A data security efficient access method and system

Abstract

The application provides a kind of data security efficient access method and system, comprising: receiving the access request of user, receiving the access request of user;With multistage recursive aggregation search algorithm, according to access request, in turn through global index, cross-device index and dynamic local index data search is carried out, and the summary search data is obtained;With the improved block encoding homomorphic encryption method combined with data reorganization rules, the search data is fed back to the user;The application is based on global index, cross-device index and dynamic local index, improved block encoding homomorphic encryption method combined with data reorganization rules, composed of the micro-service architecture of power grid, through micro-service architecture, based on access request, carries out layer by layer search, realizes the safe and efficient access of power equipment data, effectively solves the efficient retrieval of data across regions and across devices in distributed power system, solves the problem of safe transmission of data.

Description

Technical Field

[0001] This invention patent application belongs to the field of data security and distributed network technology, specifically relating to a data security and efficient access method and system. Background Technology

[0002] With the continuous intelligent development of power systems, power equipment is becoming increasingly widespread, and the amount of data generated is growing exponentially. Traditional power data management methods typically rely on centralized storage and management. While this approach can ensure data consistency, it is prone to access bottlenecks and latency when faced with massive amounts of data.

[0003] Centralized management methods struggle to meet the demands of modern power systems for efficient data access across regions and devices, as their processing efficiency and real-time performance fall short. Furthermore, with the increasing need for interconnectivity among power equipment, the transmission and sharing of data across different devices and regions faces ever-increasing security risks. Ensuring data security during transmission has become a major challenge for power system data management. Summary of the Invention

[0004] To overcome the shortcomings of the prior art, this patent application proposes a secure and efficient data access method, comprising:

[0005] Receive user access requests;

[0006] A multi-level recursive aggregation search algorithm is adopted. Based on the access request, data is searched sequentially through a pre-established global index, a cross-device index, and a dynamic local index to obtain the summarized search data.

[0007] An improved block-based homomorphic encryption method combined with data recombination rules is used to feed the search data back to the user;

[0008] The global index is constructed based on the local clustering results of electronic data and the multi-dimensional features of electronic devices, combined with a global index structure; the cross-device index is constructed based on the correlation of multi-dimensional features of electronic devices, combined with a hierarchical index structure; and the dynamic local index is constructed based on power data, combined with a dynamic hierarchical index structure.

[0009] Preferably, the multi-level recursive aggregation search algorithm is used to search for data sequentially through a pre-established global index, a cross-device index, and a dynamic local index according to the access request, to obtain summarized search data, including:

[0010] Extract key query parameters from the access request;

[0011] The query key parameters are mapped to the global cluster labels of the global index to initially determine the global cluster center, global cluster, and related device list containing the target data;

[0012] Based on the global cluster center, global cluster, and related device list containing the target data, the most relevant global cluster is selected on the global cluster index layer of the global index. After searching for the target data in the most relevant global cluster, the most relevant power equipment is selected on the power equipment index layer of the global index according to the related device list and the calculated global priority of electronic equipment.

[0013] Using the cross-device index, based on the calculated multi-dimensional feature correlation between the most relevant power equipment and each electronic device group, electronic devices that exceed the multi-dimensional feature correlation threshold are filtered out, and target data is searched by accessing the electronic devices that exceed the multi-dimensional feature correlation threshold;

[0014] Using the local dynamic index of the electronic device that exceeds the multi-dimensional feature relevance threshold, and based on the query key parameters, after locating the most relevant local cluster center, the target data is searched according to the index priority;

[0015] All target data is aggregated, and the aggregated target data is processed to serve as the final search data.

[0016] The key parameters for querying include one or more of the following: query keywords, query time range, and required data type; the data processing includes: target data sorting, target data merging, target data deduplication, and target data filtering.

[0017] Preferably, the dynamic local index structure includes a cluster center layer and a data point ranking layer; the construction of the dynamic local index includes:

[0018] Local clustering is performed on the power data of all power equipment to obtain local clustering results; the local clustering results include: multiple local clusters and multiple local cluster centers;

[0019] Based on the power data points after local clustering and the local cluster centers obtained from the local clustering results, a multidimensional weighted distance is calculated. Based on the multidimensional weighted distance, the power data points corresponding to the multidimensional weighted distance are sorted to generate a power data point sequence.

[0020] For each local cluster, a corresponding local cluster index is designed to complete the construction of the cluster center layer;

[0021] The power data points are placed into the corresponding local clustering index. Based on the power data point sequence and according to the index priority calculation formula, the index position of each power data point in the local clustering index is determined, thus completing the construction of the data point sorting layer.

[0022] Preferably, the step of performing local clustering of power data on all power equipment to obtain local clustering results includes:

[0023] Extracting multidimensional data features from multiple power data from all power equipment;

[0024] After mapping the multidimensional data features to a weighted feature space, the multidimensional data features are weighted to obtain power data points;

[0025] An adaptive dynamic clustering update method combined with a boundary fuzzing processing mechanism is used to dynamically update the local cluster centers and optimize the boundary data of the power data points, generating local clustering results.

[0026] When the local clustering result meets the clustering termination condition, the adaptive dynamic clustering update method ends, and the local clustering result is used as the power data clustering result.

