A network data processing method, device and equipment

Abstract

The application provides a network data processing method, device and equipment, belongs to the technical field of network security compliance data processing, and solves the problem of leakage risk of a large amount of data in the transmission and use process when cross-organization shared data. The method comprises the following steps: receiving target data to be transmitted; classifying the target data to be transmitted to obtain at least one type of first data; respectively encrypting the at least one type of first data to obtain at least one second data; performing homomorphic encryption processing on the at least one second data to obtain third data; performing differential privacy processing on the third data to obtain fourth data; and transmitting the fourth data. The scheme enhances the security and privacy of data in the transmission, processing and storage process, can meet the requirements of network security compliance, realizes fine-grained access control, significantly improves the data processing efficiency by reducing the data decryption scene, and realizes efficient processing and sharing of data.

Description

A network data processing method, apparatus and equipment Technical Field

[0001] This invention relates to the field of data processing technology for network security compliance, and in particular to a network data processing method, apparatus, and device. Background Technology

[0002] With the rapid development of network technology, data security has become a core issue in the field of network compliance. Currently, most data security protection methods rely on traditional encryption technologies and access control mechanisms. However, when processing large amounts of sensitive data, these methods often suffer from inefficiency and insufficient security. While current encryption technologies can provide a certain level of data protection, decryption operations during data sharing and computation risk exposing sensitive information; access control mechanisms struggle to provide fine-grained control over data usage and are easily bypassed in certain scenarios; routine encryption and decryption can lead to inefficient data processing, especially when handling large-scale data, where encryption and decryption operations become performance bottlenecks. In particular, ensuring data privacy during transmission and use when data needs to be shared across organizations is a pressing technical challenge that needs to be addressed. Summary of the Invention

[0003] This invention provides a network data processing method, apparatus, and device, which solves the problem of leakage risk during the transmission and user access of large amounts of network data when sharing data across organizations.

[0004] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0005] This invention provides a network data processing method, including:

[0006] Receive target data to be transmitted;

[0007] The target data to be transmitted is classified to obtain at least one type of first data;

[0008] Encrypt each of the at least one type of first data to obtain at least one type of second data;

[0009] The at least one second data is homomorphically encrypted to obtain the third data;

[0010] Differential privacy processing is performed on the third data to obtain the fourth data;

[0011] Transmit the fourth data.

[0012] Optionally, the target data to be transmitted is classified to obtain at least one type of first data, including:

[0013] Obtain category labels, which include data sensitivity information and sensitivity level;

[0014] Based on the classification labels, the transmitted target data is classified to obtain at least one type of first data.

[0015] Optionally, the at least one type of first data is encrypted to obtain at least one type of second data, including:

[0016] Determine the encryption algorithm for the target type corresponding to each of the at least one type of first data;

[0017] The first data is encrypted using the encryption algorithm of the target type to obtain at least one second data.

[0018] Optionally, homomorphic encryption is performed on the at least one second data to obtain third data, including:

[0019] Generate a public-private key pair;

[0020] Using the public key from a public-private key pair, encrypt the at least one second data to generate at least one encrypted second data ciphertext;

[0021] Perform at least one encryption operation on the at least one encrypted second data ciphertext to obtain third data;

[0022] The private key is securely transmitted either through offline physical transmission or through dynamic negotiation between the two parties.

[0023] Optionally, differential privacy processing is performed on the third data to obtain fourth data, including:

[0024] Set privacy budget parameters;

[0025] according to Determine the scale parameters of the noise;

[0026] The third data is subjected to noise addition based on the noise scale parameter to obtain the fourth data;

[0027] in, For privacy budget parameters, For the sensitivity of the query function, This is the scale parameter for noise.

[0028] Optionally, the method further includes:

[0029] Receive user data access requests;

[0030] Based on the data access request, the user's permissions are authenticated, and the authentication result is obtained;

[0031] When the authentication result indicates that the authentication is successful, the stored fourth data is operated on according to the preset access rules.

[0032] Optionally, the method further includes:

[0033] Obtain the user's access operation records and record the access operation records in the log;

[0034] The logs were audited to obtain the audit results;

[0035] Output the audit results.

[0036] This invention also provides a network data processing apparatus, comprising:

[0037] The receiving module is used to receive the target data to be transmitted;

[0038] The processing module is configured to classify the target data to be transmitted to obtain at least one type of first data; encrypt the at least one type of first data to obtain at least one type of second data; perform homomorphic encryption on the at least one type of second data to obtain third data; and perform differential privacy processing on the third data to obtain fourth data.

[0039] A transmission module is used to transmit the fourth data.

[0040] This invention also provides a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when run by the processor, executes the above-described method.

[0041] This invention also provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described method.

[0042] The technical solution of the present invention has at least the following effects:

[0043] The above-described solution of the present invention involves receiving target data to be transmitted; classifying the target data to obtain at least one type of first data; encrypting each of the at least one type of first data to obtain at least one type of second data; performing homomorphic encryption on the at least one type of second data to obtain third data; performing differential privacy processing on the third data to obtain fourth data; and transmitting the fourth data. This enhances the security and privacy of data during transmission, processing, and storage, meets network security compliance requirements, achieves fine-grained access control, and significantly improves data processing efficiency by reducing data decryption scenarios, thus enabling efficient data processing and sharing. Attached Figure Description

[0044] Figure 1 is a flowchart of the network data processing method provided in an embodiment of the present invention;

[0045] Figure 2 is a structural diagram of the network data processing device provided in an embodiment of the present invention;

[0046] Figure 3 is a schematic diagram of the structure of the computing device provided in an embodiment of the present invention. Detailed Implementation

[0047] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0048] As shown in Figure 1, an embodiment of the present invention proposes a network data processing method, including:

[0049] Step 11, Receive the target data to be transmitted; the target data here can be the target data to be transmitted generated by various data sources in the field of network security compliance. The target data has high requirements for data security during the transmission process.

