A cross-domain data product generation method based on data grid

Abstract

The present invention relates to the field of data processing technology, and in particular to a cross-domain data product generation method based on a data grid. According to the construction guiding principles of the data grid and based on the general field, the registration of a joint field is completed. Multiple fields develop their own data products and register virtual data products to the joint field. The privacy and security of multiple party data are guaranteed based on a double-layer encryption mechanism and a trusted computing platform, and finally the joint data product release based on a self-service data platform is realized, providing a safe and efficient implementation method for multi-party field joint privacy computing based on a data grid.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method for generating cross-domain data products based on a data grid. Background Art

[0002] Data grid is an emerging distributed data architecture design and development method. Different from traditional data architecture, data grid follows four basic principles: domain-based ownership, data as a product, self-service data service platform, and joint computing governance. Data grid separates the responsibilities between functional data domains that focus on creating data products and platform teams that focus on technical capabilities, so that data is managed by the domain closest to the business side and related data products are built, accelerating business delivery and providing more insightful data products.

[0003] The realization of data products will further refine the direction of analytical data, that is, each data product conducts in-depth analysis for business insights, and the interaction between different fields gradually becomes more flexible and efficient. However, although data grids have made significant progress in promoting innovation and business agility in the data field, there are still some important challenges in cross-domain data processing. Because data is oriented to domain autonomy, the semantics and structure of cross-domain data between different fields are inconsistent, and the quality is uneven; in addition, there is no unified methodology for the design of data privacy protection solutions for cross-domain joint computing; the final presentation form for joint computing needs needs to be specifically defined and implemented. Summary of the invention

[0004] The present invention provides a cross-domain data product generation method based on data grid, realizes the joint calculation of cross-domain privacy data in data grid, expands the single data providing function of data products, and realizes specific privacy protection and calculation according to the joint calculation requirements.

[0005] In order to achieve the purpose of the present invention, the technical solution adopted is: a cross-domain data product generation method based on a data grid, comprising the following steps:

[0006] S1: Multi-party domain initialization, which is completed by performing normal domain registration of joint computing participants in the data grid and recording necessary ID information;

[0007] S2: Build a joint computing domain, build a trusted domain that ensures that the data privacy of parties in each domain is not leaked and shared by multiple parties;

[0008] S3: Perform cross-domain data joint computing based on trusted domains;

[0009] S4: Based on the joint computing data product release of the self-service data platform, the joint computing data product release of the data grid is carried out.

[0010] As an optimization solution of the present invention, in step S1, the registration information included in the domain registration is domain ID, person in charge ID, engineer ID, domain division, domain name, and domain description.

[0011] As an optimization solution of the present invention, in step S2, the trusted domain is registered in the data grid as a domain node in the data grid, providing a trusted execution environment, on this basis, confirming the multi-party participation domain, and building a multi-party contract, and performing joint calculation of multi-party private data based on the contract confirmed by multi-party signatures. The specific implementation method is:

[0012] S2-1. A certain field is registered as a joint field. The basic nature of the joint field is the same as that of the general field. The field member A becomes the sole person in charge of the field.

[0013] S2-2, domain member A initiates an invitation to multiple domains with joint attribute data;

[0014] S2-3. The invited area agrees or refuses. If it agrees, it becomes the co-leader B, C or D of the joint area;

[0015] S2-4. The domain is managed by multiple participants in a collaborative way, forming a small federal governance. Any participant can propose modifications to the domain, but they will take effect only after confirmation by the confirming party.

[0016] S2-5. Participants in a field may choose to withdraw from the joint field at any time.

[0017] As an optimization solution of the present invention, joint computing constraints are implemented through a multi-party contract with principles and rules that are jointly observed and executed by multiple parties in the joint field. The multi-party contract includes a declaration of the joint field and a declaration of the joint computing model.