[0027] The multidimensional data features include one or more of the following: data type, data size, access frequency, data relevance, timeliness, and access pattern; the clustering termination condition is that the movement of the local cluster centers meets a preset movement threshold, and the change in the affiliation of the power data points meets a preset change threshold.

[0028] Preferably, the step of using an adaptive dynamic clustering update method combined with a boundary fuzzing mechanism to dynamically update the local cluster centers and optimize the boundary data of the power data points to generate local clustering results includes:

[0029] From the power data points, select the power data point with the widest coverage as the initial center;

[0030] Select the power data point that is farthest from the initial center from the power data points, and use it as the initial local cluster center;

[0031] Calculate the distance metric in each clustering process, and perform local clustering on the power data points based on the initial local cluster centers and the distance metric to obtain multiple local clusters;

[0032] A local cluster center update calculation formula is used to dynamically adjust the local cluster centers of each local cluster to obtain dynamically adjusted local cluster centers.

[0033] By employing a fuzzy attribution metric, the multidimensional data features at the fuzzy boundaries of the local clusters are adjusted for multi-cluster attribution to obtain the local clustering results.

[0034] Preferably, the construction of the cross-device index includes:

[0035] The multi-dimensional feature correlation of all electronic devices is calculated using the electronic device feature correlation calculation formula.

[0036] Based on the correlation of the multi-dimensional features, the electronic devices are combined to generate multiple electronic device groups;

[0037] A hierarchical index structure is adopted to hierarchically classify electronic data and electronic devices in each electronic device group, generating a provisional cross-device index hierarchy.

[0038] The provisional cross-device index hierarchy is optimized using the pre-obtained power equipment score sequence to obtain the final cross-device index hierarchy;

[0039] The optimal cross-device retrieval path for the cross-device index hierarchy is determined by using a cross-device retrieval path calculation formula, thus completing the construction of the cross-device index.

[0040] Preferably, the process of generating the power equipment score sequence includes:

[0041] Calculate the correlation between the multi-dimensional features of each power device and each local cluster center;

[0042] Based on the relevant parameters for power equipment ranking and the correlation degree, the ranking score of each power equipment is calculated using the power equipment ranking score calculation formula, the power equipment scores are ranked, and a provisional power equipment score sequence is generated.

[0043] A dynamic adaptive sorting model is adopted to dynamically adjust each electronic device in the provisional power equipment score sequence based on its current status and historical performance, thereby generating the current power equipment score sequence.

[0044] The multi-dimensional characteristics of the power equipment include one or more of the following: equipment computing power, equipment storage capacity, equipment location effect, and the matching degree between the power data stored on the equipment and the local cluster centers.

[0045] Preferably, the global index structure includes a global clustering index layer and a power equipment index layer; the construction of the global index includes:

[0046] Perform global standardization on local cluster centers and multidimensional data features;

[0047] The local cluster centers after global standardization are then subjected to global clustering to obtain global clustering results and global clustering labels; the global clustering results include: multiple global clusters and multiple global cluster centers;

[0048] For each global cluster, a corresponding global cluster index node is designed to complete the construction of the global cluster index layer; the global cluster index node includes: global cluster center, related device list, and global priority;

[0049] For each electronic device under a global cluster, a corresponding device index node is designed to complete the construction of the power equipment index layer; the device index node includes: the unique ID of the electronic device, the range of electronic data, the access frequency, and the distance between the electronic device and the local cluster center.

[0050] Preferably, the data reorganization rules include a cross-block association verification mechanism and a dynamic reorganization optimization algorithm; the step of using an improved block-based homomorphic encryption method combined with the data reorganization rules to feed the search data back to the user includes:

[0051] The search data is segmented to generate multiple data blocks, and each data block is marked with a segmentation point.

[0052] Each data block and the cut-off point marker are homomorphically encrypted, and the homomorphically encrypted data block is embedded with the data block number to generate an encrypted encoded data block.

[0053] A cross-block association verification mechanism combined with a dynamic reorganization optimization algorithm is used to reorganize the encrypted data block according to the embedded data block number to generate a reorganized data block;

[0054] Send the reconstructed data block to the user;

[0055] The cross-block association verification mechanism verifies the hash values ​​of adjacent data blocks by performing an XOR operation; the dynamic reassembly optimization algorithm dynamically adjusts the order and strategy of data reassembly based on the packet loss rate and the difficulty of data block reassembly during transmission.

[0056] Based on the same inventive concept, this patent application also provides a data security and efficiency access system, including: an access request receiving module, a search data aggregation module, and a search data feedback module;

[0057] The access request receiving module is used to receive user access requests;

[0058] The search data aggregation module is used to perform data search by using a multi-level recursive aggregation search algorithm according to the access request, sequentially through a pre-established global index, a cross-device index, and a dynamic local index, to obtain aggregated search data.

[0059] The search data feedback module is used to feed back the search data to the user using an improved block-coded homomorphic encryption method combined with data recombination rules.

[0060] The global index is constructed based on the local clustering results of electronic data and the multi-dimensional features of electronic devices, combined with a global index structure; the cross-device index is constructed based on the correlation of multi-dimensional features of electronic devices, combined with a hierarchical index structure; and the dynamic local index is constructed based on power data, combined with a dynamic hierarchical index structure.

