encrypted data and key in the encrypted data storage component and key storage component of Bureau A's business-specific all-in-one machine.

[0092] When formulating the lake environmental protection and governance work plan for the following year, the management department of Bureau A requests data sets from the trusted space integrated machines of Bureaus A and B respectively, including monitoring point information, water quality sampling information, pollution source information, inspection information, remediation measure information, and remediation result information for the current year, through the dedicated business integrated machine of Bureau A; and applies to the R&D department of Bureau A for the use of AI models through the dedicated business integrated machine of Bureau A; after obtaining all authorizations, the management department of Bureau A operates the dedicated business integrated machine of Bureau A to pull all datasets (including keys) and AI models (including keys) into the closed computing cabin of the dedicated business integrated machine of Bureau A and runs them. The AI ​​model provides suggestions for the overall work plan for lake environmental protection and governance for the following year, which are then broken down and implemented by the two bureaus.

[0093] Example 5

[0094] The use cases for interconnectivity components can be industry roots and data flow between various units within the industry, specifically using: Trusted Space Platform + Trusted Space All-in-One Machine.

[0095] For example, in recent years, meteorological departments have actively adapted to technological development trends and explored the application of artificial intelligence (AI) technology in meteorological operations. By introducing advanced AI models, meteorological departments have achieved a dual improvement in efficiency and quality in areas such as data processing, weather forecasting, and disaster early warning. In practice, the main technical solution adopted is "pre-training + transfer learning." First, data from local meteorological bureaus is aggregated into the industry's root trust space platform. This is combined with industry-wide public data stored in the industry's root trust space platform to form massive amounts of meteorological data. Then, the high computing power provided by the industry's root trust space platform is used to conduct pre-training of meteorological AI models. The generated pre-trained models are then distributed to local meteorological bureaus for transfer learning based on local data, ultimately forming AI models suitable for local conditions and specific meteorological operations. The model pre-training is completed by the R&D department of the root trust space platform, while the transfer learning is completed by the R&D departments of the local meteorological bureaus themselves. The main implementation steps are as follows:

[0096] 3a. The R&D department of the industry's root trusted space platform selects the data to be used, including local meteorological datasets from the trusted space all-in-one machines of local meteorological bureaus, public meteorological datasets from the root trusted space platform, and AI algorithm data to be used. The data resource information from the trusted space all-in-one machines of local meteorological bureaus has already been transferred to the root trusted space platform via a data catalog.

[0097] 3b. The industry-root trusted space platform recommends closed computing modules within the current trusted data space network that meet the usage requirements based on the selected data and algorithms. This includes the preferred industry-root trusted space platform and other supercomputing center trusted space platforms as alternatives. The recommendations are based on factors including, but not limited to, computing resources (idle CPU / GPU configuration, unit computing power price, etc.), storage resources (idle storage capacity, storage reliability, etc.), security level (security compliance level, data security management certification, etc.), network conditions (network bandwidth, geographical location, etc.), and other custom factors (opening hours, preferred partners, reputation, etc.).

[0098] 3c. The R&D department of the Industry Root Trusted Space Platform, based on system recommendations, selected to complete the computational tasks within the closed computing module of the Industry Root Trusted Space Platform. Subsequently, it applied to the data ownership departments of local meteorological bureaus and the data ownership department of the Root Trusted Space Platform for their respective local meteorological datasets and the platform's public meteorological datasets. The applications specified that these data would be used in the closed computing module of the Industry Root Trusted Space Platform, and authorization was granted upon the consent of the relevant ownership departments. Specifically, the applications for local meteorological data resources were submitted by forwarding messages from the Industry Root Trusted Space Platform to the Trusted Space All-in-One Machine of local meteorological bureaus.

[0099] The R&D department of the 3D industry-specific Root Trusted Space Platform operates the platform to retrieve authorized meteorological datasets and keys from the trusted space integrated machines of meteorological bureaus across various regions. These, along with authorized public meteorological datasets, AI algorithms, and keys from the Root Trusted Space Platform itself, are imported into the platform's closed computing chamber and decrypted, forming a massive meteorological dataset. The AI ​​algorithms are also stored in the Root Trusted Space Platform as encrypted data and keys, but belong to the platform's R&D department and can be directly imported without application.