[0018] As an optimization solution of the present invention, for multi-party contracts, when the invited domain user agrees to enter the joint domain, the information of the participating parties is automatically updated; when a participating party withdraws, the corresponding user information is automatically deleted. The declaration part of the joint computing model is open to all participating parties, and each participating party can add new rules. The specific execution method is as follows:

[0019] 1) A participant proposes a new rule or model for joint computing in the system;

[0020] 2) The system broadcasts the new rules to all participants;

[0021] 3) All participants give their consent or rejection options within a specified time, such as 24 hours, and rejection is assumed within the time limit;

[0022] 4) If more than half of the parties agree, the rule will be added to the multi-party contract;

[0023] 5) The attributes and models in a multi-party contract are models that can be jointly called by multiple participants;

[0024] 6) If any party objects to the adopted model, they may choose to withdraw from the joint field.

[0025] As an optimization solution of the present invention, in step S3, cross-domain data joint computing is performed based on the trusted domain to ensure that multiple parties complete the joint computing without leaking private data, and cross-domain joint training based on the data grid is performed. The specific steps are:

[0026] S3-1. The joint domain automatically generates asymmetric system keys, including the public key Sys_PublicKey and the private key Sys_PrivateKey;

[0027] S3-2, the trusted computing environment is started, and the system public key and private key are synchronized;

[0028] S3-3, system attribute base function startup, including user access control and secret key generation for users;

[0029] S3-4. The participant registers and logs in, and generates the corresponding operation key ABF_Key through ABF verification;

[0030] S3-5. Participants build their own data products in the joint field, and the output of the data products is the joint cross-domain privacy data of all parties;

[0031] S3-6. The participant requests to perform real-time computing and return results through the trusted computing environment.

[0032] As an optimization scheme of the present invention, the attribute base function mainly completes the user's access control and key generation. The access control is implemented based on the joint computing governance principle of the data grid. The access control includes registration and login. For the joint domain of the data grid, the attribute base function strategy registered by the participating party includes at least the role and domain ID. When the registered user logs in to the joint domain later, ABF generates a secret key session for this login, which is used for data encryption and verification of multi-party communication.

[0033] As an optimization solution of the present invention, for the data grid, the data provided by multiple participants in the joint field is in the form of registered data products. The data products provide the joint data required by multiple parties, that is, the data of each party is still controlled by each participant. The output of the data product is the data encrypted or feature extracted by each party according to the attribute rules in the multi-party contract. The original data will not flow out, ensuring privacy and security.

[0034] As an optimization solution of the present invention, a specific method for cross-domain privacy data joint calculation is that participant B needs a certain ID data of participant A for joint calculation, and the implementation method is:

[0035] (1) The system public key encrypts the data product output of participant A;

[0036] (2) Encrypt the data using the ABF public key of Party B;

[0037] (3) Send the data to Party B and decrypt the data using Party B’s ABF private key;

[0038] (4) Use the system public key to encrypt Party B’s own data;

[0039] (5) Use the ABF private key of Party B to encrypt the data of both parties;

[0040] (6) The encrypted data is sent to the trusted computing platform, which decrypts the data;

[0041] (7) The trusted computing platform completes joint computing;

[0042] (8) The calculation result is encrypted with the system private key and the participant B’s public key and sent to the participant B for decoding.

[0043] As an optimization solution of the present invention, in step S4, the self-service data service platform is functionally adapted for the joint field, and the result of the joint calculation is still presented in the form of a data product. Specifically, the method for publishing the joint calculation data product of the data grid is as follows:

[0044] S4-1. Cooperate with the demand side in the field to determine the data attribute requirements and calculation model;

[0045] S4-2, data attribute requirements and calculation models are broadcast to the requesting party for confirmation;

[0046] S4-3. The demander develops data products in this field and registers them in the joint field;

[0047] S4-4. The joint field ABF verification is passed and released and operates automatically.