[0061] Compared with the closest prior art, the beneficial effects of this invention patent application are as follows:

[0062] This invention patent application provides a secure and efficient data access method and system, comprising: receiving a user's access request; employing a multi-level recursive aggregation search algorithm to sequentially search data through a pre-established global index, a cross-device index, and a dynamic local index according to the access request, thereby obtaining summarized search data; and employing an improved block-based homomorphic encryption method combined with data recombination rules to feed back the search data to the user; wherein, the global index is constructed based on the local clustering results of electronic data and the multi-dimensional characteristics of electronic devices, combined with a global index structure; the cross-device index is based on the correlation of the multi-dimensional characteristics of electronic devices. The invention constructs a hierarchical index structure based on power data; the dynamic local index is constructed based on power data combined with the dynamic hierarchical index structure; the invention forms a microservice architecture for the power grid based on global index, cross-device index, dynamic local index, improved block coding homomorphic encryption method combined with data reorganization rules. Through the global index, cross-device index, and dynamic local index of the microservice architecture, layer-by-layer search is performed based on access requests, realizing secure and efficient access to power equipment data, effectively solving the problem of efficient retrieval of data across regions and devices in distributed power systems, and solving the problem of secure data transmission by adopting the improved block coding homomorphic encryption method combined with data reorganization rules. Attached Figure Description

[0063] Figure 1 A schematic diagram of a secure and efficient data access method provided in this patent application;

[0064] Figure 2 A schematic diagram illustrating the steps of a secure and efficient data access method based on electronic microservices, provided for this patent application.

[0065] Figure 3 A general framework diagram of a secure and efficient data access method based on electronic microservices provided for this invention patent application;

[0066] Figure 4A schematic diagram of a data security and efficiency access system provided for this invention patent application;

[0067] Figure 5 This is a schematic diagram of the structure of an electronic device provided in this patent application. Detailed Implementation

[0068] The specific embodiments of this patent application will be further described in detail below with reference to the accompanying drawings.

[0069] Example 1:

[0070] This invention patent application provides a secure and efficient data access method based on power microservices, such as... Figure 1 As shown, it includes:

[0071] Step 1: Receive user access requests;

[0072] Step 2: Using a multi-level recursive aggregation search algorithm, based on the access request, data is searched sequentially through a pre-established global index, a cross-device index, and a dynamic local index to obtain the summarized search data;

[0073] Step 3: Using an improved block-based homomorphic encryption method combined with data recombination rules, the search data is fed back to the user;

[0074] The global index is constructed based on the local clustering results of electronic data and the multi-dimensional features of electronic devices, combined with a global index structure; the cross-device index is constructed based on the correlation of multi-dimensional features of electronic devices, combined with a hierarchical index structure; and the dynamic local index is constructed based on power data, combined with a dynamic hierarchical index structure.

[0075] In one implementation, step 1 above involves receiving the user's access request;

[0076] For example, the user layer receives user access requests; this invention performs secure and efficient data access based on the received access requests, and the access requests are the foundation for secure and efficient data access.

[0077] In one implementation, step 2 above employs a multi-level recursive aggregation search algorithm. Based on the access request, data is searched sequentially through a pre-established global index, a cross-device index, and a dynamic local index to obtain summarized search data, including:

[0078] Extract key query parameters from the access request;

[0079] For example, extract key query parameters from the user's query request, such as query keywords, query time range, and required data type, and convert these key query parameters into a query feature vector q.f .

[0080] The query key parameters are mapped to the global cluster labels of the global index to initially determine the global cluster center, global cluster, and related device list containing the target data;

[0081] For example, based on the query key parameters and query feature vector, the query request is mapped to the global cluster labels of the global index, initially determining the global cluster centers and related devices that may contain the target data; the relevance score of the global cluster is used to evaluate the query feature vector q. f and global cluster center v j The similarity is used for matching, and the preliminary matching global cluster S is determined based on the relevance score of the global cluster. g And related equipment list D j ;

[0082] The relevance scoring expression for global clustering is as follows:

[0083]

[0084] Among them, S g相 This is the relevance score for global clustering, argmax is the argmax function, τ1 and τ2 are adjustment parameters, and q f To query the feature vector, v j As the global cluster center, Corr(q) f ,v j ) is q f With v j The correlation metric.

[0085] Based on the global cluster center, global cluster, and related device list containing the target data, the most relevant global cluster is selected on the global cluster index layer of the global index. After searching for the target data in the most relevant global cluster, the most relevant power equipment is selected on the power equipment index layer of the global index according to the related device list and the calculated global priority of electronic equipment.

[0086] For example, in the global clustering index layer of the global index, based on the global clustering label G ij and global cluster center v j The most relevant global cluster is located. Based on global priority, high-priority global clusters are retrieved first, gradually narrowing the search scope. In the power equipment index layer of the global index, the search is performed based on the global priority of the equipment and the related equipment list D. jThe most relevant power equipment is selected. The relevance of power equipment is evaluated by calculating the distance from local cluster centers to the global cluster center, and highly relevant power equipment is placed in a priority search list. The relevant equipment list is the initial set of all power equipment that may be related to the target data, while the priority search list is a high-priority subset generated after further evaluation and ranking of these equipment, arranged by priority for actual retrieval.