[0100] 3e. The R&D department of Root Trusted Space utilizes the high computing power provided by the Root Trusted Space platform to conduct model pre-training within the platform's closed computing module. During this process, the Trusted Data Space platform promptly updates the shared resource data of the closed computing module, and data key authorization and critical logs of the closed computing module's execution are uploaded to the blockchain. The pre-trained model generated is stored in the encrypted data storage component and key storage component of the Root Trusted Space platform in the form of encrypted data and keys. This pre-trained model possesses general meteorological operational processing capabilities.

[0101] 3f. Local meteorological bureaus' R&D departments, through their respective Trusted Space Integrated Machines, apply to the Root Trusted Space Platform R&D department for the use of pre-trained models and to their respective data ownership departments for the use of various local meteorological datasets, obtaining authorization. Alternatively, in this step, the computational tasks can be completed within the closed computing cabin of the industry's Root Trusted Space Platform, as recommended by the system.

[0102] 3G. Local meteorological bureaus' R&D departments operate their respective Trusted Space Integrated Machines, pulling authorized pre-trained models and keys from the root Trusted Space Platform and various authorized meteorological datasets and keys from their own Trusted Space Integrated Machines. These are then imported into the closed computing module of their Trusted Space Integrated Machines for transfer learning. During this process, the Trusted Space Integrated Machines promptly update the shared closed computing module resource data, and key logs for data key authorization and closed computing module execution are uploaded to the blockchain. Finally, the process generates an AI model suitable for local conditions and capable of handling specific local meteorological operations. This AI model can be run on each unit's dedicated operational integrated machine.

[0103] Example 6

[0104] The use case for interconnectivity components can be cross-industry data circulation scenarios, specifically using: a multi-trust space platform + a supercomputing center.

[0105] For example, geological disasters (such as landslides and debris flows) are usually triggered by the combined effects of geological conditions and meteorological factors. Therefore, effective disaster early warning requires comprehensive analysis of historical geological data and real-time meteorological data to predict the potential areas and probabilities of disasters. Through the deep integration of meteorological data, geological data, and emergency management data, AI models can significantly improve the efficiency and accuracy of geological disaster prevention and control, providing scientific support for reducing disaster losses. In practice, the emergency management industry can use the Emergency Management Root Trusted Space Platform, the Meteorological Root Trusted Space Platform, and the Land and Resources Root Trusted Space Platform to aggregate massive amounts of emergency management data, meteorological data, and geological data into a trusted space platform located in a supercomputing center. The supercomputing center's massive computing power is used to conduct pre-training of geological disaster prevention and control models. The generated pre-trained models are stored in the Emergency Management Root Trusted Space Platform and then distributed to local emergency management bureaus for transfer learning based on local conditions. The model pre-training is completed by the R&D department of the Emergency Management Root Trusted Space Platform, while the transfer learning is completed by the R&D departments of local emergency management bureaus. The final generated models provide reference for local geological disaster prevention and control work. The following are the main implementation steps of this solution:

[0106] 4a. The R&D department of the Emergency Management Root Trusted Space Platform selects the data to be used, including the emergency management dataset and AI algorithm in the Emergency Management Root Trusted Space Platform, the meteorological dataset in the Meteorological Root Trusted Space Platform, and the geological dataset in the Land and Resources Root Trusted Space Platform. The data resource information from the Meteorological and Land and Resources Root Trusted Space Platforms has been transferred to the Emergency Management Root Trusted Space Platform through a data catalog.

[0107] 4b. The Emergency Management Trusted Space Platform recommends closed computing modules within the current trusted data space network that meet usage requirements based on selected data and algorithms. This includes the preferred trusted space platform located in the supercomputing center and other alternative trusted space platforms. The recommendations are based on factors including, but not limited to, computing resources (idle CPU / GPU configuration, unit computing power price, etc.), storage resources (idle storage capacity, storage reliability, etc.), security level (security compliance level, data security management certification, etc.), network conditions (network bandwidth, geographical location, etc.), and other custom factors (opening hours, preferred partners, reputation, etc.).

[0108] 4c. Based on system recommendations, the R&D department of the Emergency Management Root Trusted Space Platform decided to complete the computational tasks within the closed computing chamber of the Supercomputing Center's Trusted Space Platform. Therefore, it applied to the respective data ownership departments of the Emergency Management, Meteorological, and Land Resources Root Trusted Space Platforms for their respective Emergency Management datasets (including AI algorithms), Meteorological datasets, and Geological datasets. The applications stated that these data would be used within the closed computing chamber of the Supercomputing Center's Trusted Space Platform, and authorization was granted upon the consent of the relevant ownership departments. Specifically, the applications for meteorological and geological data were submitted by forwarding messages from the Emergency Management Root Trusted Space Platform to the Meteorological Root Trusted Space Platform and the Land Resources Root Trusted Space Platform.

