Risk management and control method for cross-domain interaction of power data

Abstract

The invention discloses a risk management and control method for cross-domain interaction of power data. The method comprises the following steps: deploying a pre-trained sensitive information identification model at a data exit gateway to identify a sensitive information type and a risk level; dynamically matching a desensitization strategy according to the risk level, and executing hierarchical desensitization operation; performing attribute-based encryption operation on a core field of the desensitized data, wherein an access attribute of a receiver is defined by a logic combination of a role attribute and a domain attribute; data are transmitted through a cross-domain secure channel based on a block chain, and decryption can be performed only when and only when a private key binding attribute set of a receiving party meets a logic combination access strategy predefined by a sender. According to the method disclosed by the invention, the risk of sensitive information leakage in cross-domain interaction of the power data is solved, hierarchical protection, dynamic access control and transmission integrity verification of the sensitive data are realized, and the security of data sharing of the power system is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of power data security, and in particular to a risk management method for cross-domain interaction of power data. Background Technology

[0002] With the rapid development of new power systems and the energy internet, cross-domain interaction of power data has become a core requirement supporting intelligent dispatching, renewable energy consumption, and source-grid-load-storage coordination. Currently, power companies generally use grid technology to integrate heterogeneous data, solving the problem of multi-source heterogeneous data access through protocol adaptation and data abstraction, and using dynamic protocol adapters to convert data formats of different protocols. Distributed storage is also employed to improve data sharing efficiency. However, current technologies largely focus on data format compatibility and lack hierarchical protection mechanisms for sensitive information, posing a risk of data leakage. Summary of the Invention

[0003] To address the above problems, this invention provides a risk management method for cross-domain interaction of power data, the method comprising: S1. At the cross-domain interactive data export gateway, the type and risk level of sensitive information in the power data to be transmitted are identified by a pre-trained sensitive information identification model. S2. Based on the identified sensitive information type and risk level, the corresponding desensitization strategy is matched from the preset desensitization strategy library through the strategy matching engine, and the sensitive information is desensitized. S3. For the core fields of the desensitized power data, perform attribute-based encryption based on the access attributes of the recipient. The access attributes of the recipient are predefined by the sender as a logical combination access strategy that includes role attributes and domain attributes. S4. Send the desensitized and encrypted power data to the security domain of the recipient through a cross-domain secure transmission channel; S5. After receiving the power data, the receiver uses the private key bound to its own attributes to perform a decryption operation. Decryption is successful only if the set of attributes bound to the private key satisfies the logical combination access strategy predefined by the sender, and the receiver obtains the power data containing the available core fields.

[0004] Further, in step S1, the sensitive information identification model is constructed using a pre-trained language model based on the Transformer architecture, and the pre-trained language model is trained through the following steps: S1.1 Collect power data samples labeled with sensitive information types and corresponding risk levels; S1.2. Convert the structured fields and associated metadata in the sample data into a text sequence according to a preset template, wherein the text sequence retains the semantic relationship between fields; S1.3 Extract contextual features of the text sequence through the Transformer encoder and output a feature tensor with global semantic association; S1.4 Input the feature tensor into the first fully connected classification layer and the second fully connected regression layer. The first fully connected classification layer outputs the probability distribution of sensitive information types, and the second fully connected regression layer outputs the risk level value. S1.5 Construct a multi-task loss function that integrates classification loss and regression loss, and iteratively update the model weights using the Adam optimizer; S1.6. Terminate training based on the early stopping mechanism of the validation set, and select the sensitive information identification model with the best comprehensive index of the validation set to deploy to the model inference engine of the cross-domain interactive data export gateway.

[0005] Furthermore, the pre-trained sensitive information identification model is used to identify the types and risk levels of sensitive information in the power data to be transmitted: The determination of sensitive information type: Based on the probability distribution of sensitive information type output by the sensitive information identification model, the type with the maximum probability value exceeding the preset threshold is selected as the final identified sensitive information type; The risk level determination is based on the risk level value output by the sensitive information identification model, which is then mapped to three levels of risk: high, medium, and low through the power industry risk matrix. Associate and bind the sensitive information types identified by the same field with the risk level, and output the identification results including the field location, sensitive information type and corresponding risk level.

