Distributed multi-mode data sharing method and system across trust domains

Abstract

This invention provides a distributed multi-modal data sharing method and system across trust domains, relating to the field of data sharing technology. The method involves identifying the type and formatting standardization of heterogeneous data provided by multiple data source nodes; dividing data source nodes into different trust domains based on security levels, authorization scopes, and protection policies; constructing a cross-trust domain data mapping model to establish data synchronization channels; verifying access rights and recording access logs through smart contracts; synchronizing data states using a distributed consensus mechanism based on a Byzantine fault-tolerant algorithm, and resolving data conflicts through version control. This invention achieves secure sharing of heterogeneous data and maintains state consistency, improving the reliability and fault tolerance of cross-trust domain data sharing.

Description

Technical Field

[0001] The present invention relates to the technical field of data sharing, and in particular to a method and system for sharing data across trust domains based on distributed multi-modal data. Background Art

[0002] With the rapid development of distributed systems and data sharing needs, cross-organizational and cross-departmental data collaboration is becoming increasingly common. Different institutions often adopt different data storage formats and management models, resulting in heterogeneous data source nodes. These data source nodes form independent trust domains based on their respective security requirements and data protection strategies. Enabling efficient sharing of heterogeneous data across trust domains while ensuring data security and privacy has become a pressing technical challenge.

[0003] Existing technologies lack a unified management mechanism for heterogeneous data sources, and inconsistent data formats lead to inefficient sharing. Traditional centralized data sharing models struggle to adapt to multi-trust domain scenarios and present single-point failure risks. Existing data synchronization solutions fail to effectively handle node failures and malicious behavior, making it difficult to ensure data consistency. These issues severely hinder the widespread adoption of cross-trust domain data sharing.

[0004] Therefore, there is an urgent need for a data sharing method and system that can solve the problems of data standardization, secure sharing, and consistency maintenance in the process of heterogeneous data sharing across trust domains. Summary of the Invention

[0005] In order to solve the above-mentioned problems existing in the prior art, the present invention provides a distributed multi-modal data cross-trust domain data sharing method and system. By constructing a data access control mechanism based on smart contracts, a cross-domain data mapping model and a Byzantine fault-tolerant consensus mechanism, secure and reliable data sharing between multiple trust domains is achieved, while ensuring the normal operation of the system when some nodes fail or malicious behavior occurs.

[0006] The technical solutions of the present invention are as follows:

[0007] The present invention provides a distributed multi-mode data cross-trust domain data sharing method, comprising:

[0008] Receive heterogeneous data provided by multiple data source nodes, identify data types and standardize formats of the heterogeneous data; divide the multiple data source nodes into different trust domains based on the security level of the data provider, the scope of data usage authorization, and the data protection policy, each trust domain contains at least one data source node, and assigns a unique domain identifier to each trust domain; deploy smart contracts in each trust domain, and the smart contracts are used to manage data access rights and data sharing rules within the domain;

[0009] Construct a cross-trust domain data mapping model, which includes data field mapping relationships, data format conversion rules, and data consistency verification rules; based on the data mapping model, establish a data synchronization channel between different trust domains, and transmit data in an end-to-end encrypted manner; set up a data flow control module in the data synchronization channel, and dynamically adjust the data transmission rate according to the real-time network status and data priority;

[0010] When a data requester initiates a cross-domain data access request, the smart contract verifies the data requester's access rights and records data access logs; performs differential privacy processing on the authorized data and returns the differential privacy processed data to the data requester;

[0011] A distributed consensus mechanism and version control mechanism based on the Byzantine fault-tolerant algorithm are used to synchronize data between different trust domains.

[0012] As a preferred embodiment of the present invention, heterogeneous data provided by multiple data source nodes are received, and data type identification and format standardization processing are performed on the heterogeneous data; based on the security level of the data provider, the scope of data use authorization, and the data protection policy, the multiple data source nodes are divided into different trust domains, each of which contains at least one data source node, and a unique domain identifier is assigned to each trust domain; a smart contract is deployed in each trust domain, and the smart contract is used to manage data access rights and data sharing rules within the domain, including:

[0013] Receive heterogeneous data provided by multiple data source nodes, analyze the heterogeneous data through a feature extraction algorithm to obtain heterogeneous features, wherein the heterogeneous features are specifically structural features, content features, and metadata features, and identify the data type of the heterogeneous data based on the heterogeneous features; classify the heterogeneous data into structured data, semi-structured data, and unstructured data according to the results of the data type identification; construct a unified data representation model based on a graph data model, use a data conversion adapter to convert the structured data, the semi-structured data, and the unstructured data into a standardized format corresponding to the graph data model, and store the standardized format;

[0014] Analyze the security level of each data source node, where the security level includes a data security protection capability indicator, a historical credit record indicator, and a technical assurance level indicator; obtain the data usage authorization scope of each data source node, where the data usage authorization scope includes the data shareable user group, permitted usage scenarios, and time restrictions; extract the data protection policy of each data source node, where the data protection policy includes a data desensitization level, access frequency restrictions, and transmission encryption requirements; based on the security level, the data usage authorization scope, and the data protection policy, use a hierarchical clustering algorithm to divide the multiple data source nodes into different trust domains, each trust domain containing at least one data source node; and use a distributed hash algorithm to assign a unique domain identifier to each trust domain;

[0015] A smart contract is deployed in each trust domain, and the smart contract is implemented based on the Hyperledger Fabric framework. The smart contract includes an access rights management module, a data sharing rules module, and an access audit module. The access rights management module implements fine-grained access control of data within the domain based on the RBAC model. The data sharing rules module is used to configure and execute data sharing policies. The access audit module is used to record data access operations.

[0016] As a preferred embodiment of the present invention, a cross-trust domain data mapping model is constructed, wherein the data mapping model includes data field mapping relationships, data format conversion rules, and data consistency verification rules; based on the data mapping model, a data synchronization channel is established between different trust domains, and the data synchronization channel uses end-to-end encryption to transmit data, including:

[0017] Analyze the data patterns of the source trust domain and the target trust domain to extract field feature information; input the field feature information into a deep learning model to generate a field similarity matrix; construct a data field mapping relationship based on the field similarity matrix; formulate data format conversion rules based on type differences in the data field mapping relationship; establish data consistency verification rules based on the data field mapping relationship; and combine the data field mapping relationship, the data format conversion rules, and the data consistency verification rules to generate the data mapping model;

[0018] Analyzing data transmission requirements based on the data mapping model and determining communication link specifications; establishing a dedicated communication link between trust domains based on the communication link specifications; generating an asymmetric encryption key pair for the dedicated communication link; performing end-to-end encryption on the dedicated communication link using the asymmetric encryption key pair to form the data synchronization channel; and configuring transmission parameters of the data synchronization channel based on the data scale of the data mapping model;

[0019] Processing source data using the data field mapping relationship to complete field mapping; converting the format of the mapped data using the data format conversion rules; verifying the conversion result using the data consistency verification rules; using the verified data as the data to be transmitted; and transmitting the data to be transmitted to the target trust domain through the data synchronization channel;

[0020] receiving encrypted data in the target trust domain; decrypting the received data using the asymmetric encryption key pair; performing reverse data conversion according to the data mapping model; verifying the integrity and consistency of the converted data; writing the verified data to the target trust domain storage system; generating a data synchronization status report to implement data synchronization processing;

[0021] Analyze the abnormal conditions in the data synchronization status report; correct the data field mapping relationship and data format conversion rules of the data mapping model according to the abnormal conditions; monitor the performance indicators of the data synchronization channel; optimize the transmission parameters of the data synchronization channel based on the performance indicators; update the configuration of the data mapping model and the data synchronization channel, and optimize the synchronization mechanism.