[0048] The present invention has positive effects: it conducts in-depth analysis on the joint computing scenario in the data grid, and provides a specific solution for how the data grid can ensure the data privacy and security of multi-party joint computing, which has the following advantages:

[0049] (1) It does not destroy the basic architecture principles of the data grid. It is a general federated computing method with many advantages of general federated computing. It can effectively improve computing efficiency and speed and is suitable for different scenarios and devices.

[0050] (2) Maintaining the joint domain by building multi-party contracts is open and transparent, which has significant benefits for multiple participants;

[0051] (3) The joint domain is a variant of the general domain, focusing on maintaining multi-party contracts and trusted computing platforms. The multi-party data products within it are a mapping of multi-party autonomous domain products, reducing resource consumption;

[0052] (4) Add ABF attribute base function for multi-party verification and key generation, and have the scalability of data attribute level verification;

[0053] (5) Double-layer encryption technology ensures the privacy and security of data from multiple parties during the transfer process;

[0054] (6) The construction and release of joint computing products through a self-service data platform is efficient and intuitive. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative work, among which:

[0056] Figure 1 This is the main flow chart of the cross-domain data product generation method based on data grid of the present invention;

[0057] Figure 2 It is a structural diagram of several principle components of the data grid of the present invention;

[0058] Figure 3 It is a virtual mapping diagram of the data product of the present invention;

[0059] Figure 4 This is a functional component diagram of the self-service data service platform of the present invention for joint fields and joint products. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0061] like Figure 1 As shown, the present invention discloses a method for generating cross-domain data products based on a data grid, comprising the following steps:

[0062] Step 1: Initialize multiple domains;

[0063] like Figure 2 The basic architecture of the data grid shown includes the construction of grid infrastructure, the construction of a federated team, the development of a data platform, and the initialization of the division of various fields. The cross-domain data product generation method proposed in the present invention does not change the original topological design principle of the data grid. It is still based on the data ownership of the field and the development of data products for each field. Therefore, the first step is to register multiple fields with joint computing needs as a common field in the grid. Step 1 completes the initialization of the common field by performing normal field registration of the joint computing participants in the data grid and recording the necessary ID information. It is carried out without destroying the original development principles of the computing grid. Specifically, the necessary registration information required for field registration is as follows:

[0064]

[0065] Step 2: Build a joint computing domain;

[0066] Cross-domain joint computing in multiple fields needs to ensure that the data privacy of parties in each field is not leaked, and first of all, a trusted field shared by multiple parties is built. Step 2: The purpose of building a joint computing field is to build a trusted field that guarantees that the data privacy of parties in each field is not leaked and shared by multiple parties. The trusted field is registered in the data grid as a field node in order to provide a trusted execution environment. On this basis, the multi-party participation field is confirmed, and a multi-party contract is built. Joint computing of the privacy data of multiple parties is performed based on the contract confirmed by multi-party signatures.

[0067] Specifically, the implementation method is:

[0068] 1) A certain field is registered as a joint field. The basic nature of the joint field is the same as that of the general field. It becomes field member A and is the only person in charge of the field at present;

[0069] 2) Domain member A sends an invitation to multiple domains with joint attribute data;

[0070] 3) The invited field agrees or refuses. If it agrees, it becomes the co-leader B, C, D, etc. of the joint field;

[0071] 4) The domain is managed by multiple participants in a collaborative way, forming a small federal governance. Any participant can propose changes to the domain, but the changes will only take effect if more than half of the participants confirm the changes.

[0072] 5) Field participants can choose to withdraw from the joint field at any time.

[0073] The joint domain is built on the basis of a trusted computing platform with a trusted execution environment, such as Intel SGX (Software Guard Extensions) and ARM TrustZone basic hardware. The purpose is to achieve data isolation and privacy protection during joint data computing, and this domain has the same infrastructure configuration as the ordinary domain.

[0074] In addition to inheriting the federated computing governance rules of the data grid itself, the federated domain also needs to have principles and rules that multiple parties in the federated domain abide by and execute together. The present invention implements this through a multi-party contract. The federated computing constraints are implemented through a multi-party contract that has principles and rules that multiple parties in the federated domain abide by and execute together. The multi-party contract contains two parts: one is the declaration of the federated domain, and the other is the declaration of the federated computing model.