[0006] Furthermore, in step S2, the de-identification strategy library employs a dynamic hierarchical de-identification rule engine, dynamically adjusting the de-identification strength based on the risk level of sensitive information and the security authentication level of the recipient. The de-identification strength and the executed operations include: High-intensity desensitization: Perform irreversible operations on high-risk sensitive fields, including replacing the original value with a cryptographically based secure hash algorithm and adding a salt value bound to the recipient's domain attribute; Medium-intensity desensitization: Perform partial masking on sensitive fields with medium risk level, retain the first N plaintext characters according to preset position retention rules, and replace subsequent characters with mask symbols; Low-intensity desensitization: Perform format-preserving encryption on low-risk sensitive fields while maintaining the data type and length.

[0007] Further, in step S3, the attribute base encryption operation employs a ciphertext policy attribute base for encryption, specifically including: S3.1 The system master key and public parameters are generated by the attribute authorization center within the security domain to which the sender belongs; S3.2 Construct an access control tree based on the recipient's access attributes, wherein: Leaf nodes are defined as role attributes and domain attributes; Non-leaf nodes are set as AND / OR logical threshold structures to express logical combination access strategies; S3.3. Based on the access control tree and public parameters, perform encryption operations on the core fields of the de-identified power data to generate ciphertext data containing the access control tree structure; S3.4 The attribute authorization center allocates an attribute set to the receiver and generates a corresponding private key. The private key is generated using the system master key and the receiver's authenticated attribute set, and satisfies the condition that the private key can decrypt ciphertext data if and only if the receiver's attribute set matches the logical combination access policy defined by the access control tree.

[0008] Furthermore, in step S4, the cross-domain secure transmission channel adopts a lightweight data integrity verification mechanism based on blockchain, specifically including: S4.1 Divide the desensitized and encrypted power data to be transmitted into K data fragments, and generate M redundant check fragments based on the Reed-Solomon erasure coding algorithm; S4.2 Generate a lightweight Merkle tree based on all data shards. The leaf nodes of the Merkle tree contain the hash values ​​of each shard content and the corresponding shard index. The root hash of the Merkle tree is written into the smart contract of the cross-domain consortium chain as a global verification anchor point for storage. S4.3 After receiving all data fragments and redundancy check fragments through the secure channel, the receiver reconstructs the Merkle tree based on the fragment index, calculates the local Merkle tree root hash, and simultaneously calls the smart contract to obtain the global verification anchor point and compares it with the reconstructed local Merkle tree root hash: When the hash values ​​match, the received data fragments are directly combined. When hash values ​​are inconsistent, the Merkle tree is traversed based on the shard index of the consortium blockchain to locate the abnormal shard. The redundant verification shard is used to perform RS decoding to reconstruct the original data. If the reconstruction fails, a cross-domain audit request is initiated to the smart contract, which triggers the sender to retransmit the abnormal shard and update the security domain credit score of the receiver.

[0009] The beneficial effects of this invention are: This invention utilizes a Transformer-based multi-task model combined with a power industry risk matrix to achieve joint identification of sensitive types and risk levels, avoiding false positives and false negatives. It employs attribute-based encryption to define access policies as logical combinations of role and domain attributes, allowing the receiver to decrypt only when the attributes meet the policy requirements. This avoids the ambiguity of responsibilities caused by reliance on manual supervision in traditional protocols. Furthermore, it employs a lightweight blockchain-based verification mechanism, sharding data and storing it on the blockchain using Merkle root hashes. The receiver can quickly locate tampered shards through on-chain anchors and recover data using redundant shards, achieving automated handling and tracing of transmission anomalies and ensuring data integrity. Attached Figure Description

