A trusted data space privacy computing collaboration method and system
Patent Information
- Application Number
- CN202611197658.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-07
- Publication Date
- 2026-10-16
AI Technical Summary
[0004]本发明实施例提供了一种可信数据空间内隐私计算协同方法及系统，用于解决如下技术问题：现有数据联合计算技术中异构节点能力无法统一量化调度，静态策略管控安全性不足，隐私计算安全与性能难以兼顾
本发明通过接入连接器的多协议自适应适配机制，可屏蔽不同厂商、不同技术路线隐私计算组件的技术差异，实现异构隐私计算节点与可信数据空间的即插即用式对接，构建统一的分布式计算网络，无需针对不同节点进行定制化改造，大幅降低了跨域多方协同的对接成本与部署周期，提升了系统的兼容性与扩展性。
Smart Images
Figure CN122932941A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of data flow, in particular to a privacy computing collaboration method and system in a trusted data space. BACKGROUND
[0002] With the acceleration of the marketization process of data elements, the demand for cross-institution and cross-domain data collaboration in the fields of medical treatment, finance, satellite remote sensing and the like is increasingly urgent. As the core infrastructure for data compliance flow, the trusted data space realizes the interconnection and intercommunication of multiple data domains through a standardized access system, and provides a basic framework for cross-domain data interaction. The privacy computing technology can complete joint computing under the premise that the original data does not leave the domain through cryptography and distributed computing means.
[0003] However, in the process of deep landing application of the existing data joint computing technology, some technical problems are still found. First, the heterogeneous privacy computing nodes lack a unified capability quantification model, and only basic predefined protocol conversion can be achieved. The node capabilities of different manufacturers and different technical routes cannot be standardized evaluated and matched, and it is difficult to build a large-scale distributed computing network. Second, the cross-domain joint computing scheduling mostly adopts static rules or simple load balancing, and cannot structurally decompose and globally optimize complex computing tasks, so it is difficult to adapt to the scenes of cross-domain network fluctuation and dynamic changes of computing power, and the time delay of multi-party collaborative computing is high and the resource utilization rate is low. Third, the data use control relies on static rule checking at the software level, and there is a risk of bypassing the permission, and the strategy cannot be dynamically adjusted according to the scene risk, so the flexibility and security of control cannot be considered. Finally, the block chain storage only stays at the level of hash storage, and can only prove the existence of data, and cannot verify the compliance of the computing process and the correctness of the result, so the audit depth is insufficient. SUMMARY
[0004] The present application provides a privacy computing collaboration method and system in a trusted data space, which is used to solve the following technical problems: the heterogeneous node capabilities in the existing data joint computing technology cannot be uniformly quantified and scheduled, the static strategy control has insufficient security, and the privacy computing security and performance are difficult to balance.
[0005] The present application embodiment adopts the following technical scheme: On the one hand, the present application embodiment provides a privacy computing collaboration method in a trusted data space, and the method comprises: connecting the privacy computing nodes of each data domain to the trusted data space control nodes, and performing identity verification and computing capability registration to build a cross-domain privacy computing network; The trusted data space management node configures a data use policy, and stores metadata of the data use policy and a hash value corresponding to each other to a block chain node to complete notarization; wherein the data use policy at least includes constraint rules of a field permission, a use frequency, a valid time limit, and an applicable scenario; Receiving a joint computing task initiated by a data user and analyzing task requirements, generating a node cooperative computing scheme according to the task requirements and real-time loads of each privacy computing node; According to the node cooperative computing scheme, joint computing of multiple privacy computing nodes is performed, and it is checked in real time whether the current computing task meets the constraint rules of the data use policy; After the joint computing is completed, metadata in the computing process and a computing result hash value are stored corresponding to each other to a block chain node to complete notarization, and the computing result is returned to the data user.