[0075] For detailed description, the present invention uses YAML format to display the key part configuration, and the actual implementation is not limited to YAML or other arbitrary computer languages:

[0076]

[0077]

[0078] In actual development, the ID of the participant can be hidden from other participants to prevent the possibility of joint fraud and theft of data, but it needs to be recorded at the system level as a basis for joint computing governance monitoring and traceability at the data grid level.

[0079] For multi-party contracts, when the invited domain user agrees to enter the joint domain, the information of the participants is automatically updated; when a participant withdraws, the corresponding user information is automatically deleted. The joint calculation declaration part is open to all participants, and each participant can add new rules. The specific execution method is as follows:

[0080] 1) A participant proposes a new rule or model for joint computing in the system (the system generally refers to the system where the joint domain is located in the data grid, specifically the joint domain new rule and model submission function module provided by the self-service data platform);

[0081] 2) The system broadcasts the new rules to all participants;

[0082] 3) All participants give their consent or rejection options within a specified time, such as 24 hours, and rejection is assumed within the time limit;

[0083] 4) If more than half of the parties agree, the rule will be added to the multi-party contract;

[0084] 5) The attributes and models in a multi-party contract are models that can be jointly called by multiple participants;

[0085] 6) If any party objects to the adopted model, they may choose to withdraw from the joint field.

[0086] It can be seen that the joint computing attributes or models proposed by a participant must benefit multiple parties in order to be unanimously passed, because the disagreeing participants can choose to withdraw and refuse to share data, which is a loss for multiple parties.

[0087] Step 3: Perform cross-domain data joint computing based on trusted domains;

[0088] The construction and initialization of the joint domain is only the beginning of multi-party joint computing. What is important is how to ensure that multiple parties complete the joint computing without disclosing private data. Step 3 performs cross-domain data joint computing based on the trusted domain to ensure that multiple parties complete the joint computing without disclosing private data. For example, if Party A has a set of attributes of a certain ID data as [attribute 1, attribute 2], and Party B has the attributes of the same ID data as [attribute 3, attribute 4], Party A needs to combine its own data attributes with Party B's data attributes to jointly calculate and obtain a new result such as [attribute 5] for downstream tasks.

[0089] The following is a cross-domain joint training method based on data grid:

[0090] 1) The joint domain automatically generates system asymmetric keys, including the public key Sys_PublicKey and the private key Sys_PrivateKey;

[0091] 2) The trusted computing environment is started and the system public key and private key are synchronized;

[0092] 3) System attribute-based function (ABF) starts, including user access control and key generation for the passed user;

[0093] 4) The participant registers and logs in, and generates the corresponding operation key ABF_Key through ABF verification;

[0094] 5) Participants build their own data products in the joint field, and the output of the data products is the joint cross-domain privacy data of all parties;

[0095] 6) The participants request to perform real-time computing and return results through the trusted computing environment.

[0096] The attribute base function mainly completes the user's access control and key generation. Access control is implemented based on the federated computing governance principle of the data grid. Access control includes registration and login. For the federated domain of the data grid, the attribute base function strategy registered by the participants includes at least:

[0097]

[0098]

[0099] Other policies are added based on actual conditions, such as accessible time.

[0100] When a registered user subsequently logs into the joint domain, ABF generates a secret key session for this login, which is used for data encryption and verification of multi-party communication.

[0101] For data grid, the data provided by multiple participants in the joint field is in the form of registered data products. Data products provide joint data required by multiple parties, which means that the data of each party is still controlled by each participant. The output of the data product is the data encrypted or feature extracted by each party according to the attribute rules in the multi-party contract. The original data will not be leaked, ensuring privacy and security. Figure 3 As shown, in the joint field, the data products provided by multiple parties are virtual mapping registrations after development in the autonomous field. On the one hand, it reduces secondary development, and on the other hand, it minimizes the migration of data infrastructure to ensure that the original data never leaves the field.

