Intelligent label construction and rapid search method and system for distributed data privacy protection

Abstract

The invention belongs to the technical field of network security, and discloses an intelligent label construction and rapid search method for distributed data privacy protection. According to the method, the correlation interpolation function is constructed, the randomization of the main-sub label mapping is ensured, the scale of the number of function items is controlled, and the main-sub label mapping is segmented and stored in the two servers in an addition secret sharing form, so that the safety of label mapping storage is improved, and the balance between the safety and efficient storage is achieved. Besides, the search algorithm converts the query problem of the sub-tags into a function value solving problem, the tedious process of data comparison in the query process is avoided, the higher query speed can be obtained, and the requirement for large-scale data split storage can be met.

Description

Technical Field

[0001] The present invention belongs to the technical field of network security, and in particular relates to an intelligent tag construction and rapid search method and system for distributed data privacy protection. Background Art

[0002] Data splitting technology is an important means to solve data privacy protection and release data value in the cloud. Its core ideas are: 1) the sensitivity of personal data varies; 2) after separating personal data from personal identification information, the data is no longer sensitive. Therefore, most personal data is non-sensitive after being split and stored, and can be directly processed and circulated without the need to use encryption and other technologies to protect its privacy. Compared with security technologies such as encryption and trusted hardware, data splitting has great advantages in storage and implementation efficiency. In recent years, Luo Kailun studied the automatic verification of agent abstraction in the paper "Automated Verification of Propositional Agent Abstraction for Classical Planning via CTLK Model Checking" (AAAI 2023), which can be used for automatic verification of data splitting effects. Wang Jianxu et al. implemented the protection of split data search patterns in the paper "Research on String Pattern Matching Algorithm for Protecting Search Content" (Journal of Network and Information Security vol. 10, 2024), but the split data blocks need to establish corresponding labels to support queries and reconstruct the original data to meet the needs of sophisticated data analysis. Currently, the existing methods for constructing split data labels mainly include methods of generating labels using pseudo-random functions and storing them in encrypted (or unencrypted) form.

[0003] However, the prior art has the following problems when solving split data label queries: There is a problem of multi-user query efficiency: If encryption is used to protect labels, the data needs to be decrypted first during the query process. When multiple users issue label query requests at the same time, the efficiency of decrypting data will directly affect the user query efficiency. Additional trusted servers are required: Some solutions in the prior art rely on additional trusted servers to perform storage and query tasks to protect the privacy of data. However, the introduction of trusted servers will increase the latency and usage cost of network queries, which is not very practical for scenarios with large-scale data storage and query. Only computational security: Whether the prior art uses data encryption or pseudo-random functions, its security is constructed based on the problem of computational difficulty. When the adversary has a large amount of computing resources, the security of the label is still challenged. The above problems will limit the flexibility of data reconstruction and query, especially in scenarios where frequent queries or data operations are required.

[0004] Through the above analysis, the problems and defects of the prior art are as follows:

[0005] (1) There is a problem of multi-user query efficiency: If encryption is used to protect tags, the data must be decrypted first during the query process. When multiple users send tag query requests at the same time, the efficiency of decrypting data will directly affect the user query efficiency.

[0006] (2) Additional trusted servers are required: Some existing solutions rely on additional trusted servers to perform storage and query tasks to protect data privacy. However, the introduction of trusted servers will increase the latency and usage cost of network queries, which is not very practical for large-scale data storage and query scenarios.

[0007] (3) Only computational security: Whether the existing technology uses data encryption or pseudo-random functions, its security is based on the problem of computational difficulty. When the adversary has a large amount of computing resources, the security of the label is still challenged. The above problems will limit the flexibility of data reconstruction and query, especially in scenarios where frequent queries or data operations are required. Summary of the invention

[0008] In view of the problems existing in the prior art, the present invention provides an intelligent tag construction and fast search method for distributed data privacy protection.

[0009] The present invention is implemented as follows: an intelligent tag construction and fast search method for distributed data privacy protection includes:

[0010] Step 1, label construction:

[0011] Specify several random values ​​as the main labels and sub-labels of some split data, regard them as function input values ​​and output values, and construct an interpolation function; then solve more output values ​​according to different input values ​​of the interpolation function. If the calculated output value already exists, randomly give the output value and use the interpolation formula to add the node to the function until all label mapping relationships in the space are constructed;

[0012] Step 2, tag search:

[0013] Determine the main label as the input of the interpolation function, perform the function input calculation under secret sharing in two independent and non-colluding servers, and then merge the calculation results to finally get the value of the sub-label.