[0006] In a feasible implementation, the privacy computing nodes of each data domain are connected to the trusted data space management node, and identity verification and computing capability registration are performed, and a cross-domain privacy computing network is constructed, specifically including: Through the access connector deployed at the boundary of the data domain, identity credentials and capability description information of the privacy computing nodes in the domain are collected, and a registration application is initiated to the trusted data space management node; The identity credentials and capability description information are verified by the trusted data space management node, and after the audit is passed, the privacy computing nodes are recorded in a distributed computing network directory; According to the distributed computing network directory, a cross-domain privacy computing network is constructed; Through the access connector, the computing power load and network state of the privacy computing nodes are synchronized to the trusted data space management node in real time.
[0007] In a feasible implementation, the trusted data space management node configures a data use policy, and stores metadata of the data use policy and a hash value corresponding to each other to a block chain node to complete notarization, specifically including: Through a policy configuration interface of the trusted data space management node, a multi-dimensional use policy is configured for a specified data set; wherein the multi-dimensional use policy at least includes: field use permission, data use frequency permission, policy valid time, and applicable scenario; After the configuration is completed, the original text of the multi-dimensional use policy is synchronized to a policy execution engine of the corresponding domain, and a policy original text hash value and metadata are uploaded to a block chain to complete notarization.
[0008] In one feasible implementation, a collaborative computing task initiated by a data user is received and its requirements are parsed. Based on the task requirements and the real-time load of each privacy computing node, a node collaborative computing scheme is generated, specifically including: When the trusted data space management node receives a joint computing task application initiated by a data user, it parses the specific task requirements; wherein, the specific task requirements include at least: task type, target dataset, business scenario description, data dependencies, and identity credentials; The identity and permissions of the data user are verified based on the identity credentials. After confirming that the access conditions are met, the dynamic computing power scheduling engine is invoked to match several backup privacy computing nodes with corresponding computing capabilities according to the task type. Based on the current load and cross-domain network latency of the aforementioned backup privacy computing nodes, computing tasks are allocated to generate a node collaborative computing scheme.
[0009] In one feasible implementation, computing tasks are allocated based on the current load and cross-domain network latency of the plurality of backup privacy computing nodes, and a node collaborative computing scheme is generated. Specifically, A multi-objective optimization model is constructed in a deep reinforcement learning scheduling engine; wherein, the input data of the multi-objective optimization model is the load data of privacy computing nodes and network latency, and the output data is a node collaborative computing scheme that meets the preset optimization objective; Based on the data dependencies of the joint computing task, the task is broken down into several sub-tasks with dependencies. Using the current load and cross-domain network latency of each backup privacy computing node as input data, and with the optimization objectives of minimizing the average latency of subtasks, minimizing communication overhead, and achieving the required security level, the deep reinforcement learning scheduling engine generates a globally optimal node collaborative computing scheme.
[0010] In one feasible implementation, joint computation of multiple privacy-preserving computing nodes is performed according to the node collaborative computing scheme, and the current computing task is verified in real time to ensure it complies with the constraints of the data usage strategy. Specifically, this includes: According to the node collaborative computing scheme, subtasks are assigned to each backup privacy computing node. The local policy execution engine of the backup privacy computing node verifies whether the current subtask matches all the constraint rules of the data usage policy. If the verification is successful, the computing task is started. During the calculation process, the data usage is continuously monitored through the policy engine. If any calculation behavior exceeds the constraints, the calculation is stopped and reported to the trusted data space management node.
[0011] In one feasible implementation, after initiating the computing task, the method further includes: The corresponding security level is determined based on the business type of the joint computing task; If the joint computing task is a high-security task, then a pure cryptographic computing mode is adopted, and multiple cryptographic computing methods are combined for encrypted computing. If the joint computing task is a low-to-medium security level task, a hybrid computing mode is adopted, in which sensitive computing logic is encrypted and non-sensitive computing logic is executed in a plaintext environment.
[0012] In one feasible implementation, after the joint computation is completed, the metadata from the computation process and the hash value of the computation result are stored in the blockchain node to complete the notarization, and the computation result is returned to the data user, specifically including: After the joint computation is completed, each privacy-preserving computation node that participated in the joint computation generates computation result fragments, which are then aggregated to obtain the final computation result and returned to the data user through a trusted link. Extract key information from the entire task chain, package it, and write it to the blockchain using a consensus algorithm; the key information from the entire task chain includes at least metadata, task ID, participant information, execution parameters, permission verification records, and result hash value.