[0102] For real-time computing requests, assuming that Party B needs a certain ID data of Party A for joint computing, the implementation method is described as follows:

[0103] 1) Use the system public key to encrypt the data product output of participant A;

[0104] The data of participant A is Data_A_to_B. The system's public key Sys_PublicKey is used to encrypt the data to generate the ciphertext Encrypted_Data_A_to_B. The implementation process is as follows:

[0105] Encrypted_Data_A_to_B=Encrypt(Data_A_to_B,Sys_PublicKey)

[0106] 2) Encrypt the data using the ABF public key of Party B;

[0107] In order to ensure that only participant B can decrypt the data, the data is encrypted again using participant B's ABF public key ABF_PublicKey_B to generate the ciphertext Encrypted_Data_A_to_B_ABFE. The implementation process is as follows:

[0108] Encrypted_Data_A_to_B_ABFE=Encrypt(Encrypted_Data_A_to_B,ABF_PublicKey_B)

[0109] 3) Send the data to Party B and decrypt the data using Party B’s ABF private key;

[0110] Send encrypted data Encrypted_Data_A_to_B_ABFE to Party B

[0111] Use the ABF private key ABF_PrivateKey_B of participant B to decrypt the received Encrypted_Data_A_to_B_ABFE to obtain the intermediate data Encrypted_Data_A_to_B. The implementation process is as follows:

[0112] Encrypted_Data_A_to_B=Decrypt(Encrypted_Data_A_to_B_ABFE,ABF_PrivateKey_B)

[0113] 4) Use the system public key to encrypt the data of participant B;

[0114] The data of participant B is Data_B. The system's public key Sys_PublicKey is used to encrypt the data to generate the ciphertext Encrypted_Data_B. The implementation process is as follows:

[0115] Encrypted_Data_B=Encrypt(Data_B,Sys_PublicKey)

[0116] 5) Use the ABF private key of Party B to encrypt the data of both parties;

[0117] Use the ABF private key ABF_PrivateKey_B of participant B to encrypt the data of both parties twice. The implementation process is as follows:

[0118] Encrypted_Data_B_ABFE=Encrypt(Encrypted_Data_B,ABF_PrivateKey_B)

[0119] Encrypted_Data_A_to_B_ABFE=Encrypt(Encrypted_Data_A_to_B,ABF_PrivateKey_B)

[0120] 6) The encrypted data is sent to the trusted computing platform, which decrypts the data;

[0121] Use the ABF public key ABF_PublicKey_B of participant B to decrypt the data to obtain the intermediate data, as follows:

[0122] Encrypted_Data_A_to_B=Decrypt(Encrypted_Data_A_to_B_ABFE,ABF_PublicKey_B)

[0123] Encrypted_Data_B=Decrypt(Encrypted_Data_B_ABFE,ABF_PublicKey_B)

[0124] Use the system's private key Sys_PrivateKey to decrypt the intermediate data and obtain the original data. The implementation process is as follows:

[0125] Data_A_to_B=Decrypt(Encrypted_Data_A_to_B,Sys_PrivateKey)

[0126] Data_B=Decrypt(Encrypted_Data_B,Sys_PrivateKey)

[0127] 7) Trusted computing platform completes joint computing;

[0128] Perform joint calculations on the calculation rules in the joint contract to obtain the calculation results. Note that the "original data" used in the calculation here is the data that has been homomorphically encrypted or differentially private based on the joint contract.