[0010] Figure 1 This is a schematic diagram of the overall process of a risk management method for cross-domain interaction of power data provided by the present invention. Detailed Implementation

[0011] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0012] Example 1 A risk management method for cross-domain interaction of power data, the method comprising: S1. At the cross-domain interactive data export gateway, the type and risk level of sensitive information in the power data to be transmitted are identified by a pre-trained sensitive information identification model. S2. Based on the identified sensitive information type and risk level, the corresponding desensitization strategy is matched from the preset desensitization strategy library through the strategy matching engine, and the sensitive information is desensitized. S3. For the core fields of the desensitized power data, perform attribute-based encryption based on the access attributes of the recipient. The access attributes of the recipient are predefined by the sender as a logical combination access strategy that includes role attributes and domain attributes. S4. Send the desensitized and encrypted power data to the security domain of the recipient through a cross-domain secure transmission channel; S5. After receiving the power data, the receiver uses the private key bound to its own attributes to perform a decryption operation. Decryption is successful only if the set of attributes bound to the private key satisfies the logical combination access strategy predefined by the sender, and the receiver obtains the power data containing the available core fields.

[0013] As a preferred embodiment: In step S1, the sensitive information identification model is constructed using a pre-trained language model based on the Transformer architecture, and the pre-trained language model is trained through the following steps: S1.1 Collect power data samples labeled with sensitive information types and corresponding risk levels; S1.2. Convert the structured fields and associated metadata in the sample data into a text sequence according to a preset template, wherein the text sequence retains the semantic relationship between fields; S1.3 Extract contextual features of the text sequence through the Transformer encoder and output a feature tensor with global semantic association; S1.4 Input the feature tensor into the first fully connected classification layer and the second fully connected regression layer. The first fully connected classification layer outputs the probability distribution of sensitive information types, and the second fully connected regression layer outputs the risk level value. S1.5 Construct a multi-task loss function that integrates classification loss and regression loss, and iteratively update the model weights using the Adam optimizer; S1.6. Terminate training based on the early stopping mechanism of the validation set, and select the sensitive information identification model with the best comprehensive index of the validation set to deploy to the model inference engine of the cross-domain interactive data export gateway.

[0014] As a preferred embodiment, the pre-trained sensitive information identification model is used to identify the type and risk level of sensitive information in the power data to be transmitted. The determination of sensitive information type: Based on the probability distribution of sensitive information type output by the sensitive information identification model, the type with the maximum probability value exceeding the preset threshold is selected as the final identified sensitive information type; The risk level determination is based on the risk level value output by the sensitive information identification model, which is then mapped to three levels of risk: high, medium, and low through the power industry risk matrix. Associate and bind the sensitive information types identified by the same field with the risk level, and output the identification results including the field location, sensitive information type and corresponding risk level.

[0015] Specifically, sensitive fields such as user identity and geographical location in power data require differentiated protection. This embodiment adopts a multi-task Transformer model, which captures semantic dependencies between fields through a self-attention mechanism, identifies the types of sensitive information and their corresponding risk levels in the power data to be transmitted, and converts the structured power data into a text sequence through a preset template. The conversion template is "<field name 1>: value 1; <field name 2>: value 2; ...", where field names and values ​​have a one-to-one correspondence. The text sequence is input into the Transformer encoder, and the output layer outputs the probability distribution of sensitive information types and risk scores through the first fully connected classification layer and the second fully connected regression layer, respectively. The sensitive information type is determined based on the probability distribution of sensitive information types, and the risk score is mapped to the risk level through the power industry risk matrix. The sensitive information type and risk level of the same field are bound together, and the four-element recognition result of [field name, value, sensitive information type, risk level] is output. The intrinsic relationship between "sensitive information type and risk intensity" is learned through multi-task joint training, thereby improving the accuracy of sensitive information recognition.