[0050] Step 12: Classify the target data to be transmitted to obtain at least one type of first data; here, the target data to be transmitted can be classified according to the classification labels of data sensitivity and security requirements. For example, personal identity information, financial information or trade secrets are usually regarded as highly sensitive data.

[0051] Step 13: Encrypt the at least one type of first data to obtain at least one type of second data;

[0052] Step 14: Perform homomorphic encryption on the at least one second data to obtain the third data;

[0053] Step 15: Perform differential privacy processing on the third data to obtain the fourth data;

[0054] Step 16: Transmit the fourth data.

[0055] In this embodiment, the system or device acts as a data receiver, responsible for acquiring the target data to be transmitted from the data source. The data can originate from various sources, such as data collected by sensors, user input, or data transmitted from other systems. During the reception process, it is necessary to ensure the integrity and accuracy of the data to prevent data loss or corruption. Various verification mechanisms (such as checksums and checksum hashes) can be used to verify whether the received data is consistent with the original data.

[0056] After receiving the target data, the system will classify the data. The classification criteria can be determined based on various factors such as the characteristics, purpose, source, and sensitivity of the data. Through classification, the original data is divided into at least one category of primary data, thereby providing a more targeted strategy for subsequent encryption processing and improving encryption efficiency and security.

[0057] For each category of first data after classification, the system employs a corresponding encryption algorithm. The appropriate encryption algorithm is selected based on computing resources, data sensitivity, and actual security requirements. For highly sensitive data, a stronger and more complex encryption algorithm is chosen; for relatively less sensitive data, a more efficient but slightly less secure algorithm is selected to balance performance and security needs. The encrypted data becomes unreadable and can only be decrypted and accessed with the corresponding key. The encrypted data is stored in secure storage media, such as encrypted hard drives, databases, or cloud storage. By encrypting each category of first data separately, at least one second data item is obtained, ensuring that even if the data is intercepted during transmission, attackers cannot easily obtain the plaintext content of the original data.

[0058] Homomorphic encryption allows specific computational operations to be performed on encrypted data without first decrypting it. The purpose of homomorphic encryption is to further enhance data security while allowing necessary computational operations to be performed on the encrypted data. For example, in a cloud computing environment, cloud service providers can perform statistical analysis on encrypted data without obtaining the plaintext user data. After obtaining the second encrypted data, the system performs homomorphic encryption on this data. After homomorphic encryption, the third data is obtained, which, while maintaining its encrypted state, possesses operability and computability.

[0059] In the differential privacy encryption step, noise is added to the data to prevent attackers from inferring individual information from the data. After obtaining the third data after homomorphic encryption, the system performs differential privacy processing on these data. Specifically, random noise conforming to a specific distribution is added to the data. The magnitude and distribution of the noise are selected according to the sensitivity and privacy requirements of the data. Through differential privacy processing, a fourth data is obtained. The fourth data protects the privacy information in the data to the greatest extent while ensuring data availability.

[0060] After receiving the fourth data, the system will transmit it to the target receiving end. This transmission can be performed using various network communication protocols. During transmission, secure transmission protocols can be further employed to ensure data security in the transmission channel, preventing data theft or tampering during transmission. Upon receiving the fourth data, the target receiving end can perform corresponding decryption and processing operations as needed to obtain usable information from the original data.

[0061] This solution ensures data privacy and security during transmission and use through homomorphic encryption and differential privacy, effectively reducing the risk of data being stolen, tampered with, or leaked during transmission, and improving the efficiency of encrypted data processing in massive data scenarios.

[0062] In an optional embodiment of the present invention, step 12 may include:

[0063] Step 121: Obtain classification labels, which include data sensitivity and sensitivity level;

[0064] Step 122: Classify the transmitted target data according to the classification label to obtain at least one type of first data.

[0065] In this embodiment, it is necessary to obtain classification labels, which contain important information about the data, covering two core elements: data sensitivity and sensitivity level.

[0066] Data sensitivity is used to describe whether data involves sensitive information or the specific degree of sensitivity of the content. Different types of data have different levels of sensitivity. For example, in the medical field, patient medical records and genetic data are considered highly sensitive data because this information is directly related to an individual's health privacy. Once leaked, it can have serious negative impacts on patients, such as discrimination and fraud.

[0067] In some publicly available market research data, the statistical information about consumer purchasing preferences is relatively less sensitive because this information usually does not involve personal privacy and can be publicly used after appropriate processing.

[0068] Sensitive data levels are a quantitative or hierarchical representation of the sensitivity of data. They are classified according to the potential risks and impact of the data. For example, sensitive data can be divided into four levels: high, medium, low, and none. High-level sensitive data includes data involving national security, trade secrets, or core personal privacy, such as military secrets, core technologies of enterprises, and personal bank account information. Medium-level sensitive data includes data with certain commercial value or personal privacy information, such as customer lists of enterprises and social relationship data of individuals. Low-level sensitive data includes some public information or data that has been anonymized and has little impact on individuals and enterprises. No sensitive information includes some data that can always remain in public and has no impact on groups or individuals.