[0014] Furthermore, the number of generating functions t is specified as a constant, and the generating functions F1, F2, ..., F t It can be first constructed by a small number of nodes, and then the generating function can be continuously updated by adding nodes.

[0015] Further, solve the generating functions F1, F2, ..., Ft When selecting the label value set {I1,I2,...,I n A subset of} And m<<n,m=t·s(s> 2).

[0016] Let the set {I1→θ1,I2→θ2,…,I m →θ m}, give the main label subset {I1,I2,...,I m} assigns the corresponding random sub-index link value {θ1,θ2,...,θ m}, divide the set into t subsets, and construct F1, F2, ..., F based on the interpolation formula t The specific form of; order:

[0017] F1=Interp({I1,θ1},…,{I s ,θ s}),

[0018] F2=Interp({I s+1 ,θ s+1},…,{I 2s ,θ 2s}), ...

[0020] F t =Interp({I (t-1)s+1 ,θ (t-1)s+1},…,{I m ,θ m});

[0021] For any primary index I x ∈{I1,I2,...,I n} / {I1,I2,...,I m}, calculate F1(I x )=θ x , if θ x ∈{θ1,θ2,...,θ s}, we can directly use θ x As I x The index link value, otherwise calculate the remaining function value F2(I x ),F3(I x ),...,F t (I x ), until a suitable θ is constructed x Until; if after traversing all functions, still can not get the satisfying θ x , then specify a random value For I xThe sub-index connection value is used based on the Newton interpolation formula to convert the node (I x ,θ x ) added to F1,F2,...,F t Finally, repeat the above steps until the corresponding θ is found for all main indices.

[0022] Furthermore, the idea of ​​function secret sharing is used to split the generated function F into two sub-functions A1(F) and A2(F), namely Share(F)→(A1(F), A2(F)), satisfying F=A1(F)+A2(F); the sub-functions A1(F) and A2(F) are stored separately by two independent and non-colluding servers; when the main label I is given as the function input, the server calculates ω1=A1(F)(I) and ω2=A2(F)(I) respectively and sends the results to a trusted third party, which then calculates ω1+ω2=A1(F)(I)+A2(F)(I)=F(I)=θ; finally, it is split into sub-labels of data.

[0023] Another object of the present invention is to provide an intelligent tag construction and fast search system for distributed data privacy protection, comprising:

[0024] The label construction module is used to specify several random values ​​as the main labels and sub-labels of some split data, regard them as function input values ​​and output values, and construct an interpolation function; then solve more output values ​​according to different input values ​​of the interpolation function. If the calculated output value already exists, the output value is randomly given, and the interpolation formula is used to add the node to the function until all label mapping relationships in the space are constructed;

[0025] The tag search module is used to determine the main tag as the input of the interpolation function, perform the function input calculation under secret sharing in two independent and non-colluding servers, and then merge the calculation results to finally obtain the value of the sub-tag.

[0026] Another object of the present invention is to provide a computer device, characterized in that the computer device includes a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the intelligent label construction and rapid search method for distributed data privacy protection.

[0027] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to execute the steps of the intelligent tag construction and rapid search method for distributed data privacy protection.

[0028] Another object of the present invention is to provide an information data processing terminal, which is used to implement the intelligent label construction and rapid search system for distributed data privacy protection.

[0029] In combination with the above technical solutions and the technical problems solved, please analyze the advantages and positive effects of the technical solutions to be protected by the present invention from the following aspects:

[0030] First, the technical solution of the present invention solves the following key problems existing in the prior art in industrial applications and has achieved significant technical progress:

[0031] Traditional data label construction and search methods usually rely on linear scanning or simple hash mapping, which cannot efficiently handle the label construction and search requirements of large-scale data sets, especially in distributed or privacy-protected scenarios, where the performance is particularly insufficient.

[0032] In existing tag search methods, scenarios involving sensitive data often face the risk of data leakage. Even in a distributed environment, multi-party collaborative computing is prone to collusion or other security vulnerabilities, which may lead to data privacy and security not being guaranteed.

[0033] In many practical applications, the distribution and mapping of data labels are often dynamic and complex. Existing technologies are difficult to flexibly adapt to the needs of data sets of different sizes and forms, and cannot effectively deal with conflicts and overlaps in label mapping.