[0109] 4d. The R&D department of the Emergency Management Root Trusted Space Platform imports and decrypts various authorized datasets, AI algorithms, and keys from the Emergency Management, Meteorological, and Land Resources Root Trusted Space Platforms into the closed computing chamber of the Supercomputing Center's Trusted Space Platform, forming multi-source fused data. The AI ​​algorithms are also stored in the Emergency Management Root Trusted Space Platform in the form of encrypted data and keys, but belong to the R&D department of the Emergency Management Root Trusted Space Platform; no application is required, and they can be directly imported.

[0110] 4e. The R&D department of the Emergency Management Root Trusted Space Platform utilizes the supercomputing power provided by the supercomputing center's trusted space platform to conduct model pre-training within the platform's closed computing bay. During this process, the trusted data space platform promptly updates the shared resource data of the closed computing bay, and data key authorization and critical logs of the closed computing bay's execution are uploaded to the blockchain. The pre-trained model generated is stored back in the encrypted data storage component and key storage component of the Emergency Management Root Trusted Space Platform in the form of encrypted data and keys. This pre-trained model possesses general capabilities related to geological disaster prevention and control, such as geological disaster early warning and emergency resource scheduling.

[0111] 4f. Local emergency management bureaus' R&D departments apply to the emergency management root trusted space platform R&D department for the use of pre-trained models through their own trusted space integrated machines. This application involves forwarding messages to the emergency management root trusted space platform via their own trusted space integrated machines.

[0112] 4G. Local emergency management bureaus' R&D departments apply to their respective data ownership departments for various types of local emergency management data through their own trusted space integrated machines; they apply to their respective data ownership departments for various types of local meteorological data through their local meteorological bureau's trusted space integrated machines; and they apply to their respective data ownership departments for various types of geological data through their local land and resources bureau's trusted space integrated machines. Applications to the local meteorological bureau and land and resources bureau are both made by forwarding messages from one unit's trusted space integrated machine to the other's trusted space integrated machine.

[0113] 4 hours later, the R&D departments of local emergency management bureaus operate their respective Trusted Space Appliances to retrieve authorized pre-trained models and keys from the root Trusted Space Platform for emergency management. They also retrieve authorized meteorological, geological, and emergency management datasets and keys from the local meteorological bureau, local land and resources bureau, and their own Trusted Space Appliances. These are then imported into the closed computing module of their Trusted Space Appliances for transfer learning. During this process, the Trusted Space Appliances promptly update the shared closed computing module resource data, and key logs for data key authorization and closed computing module execution are uploaded to the blockchain. Finally, the process generates AI models suitable for local conditions and capable of handling specific local emergency management tasks. In this step, the AI ​​models generated by local emergency management bureaus for specific emergency management tasks are run and used on their respective dedicated business appliances.

[0114] In steps 4f and 4g, the system may recommend a closed computing module.

Claims

1. A system for trusted data space interconnection, characterized in that: It includes multiple interconnected components deployed in different trusted data space systems. Each interconnected component consists of three parts: a key storage component, an encrypted data storage component, and a closed computing module. The key storage component is used to store the key corresponding to the encrypted data, and replaces the original key with a new key after the key is obtained, so as to realize the one-time key function. The encrypted data storage component is used to store the encrypted dataset and, in conjunction with the new key generated in the key storage component, updates the original encrypted dataset to a new encrypted dataset to achieve the one-time key function. The closed computing chamber is used to provide a closed computing environment and performs the following actions: opening the data transmission channel, obtaining the encrypted datasets that need to be interconnected, closing the data transmission channel, decrypting each dataset, executing computing tasks, destroying plaintext data, ciphertext data and keys, opening the data transmission channel, and exporting the computing results.

2. The system for trusted data space interconnection and interoperability according to claim 1, characterized in that: The key storage component is controlled by the data owner or their authorized administrator, and the key owner has the authority to authorize other accounts; when there is no authorization, only the key owner can obtain the key; when there is authorization, only the key owner and the authorized account can obtain the key; the key contains a new key after being converted to a one-time key; the encrypted data storage component is responsible for storing the ciphertext data generated by encryption using the key controlled by the data owner or their authorized administrator, and the ciphertext data is only decrypted and used in the closed computing module.

3. The system for trusted data space interconnection and interoperability according to claim 1, characterized in that: The encrypted data storage component can store data including, but not limited to, algorithms, models, tools, and image files.