[0129] 8) The calculation result is encrypted with the system private key and the public key of participant B;

[0130] The Result is encrypted using the system private key, and then re-encrypted using the public key of participant B. The specific process is as follows:

[0131] Encrypted_Result=Encrypt(Result,Sys_PrivateKey)

[0132] Encrypted_Result_ABFE=Encrypt(Encrypted_Result, ABF_PublicKey_B)

[0133] 9) Send the result to participant B for decoding;

[0134] The participants receive the encrypted data, perform double decryption, and obtain the final result. The specific process is as follows:

[0135] Encrypted_Result=Decrypt(Encrypted_Result_ABFE,ABF_PrivateKey_B)

[0136] Result=Decrypt(Encrypted_Result,Sys_PublicKey)

[0137] The above process is a specific method for cross-domain privacy data joint calculation proposed by the present invention. Optionally, a secure multi-party computing protocol can be used for ciphertext joint calculation.

[0138] A certain degree of data privacy and protection can be achieved by only using the system public key to encrypt the data of both parties once and then sending it to the trusted computing platform and decrypting it with the system private key. However, the method proposed in the present invention focuses on the computing needs of the participant B to ensure that only B can decode the data. The ABF function is not only a simple layer of encryption at the data level, but also can further perform fine-grained encryption according to the attributes of the data as required. Therefore, the double-layer encryption method proposed in the present invention is more general and more secure.

[0139] Step 4: Release of joint computing data products based on the self-service data platform.

[0140] Federated computing is completed based on the joint domain. For data grids, the results of federated computing still need to be presented in the form of data products if they want to be circulated. Therefore, a method is needed to form each federated computing demand into a data product. The life cycle of data products is based on the self-service data platform. Therefore, the self-service data service platform needs to be functionally adapted for the joint domain, such as Figure 4 As shown in the figure, the method for publishing joint computing data products based on data grid is as follows:

[0141] 1) Cooperate with the demand side in the field to determine the data attribute requirements and calculation models;

[0142] 2) The demand and calculation model are broadcast to the demandee for confirmation;

[0143] 3) The demander develops data products in this field and registers them in the joint field;

[0144] 4) The joint field ABF verification is passed and released and runs automatically.

[0145] The cross-domain data product generation method based on data grid described in the present invention fully respects the construction guidelines of data grid, and completes the registration of joint domains based on general domains. Multiple domains develop their own data products and register virtual data products to the joint domains. The privacy and security of multi-party data are guaranteed based on a double-layer encryption mechanism and a trusted computing platform, and finally the joint data product release based on a self-service data platform is realized, providing a safe and efficient implementation method for multi-domain joint privacy computing based on data grids.

[0146] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A cross-domain data product generation method based on a data grid, characterized in that: The steps include: S1: Multi-party domain initialization, which is completed by performing normal domain registration of joint computing participants in the data grid and recording necessary ID information; S2: Build a joint computing domain, build a trusted domain that ensures that the data privacy of parties in each domain is not leaked and shared by multiple parties; S3: Perform cross-domain data joint computing based on trusted domains; S4: Joint computing data product release based on the self-service data platform, and joint computing data product release of the data grid; In step S2, the trusted domain is registered in the data grid as a domain node, providing a trusted execution environment. On this basis, the multi-party participating domains are confirmed, and a multi-party contract is constructed. The joint calculation of the multi-party private data is performed based on the contract confirmed by the multi-party signatures. The specific implementation method is as follows: S2-1. A certain field is registered as a joint field. The basic nature of the joint field is the same as that of the general field. The field member A becomes the sole person in charge of the field. S2-2, domain member A initiates an invitation to multiple domains with joint attribute data; S2-3. The invited area agrees or refuses. If it agrees, it becomes the co-leader B, C or D of the joint area; S2-4. The domain is managed by multiple participants in a collaborative way, forming a small federal governance. Any participant can propose modifications to the domain, but they will take effect only after confirmation by the confirming party. S2-5. Participants in the field can choose to withdraw from the joint field at any time; In step S3, cross-domain data joint computing is performed based on the trusted domain to ensure that multiple parties complete the joint computing without leaking private data, and cross-domain joint training based on the data grid is performed. The specific steps are: S3-1, Joint domain automatically generates asymmetric keys for the system, including Public key Sys_PublicKey and private key Sys_PrivateKey; S3-2, the trusted computing environment is started, and the system public key and private key are synchronized; S3-3, system attribute base function startup, including user access control and secret key generation for users; S3-4. The participant registers and logs in, and generates the corresponding operation key ABF_Key through ABF verification; S3-5. Participants build their own data products in the joint field, and the output of the data products is the joint cross-domain privacy data of all parties; S3-6. The participant requests to perform real-time computing and return results through the trusted computing environment.