[0034] The present invention introduces a label construction method based on interpolation function, regards the main label and sub-label of the partially split data as the input value and output value of the function, thereby generating an interpolation function, and using the interpolation formula to expand the label mapping relationship. This method not only improves the efficiency of label construction, but also can quickly locate sub-labels through function calculation during the search process, avoiding the inefficiency of traditional linear search.

[0035] The present invention uses the idea of ​​secret function sharing to split the generated function into two sub-functions, which are stored and calculated by two independent and non-collusive servers. This design ensures that the privacy and security of data can be effectively protected even if a single server is attacked during tag search. In addition, through secret sharing calculation and the collaboration of a trusted third party, the final result can be calculated without exposing the original data.

[0036] In the process of label construction, the present invention adopts dynamic interpolation and node addition to flexibly handle the allocation and mapping of label values. Even if there is a conflict or overlap in label values, the system can ensure that all main indexes have unique corresponding sub-labels by supplementing random values ​​and re-interpolating. This method shows extremely high adaptability when dealing with complex data structures.

[0037] ###Significant technological advances achieved:

[0038] 1. **Improved data processing efficiency: **By introducing the construction of interpolation functions and the calculation method of secret sharing, the present invention significantly improves the efficiency in label construction and search of large-scale data sets, especially in distributed environments.

[0039] 2.**Improved the security and privacy protection capabilities of the system: **Through secret sharing and distributed computing technology, the present invention effectively prevents the risk of data leakage and ensures data privacy in an untrusted environment, and is suitable for industrial applications with high security requirements.

[0040] 3.**Enhanced the flexibility and scalability of the system: **The label construction method of the present invention is highly flexible and can dynamically adjust the label distribution and mapping strategies according to different application scenarios, adapt to data processing requirements of various scales and complexities, and promote the development of data label management technology.

[0041] The present invention has achieved remarkable technical progress in the field of data tag construction and search technology, especially in improving efficiency, enhancing security and flexibility, and has provided strong technical support for industrial applications.

[0042] Second, traditional data labeling and search methods often have data leakage and privacy risks when processing sensitive data. Existing technologies cannot guarantee complete privacy and security of data during transmission and storage.

[0043] The present invention uses an intelligent tag construction and fast search method for distributed data privacy protection, and adopts secret sharing and interpolation functions to split data tags and store them in a distributed manner on independent and non-collusive servers. In this way, even if one of the servers is compromised, the integrity and privacy of the data can still be effectively protected. This method significantly improves data security and reduces the risk of data leakage.

[0044] 2. Cloud computing

[0045] In a cloud computing environment, data synchronization and management between different servers is a complex task. In the prior art, the distributed management and search efficiency of tag data is low and there are data consistency issues.

[0046] The method of the present invention allows for distributed management and synchronous search of tag data between multiple servers without sharing the original data. By constructing and updating the interpolation function, efficient search and management of data tags are achieved. At the same time, by using secret sharing technology, each server can independently process part of the data tags, improving the overall performance and reliability of the system. This solution solves the data consistency and synchronization problems in the cloud computing environment and significantly improves the data management efficiency of the distributed system.

[0047] 3. Data label construction and search efficiency

[0048] Traditional data label construction and search methods are often inefficient when processing large amounts of data and are unable to quickly respond to and process the construction and search needs of massive data labels.

[0049] The present invention realizes efficient construction of data labels through a label construction method based on interpolation functions and random values. The search process of data labels is optimized by using Newton interpolation formula and secret sharing technology. Under the premise of ensuring security, this method significantly improves the efficiency of data label construction and search, can quickly process and respond to the needs of massive data, and meets the requirements of modern industry for efficient data processing.

[0050] 4. Extensive application scenarios

[0051] Existing data labeling and search methods are often limited to specific application scenarios, lack versatility and flexibility, and are difficult to adapt to diverse industry needs.

[0052] The method of the present invention has a wide range of application scenarios, not only in the fields of data security and cloud computing, but also in other scenarios that require efficient data tag management and search, such as big data analysis, the Internet of Things, financial data processing, etc. Its versatility and flexibility enable it to adapt to the diverse needs of different industries, significantly improving the application value of the technology.

[0053] The present invention solves many technical problems in the prior art through innovative technical solutions, achieves significant technical progress in data security, cloud computing, data processing efficiency and wide application, and has important industrial application value.