[0069] The methods for obtaining the classification labels include: generating them through predefined rules and standards. For example, enterprises can formulate a set of detailed data classification rules based on their own business needs and data security policies, and clarify the sensitivity of different types of data and the sensitivity level of the data; and automatically identifying and classifying them through machine learning algorithms. This method involves learning and analyzing a large amount of historical data, and the machine learning model can automatically identify the characteristics of the data and assign corresponding classification labels to the data based on these characteristics.

[0070] The system categorizes the target data based on classification labels, grouping data with similar sensitivity levels into one category so that different processing strategies can be adopted for different categories of data.

[0071] The specific process is as follows: The system checks each data item in the target data one by one, and matches it with the pre-acquired classification labels according to its characteristics. For example, for a piece of data containing a personal ID number, the system will determine that it belongs to high-sensitivity data according to the classification label; for a piece of publicly available weather forecast data, the system will determine that it belongs to non-sensitive data, because this type of data can always be fully disclosed to individuals and enterprises.

[0072] After classification, the target data to be transmitted will be divided into at least one category of primary data. Each category of primary data has the same sensitivity and sensitivity level. This classification method makes subsequent data processing more targeted and efficient.

[0073] In an optional embodiment of the present invention, step 13 may include:

[0074] Step 131: Determine the encryption algorithm for the target type corresponding to the at least one type of first data;

[0075] Step 132: Use the encryption algorithm of the target type to encrypt the corresponding first data to obtain at least one second data.

[0076] In this embodiment, a suitable encryption algorithm for the target data type can be determined manually or automatically, thereby ensuring data security. Since different categories of data differ in sensitivity, purpose, storage, and transmission requirements, highly adaptable encryption algorithms are needed for different data types to ensure the confidentiality, integrity, and availability of data during subsequent processing and transmission. Data sensitivity is the primary consideration in selecting an encryption algorithm. For highly sensitive data, encryption algorithms with extremely high encryption strength and rigorously verified security should be selected, such as the Advanced Encryption Standard (AES) algorithm, which has high encryption efficiency and security and can effectively resist various common attack methods. For relatively insensitive data, algorithms with slightly lower encryption strength but higher computational efficiency can be selected to balance security and processing performance. Furthermore, enterprises or organizations can pre-define a set of encryption algorithm selection rules based on data classification labels, sensitivity levels, and other factors according to their own business needs, data security policies, and industry standards, thereby clarifying the encryption algorithm types corresponding to different categories of data.

[0077] After determining the encryption algorithm, the system will use the algorithm to encrypt the corresponding first data, thereby converting the original plaintext data into ciphertext data and obtaining at least one second data. The specific process is as follows: Key generation: The key is generated using a specialized key generation algorithm or hardware device, and must adhere to strict randomness and security requirements; Encryption execution: Based on the selected target type encryption algorithm, the first data and the generated key are used as input to execute the encryption algorithm's calculation process. The encryption algorithm performs a series of mathematical transformations on the original data, such as substitution, permutation, and XOR operations, thereby converting the plaintext data into ciphertext data; Data verification: To ensure that the data has not been tampered with during the encryption process, a data integrity check value, such as a hash value, is also calculated simultaneously. The hash value is transmitted or stored along with the encrypted data. After receiving and decrypting the data, the hash value is recalculated and compared with the received hash value to verify the data integrity.

[0078] In an optional embodiment of the present invention, step 14 may include:

[0079] Step 141: Generate a public-private key pair;

[0080] Step 142: Use the public key in a public-private key pair to encrypt the at least one second data to generate at least one encrypted second data ciphertext;

[0081] Step 143: Perform at least one encryption operation on the at least one encrypted second data ciphertext to obtain third data; wherein the private key is securely transmitted using offline physical transmission or by dynamic negotiation of the key between the two parties.

[0082] In this embodiment, homomorphic encryption is used to protect data privacy while supporting the processing and analysis of encrypted data. In step 141, a public-private key pair is generated. The public key is used to encrypt data, and the private key is used to decrypt data or decrypt data after performing homomorphic operations.

[0083] The process of generating the private key includes:

[0084] Randomly select a number, which is between 1 and n-1;

[0085] In the finite field GF(2) 256 The equation of the elliptic curve on y is: 2 =x 3 +ax+b; where the coefficients are: a is the first preset value, and b is the second preset value; the coordinates of the base point G on the curve are: x G =x1, y G =y1;

[0086] The order of the curve order G is the smallest positive integer n such that nG = infinity; this value of n is used as the private key.

[0087] The private key is serialized in a standard encoding format (such as PEM format).

[0088] The public key can be obtained by using `public_key = private_key.public_key()`.

[0089] The serialized public key is in a standard encoding format (such as PEM format).

[0090] In step 142, the second data is encrypted using a public key to generate ciphertext. The ciphertext can be processed without exposing the original data. After the processing, the result ciphertext, namely the third data, is obtained. The third data contains the result after homomorphic operation and ensures the privacy and security of the data through secure private key management.