[0022] As a preferred embodiment of the present invention, a data flow control module is provided in the data synchronization channel, and the data flow control module dynamically adjusts the data transmission rate according to the real-time network status and data priority, including:

[0023] A data flow control module is provided in the data synchronization channel. The data flow control module includes a network monitoring unit, a data classification unit, and a rate adjustment unit. The network monitoring unit deploys network detection points at both the transmitting and receiving ends of the data synchronization channel. The network detection points periodically send detection packets to collect round-trip delay, packet loss rate, available bandwidth, and link jitter data. The round-trip delay, packet loss rate, available bandwidth, and link jitter data are smoothed using a sliding time window to generate a network status assessment matrix.

[0024] The data classification unit divides the data into core business data, general business data, and non-critical data according to the business importance of the data, and divides the data into real-time data, quasi-real-time data, and offline data according to the timeliness requirement of the data; orthogonally combines the data divided according to the business importance and the data divided according to the timeliness requirement to form a data priority matrix; and sets a resource allocation weight for each priority in the data priority matrix;

[0025] The rate adjustment unit calculates the current network load coefficient based on the network status assessment matrix, where the network load coefficient is determined by the round-trip delay change rate, the packet loss rate, and the bandwidth utilization rate; calculates the reference transmission rate based on the network load coefficient using an improved addition increase and multiplication decrease algorithm; and allocates the reference transmission rate to data streams of different priorities according to the resource allocation weight to achieve dynamic adjustment of the data transmission rate.

[0026] As a preferred embodiment of the present invention, when a data requester initiates a cross-domain data access request, the smart contract verifies the data requester's access rights and records data access logs; performs differential privacy processing on the authorized data, and returns the differential privacy processed data to the data requester, including:

[0027] Receive a cross-domain data access request initiated by a data requester, wherein the cross-domain data access request includes an access target, an identity certificate of the requester, and a purpose of data use; verify the digital signature of the identity certificate of the requester using an asymmetric encryption algorithm to obtain a verification result;

[0028] Matching the verification result with a preset access control policy matrix, which includes permission level matching rules and usage scenario compliance rules; based on the matching result, determining whether the access permission level of the data requester meets the access requirements of the access target, and whether the data usage purpose complies with the usage scenario compliance rules;

[0029] When the access permission level meets the access requirements and the data usage purpose complies with the usage scenario compliance rules, the historical access records of the data requester are obtained and a credit score is calculated based on the historical access records; an access transaction identifier is generated, and the requester's identity credentials, access time, access target, data usage purpose, and credit score are recorded in an access log. The access log adopts a chain storage structure and is encrypted and stored using a cryptographic hash function;

[0030] When the credit score is higher than a preset credit threshold, obtaining the to-be-accessed data corresponding to the access target, identifying sensitive fields in the to-be-accessed data, and setting differential privacy budget parameters based on the sensitivity of the sensitive fields and the purpose of data use;

[0031] Select a differential privacy processing method based on the data type of the data to be accessed, add random noise that conforms to the Laplace distribution to numerical data, use an exponential mechanism to randomize categorical data, and use a semantically preserving desensitization algorithm to replace sensitive information in text data to generate privacy-processed data;

[0032] Obtaining historical data distribution characteristics of the data to be accessed, constructing a noise compensation model based on the historical data distribution characteristics, optimizing the noise in the privacy-processed data to obtain optimized data; and calculating the privacy protection strength and data utility index of the optimized data;

[0033] When both the privacy protection strength and the data utility index meet the preset privacy threshold requirements, the optimized data is returned to the data requester as an access result, and the return status of the access result is recorded in the access log.

[0034] As a preferred embodiment of the present invention, a distributed consensus mechanism and version control mechanism based on the Byzantine fault tolerance algorithm is adopted to synchronize data between different trust domains, including:

[0035] Build a distributed consensus network across trust domains, deploy state synchronization nodes in different trust domains, and divide the state synchronization nodes into consensus node groups and verification node groups; calculate trust scores based on node historical behavior, and select the node with the highest trust score as the master node; the distributed consensus network is implemented based on the Byzantine fault tolerance algorithm;

[0036] Receive a data state change request within the trust domain, the data state change request including a data content hash value, a change timestamp, an operation sequence, and a version number; the master node verifies the validity of the data state change request and generates a state synchronization proposal; broadcasts the state synchronization proposal to the consensus node group and the verification node group;

[0037] The consensus node group votes on the state synchronization proposal based on the Byzantine fault tolerance algorithm; when the number of votes exceeds the fault tolerance threshold, a pre-commit message is broadcast to the state synchronization nodes of the trust domain outside the current trust domain; and a confirmation message is received from the trust domain outside the current trust domain, wherein the confirmation message includes a node signature and a state verification result;

[0038] A version control mechanism is established between each trust domain. The version control mechanism includes a version tree structure that records the historical trajectory of data status changes. When cross-domain concurrent modification is detected, a branch version is created in the version tree structure. The main version is determined through Byzantine fault-tolerant consensus.

[0039] Analyze concurrent operation sequences across different trust domains to identify compatible and conflicting operations. For compatible operations, merge them into the master version. For conflicting operations, determine the final operation sequence through cross-domain Byzantine fault-tolerant consensus. Submit the merged state update to each trust domain for synchronization.

[0040] Monitor the operating status of nodes in each trust domain, including response latency, message integrity, and consensus participation; when node failure or malicious behavior is detected, remove the abnormal node from the consensus process; recalculate the fault tolerance threshold to ensure that the number of remaining normal nodes meets the Byzantine fault tolerance requirements.