[0054] Third, the technical solution of the present invention fills the technical gap in the industry at home and abroad: the present invention considers for the first time the problem of how to prevent the illegal reconstruction of split data under the premise of collusion of multiple servers, creates a data label management mechanism based on fitting functions, and develops original key technologies such as label construction, addition, and deletion, filling the technical gap at home and abroad.

[0055] The innovative achievements of this project are as follows: 1. Utilizing the distributed concept, the one-way function construction that supports tag management is transformed into a common high-order function construction problem, thus avoiding the problem of excessive function complexity; 2. By limiting the degree of function randomization, the function scale is controlled, which significantly reduces the space required for function storage and makes the system more practical.

[0056] The technical solution of the present invention overcomes technical bias: In this field, researchers usually assume that split data is distributed and stored by multiple non-colluding servers, and simply assign the same data label to the split data for data reconstruction. Researchers in this field have no motivation to set up a main label-sub-label management mechanism to prevent colluding servers from illegally reconstructing data. Therefore, not considering server collusion is a technical bias in this field. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a flow chart of an intelligent tag construction and rapid search method for distributed data privacy protection provided by an embodiment of the present invention.

[0058] Figure 2 It is a structural block diagram of an intelligent tag construction and fast search system for distributed data privacy protection provided by an embodiment of the present invention.

[0059] Figure 3 It is a detailed flowchart of the split data label construction and search method provided by an embodiment of the present invention.

[0060] Figure 4 This is a specific implementation diagram of split data label search provided by an embodiment of the present invention;

[0061] Figure 5 is the relationship between the sub-index generation time and the number of generation functions provided by the embodiment of the present invention;

[0062] Figure 6 This is the relationship between the sub-index generation time and the number of generated function items provided by the embodiment of the present invention. DETAILED DESCRIPTION

[0063] 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 the 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.

[0064] Example 1: Data security

[0065] In financial institutions, medical institutions and other fields, the storage and transmission of sensitive data requires a high degree of security, and traditional data labeling and search methods have the risk of data leakage.

[0066] A financial institution needs to manage and search a large amount of customer data in a tagged manner in order to conduct precision marketing and risk control. Using the method of the present invention, the financial institution can tag customer data and store the tagged data in a distributed manner on two independent and non-collusive servers through an intelligent tag construction and fast search method for distributed data privacy protection.

[0067] 1. Label construction: Specify several random values ​​as the main labels and sub-labels of some customer data, construct an interpolation function, and solve more output values ​​based on the interpolation function to complete the construction of the label mapping relationship.

[0068] 2. Tag search: When querying customer data, determine the input value of the main tag, perform function input calculations under secret sharing in two independent servers, and then merge the calculation results to obtain the value of the sub-tag, thereby achieving safe and efficient tag data search.

[0069] This method ensures the privacy and security of customer data during storage and query, significantly reduces the risk of data leakage, and improves the efficiency of data tag management and search.

[0070] Example 2: Cloud computing

[0071] In a cloud computing environment, data synchronization and management between multiple servers is a complex task, especially when dealing with large-scale data. Traditional methods are inefficient and have data consistency issues.

[0072] A large cloud computing platform needs to perform tag management and search on a large amount of data uploaded by users. The method of the present invention can realize efficient data tag management and search among multiple servers of the cloud computing platform.

[0073] 1. Label construction: By specifying random values ​​as the main label and sub-label, constructing the interpolation function, and adding nodes through the interpolation formula, the label mapping relationship is constructed.

[0074] 2. Tag search: When a user queries, the main tag is determined as the input of the interpolation function, and the function input calculation under secret sharing is performed in two independent servers respectively. The calculation results are merged to obtain the value of the sub-tag, realizing distributed tag data search.

[0075] This method realizes distributed management and synchronous search of data tags without sharing the original data, solves the data consistency and synchronization problems in the cloud computing environment, and improves the overall performance and reliability of the system.

[0076] Through the above two specific embodiments, the method of the present invention demonstrates its wide application value and significant technological progress in the fields of data security and cloud computing.

[0077] like Figure 1 As shown, an intelligent tag construction and fast search method for distributed data privacy protection provided by an embodiment of the present invention includes the following steps:

[0078] S1, label construction:

[0079] Specify several random values ​​as the main labels and sub-labels of some split data, regard them as function input values ​​and output values, and construct an interpolation function; then solve more output values ​​according to different input values ​​of the interpolation function. If the calculated output value already exists, randomly give the output value and use the interpolation formula to add the node to the function until all label mapping relationships in the space are constructed;

[0080] S2, tag search:

[0081] Determine the main label as the input of the interpolation function, perform the function input calculation under secret sharing in two independent and non-colluding servers, and then merge the calculation results to finally get the value of the sub-label.