[0016] As a preferred embodiment: In step S2, the desensitization strategy library employs a dynamic hierarchical desensitization rule engine, dynamically adjusting the desensitization strength based on the risk level of sensitive information and the security authentication level of the recipient. The desensitization strength and the executed operations include: High-intensity desensitization: Perform irreversible operations on high-risk sensitive fields, including replacing the original value with a cryptographically based secure hash algorithm and adding a salt value bound to the recipient's domain attribute; Medium-intensity desensitization: Perform partial masking on sensitive fields with medium risk level, retain the first N plaintext characters according to preset position retention rules, and replace subsequent characters with mask symbols; Low-intensity desensitization: Perform format-preserving encryption on low-risk sensitive fields while maintaining the data type and length.

[0017] Specifically, traditional static desensitization cannot adapt to the differences in the security level of the recipient in cross-domain scenarios. This embodiment introduces a dynamic rule engine, which combines the risk level of sensitive fields with the security authentication level of the recipient to match the desensitization strength and the corresponding desensitization operation. The recipient's security level is transmitted through a digital certificate. The desensitization strategy library is shown in Table 1: Table 1. Desensitization Strategy Library Risk level Receiver security level Desensitization intensity Desensitization procedure High risk Any level High-intensity desensitization The HMAC-SHA256 algorithm is used to generate irreversible hash values, a salt value is added to bind the receiver's domain attributes, and a fixed-length hash string is output. Medium risk AAA level Low-intensity desensitization Format Preservation Encryption (FPE) is performed to maintain the original data type and length, and reversible encryption is achieved using the AES-FF3 algorithm. AA grade and below Medium-intensity desensitization Retain the first N characters of plaintext (N = field length × 20%), and replace subsequent characters with a mask (*). Low risk AAA level No desensitization Preserve the original data in plaintext transmission AA grade and below Low-intensity desensitization Format Preservation Encryption (FPE) is performed to maintain the original data type and length, and reversible encryption is achieved using the AES-FF3 algorithm. This embodiment uses a dynamic de-identification strategy to enable low-risk data to retain higher value within the trusted domain, while effectively preventing the leakage of high-risk information.

[0018] Example 2 As a preferred embodiment: in step S3, the attribute base encryption operation employs a ciphertext policy attribute base for encryption, specifically including: S3.1 The system master key and public parameters are generated by the attribute authorization center within the security domain to which the sender belongs; S3.2 Construct an access control tree based on the recipient's access attributes, wherein: Leaf nodes are defined as role attributes and domain attributes; Non-leaf nodes are set as AND / OR logical threshold structures to express logical combination access strategies; S3.3. Based on the access control tree and public parameters, perform encryption operations on the core fields of the de-identified power data to generate ciphertext data containing the access control tree structure; S3.4 The attribute authorization center allocates an attribute set to the receiver and generates a corresponding private key. The private key is generated using the system master key and the receiver's authenticated attribute set, and satisfies the condition that the private key can decrypt ciphertext data if and only if the receiver's attribute set matches the logical combination access policy defined by the access control tree.

[0019] Specifically, traditional encryption methods, which rely solely on role-based access control, cannot express cross-domain composite strategies. This embodiment employs ciphertext policy attribute-based encryption, which is based on role attributes and domain attributes. The access strategy is encoded as a logic tree and embedded in the ciphertext for encryption. The logical combination strategy supports the principle of least privilege across domains (e.g., the new energy domain can only decrypt data that matches its attributes), thus avoiding the permission generalization problem of traditional encryption methods.