[0087] Using the cross-device index, based on the calculated multi-dimensional feature correlation between the most relevant power equipment and each electronic device group, electronic devices that exceed the multi-dimensional feature correlation threshold are filtered out, and target data is searched by accessing the electronic devices that exceed the multi-dimensional feature correlation threshold;

[0088] For example, using the relevance metric calculated in cross-device indexes, the formula Corr is used. AB The process involves evaluating the relevance of the most relevant power devices selected to the multi-dimensional features of each electronic device group. For highly relevant electronic device groups, the optimal path formula (Path) is used for path optimization to determine the optimal device retrieval order. Within the optimized device path, highly relevant devices are accessed sequentially, progressively deepening the search for target data. If partial data is found within the current electronic device, the search path is recorded, and the search scope is further narrowed down to similar related electronic devices.

[0089] Using the local dynamic index of the electronic device that exceeds the multi-dimensional feature relevance threshold, and based on the query key parameters, after locating the most relevant local cluster center, the target data is searched according to the index priority;

[0090] For example, based on the local dynamic index within the electronic device, using query key parameters and query feature vectors, the most relevant local cluster center μ is located. j According to the index priority P(x) i Prioritize retrieving data points with higher priority. During the search process, if the access frequency of certain data points suddenly increases, dynamically adjust the hierarchical position of these data points in the local index. Move high-priority data points to higher levels so that they can be retrieved faster in subsequent queries.

[0091] All target data is aggregated, and the aggregated target data is processed to serve as the final search data.

[0092] For example, the system aggregates target data from all electronic devices, sorts and merges it according to the degree of matching between the target data and the user's query request, including evaluation dimensions such as keyword matching, time range, and data type (relevance requirements of the user's query). The target data is deduplicated and filtered, and the final search data is returned to the user, ensuring accurate and efficient results. Based on the access frequency and query pattern of this search, the priority of local dynamic indexes and global indexes is updated. Relevance metrics in cross-device indexes are adjusted as needed to ensure more efficient future searches.

[0093] The key parameters for querying include one or more of the following: query keywords, query time range, and required data type; the data processing includes: target data sorting, target data merging, target data deduplication, and target data filtering.

[0094] In one implementation, the dynamic local index structure includes a cluster center layer and a data point ranking layer; the construction of the dynamic local index includes:

[0095] Local clustering is performed on the power data of all power equipment to obtain local clustering results; the local clustering results include: multiple local clusters and multiple local cluster centers;

[0096] For example, power data is stored in power equipment at the equipment management layer. First, the characteristics of the power data are extracted, including: data type, data size, access frequency, data relevance, timeliness, and access pattern. The extracted power data is then weighted according to its features, and clustered based on the weighted feature space. An adaptive dynamic clustering update method is used to dynamically adjust the data point assignments, and fuzzy boundary processing and cluster center convergence algorithms are used to improve the clustering accuracy. Through these operations, the data stored in all power equipment is clustered and archived, yielding local clustering results.

[0097] Based on the power data points after local clustering and the local cluster centers obtained from the local clustering results, a multidimensional weighted distance is calculated. Based on the multidimensional weighted distance, the power data points corresponding to the multidimensional weighted distance are sorted to generate a power data point sequence.

[0098] Example: Set of electricity data points after local clustering of data:

[0099] D = {x1, x2, ..., x} n}

[0100] Wherein, D is the set of power data points, x n This is the nth power data point;

[0101] After clustering is completed, each power data point belongs to its corresponding local cluster, and the final local cluster has a corresponding final local cluster center μ. j To reflect the importance and differences of electricity data across various dimensions, a multidimensional weight α is introduced. ik Define the multidimensional weighted distance formula:

[0102]

[0103] Where, d w (x i ,μ j ) is x i With μ j Multidimensional weighted distance, x i For the i-th power data point, μ j Let x be the j-th local cluster center after clustering, and also the j-th local centroid. ik Let μ be the coordinate of the i-th power data point in the k-th dimension. ik Let α be the coordinates of the i-th local cluster center in the k-th dimension. ik Let m be the importance weight of the i-th power data point in the k-th dimension, where m is the number of dimensions.

[0104] For each power data point x i Based on multidimensional weighted distance, the power data points x i Sort the power data points x according to their relationship with local cluster centers, in ascending order of distance. i Sort the data points so that the smaller the distance, the closer they are to the top layer of the index.

[0105] For each local cluster, a corresponding local cluster index is designed to complete the construction of the cluster center layer;

[0106] For example, a dynamic hierarchical index structure is introduced, based on power data point x. i The priority of the index is dynamically adjusted based on the access frequency and query relevance.

[0107] The local index structure is designed as follows:

[0108] First layer (cluster center layer): contains the local cluster centers μ after clustering. j The index nodes at this level each correspond to a local cluster. This layer of indexing is primarily used to quickly locate the position of each local cluster, ensuring that when a query arrives, the local cluster relevant to the query can be found quickly.

[0109] The power data points are placed into the corresponding local clustering index. Based on the power data point sequence and according to the index priority calculation formula, the index position of each power data point in the local clustering index is determined, thus completing the construction of the data point sorting layer.