[0091] In this embodiment, to ensure that information is not stolen by attackers, a more secure protection method for the private key is required. Therefore, the private key needs to be securely transmitted using offline physical transmission or dynamic key negotiation between the two parties. The dynamic key negotiation between the two parties refers to using a key exchange protocol, which allows the communicating parties to negotiate a shared key by exchanging public information without directly transmitting the private key. This key can be used for subsequent private key transmission or encrypted data communication. At the same time, the private key also needs to be changed regularly to reduce the risk of the private key being cracked over a long period of time and to ensure that it is not eavesdropped or tampered with during transmission.

[0092] In step 143, at least one encryption operation is performed on the at least one encrypted second data ciphertext to obtain the third data. For example, for two ciphertexts (C_1) and (C_2), (C_3=C_1+C_2) or (C_4=C_1×C_2) can be calculated directly without decryption first.

[0093] Finally, the private key is used to decrypt the ciphertext result after the homomorphic operation to obtain the plaintext result.

[0094] In an optional embodiment of the present invention, step 15 may include:

[0095] Step 151: Obtain the preset privacy budget parameters;

[0096] Step 152, according to Determine the scale parameters of the noise; where, Preset privacy budget parameters; For the sensitivity of the query function; The scale parameter for noise;

[0097] Step 153: Add noise to the third data according to the noise scale parameter to obtain the fourth data.

[0098] In this embodiment, differential privacy is used to process third-party data. The essence of this process is to establish a provable mathematical balance between "privacy protection" and "data usability." Through rigorous noise mechanism design, it ensures that the data analysis results do not disclose any individual's sensitive information. This requires both protecting data privacy by adding noise and minimizing the impact on data usability. The main process includes:

[0099] Define a dataset and its neighboring datasets, where each dataset contains n records, each representing information about an individual, denoted as D and . ;

[0100] If two datasets D and Only one record differs (i.e.) =D±{x}, where x is a record), and the two are called "neighboring datasets". It is necessary to ensure that the algorithm's output distribution across neighboring datasets has minimal differences, thus hiding the influence of individual records.

[0101] Determine the query function, which is an analytical operation on the dataset, denoted as f(), for example: statistical query: calculate the number of records in the dataset that meet a certain condition; aggregation query: calculate the mean or sum of data.

[0102] Calculate the sensitivity of the query function, where sensitivity measures the maximum impact of a single record on the query result, defined as the maximum difference in query results across adjacent datasets, denoted as Δf. .

[0103] Different queries have different sensitivities. Among them, the sensitivity of the counting query is 1 (adding or deleting a record changes the count result by a maximum of 1).

[0104] Sensitivity of summation queries: If the data values ​​are bounded (e.g., age range [0, 120]), the sensitivity is the difference between the maximum and minimum values ​​of the data (e.g., 120). Depending on the definition of norm, the sensitivity is calculated differently, and can be divided into L1 sensitivity and L2 sensitivity. L1 sensitivity can be obtained by summing the absolute values; L2 sensitivity can be obtained by taking the square root of the square.

[0105] Once the sensitivity is determined, an appropriate privacy budget parameter needs to be set. The privacy budget parameter is used in differential privacy to measure the degree of privacy protection. The smaller the value of the privacy budget parameter, the higher the level of privacy protection, but the data availability will be reduced accordingly. The larger the value of the privacy budget parameter, the higher the data availability, but the degree of privacy protection will be weakened. Therefore, different privacy budget parameters need to be set for different use cases.

[0106] Set appropriate privacy budget parameters based on the actual situation. The process of setting appropriate privacy budget parameters includes:

[0107] Initially, a privacy budget parameter is set. The data after adding noise is checked to see if it meets the privacy and usability requirements. The privacy budget parameter is dynamically adjusted until a balance point between privacy and usability is found, which is the appropriate privacy budget parameter for the current scenario. Typically, ε is between 0.1 and 10.

[0108] Once the privacy budget parameters are determined, according to The appropriate noise scale parameters are determined to implement the following computational mechanism for privacy verification:

[0109] Noise obeys The probability density function is The output is f(D) + η; where, For privacy budget parameters; η represents the sensitivity of the query function, which reflects how sensitive the query results are to changes in the data records; x represents the possible values ​​of the noise; η represents the actual noise value, i.e., an instance of x. This represents the noise scale parameter. The noise scale parameter determines the intensity of the noise added to the data. A larger noise scale parameter value indicates a stronger noise intensity, thus providing stronger privacy protection, but it will also further reduce data availability. If the sensitivity Δf of the count query is 1, the privacy budget parameter... If the noise level is 1, the mean of the noise added by the Laplacian mechanism is 0. This mean is always 0 and is independent of the noise scale parameter, which is 1 / 1 = 1, meaning the noise follows a Lap(0,1) pattern. Depending on the type of query function and the characteristics of the data, for scalar query results, the generated random noise can be directly added to the query result and the noisy result output, resulting in the fourth data. If the third data is multidimensional (such as a vector or matrix), noise needs to be added to each dimension separately to obtain the fourth data.

[0110] After adding noise, it is necessary to verify through theoretical analysis or simulation experiments whether the noise-added data meets the definition of differential privacy, and to evaluate the impact of the noise-added data on subsequent analysis and applications. This can be primarily measured by calculating metrics such as data error and accuracy to assess data usability. Privacy verification is then performed on the output data to ensure that for any two adjacent datasets D... And any possible set of output results S, satisfying Wherein, Mechanism() is a randomized algorithm; D and S represents adjacent datasets; S is any subset of the outputs, covering all possible privacy-disclosure scenarios. Here, Pr[Mechanism(D)∈S] is the privacy budget parameter; Pr[Mechanism( [)∈S] is the mechanism in the dataset The output is the probability of belonging to S; this inequality ensures that the upper limit of the impact of the existence or non-existence of a single record on the result is controlled by the privacy budget parameter.