[0082] The embodiment of the present invention provides that the number of generating functions t is a fixed value, and the generating functions F1, F2, ..., F t It can be first constructed by a small number of nodes, and then the generating function can be continuously updated by adding nodes.

[0083] The solution generation functions F1, F2, ..., F provided in the embodiment of the present invention t When selecting the label value set {I1,I2,...,I n A subset of} And m<<n,m=t·s(s> 2).

[0084] Let the set {I1→θ1,I2→θ2,…,I m →θ m}, give the main label subset {I1,I2,...,I m} assigns the corresponding random sub-index link value {θ1,θ2,...,θ m}, divide the set into t subsets, and construct F1, F2, ..., F based on the interpolation formula t The specific form of; order:

[0085] F1=Interp({I1,θ1},…,{I s ,θ s}),

[0086] F2=Interp({I s+1 ,θ s+1},…,{I 2s ,θ 2s}),...

[0087] F t =Interp({I (t-1)s+1 ,θ (t-1)s+1},…,{I m ,θ m});

[0088] For any primary index I x ∈{I1,I2,...,I n}\{I1,I2,...,I m}, calculate F1(I x )=θ x , if θ x ∈{θ1,θ2,...,θ s}, we can directly use θ x As I x The index link value, otherwise calculate the remaining function value F2(I x ),F3(I x ),...,F t (I x ), until a suitable θ is constructed x Until; if after traversing all functions, still can not get the satisfying θ x , then specify a random value For I x The sub-index connection value is used based on the Newton interpolation formula to convert the node (I x ,θ x ) added to F1,F2,...,F t Finally, repeat the above steps until the corresponding θ is found for all main indices.

[0089] The embodiment of the present invention provides a method for splitting a generated function F into two sub-functions A1(F) and A2(F) using the idea of ​​function secret sharing, namely Share(F)→(A1(F), A2(F)), satisfying F=A1(F)+A2(F); the sub-functions A1(F) and A2(F) are respectively stored by two independent and non-colluding servers; when a main label I is given as a function input, the server calculates ω1=A1(F)(I) and ω2=A2(F)(I) respectively and sends the results to a trusted third party, which then calculates ω1+ω2=A1(F)(I)+A2(F)(I)=F(I)=θ; and finally splits the function into sub-labels of data.

[0090] like Figure 2 As shown, an intelligent tag construction and fast search system for distributed data privacy protection provided by an embodiment of the present invention includes:

[0091] The label construction module is used to specify several random values ​​as the main labels and sub-labels of some split data, regard them as function input values ​​and output values, and construct an interpolation function; then solve more output values ​​according to different input values ​​of the interpolation function. If the calculated output value already exists, the output value is randomly given, and the interpolation formula is used to add the node to the function until all label mapping relationships in the space are constructed;

[0092] The tag search module is used to determine the main tag as the input of the interpolation function, perform the function input calculation under secret sharing in two independent and non-colluding servers, and then merge the calculation results to finally obtain the value of the sub-tag.

[0093] Another object of the present invention is to provide a computer device, characterized in that the computer device includes a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the intelligent label construction and rapid search method for distributed data privacy protection.

[0094] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to execute the steps of the intelligent tag construction and rapid search method for distributed data privacy protection.

[0095] Another object of the present invention is to provide an information data processing terminal, which is used to implement the intelligent label construction and rapid search system for distributed data privacy protection.

[0096] The present invention is specifically implemented:

[0097] The label construction step specifically includes the following contents:

[0098] Initialize the personal data primary index set {I1,I2,...,I n} is the input of the generating function, select a subset And m<<n,m=t·s(s> 2), specify random value M: = {θ1, θ2, ..., θ m} constitutes a mapping set {I1→θ1,I2→θ2,…,I m →θ m}; By random value θ i It is unique and represents the link value of the sub-label, that is: θ i =I i,1 ||I i,2 ||...||I i,k ,i=1,2,...,n, where “||” represents the juxtaposition and concatenation of symbols.