[0013] In one feasible implementation, after returning the calculation results to the data user, the method further includes: The usage metering ledger of the corresponding dataset is updated according to each completed joint computing task, and the usage metering ledger is displayed in the management terminal for data providers to view; wherein, the usage metering ledger includes at least data usage records and the remaining usage quota of various types of data.
[0014] On the other hand, embodiments of the present invention also provide a privacy computing collaboration system within a trusted data space, the system comprising: The network construction module is used to connect privacy computing nodes from various data domains to the trusted data space management and control node, and to perform identity verification and computing capability registration to build a cross-domain privacy computing network. The collaborative computing module is used to configure data usage policies through the trusted data space management node, and store the metadata of the data usage policies and their corresponding hash values to the blockchain node for notarization. The data usage policies at least include constraints on field permissions, usage frequency, validity period, and applicable scenarios. It receives joint computing tasks initiated by data users, parses the task requirements, and generates a node collaborative computing scheme based on the task requirements and the real-time load of each privacy computing node. It performs joint computing across multiple privacy computing nodes according to the node collaborative computing scheme and verifies in real-time whether the current computing task conforms to the constraints of the data usage policy. The data storage module is used to store the metadata, execution parameters, verification records, and hash values of the calculation results in the blockchain node after the joint calculation is completed, and then return the calculation results to the data user.
[0015] Compared with the prior art, the privacy computing collaboration method and system in a trusted data space provided by the embodiments of the present invention have the following beneficial effects: This invention utilizes a multi-protocol adaptive adaptation mechanism in the access connector to shield the technical differences between privacy computing components from different manufacturers and using different technical approaches. This enables plug-and-play connection between heterogeneous privacy computing nodes and trusted data spaces, building a unified distributed computing network. It eliminates the need for customized modifications to different nodes, significantly reducing the connection cost and deployment cycle for cross-domain multi-party collaboration, and improving the system's compatibility and scalability.
[0016] Furthermore, this invention supports configuring data usage policies based on four dimensions: data fields, usage frequency, validity period, and business scenario. It also enforces permission boundaries through a policy execution engine embedded in connectors and computing nodes. Data providers can precisely control the scope, frequency, and lifecycle of data usage, constraining data usage behavior from the source. Simultaneously, it includes metering and ledger capabilities, enabling statistical and traceable data usage, upgrading data management from broad access authorization to refined full lifecycle management.
[0017] This invention leverages the distributed consensus and immutability of blockchain to store key information across the entire data flow and computation process on the blockchain, including data usage strategies, task application information, collaborative computing parameters, permission verification records, and computation result hashes. This ensures that the entire process is traceable, auditable, and verifiable. It provides a reliable basis for defining the rights and responsibilities of all participating parties and offers regulatory agencies a compliance auditing channel, effectively enhancing the overall credibility and compliance of data circulation. Attached Figure Description
[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 A flowchart of a collaborative method for privacy computing in a trusted data space provided by an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a privacy computing collaborative system in a trusted data space, provided as an embodiment of the present invention. Detailed Implementation
[0019] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.
[0020] This invention provides a collaborative method for privacy computing within a trusted data space, such as... Figure 1 As shown, the privacy-preserving computation collaborative method within the trusted data space specifically includes steps S101-S105: S101. Connect the privacy computing nodes of each data domain to the trusted data space management node, and perform identity verification and computing capability registration to build a cross-domain privacy computing network.
[0021] Specifically, by using access connectors deployed at the data domain boundary, the identity credentials and capability descriptions of privacy computing nodes within the domain are collected, and a registration application is initiated to the trusted data space management node.
[0022] Furthermore, the identity credentials and capability description information are verified through the trusted data space control node. After the verification is approved, the privacy computing node is entered into the distributed computing network directory, and a cross-domain privacy computing network is constructed based on the distributed computing network directory.
[0023] Furthermore, the computing load and network status of the privacy computing node are synchronized in real time to the trusted data space management node through the access connector.