[0111] In an optional embodiment of the present invention, the data access control process includes:

[0112] Step 171: Receive the user's data access request;

[0113] Step 172: Based on the data access request, perform user authorization authentication and obtain the authentication result;

[0114] Step 173: When the authentication result indicates that the authentication is successful, the stored fourth data is operated on according to the preset access rules; wherein, the user can perform specific operations with different permission levels according to the preset access rules.

[0115] In this embodiment, fine-grained access control policies can be implemented to ensure that only authorized users can access specific data. Before a user attempts to access data, the system first verifies the user's identity through a multi-factor authentication mechanism, ensuring that only legitimate users can proceed to the next permission verification process, i.e., obtain an authentication result. After successful user authentication, the system dynamically generates a user access permission list based on the user's role, position, department, and other attributes, as well as preset access control rules (i.e., which roles, positions, departments, etc., can perform which level of operation). Data access permissions are further subdivided into multiple levels, such as read, write, execute, modify, and delete, with each level corresponding to different operation permissions. System administrators can assign different permission levels to different users or user groups through the management interface, achieving flexible permission management. For highly sensitive data, the system can also set up additional approval processes, such as two-factor authentication or leadership approval, to ensure the legitimacy of data access.

[0116] During user data access, the system monitors user behavior in real time and verifies permissions based on the user's permission list. If a user attempts to perform an operation beyond their authorized scope, the system will immediately block and record the behavior, and send an alert to the system administrator. All user access behavior is recorded in detail, including access time, location, and content, for subsequent auditing and tracking. As business develops and user roles change, the access control module supports dynamic adjustment of user access permissions. System administrators can also periodically or as needed review and update user permissions to ensure the rationality and effectiveness of permission allocation.

[0117] In an optional embodiment of the present invention, the data auditing process includes:

[0118] Step 181: Obtain the user's access operation records and record the access operation records in the log;

[0119] Step 182: Audit the logs to obtain the audit results;

[0120] Step 183: Output the audit results.

[0121] In this embodiment, the system records all data operation behaviors in a log, including data access, modification, and deletion, for subsequent tracking and auditing. Deep integration of the auditing process with the system ensures that every data operation event can be captured. For each data operation event, key information is extracted from the log and recorded, including operation time, operation type (e.g., access, modification, deletion), operation object (specific data item or dataset), and operator (user or system performing the operation). Complete records are saved in log form, ensuring information integrity and traceability. The system reviews, analyzes, and evaluates the log records to verify the system's compliance, security, and effectiveness, generating an audit report. This audit report may include an overview of the operation behavior, detailed operation records, and anomaly analysis, facilitating management personnel's understanding of data usage and enabling necessary compliance reviews. When system risks are identified in the audit report through automated means, a warning message is sent to the administrator via SMS or email to provide early warning and timely detection and correction of problems. Simultaneously, management personnel can review the audit report to further track data operation behaviors, helping to ensure the compliance of data usage.

[0122] A specific embodiment of the data processing method provided in this invention is as follows:

[0123] Step 1: Receive the target data to be transmitted. The target data can come from multiple different data sources such as databases, file systems, and external systems.

[0124] Step 2: Obtain classification labels using predefined rules and standards. These labels contain important information about the data, covering two core elements: data sensitivity and sensitivity level. Data sensitivity describes whether the data involves sensitive information or the specific degree of sensitivity of the content; different data will have different sensitivities. Classify the target data according to the classification labels to obtain different categories of first data.

[0125] Step 3: Based on the different categories of first data divided by the classification labels, automatically or manually select an appropriate encryption algorithm and encrypt the original data using various encryption algorithms such as Advanced Encryption Standard (AES) and Elliptic Curve Cryptography (ECC) to obtain the encrypted second data.

[0126] Step 4 involves obtaining the second data and performing homomorphic encryption on it to obtain the second data ciphertext. This includes: selecting a suitable homomorphic encryption algorithm, such as partial homomorphic encryption, finite homomorphic encryption, or full homomorphic encryption; performing preprocessing operations on the data, such as standard format conversion and data cleaning; generating a key pair, where the public key is used for encryption and the private key is used for decryption; and using the public key in the key pair to encrypt the second data to obtain the second data ciphertext. Homomorphic operations and data computation tasks, including data analysis, data mining, and machine learning model training, can then be performed on the encrypted second data ciphertext to obtain the encrypted result ciphertext, i.e., the third data. To prevent information from being stolen by attackers, more secure protection methods are needed for private keys. Therefore, private keys should be securely transmitted using offline physical transmission or dynamic key negotiation between the two parties. Dynamic key negotiation between the two parties refers to using a key exchange protocol, which allows the communicating parties to negotiate a shared key by exchanging public information without directly transmitting the private key. This key can be used for subsequent private key transmission or encrypted data communication. At the same time, it is also necessary to change the private key regularly to reduce the risk of the private key being cracked over a long period of time and to ensure that it is not eavesdropped or tampered with during transmission.