[0110] Example, second layer (data point sorting layer): based on power data point x i Multidimensional weighted distance, dividing power data points x i They are placed into the index of their respective local clusters according to priority. This layer contains all the power data points x in each local cluster. i Based on power data point x i With the local cluster centers μ after clustering j The data points are sorted by multi-dimensional weighted distance. The closer the data point x is, the higher its access frequency. i It is placed in a higher index position. This means that during a query, the higher priority data points are retrieved first, and the results are returned quickly.

[0111] The index priority formula is:

[0112]

[0113] Wherein, P(x i ) represents power data point x i Index priority, β i For power data point x i The access weight takes into account access frequency and latency factors. Data points with smaller distances and higher access weights are assigned higher priority. w (x i ,μ j ) is the power data point x i To the local cluster centers μ after clustering j The weighted distance.

[0114] According to P(x) i The value of ) dynamically adjusts the index hierarchy, placing high-priority data points at higher levels.

[0115] To better describe the boundary regions of each local cluster, the relative distance intervals between power data points are defined:

[0116] B j =[min(x ik -x il ),max(x ik -x il )]

[0117] Among them, B j x represents the relative distance range between power data points. ik Let x be the coordinate of the i-th power data point in the k-th dimension. il Let x be the coordinate of the i-th power data point in the l-th dimension. ik x il ∈Cj C j For the j-th local cluster;

[0118] B j The relative distances between local cluster boundary points are described for boundary delineation of the second-level index. This invention establishes cross-device indexes and dynamic local indexes at the device management level. Local indexes are created for locally clustered electronic data on power equipment. Local indexes can effectively reduce the query scope and accelerate the retrieval process. The index is created based on the local cluster centers of electronic data, enabling rapid data location during searches. Based on the correlation between electronic data, cross-device indexes are created for similar data across different power equipment, enabling rapid cross-device retrieval during searches.

[0119] In one implementation, the step of performing local clustering of power data on all power devices to obtain local clustering results includes:

[0120] Extracting multidimensional data features from multiple power data from all power equipment;

[0121] After mapping the multidimensional data features to a weighted feature space, the multidimensional data features are weighted to obtain power data points;

[0122] For example, the multidimensional data features of electricity data are mapped to a weighted feature space to obtain electricity data points; each electricity data point is assumed to have m-dimensional features, represented as a feature vector:

[0123] F(x i )=[f{i1},f{i2},…,f{i m},]

[0124] Where F(x) i Let f{i} be the feature vector of the i-th power data point. m} represents the m-dimensional feature vector of the i-th power data point;

[0125] In the weighted feature space, each feature dimension has a corresponding weight; initially, the weight can be set based on empirical values.

[0126] The formula for calculating the feature-weighted mapping is:

[0127] F′(x i )=[w1·f i1 ,w2·f i2 ,…,w m ·f im ]

[0128] Where, F′(x i Let w be the feature weighting vector of the i-th power data point.m f represents the weights of the m-dimensional eigenvectors. im The feature vector of the i-th power data point in m dimensions;

[0129] The weights can be initialized based on experience. After the weights are initialized, all electronic data points are mapped to a weighted feature space. Multidimensional data features are used to weight the power data points to obtain the power data points. Each electronic data point is assigned to the corresponding cluster through an adaptive dynamic clustering update algorithm, and fuzzy boundary processing and cluster center convergence algorithms are used to improve the clustering accuracy.

[0130] An adaptive dynamic clustering update method combined with a boundary fuzzing processing mechanism is used to dynamically update the local cluster centers and optimize the boundary data of the power data points, generating local clustering results.

[0131] When the local clustering result meets the clustering termination condition, the adaptive dynamic clustering update method ends, and the local clustering result is used as the power data clustering result.

[0132] For example, clustering convergence is determined by calculating the movement of local cluster centers and the change in the affiliation of electronic data points. Convergence is achieved when the clustering termination condition is met. The convergence condition is as follows:

[0133]

[0134] in, The cluster centers are the cluster centers after iteration at time t+1 in the clustering process. Let max be the cluster center after iteration t in the clustering process. k The maximum movement of local cluster centers. For the i-th power data point at time t+1 in C k Attribution quantity, Let the i-th power data point at time t be located in C. k The upper-level assignment is defined as follows: ε is the threshold for the movement of local cluster centers during the clustering process, δ is the threshold for the change in the assignment of power data points, N is the number of power data points, and K is the number of local cluster centers.

[0135] When all local cluster centers and electronic data point assignments satisfy the convergence condition, the adaptive dynamic clustering update method terminates and outputs the final local clustering result as the power data clustering result. The power data clustering result includes: the final local cluster center μ. j .

[0136] The multidimensional data features include one or more of the following: data type, data size, access frequency, data relevance, timeliness, and access pattern; the clustering termination condition is that the movement of the local cluster centers meets a preset movement threshold, and the change in the affiliation of the power data points meets a preset change threshold.