2. The method for generating cross-domain data products based on data grid according to claim 1, characterized in that: In step S1, the registration information of the domain registration includes domain ID, person in charge ID, engineer ID, domain division, domain name, and domain description.

3. The method for generating cross-domain data products based on data grid according to claim 2, characterized in that: The joint computing constraints are implemented through a multi-party contract with principles and rules that are jointly observed and executed by multiple parties in the joint field. The multi-party contract contains a declaration of the joint field and a declaration of the joint computing model.

4. The method for generating cross-domain data products based on data grid according to claim 3, characterized in that: For multi-party contracts, the information of the parties is automatically updated when the invited domain user agrees to enter the joint domain; When a participant withdraws, the corresponding user information is automatically deleted. The declaration part of the joint computing model is open to all participants. Each participant can add new rules. The specific execution method is as follows: 1) A participant proposes a new rule or model for joint computing in the system; 2) The system broadcasts the new rules to all participants; 3) All participants give their consent or rejection options within 24 hours of the specified time, and rejection is assumed after the expiration of the time limit; 4) If more than half of the parties agree, the rule will be added to the multi-party contract; 5) The attributes and models in a multi-party contract are models that can be jointly called by multiple participants; 6) If any party objects to the adopted model, they may choose to withdraw from the joint field.

5. The method for generating cross-domain data products based on data grid according to claim 4, characterized in that: The attribute base function mainly completes the user's access control and key generation. Access control is implemented based on the joint computing governance principle of the data grid. Access control includes registration and login. For the joint domain of the data grid, the attribute base function strategy registered by the participants includes at least the role and domain ID. When the registered user logs in to the joint domain later, ABF generates a secret key session for this login, which is used for data encryption and verification of multi-party communication.

6. The method for generating cross-domain data products based on data grid according to claim 5, characterized in that: For the data grid, the data provided by multiple participants in the joint field is in the form of registered data products. The data products provide the joint data required by multiple parties. That is to say, the data of each party is still controlled by each participant. The output of the data product is the data encrypted or feature extracted by each party according to the attribute rules in the multi-party contract. The original data will not flow out, ensuring privacy and security.

7. The method for generating cross-domain data products based on data grid according to claim 6, characterized in that: The specific method of cross-domain privacy data joint calculation is that participant B needs the ID data of participant A for joint calculation. The implementation method is: (1) The system public key encrypts the data product output of participant A; (2) Encrypt the data using the ABF public key of Party B; (3) Send the data to Party B and decrypt the data using Party B’s ABF private key; (4) Use the system public key to encrypt Party B’s own data; (5) Use the ABF private key of Party B to encrypt the data of both parties; (6) The encrypted data is sent to the trusted computing platform, which decrypts the data; (7) The trusted computing platform completes joint computing; (8) The calculation result is encrypted with the system private key and the participant B’s public key and sent to the participant B for decoding.

8. The method for generating cross-domain data products based on data grid according to claim 7, characterized in that: In step S4, the self-service data service platform is functionally adapted for the joint domain, and the results of the joint computing are still presented in the form of data products. The method for publishing the joint computing data products of the data grid is as follows: S4-1. Cooperate with the demand side in the field to determine the data attribute requirements and calculation model; S4-2, data attribute requirements and calculation models are broadcast to the requesting party for confirmation; S4-3. The demander develops data products in this field and registers them in the joint field; S4-4. The joint field ABF verification is passed and released and operates automatically.

Images