[0041] The present invention also provides a distributed multi-mode data cross-trust domain data sharing system, which includes a data access and trust management unit, a cross-domain data mapping and synchronization unit, an access control and privacy protection unit, and a consensus and version management unit, wherein:

[0042] The data access and trust management unit receives heterogeneous data provided by multiple data source nodes, performs data type identification and format standardization on the heterogeneous data; divides the multiple data source nodes into different trust domains based on the security level of the data provider, the scope of data use authorization, and the data protection policy, each of which contains at least one data source node, and assigns a unique domain identifier to each trust domain; deploys a smart contract in each trust domain, and the smart contract is used to manage data access rights and data sharing rules within the domain;

[0043] The cross-domain data mapping and synchronization unit has a cross-trust domain data mapping model built in it, and the data mapping model includes data field mapping relationships, data format conversion rules, and data consistency verification rules; based on the data mapping model, a data synchronization channel is established between different trust domains, and the data synchronization channel uses end-to-end encryption to transmit data; a data flow control module is set in the data synchronization channel, and the data flow control module dynamically adjusts the data transmission rate according to the real-time network status and data priority;

[0044] When a data requester initiates a cross-domain data access request, the access control and privacy protection unit verifies the access rights of the data requester through the smart contract and records the data access log; performs differential privacy processing on the data authorized for access and returns the differential privacy processed data to the data requester;

[0045] The consensus reaching and version management unit adopts a distributed consensus mechanism and version control mechanism based on the Byzantine fault tolerance algorithm to synchronize data between different trust domains.

[0046] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the program, the distributed multimodal data cross-trust domain data sharing method as described in any one of the embodiments is implemented.

[0047] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the distributed multimodal data cross-trust domain data sharing method as described in any one of the embodiments.

[0048] The beneficial effects of this application are as follows:

[0049] 1. Through technical means such as dividing trust domains, deploying smart contracts, end-to-end encrypted transmission and differential privacy processing, sensitive data can be effectively protected, data leakage can be prevented, and the security of data sharing can be enhanced.

[0050] 2. By building a data mapping model and data synchronization channel across trust domains, the standardization and efficient transmission of heterogeneous data are achieved, and the data transmission rate is dynamically adjusted through the data flow control module, thereby improving the efficiency of cross-domain data sharing.

[0051] 3. Adopting distributed consensus mechanism and version control mechanism to ensure data status synchronization and data consistency between different trust domains, even in the event of failure or malicious behavior of some nodes, it can guarantee the stable operation of the system and data reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 Schematic diagram of the process of a distributed multi-mode data cross-trust domain data sharing method according to an embodiment of the present invention;

[0053] Figure 2 This is a structural diagram of a distributed multi-mode data cross-trust domain data sharing system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0055] The following specific embodiments are used to describe the technical solution of the present invention in detail. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0056] Example 1:

[0057] Figure 1 FIG. 1 is a flow chart of a distributed multi-mode data cross-trust domain data sharing method according to embodiment 1 of the present invention. Figure 1 As shown, the method includes:

[0058] S1. Receive heterogeneous data provided by multiple data source nodes, identify data types and standardize formats of the heterogeneous data; divide the multiple data source nodes into different trust domains based on the data provider's security level, data usage authorization scope, and data protection policy, each of which contains at least one data source node, and assign a unique domain identifier to each trust domain; deploy a smart contract within each trust domain, wherein the smart contract is used to manage data access rights and data sharing rules within the domain;

[0059] S11, receiving heterogeneous data provided by multiple data source nodes, for example, a medical data sharing platform that receives patient electronic medical records, medical imaging data, and clinical research data from different hospitals;

[0060] Analyzing the heterogeneous data using a feature extraction algorithm to obtain heterogeneous features, wherein the heterogeneous features specifically include structural features, content features, and metadata features (e.g., data tags, data descriptions, etc.), and identifying the data type of the heterogeneous data based on the heterogeneous features; classifying the heterogeneous data into structured data (e.g., tabular data in a relational database), semi-structured data (e.g., data in JSON or XML formats), and unstructured data (e.g., text, images, audio, and video data) based on the data type identification results. For example, determining whether the data is structured data by analyzing whether it has fixed patterns and fields;

[0061] A unified data representation model is constructed based on the graph data model. A data conversion adapter is used to convert the structured data, semi-structured data, and unstructured data into a standardized format corresponding to the graph data model, and the standardized format is stored. Specifically, different types of data are converted into nodes and edges for storage to facilitate subsequent processing and analysis. For example, various indicators in a patient's electronic medical record are used as nodes, and the relationships between indicators as edges to construct a knowledge graph of the patient's health status.

[0062] S12. Analyze the security level of each of the data source nodes. The data source node is essentially an abstraction of the data provider as a node in the graph data model. The two essentially describe the same object, but use different expressions in different scenarios. Therefore, the analysis of the security level of each of the data source nodes here is to analyze the security level of the data provider in step S1; the security level includes data security protection capability indicators (such as whether it has security certification, the degree of perfection of the security policy), historical credit record indicators (such as the history of data leakage, the reliability of data quality) and technical guarantee level indicators (such as the security of data storage and transmission, data backup and recovery mechanism); obtain the data use authorization scope of each of the data source nodes, the data use authorization scope includes the data shareable user group, the allowed usage scenarios (such as scientific research, medical diagnosis) and time restrictions; extract the data protection policy of each of the data source nodes, the data protection policy includes the data desensitization level (such as removing sensitive information, adding noise), access frequency restrictions and transmission encryption requirements;

[0063] The content of the above step S12 is further described with examples as follows:

[0064] Hospital A has strong data security capabilities, a good historical credit record, and a high level of technical support. Its data use authorization is limited to research institutions, and its data protection policy requires data desensitization. Hospital B has a relatively low data security level, a wider range of data use authorizations, and a data protection policy requiring encrypted data transmission.

[0065] S13. Based on the security level, the data usage authorization scope, and the data protection policy, use a hierarchical clustering algorithm to divide the multiple data source nodes into different trust domains, where each trust domain contains at least one data source node, and the nodes within the domain have similar security levels, data usage authorization scopes, and data protection policies. For example, Hospital A and hospitals with similar security levels may be divided into one trust domain, and Hospital B and hospitals with similar security levels may be divided into another trust domain.

[0066] Using a distributed hash algorithm to assign a unique domain identifier to each of the trust domains, such as "Domain_1" or "Domain_2";

[0067] S14. Deploy a smart contract within each trust domain. The smart contract is implemented based on the Hyperledger Fabric framework. The smart contract includes an access rights management module, a data sharing rules module, and an access audit module. The access rights management module implements fine-grained access control of data within the domain based on the RBAC (role-based access control) model. The data sharing rules module is used to configure and execute data sharing policies, such as specifying which users can access which data and how the data is used. The access audit module is used to record data access operations for easy tracking and auditing. For example, the smart contract deployed within "Domain_1" stipulates that only authorized scientific research institution users can access data within the domain, and only for scientific research purposes.