[0137] In one implementation, the method of using an adaptive dynamic clustering update method combined with a boundary fuzzing processing mechanism to dynamically update the local cluster centers and optimize the boundary data of the power data points, generating local clustering results, includes:

[0138] From the power data points, select the power data point with the widest coverage as the initial center;

[0139] For example, select the power data points with the widest coverage among all power data points as the initial centers. The formula for calculating the initial centers is defined as follows:

[0140]

[0141] Where C is the initial center, argmax is the argmax function, and x i For the i-th power data point, x j Let d(F′(x) be the j-th power data point, D be the set of power data points, and d(F′(x)) be the power data point set. i ),F′(x j F′(x) is the weighted feature space. i ) and F′(x j The Euclidean distance F′(x) i F′(x) is the feature weighting vector of the i-th power data point. j ) is the feature weighting vector of the j-th power data point, ε is the preset distance threshold, and δ is the indicator function, which returns 1 or 0.

[0142] Select the power data point that is farthest from the initial center from the power data points, and use it as the initial local cluster center;

[0143] For example, based on the initial center C, a set of electronic data points farthest from C is selected from all electronic data points as the initial local cluster centers C. k This ensures that the initial center has a large coverage area.

[0144] Calculate the distance metric in each clustering process, and perform local clustering on the power data points based on the initial local cluster centers and the distance metric to obtain multiple local clusters;

[0145] With the addition of new electronic data points, dynamic optimization of data distribution is achieved through dynamic adjustment of local cluster centers and real-time updates of electronic data point affiliation, in order to adapt to the ever-changing power system data environment.

[0146] For example, the distance metric is dynamically adjusted during each clustering iteration to enhance the influence of different features. The adjusted distance metric is:

[0147]

[0148] Wherein, d(F′(x) i ),C k ) is F′(x i ) and C k Distance metric, F′(x) i Let C be the feature weighting vector of the i-th power data point. k As the initial local cluster centers, α j w is a dynamically adjusted coefficient used to enhance or weaken the influence of a certain feature. j f represents the weights of the eigenvectors in dimension j. ij Let c be the eigenvector of the i-th electronic data point in dimension j. kj The initial local cluster center C k The value of the feature in the j-th dimension, where m is the number of dimensions;

[0149] Power data point x i It is assigned to the initial local cluster center C that is closest to it. k Specifically, if power data point x i The power data point x is the one whose distance to the local cluster center in a certain clustering process is the smallest. i It is classified into a local cluster center in the clustering process.

[0150] A local cluster center update calculation formula is used to dynamically adjust the local cluster centers of each local cluster to obtain dynamically adjusted local cluster centers.

[0151] For example, the initial local cluster centers after classification need to be dynamically adjusted; in each iteration, the local cluster centers are recalculated based on the positions of all the electronic data points to which they belong, so that the local cluster centers gradually approach the average position of the electronic data points during the clustering process. For each cluster center, the weighted average on each dimension is calculated, and the formula for updating the local cluster centers is as follows:

[0152]

[0153] in, Let x be the cluster center after iteration at time t+1 in the clustering process.i For the i-th power data point, cluster k α represents the number of clusters. j For dynamically adjusted coefficients, w j Let F′(x) be the weight of the eigenvector in dimension j. i ) is the feature weighting vector of the i-th power data point;

[0154] After the cluster centers are updated during the clustering process, the distance between the data points and each center is recalculated based on the new cluster centers, and the data points are assigned to different clusters. Through multiple clustering iterations, local clustering results are obtained.

[0155] By employing a fuzzy attribution metric, the multidimensional data features at the fuzzy boundaries of the local clusters are adjusted for multi-cluster attribution to obtain the local clustering results.

[0156] For example, a boundary fuzzy processing mechanism is introduced, allowing electronic data points at the boundary to belong to multiple local clusters. The fuzzy affiliation metric is calculated as follows:

[0157]

[0158] Among them, P ik For electronic data point x i The probability of belonging to local cluster k, d(F′(x) i ),C k ) is F′(x i ) and C k Distance metric, F′(x) i Let C be the feature weighting vector of the i-th power data point. k β is the initial local cluster center, θ is the boundary ambiguity adjustment parameter, and θ is the boundary determination threshold.

[0159] For electronic data points at fuzzy boundaries, their assignment probability P is used as a basis. ik Perform multi-local clustering adjustment and calculate the contribution of multi-attributed electronic data points to each local cluster center.

[0160] In one implementation, the construction of the cross-device index includes:

[0161] The multi-dimensional feature correlation of all electronic devices is calculated using the electronic device feature correlation calculation formula.

[0162] For example, the formula for calculating the correlation of electronic device features is defined, taking into account differences in data feature distribution, access paths, and network conditions, as follows:

[0163]

[0164] Among them, CorAB For the correlation of multidimensional features on electronic devices A and B, A ik Let B be the value of the i-th electronic device in the k-th dimension feature. jk Let σ be the value of the j-th electronic device in the k-th dimension, m be the number of dimensions, and σ be the value of the k-th electronic device. k Let dis(A) be the standard deviation of the k-th feature. i B j The distance or network latency between electronic device A and electronic device B is represented by the formula, which is used to evaluate the correlation of similar data on different devices.

[0165] Based on the correlation of the multi-dimensional features, the electronic devices are combined to generate multiple electronic device groups;

[0166] For example, the multi-dimensional feature correlation between all electronic devices is calculated, and multiple device groups are formed based on the level of multi-dimensional feature correlation. The number of devices in each group is dynamic and has no fixed limit.

[0167] A hierarchical index structure is adopted to hierarchically classify electronic data and electronic devices in each electronic device group, generating a provisional cross-device index hierarchy.