[0020] As a preferred embodiment: In step S4, the cross-domain secure transmission channel adopts a lightweight data integrity verification mechanism based on blockchain, specifically including: S4.1 Divide the desensitized and encrypted power data to be transmitted into K data fragments, and generate M redundant check fragments based on the Reed-Solomon erasure coding algorithm; S4.2 Generate a lightweight Merkle tree based on all data shards. The leaf nodes of the Merkle tree contain the hash values ​​of each shard content and the corresponding shard index. The root hash of the Merkle tree is written into the smart contract of the cross-domain consortium chain as a global verification anchor point for storage. S4.3 After receiving all data fragments and redundancy check fragments through the secure channel, the receiver reconstructs the Merkle tree based on the fragment index, calculates the local Merkle tree root hash, and simultaneously calls the smart contract to obtain the global verification anchor point and compares it with the reconstructed local Merkle tree root hash: When the hash values ​​match, the received data fragments are directly combined. When hash values ​​are inconsistent, the Merkle tree is traversed based on the shard index of the consortium blockchain to locate the abnormal shard. The redundant verification shard is used to perform RS decoding to reconstruct the original data. If the reconstruction fails, a cross-domain audit request is initiated to the smart contract, which triggers the sender to retransmit the abnormal shard and update the security domain credit score of the receiver.

[0021] Specifically, traditional data transmission channels cannot verify whether data has been tampered with or lost during transmission. This embodiment achieves shard-level integrity verification through a lightweight Merkle tree and consortium blockchain notarization, while erasure coding ensures data availability. The data to be transmitted is divided into multiple data shards, and multiple redundant shards are generated through Reed-Solomon encoding. A Merkle tree is constructed, and the Merkle tree root hash is written into the cross-domain consortium blockchain smart contract as a global verification anchor for notarization. The local system receives all transmitted data shards and redundant shards, reconstructs the Merkle tree based on the shard index, and calculates the local Merkle tree root hash. The smart contract is called to obtain the hash value of the on-chain global verification anchor and compare it with the local hash value. When the hash values ​​are inconsistent, the Merkle tree is traversed to locate the abnormal shard index, and the data is reconstructed using the redundant shards. If reconstruction fails, the smart contract automatically deducts the recipient's domain credit score and triggers retransmission. This embodiment improves the efficiency of data integrity verification, and blockchain notarization provides non-repudiable audit traceability for cross-domain data transmission.

[0022] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0023] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0024] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0025] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0026] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0027] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A risk management method for cross-domain interaction of power data, characterized in that, The method includes: S1. At the cross-domain interactive data export gateway, the type and risk level of sensitive information in the power data to be transmitted are identified by a pre-trained sensitive information identification model. S2. Based on the identified sensitive information type and risk level, the corresponding desensitization strategy is matched from the preset desensitization strategy library through the strategy matching engine, and the sensitive information is desensitized. S3. For the core fields of the desensitized power data, perform attribute-based encryption based on the access attributes of the recipient. The access attributes of the recipient are predefined by the sender as a logical combination access strategy that includes role attributes and domain attributes. S4. Send the desensitized and encrypted power data to the security domain of the recipient through a cross-domain secure transmission channel; S5. After receiving the power data, the receiver uses the private key bound to its own attributes to perform a decryption operation. Decryption is successful only if the set of attributes bound to the private key satisfies the logical combination access strategy predefined by the sender, and the receiver obtains the power data containing the available core fields.

2. The risk management method for cross-domain interaction of power data according to claim 1, characterized in that: In step S1, the sensitive information identification model is constructed using a pre-trained language model based on the Transformer architecture. The pre-trained language model is trained through the following steps: S1.1 Collect power data samples labeled with sensitive information types and corresponding risk levels; S1.

2. Convert the structured fields and associated metadata in the sample data into a text sequence according to a preset template, wherein the text sequence retains the semantic relationship between fields; S1.3 Extract contextual features of the text sequence through the Transformer encoder and output a feature tensor with global semantic association; S1.4 Input the feature tensor into the first fully connected classification layer and the second fully connected regression layer. The first fully connected classification layer outputs the probability distribution of sensitive information types, and the second fully connected regression layer outputs the risk level value. S1.5 Construct a multi-task loss function that integrates classification loss and regression loss, and iteratively update the model weights using the Adam optimizer; S1.