[0127] Step 5: Perform differential privacy processing on the third dataset to obtain the number of users with diabetes in the statistical dataset. The dataset contains 1000 user records. Obtain the query function based on f(D) = count(x∈D|diabetes=True). Since the addition or subtraction of a single record in the count query will at most change the result by 1, the sensitivity Δf = 1, where D is the dataset, f() is the query function, and x is each data record. Set an appropriate privacy budget. =1; where, For privacy budget parameters; The sensitivity of a query function reflects how sensitive the query results are to changes in the data records. The scale parameter of the noise determines the intensity of the noise added to the data. A larger scale parameter value indicates a stronger noise intensity, thus providing stronger privacy protection, but also further reducing data usability. After determining the privacy budget parameter, based on... Calculate the corresponding noise scale parameters to obtain This adds noise of corresponding intensity to the query results, and the actual noise after addition follows a set pattern. Where η is the actual noise value, a random actual noise value η=2 is obtained, and according to f(D)+η=120+2=122, the fourth data after differential privacy processing is obtained. Even if a user deletes their own record (the dataset changes from D to ...), ... ), f( )=119, after adding noise, the output 122 must meet the following requirements. ,but =3, at which point the probability density ratio is 3. By satisfying differential privacy, the impact of individual users can be masked by noise, making changes to individual records insensitive during data processing. In other words, data privacy is guaranteed while ensuring data availability.

[0128] Step 6: Transmit the fourth data to the device or system that needs to be used.

[0129] Step 7: Before a user attempts to access data, the system first verifies the user's identity through a multi-factor authentication mechanism, ensuring that only legitimate users can proceed to the next permission verification process. After successful user authentication, the system dynamically generates the user's access permission list based on attributes such as the user's role, position, and department, as well as preset access control rules. Data access permissions are subdivided into multiple levels, such as read, write, execute, and delete, with each level corresponding to different operation permissions. For particularly sensitive data, the system can also set up additional approval processes, such as dual authentication or leadership approval, to ensure the legitimacy of data access. During the user's data access process, the system monitors the user's operational behavior in real time and verifies permissions based on the user's permission list. If a user attempts to perform an operation beyond their authorized scope, the system will immediately block and record the behavior, and send an alert to the system administrator. All user access behaviors are recorded in detail, including access time, access location, and access content, for subsequent auditing and tracking. As business develops and user roles change, the access control module supports dynamically adjusting user access permissions. System administrators can assign different permission levels to different users or user groups through the management interface, enabling flexible permission management. System administrators can periodically or as needed review and update user permissions to ensure the rationality and effectiveness of permission allocation, thereby implementing fine-grained access control policies to ensure that only authorized users can access specific data.

[0130] Step 8: The system records all data operation behaviors in the log, including data access, modification, and deletion, for subsequent tracking and auditing. Through deep integration of the auditing process with the system, every data operation event is captured. For each data operation event, key information is extracted from the log and recorded, including operation time, operation type (e.g., access, modification, deletion), operation object (specific data item or dataset), and operator (user or system performing the operation). Complete records are saved in log form, ensuring information integrity and traceability. The system reviews, analyzes, and evaluates the log records to verify the system's compliance, security, and effectiveness, generating an audit report. This audit report may include an overview of the operation behavior, detailed operation records, and anomaly analysis, facilitating management personnel's understanding of data usage and enabling necessary compliance reviews. When system risks are identified in the audit report through automated means, warning information is sent to the administrator via SMS or email to provide early warnings and timely detection and correction of problems. Simultaneously, management personnel can review the audit report to further track data operation behaviors, helping to ensure the compliance of data usage.

[0131] The data processing method proposed in this invention integrates homomorphic encryption and differential privacy technologies through a privacy computing engine, allowing computation on encrypted data without prior decryption. The computation result after decryption is identical to the result of computation directly on plaintext data. This achieves data analysis and processing while protecting data privacy, meets the data computing needs in different scenarios, achieves a balance between data security and data usage performance, and supports the secure circulation and sharing of data in the field of network security compliance.

[0132] As shown in Figure 2, this embodiment of the invention also provides a network data processing device 20, comprising:

[0133] Receiver module 21 is used to receive target data to be transmitted;

[0134] Processing module 22 is configured to classify the target data to be transmitted to obtain at least one type of first data; encrypt the at least one type of first data to obtain at least one type of second data; perform homomorphic encryption on the at least one type of second data to obtain third data; and perform differential privacy processing on the third data to obtain fourth data.

[0135] The transmission module 23 is used to transmit the fourth data.

[0136] Optionally, the target data to be transmitted is classified to obtain at least one type of first data, including:

[0137] Obtain category labels, which include data sensitivity information and sensitivity level;

[0138] Based on the classification labels, the transmitted target data is classified to obtain at least one type of first data.

[0139] Optionally, the at least one type of first data is encrypted to obtain at least one type of second data, including:

[0140] Determine the encryption algorithm for the target type corresponding to each of the at least one type of first data;

[0141] The first data is encrypted using the encryption algorithm of the target type to obtain at least one second data.

[0142] Optionally, homomorphic encryption is performed on the at least one second data to obtain third data, including:

[0143] Generate a public-private key pair;

[0144] Using the public key from a public-private key pair, encrypt the at least one second data to generate at least one encrypted second data ciphertext;

[0145] Perform at least one encryption operation on the at least one encrypted second data ciphertext to obtain third data;

[0146] The private key is securely transmitted either through offline physical transmission or through dynamic negotiation between the two parties.