[0024] As a feasible implementation method, this invention adopts a layered and decoupled fusion architecture, consisting of a trusted data space management layer, a connector adaptation layer, a distributed privacy computing layer, a policy execution layer, and a blockchain evidence storage layer from top to bottom. Each layer interacts through standardized interfaces to enable the flow of highly sensitive data across domains. The trusted data space management layer, as the core of global scheduling and control, is responsible for node access management, unified task scheduling, policy configuration distribution, and end-to-end operational monitoring. It has a built-in dynamic computing power scheduling engine that can receive computing task requests from multiple parties. The connector adaptation layer is deployed at the boundary nodes of each data domain, serving as the interface hub between the trusted data space and privacy computing nodes within the domain. It has a built-in multi-protocol adaptive adaptation module that can automatically identify privacy computing nodes from different vendors and using different technical approaches.
[0025] As a feasible implementation method, after deploying access connectors and privacy computing nodes in each data domain, the connector initiates an access application to the trusted data space management node, submitting node identity credentials and a description of privacy computing capabilities, including supported computing types, computing power specifications, and callable data directories. After completing identity verification and capability review, the trusted data space management node includes the node in the distributed computing network directory, thus completing the construction of the cross-domain privacy computing network.
[0026] S102. Configure data usage policies through the trusted data space management node, and store the metadata of the data usage policies and their corresponding hash values to the blockchain node to complete the notarization. The data usage policy should at least include constraints on field permissions, usage frequency, validity period, and applicable scenarios.
[0027] Specifically, through the policy configuration interface of the trusted data space management node, multi-dimensional usage policies can be configured for a specified dataset; among them, the multi-dimensional usage policies include at least: field usage permissions, data usage frequency permissions, policy validity period, and applicable scenarios.
[0028] Furthermore, after configuration, the original text of the multi-dimensional usage strategy is synchronized to the strategy execution engine of the corresponding domain, and the hash value of the original strategy text and metadata are uploaded to the blockchain for evidence storage.
[0029] As a feasible implementation method, data providers can configure multi-dimensional usage policies for specified datasets through the policy configuration interface of the control layer. For example, they can select fields allowed to participate in calculations, set cumulative usage counts and periodic call limits, define the policy's effective time period, and restrict allowed business scenarios and calculation types. After configuration, the policy text is synchronized to the corresponding domain's policy execution engine, and the policy hash and metadata are uploaded to the blockchain for notarization, ensuring that the policy itself is tamper-proof and traceable.
[0030] S103. Receive the joint computing task initiated by the data user and parse the task requirements. Based on the task requirements and the real-time load of each privacy computing node, generate a node collaborative computing scheme.
[0031] Specifically, when the trusted data space management node receives a joint computing task request initiated by a data user, it parses the specific task requirements. These requirements include at least: task type, target dataset, business scenario description, data dependencies, and identity credentials. Task types primarily include: set intersection, feature query, joint modeling, and dense state statistics.
[0032] Furthermore, the identity and permissions of the data user are verified based on the identity credentials. After confirming that the access conditions are met, the dynamic computing power scheduling engine is invoked to match several backup privacy computing nodes with corresponding computing capabilities according to the task type.
[0033] Furthermore, based on the current load and cross-domain network latency of several backup privacy computing nodes, computing tasks are allocated to generate a node collaborative computing scheme. Specifically, this includes constructing a multi-objective optimization model within a deep reinforcement learning scheduling engine. The input data for this model consists of the load data and network latency of the privacy computing nodes, while the output data is a node collaborative computing scheme that meets preset optimization objectives. Based on the data dependencies of the joint computing tasks, the tasks are decomposed into several sub-tasks with dependencies. Using the current load and cross-domain network latency of each backup privacy computing node as input data, and with the optimization objectives of minimizing the average latency of the sub-tasks, minimizing communication overhead, and achieving the required security level, a globally optimal node collaborative computing scheme is generated through the deep reinforcement learning scheduling engine.