[0068] In summary, based on step S1, this embodiment achieves refined management of data of different security levels by dividing trust domains and deploying smart contracts, effectively preventing data leakage and abuse; by designing standardized data formats and unified data sharing rules, it simplifies the data sharing process and improves data sharing efficiency; this embodiment also sets an access audit module to record all data access operations, ensuring the traceability of data sources and usage processes, and improving the credibility of the data.

[0069] S2. Build a cross-trust domain data mapping model, which includes data field mapping relationships, data format conversion rules, and data consistency verification rules; based on the data mapping model, establish a data synchronization channel between different trust domains, and transmit data through the data synchronization channel using end-to-end encryption;

[0070] S21. Analyze the data patterns of the source trust domain and the target trust domain to extract field feature information. The source trust domain is the sender / provider of the data and is the security boundary environment where the original data is located. The target trust domain is the receiver / user of the data and is the security boundary environment where the data needs to be obtained and used. For example:

[0071] The source domain database table storing user information contains fields such as "username," "user ID," "email address," and "telephone number," while the target domain's user information table contains fields such as "name," "ID," "email," and "mobile number." By analyzing the field names, data types, and meanings of these two tables, we extract field feature information, such as semantic similarity between field names and data type compatibility.

[0072] Input the field feature information into a deep learning model, such as a BERT model, and use the extracted field feature information as input to generate a field similarity matrix. This matrix reflects the similarity between each field in the source and target domains. For example, the similarity between "username" and "name" is 0.9, the similarity between "user ID" and "ID" is 0.95, the similarity between "email address" and "email" is 0.92, and the similarity between "telephone number" and "mobile number" is 0.88.

[0073] Construct a data field mapping relationship based on the field similarity matrix. Specifically, map the source domain field to the target domain field based on the similarity score. Set a threshold, such as 0.8, so that only fields with a similarity score above the threshold are mapped. For example, "username" is mapped to "name", "user ID" is mapped to "ID", "email address" is mapped to "email", and "telephone number" is mapped to "mobile number";

[0074] Develop data format conversion rules based on the type differences in the data field mapping relationships. For example, the "Phone Number" field in the source domain is stored in the format of "+86-12345678901," while the "Mobile Number" field in the target domain is stored in the format of "12345678901." A conversion rule needs to be developed to remove the "+86-" prefix.

[0075] Establish data consistency verification rules based on the data field mapping relationship. For example, if the "User ID" field in the source domain is unique, the "ID" field in the target domain must also be unique. Verification rules can ensure that data remains consistent after mapping and conversion.

[0076] Combining the data field mapping relationship, the data format conversion rule, and the data consistency verification rule to generate the data mapping model;

[0077] S22. Establish a data synchronization channel:

[0078] Analyze data transmission requirements based on the data mapping model, such as data volume, transmission frequency, security requirements, etc., and determine communication link specifications, such as bandwidth, latency, encryption algorithm, etc.;

[0079] Establishing a dedicated communication link between the trust domains according to the communication link specifications, for example, establishing a point-to-point connection using VPN or SDN technology;

[0080] generating an asymmetric encryption key pair for the dedicated communication link, comprising a public key and a private key, wherein the public key is used to encrypt data and the private key is used to decrypt data;

[0081] Using the asymmetric encryption key pair to perform end-to-end encryption on the dedicated communication link to form the data synchronization channel, ensuring that only the target domain holding the private key can decrypt the data;

[0082] Configuring the transmission parameters of the data synchronization channel according to the data scale of the data mapping model, such as data packet size, transmission rate, retransmission mechanism, etc.;

[0083] S23. Execute data mapping conversion:

[0084] Process the source data using the data field mapping relationship to complete field mapping, for example, mapping the "user name" field value in the source data to the "name" field in the target data;

[0085] Use the data format conversion rule to convert the format of the mapped data, for example, convert "+86-12345678901" in the source data to "12345678901";

[0086] Apply the data consistency check rules to verify the conversion results, for example, check whether the "ID" field in the target data is unique;

[0087] Using the verified data as data to be transmitted; transmitting the data to be transmitted to the target trust domain through the data synchronization channel;

[0088] S24. Implement data synchronization processing:

[0089] receiving encrypted data in a target trust domain; decrypting the received data using the asymmetric encryption key pair; performing reverse data conversion according to the data mapping model; verifying the integrity and consistency of the converted data; and writing the verified data into a target trust domain storage system;

[0090] Generate a data synchronization status report, recording various indicators during the synchronization process, such as the amount of data transferred, transfer time, number of errors, etc.

[0091] S25. Optimize synchronization mechanism:

[0092] Analyze abnormal conditions in the data synchronization status report, such as data errors, transmission delays, etc.;

[0093] Correcting the data field mapping relationship and data format conversion rules of the data mapping model according to the abnormal situation. For example, if it is found that some field mappings are inaccurate, it is necessary to adjust the field similarity threshold or modify the mapping relationship.

[0094] Monitoring performance indicators of the data synchronization channel, such as bandwidth utilization, latency, packet loss rate, etc.;

[0095] Optimizing the transmission parameters of the data synchronization channel based on the performance indicators, such as adjusting the data packet size, transmission rate, etc.;

[0096] Updating the configuration of the data mapping model and the data synchronization channel to adapt to changing data synchronization requirements;

[0097] Based on step S2, this embodiment adopts end-to-end encryption technology to ensure that data is not stolen or tampered with during transmission, effectively protecting the security of sensitive data; through automated data mapping and conversion, it reduces manual intervention, improves data synchronization efficiency, and reduces data synchronization costs; and the designed data consistency verification rules ensure that data maintains integrity and consistency during the synchronization process, improves the quality of target domain data, and provides a reliable data foundation for subsequent data analysis and application.

[0098] S3. Setting a data flow control module in the data synchronization channel, wherein the data flow control module dynamically adjusts the data transmission rate according to the real-time network status and data priority, includes:

[0099] S31. Setting a data flow control module in the data synchronization channel, wherein the data flow control module includes a network monitoring unit, a data classification unit, and a rate adjustment unit;

[0100] S311. The network monitoring unit deploys network detection points at both ends of the data synchronization channel. The network detection points periodically (for example, every 50 milliseconds) send detection packets to collect round-trip delay, packet loss rate, available bandwidth and link jitter data. For example, the detection point sends a detection packet of 1KB in size. The receiving end immediately replies after receiving it. The sending end records the time difference between sending and receiving and calculates the round-trip delay; counts the number of detection packets successfully sent and received within a period of time and calculates the packet loss rate; calculates the available bandwidth by measuring the amount of data transmitted within a period of time; records the changes in the round-trip delay of multiple consecutive detection packets and calculates the link jitter;

[0101] A sliding time window is used to smooth the round-trip delay, packet loss rate, available bandwidth, and link jitter data to generate a network status assessment matrix. For example, a window with a length of 1 second is used to average or weighted average the collected round-trip delay, packet loss rate, available bandwidth, and link jitter data to generate a network status assessment matrix. For example, if 10 round-trip delay values ​​are collected within the last second, namely 10ms, 12ms, 11ms, 13ms, 10ms, 9ms, 11ms, 12ms, 10ms, and 11ms, the average round-trip delay is 10.9ms. Sliding time window smoothing can filter out the impact of instantaneous network fluctuations and more accurately reflect the network status.