[0147] Optionally, differential privacy processing is performed on the third data to obtain fourth data, including:

[0148] Set privacy budget parameters;

[0149] according to Determine the scale parameters of the noise;

[0150] The third data is subjected to noise addition based on the noise scale parameter to obtain the fourth data;

[0151] in, For privacy budget parameters, For the sensitivity of the query function, This is the scale parameter for noise.

[0152] Optionally, the method further includes:

[0153] Receive user data access requests;

[0154] Based on the data access request, the user's permissions are authenticated, and the authentication result is obtained;

[0155] When the authentication result indicates that the authentication is successful, the stored fourth data is operated on according to the preset access rules;

[0156] Users can perform specific operations with different permission levels according to the preset access rules.

[0157] Optionally, the processing module 22 is further configured to:

[0158] Obtain the user's access operation records and record the access operation records in the log;

[0159] The logs were audited to obtain the audit results;

[0160] Output the audit results.

[0161] It should be noted that this device is a device corresponding to the above method. All implementation methods in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0162] As shown in Figure 3, this embodiment of the invention also provides a computing device 30, including a processor 31, a memory 32, and a program or instructions stored in the memory 32 and executable on the processor 31. When the program or instructions are executed by the processor 31, they implement the various processes of the above-described network data processing method embodiments and achieve the same technical effects. To avoid repetition, further details are omitted here. It should be noted that the computing device in this embodiment of the invention includes the aforementioned mobile electronic devices and non-mobile electronic devices.

[0163] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0164] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0165] In the embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0166] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0167] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0168] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0169] Furthermore, it should be noted that in the apparatus and method of the present invention, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent solutions of the present invention. Moreover, the steps performing the above series of processes can naturally be executed in the order described, but are not necessarily required to be executed in chronological order; some steps can be executed in parallel or independently of each other. Those skilled in the art will understand that all or any step or component of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or network of computing devices, in hardware, firmware, software, or a combination thereof. This is something that those skilled in the art can achieve by using their basic programming skills after reading the description of the present invention.

[0170] Therefore, the object of the present invention can also be achieved by running a program or a set of programs on any computing device. The computing device can be a known general-purpose device. Therefore, the object of the present invention can also be achieved simply by providing a program product containing program code for implementing the method or apparatus. That is, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any known storage medium or any storage medium developed in the future. It should also be noted that in the apparatus and method of the present invention, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent to the present invention. Furthermore, the steps for performing the above series of processes can naturally be performed in the order described, but are not necessarily required to be performed in chronological order. Some steps can be performed in parallel or independently of each other.

[0171] The above are preferred embodiments of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A network data processing method, characterized in that, include: Receive target data to be transmitted; The target data to be transmitted is classified to obtain at least one type of first data; Encrypt the at least one type of first data to obtain at least one second data; perform homomorphic encryption on the at least one second data to obtain third data; perform differential privacy processing on the third data to obtain fourth data; Transmitting the fourth data; wherein, classifying the target data to be transmitted to obtain at least one category of first data includes: obtaining classification labels, wherein the classification labels include data sensitivity and sensitivity level; the data sensitivity is used to describe whether the data involves sensitive information or the specific degree of sensitivity of the content, and different data will have different sensitivities; the sensitivity level is a quantitative or graded representation of the data sensitivity, and is divided according to the potential risks and impact of the data; classifying the target data to be transmitted according to the classification labels to obtain at least one category of first data; specifically including: the system performs a classification operation on the target data to be transmitted according to the classification labels, grouping data with similar sensitivity and sensitivity levels into one category, and adopting different processing strategies for different categories of data, the specific process being: the system checks each data item in the target data to be transmitted one by one, matches its characteristics with the pre-obtained classification labels, and after the classification operation, the target data to be transmitted will be divided into at least one category of first data, each category of first data having similar characteristics. The process involves different levels of sensitivity and sensitivity of data. Specifically, encrypting at least one type of first data to obtain at least one type of second data includes: determining an encryption algorithm corresponding to a target type of the at least one type of first data; encrypting the corresponding first data using the encryption algorithm of the target type to obtain at least one type of second data; and performing homomorphic encryption on the at least one type of second data to obtain third data, including: generating a public-private key pair; encrypting the at least one type of second data using the public key in the public-private key pair to generate at least one encrypted second data ciphertext; performing at least one encryption operation on the at least one encrypted second data ciphertext to obtain third data, including: for two ciphertexts (C_1) and (C_2), directly calculating (C_3=C_1+C_2) or (C_4=C_1×C_2), and decrypting the ciphertext result after the homomorphic operation using the private key to obtain the plaintext result after the operation; and performing differential privacy processing on the third data to obtain fourth data, including: setting a privacy budget parameter; and according to... Determine the noise scale parameters; perform noise addition on the third data based on the noise scale parameters to obtain the fourth data; wherein, For privacy budget parameters, Take a value between 0.1 and 10. For the sensitivity of the query function, Let be the scale parameter of the noise; where is the sensitivity of the query function. The following process is used to determine: Define a dataset and neighboring datasets, where each dataset contains n records, each representing information about an individual, denoted as D and . If two datasets D and The difference is one record, that is =D±{x}, where x is a record, then the dataset D and Given adjacent datasets; determine a query function, denoted as f(), which is an analytical operation on the datasets; calculate the sensitivity of the query function, which measures the maximum impact of a single record on the query results. The actual noise value of the noise scale parameter follows... The probability density function is The output is f(D) + η; x represents the possible values ​​of noise; η represents the actual noise value, i.e., an instance of x; if the sensitivity Δf of the counting query is 1, the privacy budget parameter... =1, then the noise added by the Laplacian mechanism has a mean of 0. This mean is always 0 and is independent of the noise scale parameter, which is 1 / 1 = 1, meaning the noise follows Lap(0,1). Based on the type of the query function and the characteristics of the data, for scalar query results, the generated random noise is added to the query results and the noise-added result is output, resulting in the fourth data. For any adjacent dataset D and And any possible set of output results S, satisfying Wherein, Mechanism() is a randomization algorithm; Pr[Mechanism(D)∈S] is the probability that the mechanism outputs a value belonging to S on dataset D; Pr[Mechanism( [)∈S] is the mechanism in the dataset Output the probability that it belongs to S.