[0034] As a feasible implementation method, the data user initiates a joint computing task application to the trusted data space management node, submitting data such as task type, target dataset, business scenario description, data dependencies, and identity credentials. The trusted data space management node first performs an initial identity and scenario verification. After confirming that the conditions for admission are met, it calls the dynamic computing power scheduling engine to decompose the joint computing task into a set of subtasks with dependencies. Then, based on a pre-built multi-objective optimization model, it generates an optimal node collaborative computing scheme that minimizes the preset objectives. Finally, according to the personnel allocation strategy in the collaborative computing scheme, it assigns the corresponding subtasks to the corresponding privacy computing nodes.
[0035] S104. Perform joint computation of multiple privacy computing nodes according to the node collaborative computing scheme, and verify in real time whether the current computing task complies with the constraint rules of the data usage strategy.
[0036] Specifically, subtasks are assigned to each backup privacy computing node according to the node collaborative computing scheme. The local policy execution engine of the backup privacy computing node verifies whether the current subtask matches all the constraint rules of the data usage policy. If the verification is successful, the computing task is started.
[0037] As a feasible implementation method, after receiving the task instruction, each privacy-preserving computing node participating in the computation has its local policy execution engine verify whether the current task matches all constraints of the data usage policy. Computation can only begin after the verification is successful. The computation process strictly adheres to the corresponding privacy-preserving computation technical specifications; the raw data never leaves the local domain, and only encrypted intermediate parameters are exchanged. During the computation process, the policy engine continuously monitors the system. If violations such as exceeding the allowed number of attempts, exceeding the allowed number of fields, or exceeding the allowed time limit occur, the computation is immediately terminated and reported to the control node.
[0038] Furthermore, during the calculation process, the data usage is continuously monitored through the policy engine. If any calculation behavior exceeds the constraint rules, the calculation is stopped and reported to the trusted data space management node.
[0039] As a feasible implementation method, the corresponding security level is looked up in the business level table based on the business type of the joint computing task. If the joint computing task is a high-security task, a pure cryptographic computing mode is adopted, which combines various composable cryptographic computing methods from the encryption algorithm library for encryption computing. The specific combination method is not elaborated here and can be combined based on the experience of technical personnel. If the joint computing task is a medium-to-low-security task, a hybrid computing mode is adopted, in which sensitive computing logic is encrypted and non-sensitive computing logic is executed in a plaintext environment.
[0040] In one embodiment, the computation execution mode is automatically selected based on the task's security level: high-security tasks employ a pure cryptographic computation mode, utilizing cryptographic techniques such as PSI, homomorphic encryption, and federated learning throughout, with both raw data and intermediate results remaining encrypted. For medium- and low-security tasks, a hybrid computation mode is used, where sensitive core logic is encrypted and computed locally on the node, while non-sensitive computations are executed in a plaintext environment, saving significant computational power without compromising data security. The sensitive core logic includes at least gradient aggregation, intersection statistics, and parameter summarization.
[0041] S105. After the joint calculation is completed, the metadata of the calculation process and the hash value of the calculation result are stored in the blockchain node to complete the notarization, and the calculation result is returned to the data user.
[0042] Specifically, after the joint computation is completed, each privacy-preserving computation node that participated in the joint computation generates computation result fragments, which are then aggregated to obtain the final computation result and returned to the data user through a trusted link.
[0043] Furthermore, key information across the entire task chain is extracted, packaged, and written into the blockchain using a consensus algorithm. The key information across the entire task chain includes at least metadata, task ID, participant information, execution parameters, permission verification records, and result hash values.
[0044] Furthermore, the usage metering ledger of the corresponding dataset is updated according to each completed joint computing task, and the usage metering ledger is displayed in the management terminal for data providers to view; the usage metering ledger includes at least data usage records and the remaining usage quota for various types of data.