[0102] S312. The data classification unit classifies the data into core business data, general business data, and non-critical data according to the business importance of the data. For example, transaction-related order data is core business data, user browsing history is general business data, and advertising recommendation data is non-critical data.

[0103] Data is divided into real-time data, quasi-real-time data, and offline data based on the timeliness requirements of the data. For example, video conferencing data is real-time data, email is quasi-real-time data, and log data is offline data;

[0104] Orthogonally combine data divided by business importance and data divided by timeliness requirements to form a data priority matrix, such as core business real-time data, core business near-real-time data, general business real-time data, etc.

[0105] Set resource allocation weights for each priority level in the data priority matrix. For example, core business real-time data has a weight of 0.5, general business real-time data has a weight of 0.3, and non-critical data has a weight of 0.1. The larger the weight value, the higher the priority data will have in network resource competition.

[0106] S313. The rate adjustment unit calculates a current network load factor based on the network status evaluation matrix. The network load factor is determined by a round-trip delay variation rate, a packet loss rate, and a bandwidth utilization rate. For example, if the round-trip delay variation rate is high, the packet loss rate is high, and the bandwidth utilization rate is high, the network load factor is high.

[0107] The base transmission rate is calculated using an improved algorithm of increasing by addition and decreasing by multiplication based on the network load factor. When the network load factor is low, the base transmission rate is increased by a certain percentage; when the network load factor is high, the base transmission rate is decreased by a certain percentage. For example, if the initial base transmission rate is 1 Mbps, when the network load factor is lower than 0.5, the base transmission rate is increased by 0.1 Mbps each time; when the network load factor is higher than 0.7, the base transmission rate is decreased by 0.1 Mbps each time.

[0108] S32. Allocate the reference transmission rate to data flows of different priorities according to the resource allocation weights to achieve dynamic adjustment of the data transmission rate. For example, if the reference transmission rate is 1 Mbps and the core service real-time data weight is 0.5, the transmission rate of the core service real-time data is 0.5 Mbps.

[0109] Going a step further, the network load factor is calculated as follows:

[0110]

[0111] Among them, α represents the network load coefficient, w1, w2 and w3 represent the preset weight coefficients, △RTT represents the change in round-trip delay, RTT base represents the benchmark round trip delay, PLR represents the packet loss rate, BW used Indicates the used bandwidth, BW total Indicates the total available bandwidth;

[0112] The formula for calculating the baseline transmission rate is as follows:

[0113]

[0114] Among them, R(t) represents the transmission rate at time t, β represents the growth factor, γ represents the reduction factor, and α threshold Indicates the load threshold;

[0115] Based on step S3, this embodiment can fully utilize network bandwidth, avoid network congestion, and improve network resource utilization by dynamically adjusting the data transmission rate. Through data classification and resource allocation weights, the transmission of core business data and real-time data can be prioritized to avoid delays or losses in important data transmission. Through network monitoring and rate adjustment, the data transmission rate can be dynamically adjusted according to network conditions to avoid network overload and enhance network stability.

[0116] S4. When a data requester initiates a cross-domain data access request, the smart contract verifies the data requester's access rights and records data access logs. Differential privacy processing is performed on the authorized data and the differentially private processed data is returned to the data requester, ensuring data availability while preventing sensitive information from being leaked.

[0117] S41. The data requester initiates a cross-domain data access request through a unified access gateway. The gateway receives the cross-domain data access request initiated by the data requester, which includes the access target identifier, the requester's identity credentials extracted from the digital certificate, and a clear data usage purpose. The gateway verifies the digital signature of the requester's identity credentials using an asymmetric encryption algorithm to obtain a verification result. Specifically, the digital certificate is verified, the legitimacy of the digital signature is verified through the public key infrastructure, and the identity attribute information contained therein is extracted. For example, the requester is a medical research institution, the access target is a hospital's clinical dataset, and the usage purpose is new drug research and development data analysis.

[0118] The verification result is matched with a preset access control policy matrix, which includes permission level matching rules and usage scenario compliance rules. Based on the matching result, it is determined whether the access permission level of the data requester meets the access requirements of the access target, and whether the purpose of data use complies with the usage scenario compliance rules. For example, for medical data, medical institutions have a higher permission level, and new drug research and development belongs to a compliant usage scenario. The system determines whether the permission level and usage purpose of the requester meet the policy requirements.

[0119] When the access permission level meets the access requirements and the data usage purpose complies with the usage scenario compliance rules, the historical access records of the data requester are obtained, including dimensions such as access frequency and data usage compliance; a credit score is calculated based on the historical access records; for example, a higher score is obtained if there is no record of illegal use in the past year; a unique access transaction identifier is generated, and the requester's identity credentials, access time, access target, data usage purpose, and credit score are recorded in the access log on the blockchain. The access log adopts a chain storage structure and is encrypted and stored using a cryptographic hash function;

[0120] S42. When the credit score is higher than a preset credit threshold, obtain the data to be accessed corresponding to the access target, identify sensitive fields in the data to be accessed, such as the patient's ID number, diagnosis results, etc.; and set differential privacy budget parameters based on the sensitivity of the sensitive fields and the purpose of data use;

[0121] A differential privacy processing method is selected based on the data type of the data to be accessed. Random noise that conforms to the Laplace distribution is added to numerical data such as age. Categorical data such as gender is randomized using an exponential mechanism. Sensitive information is replaced with a semantically preserved desensitization algorithm for text data such as medical records. Taking outpatient medical records as an example, the original data contains basic patient information and medical records. By adding random noise that conforms to the Laplace distribution, the age field is converted into a fuzzy interval. The gender field is processed using a random response mechanism. Sensitive entities in the medical record description text (such as the patient's name, address, etc.) are desensitized and replaced.

[0122] Generate privacy-processed data;

[0123] S43. Obtain historical data distribution characteristics of the data to be accessed, including the distribution patterns of each field; construct a noise compensation model based on the historical data distribution characteristics to optimize the noise in the privacy-processed data, thereby preserving the statistical characteristics of the data as much as possible while protecting privacy, and obtain optimized data; measure the privacy protection strength by calculating ε-differential privacy, and simultaneously evaluate the data availability index;

[0124] S44. When both the privacy protection strength and the data utility index meet the preset privacy threshold requirements, the optimized data is returned to the data requester as an access result, and the return status of the access result is recorded in the access log, forming a complete access audit trail. Throughout the entire process, all access control judgments, data processing, and logging operations are executed through smart contracts, ensuring transparency and immutability of the process.