[0099] The mapping {I1→θ1,I2→θ2,…,I m →θ m} is divided into t subsets, each subset contains s elements, and the parameter I j and θ j Both are considered as finite fields elements in , and satisfy p>θ i . Then the set The mapping of each subset is regarded as the s inputs and outputs of a function f, satisfying f(I j )=θ j ,(j=k1,k2,...,k s ). When the input and output pairs are known, the specific form of the function can be calculated using the Newton interpolation formula:

[0100] F1=Interp({I1,θ1},…,{I s ,θ s}),

[0101] F2=Interp({I s+1 ,θ s+1},…,{I 2s ,θ 2s}),...,

[0102] F t =Interp({I (t-1)s+1 ,θ (t-1)s+1},…,{I m ,θ m});

[0103] For any primary index I x ∈{I1,I2,...,I n} / {I1,I2,...,I m}, calculate F1(I x )=θ x , if θ x ∈{θ1,θ2,...,θ s}, we can directly use θ x As I x The index link value, otherwise calculate the remaining function value F2(I x ),F3(I x ),...,F t (I x ), until a suitable θ is constructed x If after traversing all functions, no θ that meets the conditions can be obtained, x , then specify a random value For Ix The sub-label connection value is used based on the Newton interpolation formula to convert the node (I x ,θ x ) added to F1,F2,...,F t Finally, repeat the above steps until the corresponding θ is found for all main labels.

[0104] Use additive secret sharing to split and store functions F1, F2, ..., F t for {A1(F1),A2(F1)},{A1(F2),A2(F2)},...,{A1(F t ),A2(F t )}, store the function sequence in two independent servers that do not collude: {A1(F1),A1(F2),...,A1(F t )} or {A2(F1),A2(F2),...,A2(F t )}.

[0105] The tag search step specifically includes the following contents:

[0106] Given the input main label I j , is sent to two independent label servers c,(where c=1,2).

[0107] Tag server c uses the function {A c (F1),A c (F2),...,A c (Ft)} is calculated as I j All the outputs of the input are recorded as {A c (F1)(I j ),A c (F1)(I j ),...,A c (F t )(I j )} and returns it to the trusted server.

[0108] Trusted server, selects a specific function result A1(Fi)(I j ),A2(Fi)(I j ), and combine to calculate F i (I j ), and finally separate the sub-tags {I j,1 ,I j,2 ,...,I j,k}.

[0109] The present invention specifically relates to a split data label construction and search algorithm based on interpolation function, which specifically includes two parts: label construction and label association search. Label construction part: Use the interpolation formula to construct related functions by randomly specifying the mapping of main-sub-labels. However, in order to control the scale of the function, some nodes are used to construct the interpolation function, and then more output values ​​are solved according to different input label values ​​of the interpolation function. If the calculated output value already exists, the output value is randomly given, and the node is added to the function using the interpolation formula until all label mapping relationships in the space are constructed; label association search part: Determine the main label as the input of the interpolation function, perform the function input calculation under secret sharing in two independent and non-colluding servers, and then merge the calculation results to finally obtain the value of the sub-label.

[0110] like Figure 3 , 4 As shown, specifically including the following:

[0111] The main data tags include: {I1,I2,...,I n};

[0112] The data sub-tags include: 1,1 ||I 1,2 ||...||I 1,k},...,{I i,1 ||I i,2 ||...||I i,k},...,{I n,1 ||I n,2 ||...||I n,k}}, the connection value of a single piece of data corresponding to the sub-label is recorded as: θ i =I i,1 ||I i,2 ||...||I i,k ,i=1,2,...,n。

[0113] First, make the following settings: Set the parameters t and m to fixed values, and satisfy m<<n,m=t·s(s> 2); Interp(*) is the Newton interpolation formula; enter the main label {I1,I2,...,I n}, parameters t, m, output generation functions F1, F2, ..., F t and its additive secret sharing form.

[0114] S101: Node interpolation method, constructing a generating function by inputting a main label data set;

[0115] Specifically include: Initialization: Select a subset of personal data main labels Specify random value M: = {θ1,θ2,...,θ m} constitutes a mapping set {I1→θ1,I2→θ2,…,I m →θ m}, divide the above mapping into t subsets, each subset contains s elements, and calculate the initial form of t functions based on Newton interpolation formula;

[0116] For any primary index I x ∈{I1,I2,...,I n} / {I1,I2,...,I m}, calculate F1(I x )=θ x ,like This shows that if θ x is different from the existing θ value corresponding to F1, then it is considered that θ x It can be used as the new sub-label value, otherwise the remaining function value F2(I x ),F3(I x ),...,F t (I x );