[0168] For example, a cross-device index is built for each electronic device group. For each electronic device group, a hierarchical index structure is constructed to hierarchically divide the electronic data and electronic devices in the electronic device group, in order to optimize data retrieval efficiency and provide efficient cross-device access capabilities.

[0169] The provisional cross-device index hierarchy is optimized using the pre-obtained power equipment score sequence to obtain the final cross-device index hierarchy;

[0170] The optimal cross-device retrieval path for the cross-device index hierarchy is determined by using a cross-device retrieval path calculation formula, thus completing the construction of the cross-device index.

[0171] For example, based on the enclosing interval B j Divide the index into segments; to optimize cross-device retrieval paths, introduce a cross-device retrieval path calculation formula:

[0172]

[0173] Where Path is the optimal path for cross-device retrieval, argmin is the argmin function, and Cor AB For the correlation of multidimensional features on electronic devices A and B, dis(A) i B j ) represents the distance between the i-th electronic device A and the j-th electronic device B, and γ represents the distance between them. yY represents the path selection weight of electronic devices, and Y is the number of path selection weights for electronic devices.

[0174] The cross-device retrieval path calculation formula optimizes the path by minimizing the overall cost to determine the optimal retrieval path. This invention employs an adaptive dynamic clustering update method based on weighted feature space to accurately classify and archive electronic data, reducing the retrieval scope of irrelevant electronic data and thus accelerating data access speed. A boundary fuzziness processing mechanism is introduced, allowing electronic data points at boundaries to belong to multiple local clusters, optimizing local clustering. This invention constructs a cross-device index for cross-regional power equipment, considering factors such as network latency, data feature similarity, and correlation. This cross-device index allows for rapid location of relevant data across devices, improving data retrieval efficiency. The multi-dimensional feature correlation on electronic devices in this invention comprehensively considers differences in the feature distribution of electronic devices, access paths, and network conditions. Therefore, it can form highly correlated groups of electronic devices based on multi-dimensional feature correlation. The hierarchical index structure of this invention optimizes retrieval efficiency and provides efficient cross-device access capabilities.

[0175] In one implementation, the process of generating the power equipment score sequence includes:

[0176] Calculate the correlation between the multi-dimensional features of each power device and each local cluster center;

[0177] For example, firstly, the characteristics of power equipment are represented in a multi-dimensional way. Considering the following main characteristics and introducing the interrelationships between electronic devices, the rationality of the electronic device score ranking can be improved. The main characteristics of power equipment include: equipment computing power P. i Device storage capacity C i Equipment position effect L i Inter-device correlation R ij (Interconnectivity between devices), M i The degree of matching between the electronic data stored on the device and the final local cluster centers reflects the correlation between electronic devices and electronic data.

[0178] The feature vector of each power device is represented as X. i :

[0179] X i =[P i C i M i ,L i ]

[0180] Among them, X i Let P be the multi-dimensional feature vector of the i-th electronic device. i Let C be the computing power of the i-th electronic device.i Let M be the storage capacity of the i-th electronic device. i Let L be the matching degree between the electronic data stored on the i-th electronic device and the local cluster centers. i This refers to the positional effect of the i-th electronic device;

[0181] The correlation between electronic devices can be represented as matrix R:

[0182] R = [R ij ] N×N

[0183] Where R represents the correlation between electronic devices, R ij Let N represent the correlation between the i-th electronic device and the j-th electronic device, and N be the number of electronic devices.

[0184] R ij The calculation formula is as follows:

[0185] R ij =f(X) i ,X j )

[0186] Among them, R ij Let f be the correlation between the i-th electronic device and the j-th electronic device, and let X be the correlation degree calculation function. i Let X be the multi-dimensional feature vector of the i-th electronic device. j Let j be the multi-dimensional feature vector of the j-th electronic device;

[0187] The ranking of electronic devices needs to consider their association with the final local cluster centers. For each electronic device, the correlation μ between it and each final local cluster center is calculated. j The correlation between the electronic device and the final local cluster center is defined as:

[0188]

[0189] Where d(X) i ,μ j ) is X i With μ j The distance, X i For the multi-dimensional feature components of the i-th electronic device, μ j w represents the final local cluster centers after clustering. j Let X be the weight of the eigenvector in dimension j. ij Let μ be the value of the i-th electronic device on the j-dimensional eigenvector. kj This represents the value of the k-th local cluster center on the j-dimensional eigenvector after clustering.

[0190] The association between electronic devices and local cluster centers is defined as follows:

[0191]

[0192] Among them, R ik Let d(X) represent the association between the i-th electronic device and the k local cluster centers after clustering. i ,μ j ) is X i With μ j The distance, X i Let μ be the multi-dimensional feature vector of the i-th electronic device. j These are the final local cluster centers after clustering;

[0193] This allows electronic devices to be more closely associated with local cluster centers as they are closer together.