[0125] The requesting party must process the acquired private data in accordance with the agreed upon usage specifications. The system continuously tracks data usage and regularly updates the requester's credit score. Any violations will result in a credit score reduction and be recorded on the blockchain, impacting subsequent access requests. This approach effectively protects sensitive data while meeting data sharing requirements.

[0126] Access control and data processing parameters at each stage can be configured and adjusted based on the specific application scenario. For example, permission requirements and privacy protection can be increased for more sensitive data, while access restrictions can be relaxed for more trustworthy requesters. The entire mechanism is automated through smart contracts, providing a secure, transparent, and auditable cross-domain data sharing environment.

[0127] S5. Use a distributed consensus mechanism and version control mechanism based on the Byzantine fault tolerance algorithm to synchronize data between different trust domains;

[0128] S51. Build a distributed consensus network across trust domains, deploy state synchronization nodes in different trust domains, and divide the state synchronization nodes into a consensus node group and a verification node group. For example, assume there are three trust domains A, B, and C, and deploy three nodes in each trust domain, two of which serve as consensus nodes and one as a verification node.

[0129] A trust score is calculated based on the node's historical behavior (e.g., timeliness of message delivery, active participation in consensus, etc.). The node with the highest trust score is selected as the master node, responsible for coordinating the data state synchronization process. Assume that node A1 in domain A is selected as the master node due to its good historical behavior. The distributed consensus network is implemented based on the Byzantine Fault Tolerance algorithm, ensuring that the system can maintain normal operation even if no more than one-third of the nodes fail or behave maliciously.

[0130] S52. The distributed consensus network is implemented based on the Byzantine fault tolerance algorithm;

[0131] Receive a data state change request within a trust domain. The request includes a data content hash value, a change timestamp, an operation sequence, and a version number. When the data state within a trust domain changes, for example, if the data in domain A changes from "value1" to "value2," a data state change request is generated. This request includes a data content hash value (e.g., "hash_value2"), a change timestamp (e.g., "2024-07-27 10:00:00"), an operation sequence (e.g., "update"), and a version number (e.g., from 1 to 2).

[0132] The master node verifies the validity of the data state change request, such as checking whether the hash value matches, whether the timestamp is valid, etc. After the verification is passed, the master node generates a state synchronization proposal containing all the above information; and broadcasts the state synchronization proposal to the consensus node group and the verification node group;

[0133] The consensus node group votes on the state synchronization proposal based on the Byzantine fault tolerance algorithm; when the number of votes exceeds the fault tolerance threshold, a pre-commit message is broadcast to the state synchronization nodes of the trust domain outside the current trust domain; and a confirmation message is received from the trust domain outside the current trust domain, wherein the confirmation message includes the node signature and the state verification result. A further example is as follows:

[0134] The consensus node group votes on state synchronization proposals based on a Byzantine fault-tolerant algorithm. Assume that after receiving a proposal, the consensus node group begins voting. When the number of votes exceeds a tolerance threshold (e.g., two-thirds), the consensus node group broadcasts a pre-commit message to state synchronization nodes in other trust domains. For example, if both consensus nodes in domain A and consensus nodes in other domains vote in favor, the tolerance threshold is met.

[0135] After receiving the pre-commit message, the state synchronization nodes in the other trust domains verify the data state change request and return a confirmation message. The confirmation message includes the node's signature and the state verification result. For example, nodes in domains B and C verify the hash value and version number of "value2" and return a confirmation message after confirming that they are correct.

[0136] A version control mechanism is established between each trust domain. This mechanism includes a version tree structure that records the historical trajectory of data status changes. When concurrent cross-domain modifications are detected, a branch version is created in the version tree structure. For example, if domain A modifies the data to "value2" and domain B modifies the data to "value3", a branch version will be created in the version tree structure. Assuming "value1" is version 1, "value2" is version 2, and "value3" is version 3, the version tree will show that version 1 has two branches, version 2 and version 3.

[0137] Determine the major version through Byzantine Fault Tolerant consensus. Assume that the consensus result is to select "value2" as the major version;

[0138] Analyze concurrent operation sequences across different trust domains and identify compatible and conflicting operations. If an operation in domain A adds a field and an operation in domain B modifies another field, the two operations are compatible. If both domains A and B modify the same field, the two operations are conflicting.

[0139] For compatible operations, they are merged into the main version. For conflicting operations, the final operation sequence is determined through cross-domain Byzantine fault-tolerant consensus. For example, if domains A and B both modify the same field, the final consensus result is to adopt the modification in domain A.

[0140] Submit the merged status update to each trust domain for synchronization. Based on the above example, the data in all trust domains is updated to "value2";

[0141] Monitor the operating status of nodes in each trust domain, including response latency, message integrity, and consensus participation; when node failure or malicious behavior is detected, such as a node being continuously unresponsive, remove the abnormal node from the consensus process; recalculate the fault tolerance threshold to ensure that the number of remaining normal nodes meets the Byzantine fault tolerance requirements.

[0142] This embodiment is based on step S5 and uses distributed consensus and version control mechanisms to ensure that the data status between different trust domains remains consistent, avoiding data conflicts and inconsistencies. Based on the Byzantine fault tolerance algorithm, even if some nodes fail or engage in malicious behavior, the system can still operate normally, ensuring data security and system stability. It also supports data synchronization across multiple trust domains, facilitating system expansion and upgrades, and adapting to more complex application scenarios.

[0143] Example 2:

[0144] Figure 2 This is a structural diagram of a distributed multi-mode data cross-trust domain data sharing system according to embodiment 2 of the present invention. Figure 2 As shown, the system includes a data access and trust management unit, a cross-domain data mapping and synchronization unit, an access control and privacy protection unit, and a consensus and version management unit, wherein:

[0145] The data access and trust management unit receives heterogeneous data provided by multiple data source nodes, performs data type identification and format standardization on the heterogeneous data; divides the multiple data source nodes into different trust domains based on the security level of the data provider, the scope of data use authorization, and the data protection policy, each of which contains at least one data source node, and assigns a unique domain identifier to each trust domain; deploys a smart contract in each trust domain, and the smart contract is used to manage data access rights and data sharing rules within the domain;

[0146] The cross-domain data mapping and synchronization unit has a cross-trust domain data mapping model built in it, and the data mapping model includes data field mapping relationships, data format conversion rules, and data consistency verification rules; based on the data mapping model, a data synchronization channel is established between different trust domains, and the data synchronization channel uses end-to-end encryption to transmit data; a data flow control module is set in the data synchronization channel, and the data flow control module dynamically adjusts the data transmission rate according to the real-time network status and data priority;

[0147] When a data requester initiates a cross-domain data access request, the access control and privacy protection unit verifies the access rights of the data requester through the smart contract and records the data access log; performs differential privacy processing on the data authorized for access and returns the differential privacy processed data to the data requester;

[0148] The consensus reaching and version management unit adopts a distributed consensus mechanism and version control mechanism based on the Byzantine fault tolerance algorithm to synchronize data between different trust domains.