[0117] If after calculating all t function values, the θ that satisfies the conditions cannot be obtained x , choose any function F r , specify a random value For I x The corresponding output and calculate F r :=Interp(Fr,{I x ,θ x}},M:=M∪θ x ;

[0118] Repeat the above steps to continue calculating the output values ​​corresponding to the remaining main labels until the functions F1, F2, ..., F of all labels in the main label set are completed. t The structure of

[0119] Additive secret sharing F1,F2,...,F t And store the shared function fragments {A1(F1), A1(F2), ..., A1(F t )} and {A2(F1),A2(F2),...,A2(F t )}In two separate servers

[0120] Determine the main label as the input of the interpolation function, perform the function input calculation under secret sharing in two independent and non-colluding servers, and finally merge the calculation results to get the value of the sub-label and return the result to the user.

[0121] Specifically include:

[0122] Determine the primary tag value I of the subtag to be queried x , and sent to two independent label servers c=1,2;

[0123] The server uses the stored function sequence {A c (F1),A c (F2),...,A c (F t )}, calculate all the outputs with Ij as input, and record the result as {A c (F1)(I x ),A c (F2)(I x ),...,A c (F t )(I x )} and returns it to the trusted server;

[0124] Trusted server, selects a specific function result A1(F i )(I x ),A2(F i )(I x ), calculate A1(F i )(I x )+A2(F i )(I x )=F i (I x ), and separate the sub-tags {I j,1 ,I j,2 ,...,I j,k} and returns it to the user.

[0125] The present invention can be directly applied to a data distributed storage system and used to identify single pieces of data stored in different servers, which is beneficial to efficient retrieval and recovery of data.

[0126] In order to verify the technical effect of the present invention, an experimental environment was built on the Windows platform. The workstation CPU is Inteli9-10900K CPU 3.70GHz and the memory is 128GB. The generation function construction and sub-index generation are implemented based on Python 3.11.5. The test data set uses the H-1B VisaPetitions 2011-2016 data set, which contains 10 million data. The primary index value is 10 million, and the primary index I is taken as x Length 32 bits, θ x (The sub-index is through the xThe word length of the modulus p is 33 bits, and the number of functions t is 100. First, a subset of 100,000 primary index values ​​is selected, and the corresponding θ values ​​are randomly given to construct 100 initial generating functions. The θ values ​​of other primary indexes are calculated using the label construction method, and finally each generating function can be controlled within 1200 times.

[0127] Based on the experiment of generating function construction, we conducted an experiment on sub-index generation. For a given input main index I, the time to calculate the sub-index is affected by the number of generated functions t and the number of function items r. The case where the server uses a parallel method to perform function calculations is not considered here. Under the premise that the number of function items is fixed at 4000 and the number of split blocks is 2, when the number of functions t increases from 100 to 10000, the sub-index calculation time increases from 0.22 seconds to 20.99 seconds. When the number of generated functions t=100 and k=2, the number of function items increases from 4000 to 10000, and the sub-index calculation time increases from 0.08 seconds to 0.21 seconds. It can be seen that this method is more efficient in calculating sub-indexes under the premise of a low number of function items. The experimental results further illustrate the necessity of controlling the number of function items. Therefore, the label management method proposed in the patent of this invention has strong practicality.

[0128] Figure 5 is the relationship between the sub-index generation time and the number of generation functions provided by the embodiment of the present invention;

[0129] Figure 6 This is the relationship between the sub-index generation time and the number of generated function items provided by the embodiment of the present invention.

[0130] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. It can be understood by a person of ordinary skill in the art that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium. The device and its modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0131] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. An intelligent tag construction and fast search method for distributed data privacy protection, characterized in that: The following steps are involved: Step 1, label construction: specify several random values ​​as the main label and sub-label of the partially split data, regard them as the input value and output value of the function, and construct the interpolation function; then solve more output values ​​according to different input values ​​of the interpolation function. If the calculated output value already exists, randomly give the output value and form a function node, and use the interpolation formula to add the node to the function until all label mapping relationships in the space are constructed; Step 2, tag search: Determine the main tag as the input of the interpolation function, perform the function input calculation under secret sharing in two independent and non-colluding servers, and then merge the calculation results to finally get the value of the sub-tag.