[0045] In addition, embodiments of the present invention also provide a privacy computing collaborative system within a trusted data space, such as... Figure 2 As shown, the privacy computing collaboration system 200 within the trusted data space specifically includes: Network construction module 210 is used to connect privacy computing nodes of each data domain to the trusted data space management node, and to perform identity verification and computing capability registration to build a cross-domain privacy computing network. The collaborative computing module 220 is used to configure data usage policies through the trusted data space management node, and store the metadata of the data usage policies and their corresponding hash values to the blockchain node for notarization; wherein, the data usage policy includes at least field permissions, usage frequency, validity period, and constraint rules for applicable scenarios; receive joint computing tasks initiated by data users and parse task requirements, generate a node collaborative computing scheme based on the task requirements and the real-time load of each privacy computing node; perform joint computing of multiple privacy computing nodes according to the node collaborative computing scheme, and verify in real time whether the current computing task conforms to the constraint rules of the data usage policy; The data storage module 230 is used to store the metadata, execution parameters, verification records and hash values of the calculation results in the blockchain node after the joint calculation is completed, and return the calculation results to the data user.
[0046] The various embodiments in this invention are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.
[0047] The foregoing has described specific embodiments of the present invention. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.
[0048] The above description is merely an embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various modifications and variations can be made to the embodiments of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the embodiments of the present invention should be included within the protection scope of the present invention.
Claims
1. A collaborative method for privacy-preserving computation within a trusted data space, characterized in that, The method includes: By connecting privacy computing nodes in each data domain to the trusted data space management node and verifying their identities and registering their computing capabilities, a cross-domain privacy computing network can be built. The trusted data space management node configures the data usage policy, and stores the metadata of the data usage policy and its hash value in the blockchain node to complete the notarization; wherein, the data usage policy includes at least the constraint rules of field permissions, usage frequency, validity period and applicable scenarios; Receive joint computing tasks initiated by data users and parse the task requirements. Based on the task requirements and the real-time load of each privacy computing node, generate a node collaborative computing scheme. The node collaborative computing scheme is used to perform joint computing of multiple privacy computing nodes, and the current computing task is verified in real time to see if it complies with the constraint rules of the data usage strategy. After the joint computation is completed, the metadata of the computation process and the hash value of the computation result are stored in the blockchain node to complete the notarization, and the computation result is returned to the data user.
2. The collaborative method for privacy computing in a trusted data space according to claim 1, characterized in that, By connecting privacy-preserving computing nodes from various data domains to the trusted data space management node and performing identity verification and computing capability registration, a cross-domain privacy-preserving computing network is constructed, specifically including: By using access connectors deployed at the boundary of the data domain, the identity credentials and capability descriptions of privacy computing nodes within the domain are collected, and a registration application is initiated to the trusted data space management node. The trusted data space management node verifies the identity credentials and capability description information, and after the verification is approved, the privacy computing node is entered into the distributed computing network directory. Construct a cross-domain privacy computing network based on the aforementioned directory of distributed computing networks; The computing load and network status of the privacy computing node are synchronized in real time to the trusted data space management node through the access connector.
3. The collaborative method for privacy computing in a trusted data space according to claim 1, characterized in that, The trusted data space management node configures data usage policies, and the metadata of the data usage policies and their corresponding hash values are stored in the blockchain node to complete the notarization process, specifically including: Through the policy configuration interface of the trusted data space management node, multi-dimensional usage policies can be configured for a specified dataset; wherein, the multi-dimensional usage policies include at least: field usage permissions, data usage frequency permissions, policy validity period, and applicable scenarios; After configuration, the original text of the multi-dimensional usage strategy is synchronized to the strategy execution engine of the corresponding domain, and the hash value of the original strategy text and metadata are uploaded to the blockchain to complete the notarization.
4. The collaborative method for privacy computing in a trusted data space according to claim 1, characterized in that, The system receives collaborative computing tasks initiated by data users and parses the task requirements. Based on the task requirements and the real-time load of each privacy computing node, it generates a node collaborative computing scheme, specifically including: When the trusted data space management node receives a joint computing task application initiated by a data user, it parses the specific task requirements; wherein, the specific task requirements include at least: task type, target dataset, business scenario description, data dependencies, and identity credentials; The identity and permissions of the data user are verified based on the identity credentials. After confirming that the access conditions are met, the dynamic computing power scheduling engine is invoked to match several backup privacy computing nodes with corresponding computing capabilities according to the task type. Based on the current load and cross-domain network latency of the aforementioned backup privacy computing nodes, computing tasks are allocated to generate a node collaborative computing scheme.