4. The system for trusted data space interconnection and interoperability according to claim 1, characterized in that: The closed computing cabin consists of a group of containers, virtual machines, physical machines or other computing devices, individually or in combination; the data transmission channel of the closed computing cabin is immediately closed after obtaining the encrypted data and the corresponding key authorization, thus maintaining a closed state during the computing process.

5. A system for trusted data space interconnection and interoperability according to claim 1, characterized in that: The interconnection components are implemented in one or more combinations of trusted space all-in-one machines, trusted space platforms, and business-specific all-in-one machines. The Trusted Space All-in-One Machine is an interconnected component implemented in an all-in-one form. Its usage includes: a single machine for data circulation within a single space; multiple machines interconnected to enable data circulation between different regions within the same industry; and multiple machines interconnected to enable data circulation between different industries within the same region. The aforementioned trusted space platform is an interconnected component implemented in the form of a medium-to-large-scale cluster, serving as an industry-root trusted space platform or a supercomputing center trusted space platform. Its usage includes: interconnecting the industry-root trusted space platform with trusted space all-in-one machines of various units within the industry, acting as an industry data center to aggregate and store massive amounts of industry data; interconnecting the industry-root trusted space platform with other industry-root trusted space platforms, providing a cross-industry, multi-source data usage environment; and interconnecting the supercomputing center trusted space platform with multiple industry-root trusted space platforms, addressing the performance limitations of various industry trusted space platforms in complex scenarios with large data volumes and high computing demands. The dedicated business appliance is equipped with customized models, algorithms, or software suitable for specific business scenarios. When in use, the customized models, algorithms, or software, along with encrypted data and keys, are imported / accessed into the closed computing compartment of the dedicated business appliance for business processing to complete the work of the specific business scenario.

6. A method for implementing trusted data space interconnection based on the trusted data space interconnection system according to any one of claims 1-5, characterized in that... The steps for interconnection are as follows: Step 1: For the interoperability requirements of different trusted data spaces, the encrypted dataset and key of the data to be interoperable are stored separately in the encrypted data storage component and key storage component of the interoperability component in the corresponding application area of ​​the requirement scenario. Based on the closed computing cabin resource data and interoperability rule data shared by each trusted data space, a network transmission channel is established. Step 2: The data user submits a usage application to a dataset in one or more target domains. The usage application includes the encrypted dataset to be used and the specified data usage domain. The data usage domain defines which demand scenario the dataset to be used for and in which closed computing chamber the data will be processed. Step 3: After obtaining authorization for use, the data user uses the interconnection component of the data use domain specified in the application to pull the authorized encrypted dataset and its corresponding key stored in the interconnection component of the target domain to the designated closed computing chamber. Step 4: After the encrypted dataset and key are obtained, the key and encrypted data that have been authorized to be obtained in the interoperability component of the target domain are updated with the new key and new encrypted data; After the designated closed computing module retrieves the required encrypted dataset and key, it performs the following actions in sequence: closing the data transmission network channel, decrypting each data, executing the computing task, destroying plaintext data, ciphertext data and key, opening the data transmission channel, and exporting the computing results. At the same time, the trusted data space updates the shared closed computing module resource data in a timely manner, and the data key authorization and key logs of the closed computing module execution are uploaded to the blockchain. Step 5: Data users obtain the calculation results exported from the closed computing module, which means that the trusted data space has completed one interconnection.

7. The method for trusted data space interconnection and interoperability according to claim 6, characterized in that: The closed computing module resource data shared by each trusted data space shall include at least one of the following: data encryption algorithm, available computing power, network status, storage status, and security level. The interconnection rule data shall include at least one of the following: preferred cooperative data space routing configuration, data space reputation routing configuration, data price priority configuration, data comprehensive score priority configuration, closed computing module price priority configuration, and security level priority configuration.

8. The method for trusted data space interconnection and interoperability according to claim 6, characterized in that: Design a general data space connectivity protocol, and build a trusted data space network based on steps 1-5. In the trusted data space network, the shared and circulating closed computing module resource data and the key steps executed by the closed computing module are updated on the chain in a timely manner. The trusted data space quickly selects routes based on the blockchain node data of this space to obtain the comprehensive and optimal interconnected data and computing power, so as to realize the sharing and circulation of data.

9. The method for trusted data space interconnection and interoperability according to claim 6, characterized in that: The trusted data space network utilizes idle resources through interconnected components; these idle resources include, but are not limited to, idle data center computing power and storage space.

Images