2. The intelligent tag construction and rapid search method for distributed data privacy protection as claimed in claim 1, characterized in that: The number of generating functions t is specified as a constant, and the generating functions F1, F2, ..., F t It can be first constructed by a small number of nodes, and then the generating function can be continuously updated by adding nodes.

3. The intelligent tag construction and rapid search method for distributed data privacy protection as claimed in claim 1, characterized in that: Solve the generating functions F1, F2, ..., F t When selecting the label value set {I1,I2,...,I n A subset of} And m<<n,m=t·s(s> 2). Let the set {I1→θ1,I2→θ2,…,I m →θ m }, give the main label subset {I1,I2,...,I m } assigns the corresponding random sub-index link value {θ1,θ2,...,θ m }, divide the set into t subsets, and construct F1, F2, ..., F based on the interpolation formula t the specific form of make: F1=Interp({I1,θ1},…,{I s ,i s }), F2=Interp({I s+1 ,θ s+1 },…,{I 2s ,θ 2s }), ... F t =Interp({I (t-1)s+1 ,i (t-1)s+1 },…,{I m ,i m }); For any primary index I x ∈{I1,I2,...,I n } / {I1,I2,...,I m }, calculate F1(I x )=θ x , if θ x ∈{θ1,θ2,...,θ s }, we can directly use θ x As I x The index link value, otherwise calculate the remaining function value F2(I x ),F3(I x ),...,F t (I x ), until a suitable θ is constructed x Until; if after traversing all functions, still can not get the satisfying θ x , then specify a random value For I x The sub-index connection value is used based on the Newton interpolation formula to convert the node (I x ,θ x ) added to F1,F2,...,F t Finally, repeat the above steps until the corresponding θ is found for all main indices.

4. The intelligent tag construction and rapid search method for distributed data privacy protection as claimed in claim 1, characterized in that: The idea of ​​function secret sharing is used to split the generated function F into two sub-functions A1(F) and A2(F), namely Share(F)→(A1(F), A2(F)), satisfying F=A1(F)+A2(F); the sub-functions A1(F) and A2(F) are stored separately by two independent and non-colluding servers; when the main label I is given as the function input, the server calculates ω1=A1(F)(I) and ω2=A2(F)(I) respectively and sends the results to a trusted third party, which then calculates ω1+ω2=A1(F)(I)+A2(F)(I)=F(I)=θ; finally, the data is split into sub-labels.

5. An intelligent tag construction and rapid search system for distributed data privacy protection implementing the intelligent tag construction and rapid search method for distributed data privacy protection as described in any one of claims 1 to 4, characterized in that: The intelligent tag construction and fast search system for distributed data privacy protection includes: The label construction module is used to specify several random values ​​as the main labels and sub-labels of some split data, regard them as function input values ​​and output values, and construct an interpolation function; then solve more output values ​​according to different input values ​​of the interpolation function. If the calculated output value already exists, the output value is randomly given, and the interpolation formula is used to add the node to the function until all label mapping relationships in the space are constructed; The tag search module is used to determine the main tag as the input of the interpolation function, perform the function input calculation under secret sharing in two independent and non-colluding servers, and then merge the calculation results to finally obtain the value of the sub-tag.

6. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the intelligent label construction and rapid search method for distributed data privacy protection as described in any one of claims 1-4.

7. A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor executes the steps of the intelligent tag construction and rapid search method for distributed data privacy protection as described in any one of claims 1 to 4.

8. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the intelligent label construction and rapid search system for distributed data privacy protection as described in claim 5.

9. A method for protecting data security in a distributed system based on the method of claim 1, comprising: By specifying random numerical values ​​as the main label and sub-label of the data, and constructing an interpolation function to establish a label mapping relationship; when querying data, determining the main label as the input of the interpolation function, and performing the function input calculation under secret sharing in at least two independent and non-collusive servers respectively; merging the calculation results from different servers to obtain the value of the sub-label, thereby realizing secure search of data, wherein the method ensures data privacy and security during data storage and query.

10. A method for implementing data tag management and search in a cloud computing environment based on the method of claim 1, the method comprising: The label mapping relationship is established by constructing an interpolation function and using random values ​​as the main label and sub-label of the data; when a user queries, the main label is used as the input of the interpolation function, and the function input calculation is performed under secret sharing in at least two independent servers; The calculation results from different servers are merged to obtain the value of the sub-tag, thereby realizing distributed data tag management and search in a cloud computing environment, improving the efficiency of data management and data consistency.

Images