2. The network data processing method according to claim 1, characterized in that, The method further includes: receiving a user's data access request; performing permission authentication on the user based on the data access request to obtain an authentication result; and when the authentication result indicates that the authentication is successful, operating on the stored fourth data according to preset access rules.

3. The network data processing method according to claim 1, characterized in that, The method further includes: obtaining the user's access operation records, recording the access operation records in a log; auditing the log to obtain audit results; and outputting the audit results.

4. A network data processing device, characterized in that, include: The receiving module is used to receive the target data to be transmitted; The processing module is used to classify the target data to be transmitted to obtain at least one type of first data; Encrypt each of the at least one type of first data to obtain at least one type of second data; The at least one second data is homomorphically encrypted to obtain the third data; the third data is then differentially private to obtain the fourth data. A transmission module is used to transmit the fourth data; wherein, classifying the target data to be transmitted to obtain at least one category of first data includes: obtaining classification labels, wherein the classification labels include data sensitivity and sensitivity level; the data sensitivity is used to describe whether the data involves sensitive information or the specific degree of sensitivity of the content, and different data will have different sensitivities; the sensitivity level is a quantitative or graded representation of the data sensitivity, and is divided according to the potential risks and impact of the data; classifying the target data to be transmitted according to the classification labels to obtain at least one category of first data; specifically including: the system performs a classification operation on the target data to be transmitted according to the classification labels, grouping data with similar sensitivity and sensitivity levels into one category, and adopting different processing strategies for different categories of data, the specific process being: the system checks each data item in the target data to be transmitted one by one, matches its characteristics with the pre-obtained classification labels, and after the classification operation, the target data to be transmitted will be divided into at least one category of first data, each category of first data... Having the same sensitivity and sensitivity level; wherein, encrypting the at least one type of first data to obtain at least one type of second data includes: determining an encryption algorithm for a target type corresponding to the at least one type of first data; using the encryption algorithm of the target type to encrypt the corresponding first data to obtain at least one type of second data; wherein, homomorphic encryption processing is performed on the at least one type of second data to obtain third data includes: generating a public-private key pair; using the public key in the public-private key pair to encrypt the at least one type of second data to generate at least one encrypted second data ciphertext; performing at least one encryption operation on the at least one encrypted second data ciphertext to obtain third data includes: for two ciphertexts (C_1) and (C_2), directly calculating (C_3=C_1+C_2) or (C_4=C_1×C_2), and using the private key to decrypt the ciphertext result after the homomorphic operation to obtain the plaintext result after the operation; wherein, differential privacy processing is performed on the third data to obtain fourth data includes: setting a privacy budget parameter; according to Determine the noise scale parameters; perform noise addition on the third data based on the noise scale parameters to obtain the fourth data; wherein, For privacy budget parameters, Take a value between 0.1 and 10. For the sensitivity of the query function, Let be the scale parameter of the noise; where is the sensitivity of the query function. The following process is used to determine: Define a dataset and neighboring datasets, where each dataset contains n records, each representing information about an individual, denoted as D and . If two datasets D and The difference is one record, that is =D±{x}, where x is a record, then the dataset D and Given adjacent datasets; determine a query function, denoted as f(), which is an analytical operation on the datasets; calculate the sensitivity of the query function, which measures the maximum impact of a single record on the query results. The actual noise value of the noise scale parameter follows... The probability density function is The output is f(D) + η; x represents the possible values ​​of noise; η represents the actual noise value, i.e., an instance of x; if the sensitivity Δf of the counting query is 1, the privacy budget parameter... =1, then the noise added by the Laplacian mechanism has a mean of 0. This mean is always 0 and is independent of the noise scale parameter, which is 1 / 1 = 1, meaning the noise follows Lap(0,1). Based on the type of the query function and the characteristics of the data, for scalar query results, the generated random noise is added to the query results and the noise-added result is output, resulting in the fourth data. For any adjacent dataset D and And any possible set of output results S, satisfying Wherein, Mechanism() is a randomization algorithm; Pr[Mechanism(D)∈S] is the probability that the mechanism outputs a value belonging to S on dataset D; Pr[Mechanism( [)∈S] is the mechanism in the dataset Output the probability that it belongs to S.

5. A computing device, characterized in that, include: A processor, a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method as described in any one of claims 1 to 3.

6. A computer-readable storage medium, characterized in that, The system stores instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1 to 3.

Images