[0149] Example 3:

[0150] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the distributed multimodal data cross-trust domain data sharing method as described in any one of Embodiment 1 is implemented.

[0151] Example 4:

[0152] This embodiment also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the distributed multimodal data cross-trust domain data sharing method as described in any one of Embodiment 1 is implemented.

[0153] It is worth noting that the system, electronic device and computer-readable storage medium described in the present invention are all based on the same inventive concept as the method described in Example 1, and will not be described in detail here.

[0154] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A distributed multi-modal data sharing method across trust domains, characterized in that: include: Receive heterogeneous data provided by multiple data source nodes, and perform data type identification and format standardization processing on the heterogeneous data; Based on the data provider's security level, data usage authorization scope, and data protection policy, the multiple data source nodes are divided into different trust domains, each of which contains at least one data source node, and a unique domain identifier is assigned to each trust domain; a smart contract is deployed in each trust domain, and the smart contract is used to manage data access rights and data sharing rules within the domain; Construct a cross-trust domain data mapping model, which includes data field mapping relationships, data format conversion rules, and data consistency verification rules; based on the data mapping model, establish a data synchronization channel between different trust domains, and transmit data in an end-to-end encrypted manner; set up a data flow control module in the data synchronization channel, and dynamically adjust the data transmission rate according to the real-time network status and data priority; When a data requester initiates a cross-domain data access request, the data requester's access rights are verified through the smart contract, and a data access log is recorded; Perform differential privacy processing on the authorized data and return the differential privacy processed data to the data requester; A distributed consensus mechanism and version control mechanism based on the Byzantine fault-tolerant algorithm are used to synchronize data between different trust domains.

2. The distributed multi-mode data cross-trust domain data sharing method according to claim 1 is characterized in that: Receive heterogeneous data provided by multiple data source nodes, and perform data type identification and format standardization processing on the heterogeneous data; Based on the security level of the data provider, the scope of data use authorization, and the data protection policy, the multiple data source nodes are divided into different trust domains, each of the trust domains includes at least one data source node, and a unique domain identifier is assigned to each of the trust domains; A smart contract is deployed in each trust domain. The smart contract is used to manage data access rights and data sharing rules within the domain, including: Receive heterogeneous data provided by multiple data source nodes, analyze the heterogeneous data through a feature extraction algorithm to obtain heterogeneous features, wherein the heterogeneous features are specifically structural features, content features, and metadata features, and identify the data type of the heterogeneous data based on the heterogeneous features; classify the heterogeneous data into structured data, semi-structured data, and unstructured data according to the results of the data type identification; construct a unified data representation model based on a graph data model, use a data conversion adapter to convert the structured data, the semi-structured data, and the unstructured data into a standardized format corresponding to the graph data model, and store the standardized format; Analyze the security level of each data source node, where the security level includes a data security protection capability indicator, a historical credit record indicator, and a technical assurance level indicator; obtain the data usage authorization scope of each data source node, where the data usage authorization scope includes the data shareable user group, permitted usage scenarios, and time restrictions; extract the data protection policy of each data source node, where the data protection policy includes a data desensitization level, access frequency restrictions, and transmission encryption requirements; based on the security level, the data usage authorization scope, and the data protection policy, use a hierarchical clustering algorithm to divide the multiple data source nodes into different trust domains, each trust domain containing at least one data source node; and use a distributed hash algorithm to assign a unique domain identifier to each trust domain; A smart contract is deployed in each trust domain, and the smart contract is implemented based on the Hyperledger Fabric framework. The smart contract includes an access rights management module, a data sharing rules module, and an access audit module. The access rights management module implements fine-grained access control of data within the domain based on the RBAC model. The data sharing rules module is used to configure and execute data sharing policies. The access audit module is used to record data access operations.

3. The distributed multi-mode data sharing method across trust domains according to claim 1 is characterized in that: Constructing a cross-trust domain data mapping model, the data mapping model includes data field mapping relationships, data format conversion rules, and data consistency verification rules; based on the data mapping model, establishing a data synchronization channel between different trust domains, the data synchronization channel using end-to-end encryption to transmit data includes: Analyze the data patterns of the source trust domain and the target trust domain to extract field feature information; input the field feature information into a deep learning model to generate a field similarity matrix; construct a data field mapping relationship based on the field similarity matrix; formulate data format conversion rules based on type differences in the data field mapping relationship; establish data consistency verification rules based on the data field mapping relationship; and combine the data field mapping relationship, the data format conversion rules, and the data consistency verification rules to generate the data mapping model; Analyzing data transmission requirements based on the data mapping model and determining communication link specifications; establishing a dedicated communication link between trust domains based on the communication link specifications; generating an asymmetric encryption key pair for the dedicated communication link; performing end-to-end encryption on the dedicated communication link using the asymmetric encryption key pair to form the data synchronization channel; and configuring transmission parameters of the data synchronization channel based on the data scale of the data mapping model; Processing source data using the data field mapping relationship to complete field mapping; converting the format of the mapped data using the data format conversion rules; verifying the conversion result using the data consistency verification rules; using the verified data as the data to be transmitted; and transmitting the data to be transmitted to the target trust domain through the data synchronization channel; receiving encrypted data in the target trust domain; decrypting the received data using the asymmetric encryption key pair; performing reverse data conversion according to the data mapping model; verifying the integrity and consistency of the converted data; writing the verified data to the target trust domain storage system; generating a data synchronization status report to implement data synchronization processing; Analyze the abnormal conditions in the data synchronization status report; correct the data field mapping relationship and data format conversion rules of the data mapping model according to the abnormal conditions; monitor the performance indicators of the data synchronization channel; optimize the transmission parameters of the data synchronization channel based on the performance indicators; update the configuration of the data mapping model and the data synchronization channel, and optimize the synchronization mechanism.