5. The collaborative method for privacy computing in a trusted data space according to claim 4, characterized in that, Based on the current load and cross-domain network latency of the aforementioned backup privacy computing nodes, computing tasks are allocated to generate a node collaborative computing scheme. Specifically, A multi-objective optimization model is constructed in a deep reinforcement learning scheduling engine; wherein, the input data of the multi-objective optimization model is the load data of privacy computing nodes and network latency, and the output data is a node collaborative computing scheme that meets the preset optimization objective; Based on the data dependencies of the joint computing task, the task is broken down into several sub-tasks with dependencies. Using the current load and cross-domain network latency of each backup privacy computing node as input data, and with the optimization objectives of minimizing the average latency of subtasks, minimizing communication overhead, and achieving the required security level, the deep reinforcement learning scheduling engine generates a globally optimal node collaborative computing scheme.
6. The collaborative method for privacy computing in a trusted data space according to claim 5, characterized in that, The node collaborative computing scheme involves joint computation by multiple privacy-preserving computing nodes, and real-time verification of whether the current computing task conforms to the constraints of the data usage strategy. Specifically, this includes: According to the node collaborative computing scheme, subtasks are assigned to each backup privacy computing node. The local policy execution engine of the backup privacy computing node verifies whether the current subtask matches all the constraint rules of the data usage policy. If the verification is successful, the computing task is started. During the calculation process, the data usage is continuously monitored through the policy engine. If any calculation behavior exceeds the constraints, the calculation is stopped and reported to the trusted data space management node.
7. The collaborative method for privacy computing in a trusted data space according to claim 6, characterized in that, After starting the computation task, the method further includes: The corresponding security level is determined based on the business type of the joint computing task; If the joint computing task is a high-security task, then a pure cryptographic computing mode is adopted, and multiple cryptographic computing methods are combined for encrypted computing. If the joint computing task is a low-to-medium security level task, a hybrid computing mode is adopted, in which sensitive computing logic is encrypted and non-sensitive computing logic is executed in a plaintext environment.
8. The collaborative method for privacy computing in a trusted data space according to claim 1, characterized in that, After the joint computation is completed, the metadata from the computation process and the hash value of the computation result are stored on the blockchain node to complete the notarization, and the computation result is returned to the data user, specifically including: After the joint computation is completed, each privacy-preserving computation node that participated in the joint computation generates computation result fragments, which are then aggregated to obtain the final computation result and returned to the data user through a trusted link. Extract key information from the entire task chain, package it, and write it to the blockchain using a consensus algorithm; the key information from the entire task chain includes at least metadata, task ID, participant information, execution parameters, permission verification records, and result hash value.
9. The collaborative method for privacy computing in a trusted data space according to claim 1, characterized in that, After returning the calculation results to the data user, the method further includes: The usage metering ledger of the corresponding dataset is updated according to each completed joint computing task, and the usage metering ledger is displayed in the management terminal for data providers to view; wherein, the usage metering ledger includes at least data usage records and the remaining usage quota of various types of data.
10. A collaborative system for privacy computing within a trusted data space, characterized in that, The system includes: The network construction module is used to connect privacy computing nodes from various data domains to the trusted data space management and control node, and to perform identity verification and computing capability registration to build a cross-domain privacy computing network. The collaborative computing module is used to configure data usage policies through the trusted data space management node, and store the metadata of the data usage policies and their corresponding hash values to the blockchain node for notarization. The data usage policies at least include constraints on field permissions, usage frequency, validity period, and applicable scenarios. It receives joint computing tasks initiated by data users, parses the task requirements, and generates a node collaborative computing scheme based on the task requirements and the real-time load of each privacy computing node. It performs joint computing across multiple privacy computing nodes according to the node collaborative computing scheme and verifies in real-time whether the current computing task conforms to the constraints of the data usage policy. The data storage module is used to store the metadata, execution parameters, verification records, and hash values of the calculation results in the blockchain node after the joint calculation is completed, and then return the calculation results to the data user.