6. Terminate training based on the early stopping mechanism of the validation set, and select the sensitive information identification model with the best comprehensive index of the validation set to deploy to the model inference engine of the cross-domain interactive data export gateway.

3. The risk management method for cross-domain interaction of power data according to claim 2, characterized in that: The pre-trained sensitive information identification model is used to identify the types and risk levels of sensitive information in the power data to be transmitted. The determination of sensitive information type: Based on the probability distribution of sensitive information type output by the sensitive information identification model, the type with the maximum probability value exceeding the preset threshold is selected as the final identified sensitive information type; The risk level determination is based on the risk level value output by the sensitive information identification model, which is then mapped to three levels of risk: high, medium, and low through the power industry risk matrix. Associate and bind the sensitive information types identified by the same field with the risk level, and output the identification results including the field location, sensitive information type and corresponding risk level.

4. The risk management method for cross-domain interaction of power data according to claim 1, characterized in that: In step S2, the de-identification strategy library employs a dynamic hierarchical de-identification rule engine, dynamically adjusting the de-identification strength based on the risk level of sensitive information and the security authentication level of the recipient. The de-identification strength and the executed operations include: High-intensity desensitization: Perform irreversible operations on high-risk sensitive fields, including replacing the original value with a cryptographically based secure hash algorithm and adding a salt value bound to the recipient's domain attribute; Medium-intensity desensitization: Perform partial masking on sensitive fields with medium risk level, retain the first N plaintext characters according to preset position retention rules, and replace subsequent characters with mask symbols; Low-intensity desensitization: Perform format-preserving encryption on low-risk sensitive fields while maintaining the data type and length.

5. The risk management method for cross-domain interaction of power data according to claim 1, characterized in that: In step S3, the attribute-based encryption operation employs a ciphertext policy attribute-based encryption method, specifically including: S3.1 The system master key and public parameters are generated by the attribute authorization center within the security domain to which the sender belongs; S3.2 Construct an access control tree based on the recipient's access attributes, wherein: Leaf nodes are defined as role attributes and domain attributes; Non-leaf nodes are set as AND / OR logical threshold structures to express logical combination access strategies; S3.

3. Based on the access control tree and public parameters, perform encryption operations on the core fields of the de-identified power data to generate ciphertext data containing the access control tree structure; S3.4 The attribute authorization center allocates an attribute set to the receiver and generates a corresponding private key. The private key is generated using the system master key and the receiver's authenticated attribute set, and satisfies the condition that the private key can decrypt ciphertext data if and only if the receiver's attribute set matches the logical combination access policy defined by the access control tree.

6. The risk management method for cross-domain interaction of power data according to claim 1, characterized in that: In step S4, the cross-domain secure transmission channel employs a lightweight data integrity verification mechanism based on blockchain, specifically including: S4.1 Divide the desensitized and encrypted power data to be transmitted into K data fragments, and generate M redundant check fragments based on the Reed-Solomon erasure coding algorithm; S4.2 Generate a lightweight Merkle tree based on all data shards. The leaf nodes of the Merkle tree contain the hash values ​​of each shard content and the corresponding shard index. The root hash of the Merkle tree is written into the smart contract of the cross-domain consortium chain as a global verification anchor point for storage. S4.3 After receiving all data fragments and redundancy check fragments through the secure channel, the receiver reconstructs the Merkle tree based on the fragment index, calculates the local Merkle tree root hash, and simultaneously calls the smart contract to obtain the global verification anchor point and compares it with the reconstructed local Merkle tree root hash: When the hash values ​​match, the received data fragments are directly combined. When hash values ​​are inconsistent, the Merkle tree is traversed based on the shard index of the consortium blockchain to locate the abnormal shard. The redundant verification shard is used to perform RS decoding to reconstruct the original data. If the reconstruction fails, a cross-domain audit request is initiated to the smart contract, which triggers the sender to retransmit the abnormal shard and update the security domain credit score of the receiver.

Images