4. The distributed multi-mode data cross-trust domain data sharing method according to claim 1 is characterized in that: Setting a data flow control module in the data synchronization channel, wherein the data flow control module dynamically adjusts the data transmission rate according to the real-time network status and data priority includes: A data flow control module is provided in the data synchronization channel. The data flow control module includes a network monitoring unit, a data classification unit, and a rate adjustment unit. The network monitoring unit deploys network detection points at both the transmitting and receiving ends of the data synchronization channel. The network detection points periodically send detection packets to collect round-trip delay, packet loss rate, available bandwidth, and link jitter data. The round-trip delay, packet loss rate, available bandwidth, and link jitter data are smoothed using a sliding time window to generate a network status assessment matrix. The data classification unit divides the data into core business data, general business data, and non-critical data according to the business importance of the data, and divides the data into real-time data, quasi-real-time data, and offline data according to the timeliness requirement of the data; orthogonally combines the data divided according to the business importance and the data divided according to the timeliness requirement to form a data priority matrix; and sets a resource allocation weight for each priority in the data priority matrix; The rate adjustment unit calculates the current network load coefficient based on the network status assessment matrix, where the network load coefficient is determined by the round-trip delay change rate, the packet loss rate, and the bandwidth utilization rate; calculates the reference transmission rate based on the network load coefficient using an improved addition increase and multiplication decrease algorithm; and allocates the reference transmission rate to data streams of different priorities according to the resource allocation weight to achieve dynamic adjustment of the data transmission rate.

5. The distributed multi-mode data sharing method across trust domains according to claim 1 is characterized in that: When a data requester initiates a cross-domain data access request, the data requester's access rights are verified through the smart contract, and a data access log is recorded; Perform differential privacy processing on the authorized data and return the differential privacy processed data to the data requester, including: Receive a cross-domain data access request initiated by a data requester, wherein the cross-domain data access request includes an access target, an identity certificate of the requester, and a purpose of data use; verify the digital signature of the identity certificate of the requester using an asymmetric encryption algorithm to obtain a verification result; Matching the verification result with a preset access control policy matrix, which includes permission level matching rules and usage scenario compliance rules; based on the matching result, determining whether the access permission level of the data requester meets the access requirements of the access target, and whether the data usage purpose complies with the usage scenario compliance rules; When the access permission level meets the access requirements and the data usage purpose complies with the usage scenario compliance rules, the historical access records of the data requester are obtained and a credit score is calculated based on the historical access records; an access transaction identifier is generated, and the requester's identity credentials, access time, access target, data usage purpose, and credit score are recorded in an access log. The access log adopts a chain storage structure and is encrypted and stored using a cryptographic hash function; When the credit score is higher than a preset credit threshold, obtaining the to-be-accessed data corresponding to the access target, identifying sensitive fields in the to-be-accessed data, and setting differential privacy budget parameters based on the sensitivity of the sensitive fields and the purpose of data use; Select a differential privacy processing method based on the data type of the data to be accessed, add random noise that conforms to the Laplace distribution to numerical data, use an exponential mechanism to randomize categorical data, and use a semantically preserving desensitization algorithm to replace sensitive information in text data to generate privacy-processed data; Obtaining historical data distribution characteristics of the data to be accessed, constructing a noise compensation model based on the historical data distribution characteristics, optimizing the noise in the privacy-processed data to obtain optimized data; and calculating the privacy protection strength and data utility index of the optimized data; When both the privacy protection strength and the data utility index meet the preset privacy threshold requirements, the optimized data is returned to the data requester as an access result, and the return status of the access result is recorded in the access log.

6. The distributed multi-mode data sharing method across trust domains according to claim 1 is characterized in that: Using a distributed consensus mechanism and version control mechanism based on the Byzantine fault-tolerant algorithm, data synchronization between different trust domains includes: Build a distributed consensus network across trust domains, deploy state synchronization nodes in different trust domains, and divide the state synchronization nodes into consensus node groups and verification node groups; calculate trust scores based on node historical behavior, and select the node with the highest trust score as the master node; the distributed consensus network is implemented based on the Byzantine fault tolerance algorithm; Receive a data state change request within the trust domain, the data state change request including a data content hash value, a change timestamp, an operation sequence, and a version number; the master node verifies the validity of the data state change request and generates a state synchronization proposal; broadcasts the state synchronization proposal to the consensus node group and the verification node group; The consensus node group votes on the state synchronization proposal based on the Byzantine fault tolerance algorithm; when the number of votes exceeds the fault tolerance threshold, a pre-commit message is broadcast to the state synchronization nodes of the trust domain outside the current trust domain; and a confirmation message is received from the trust domain outside the current trust domain, wherein the confirmation message includes a node signature and a state verification result; A version control mechanism is established between each trust domain. The version control mechanism includes a version tree structure that records the historical trajectory of data status changes. When cross-domain concurrent modification is detected, a branch version is created in the version tree structure. The main version is determined through Byzantine fault-tolerant consensus. Analyze concurrent operation sequences across different trust domains to identify compatible and conflicting operations. For compatible operations, merge them into the master version. For conflicting operations, determine the final operation sequence through cross-domain Byzantine fault-tolerant consensus. Submit the merged state update to each trust domain for synchronization. Monitor the operating status of nodes in each trust domain, including response latency, message integrity, and consensus participation; when node failure or malicious behavior is detected, remove the abnormal node from the consensus process; recalculate the fault tolerance threshold to ensure that the number of remaining normal nodes meets the Byzantine fault tolerance requirements.

7. A distributed multi-mode data cross-trust domain data sharing system, used to implement the distributed multi-mode data cross-trust domain data sharing method according to any one of claims 1 to 6, characterized in that: It includes data access and trust management unit, cross-domain data mapping and synchronization unit, access control and privacy protection unit, and consensus and version management unit, among which: The data access and trust management unit receives heterogeneous data provided by multiple data source nodes, performs data type identification and format standardization on the heterogeneous data; divides the multiple data source nodes into different trust domains based on the security level of the data provider, the scope of data use authorization, and the data protection policy, each of which contains at least one data source node, and assigns a unique domain identifier to each trust domain; deploys a smart contract in each trust domain, and the smart contract is used to manage data access rights and data sharing rules within the domain; The cross-domain data mapping and synchronization unit has a cross-trust domain data mapping model built in it, and the data mapping model includes data field mapping relationships, data format conversion rules, and data consistency verification rules; based on the data mapping model, a data synchronization channel is established between different trust domains, and the data synchronization channel uses end-to-end encryption to transmit data; a data flow control module is set in the data synchronization channel, and the data flow control module dynamically adjusts the data transmission rate according to the real-time network status and data priority; When a data requester initiates a cross-domain data access request, the access control and privacy protection unit verifies the access rights of the data requester through the smart contract and records the data access log; performs differential privacy processing on the data authorized for access and returns the differential privacy processed data to the data requester; The consensus reaching and version management unit adopts a distributed consensus mechanism and version control mechanism based on the Byzantine fault tolerance algorithm to synchronize data between different trust domains.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the distributed multimodal data cross-trust domain data sharing method as described in any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the distributed multi-modal data cross-trust domain data sharing method as described in any one of claims 1 to 6 is implemented.

Images