[0194] To ensure the reasonableness of the electronic device score ranking, considering the interrelationships between electronic devices, the following weighted feature calculation formula is proposed:

[0195]

[0196] Among them, X i R' is the weighted feature vector of the electronic device, used for ranking calculation. ij For the correlation between the i-th electronic device and the j-th electronic device, X j Let be the multi-dimensional feature vector of the j-th electronic device, and N be the number of electronic devices;

[0197] The correlation between electronic devices can be measured by feature similarity and geographical location. The correlation calculation formula is R. ij It can be expressed based on the following factors:

[0198] R ij =α×similarity(X) i ,X j )+β1×proximity(L i ,L j )

[0199] Among them, R ij For the correlation between the i-th electronic device and the j-th electronic device, similarity(X) i ,X j ) is X i With X j Similarity between them, X i Let X be the multi-dimensional feature vector of the i-th electronic device. j Let proximity(L) be the multi-dimensional feature vector of the j-th electronic device. i ,L j ) is Li With L i The degree of closeness, L i Let L be the location of the i-th electronic device. j For the j-th electronic device, α and β1 are the weighting coefficients that control these two factors.

[0200] Based on the relevant parameters for power equipment ranking and the correlation degree, the ranking score of each power equipment is calculated using the power equipment ranking score calculation formula, the power equipment scores are ranked, and a provisional power equipment score sequence is generated.

[0201] Example: Power equipment ranking score S i Taking into account the characteristics of the device itself and its correlation with local cluster centers, the formula is:

[0202]

[0203] Among them, S i Let X be the ranking score of the i-th electronic device, γ be the weight balancing the characteristics of the electronic device and the characteristics of the electronic data, and X be the ranking score of the i-th electronic device. i R' is the weighted feature vector of the electronic device. ik f represents the correlation between the i-th electronic device and the k-th local cluster center after clustering. k Let K be the number of local cluster centers, representing the data characteristics corresponding to local cluster k, such as data size, access frequency, and timeliness. A dynamic adaptive ranking model is used to dynamically adjust the current power equipment score sequence for each electronic device in the provisional power equipment score sequence based on its current state and historical performance.

[0204] For example, to adapt the ranking to changes in the dynamic environment, a dynamic adaptive ranking model is introduced. This model is based on the current state and historical performance of the electronic device. At each time step t, a comprehensive ranking score for the electronic device is calculated, and the historical performance component is updated according to the time window length. The objective function of the dynamic adaptive ranking model is as follows:

[0205]

[0206] Among them, S i (t) represents the ranking score of electronic devices at time t, i.e., the current score of power equipment, where γ is the weight balancing the characteristics of electronic devices and the characteristics of electronic data, and X i R' is the weighted feature vector of the electronic device. ik P represents the association between the i-th electronic device and the k-th local cluster center after clustering. k This represents the historical performance of electronic devices in processing local cluster k data over past time steps.

[0207] The equipment sorting is dynamically adjusted over time to adapt to changes in the environment; based on the current sorting score S of the equipment. i (t) sorts all electronic devices, prioritizing those with higher scores. This sorting can be used to optimize cross-rate index hierarchies and data scheduling, ensuring efficient access to data stored on different devices.

[0208] The multi-dimensional features of the power equipment include one or more of the following: equipment computing power, equipment storage capacity, equipment location effect, and the matching degree between the power data stored on the equipment and local cluster centers. After completing electronic data clustering, this invention sorts the power equipment based on the electronic data clustering results and the power equipment features. The extracted features of the power equipment include: storage capacity, processing power, network bandwidth, energy consumption, and efficiency. Through multi-dimensional feature extraction and sorting, combined with a weighted feature model and a dynamic adaptive sorting algorithm, the priority level of power equipment retrieval is optimized, improving the response speed of data access. The weighted feature model considers the power equipment, data clustering results, and the correlation between devices. The adaptive sorting algorithm is based on the current state and historical performance of the equipment. This invention extracts multi-dimensional features of electronic equipment, including storage capacity, processing power, network bandwidth, and energy consumption. Meanwhile, the correlation between electronic devices is further calculated based on geographical location and interdependence to achieve comprehensive electronic device ranking; based on the multi-dimensional characteristics of power equipment and local clustering results, the present invention uses a dynamic ranking algorithm to prioritize the devices, ensuring that data access requests are preferentially allocated to efficient devices, optimizing the data retrieval order across devices, thereby improving response speed and overall system efficiency; the power equipment score sequence can be used to optimize cross-device indexing hierarchy and data scheduling, ensuring efficient access to data stored on different electronic devices.

[0209] In one implementation, the global index structure includes a global clustering index layer and a power equipment index layer; the construction of the global index includes:

[0210] Perform global standardization on local cluster centers and multidimensional data features;

[0211] For example, local index information of electronic data points on power equipment, including the local cluster centers μ of electronic data points on each electronic device. j Data range R j and access frequency F j Extract the local cluster centers, data ranges, and access frequencies of electronic data points from each electronic device, and then perform global standardization processing.

[0212] The local cluster centers after global standardization are then subjected to global clustering to obtain global clustering results and global clustering labels; the global clustering results include: multiple global clusters and multiple global cluster centers;

[0213] For example, in the cloud, the standardized local cluster centers are used for global clustering to obtain the global cluster center v. j and global clustering label G ij The expression for the global clustering label is as follows:

[0214]

[0215] Among them, G ij The global cluster label is argmin, and μ is the argmin function. jk The local cluster centers μ after clustering j In a k-dimensional vector, v jk v is the global cluster center j A vector with k dimensions, where m is the number of dimensions;

[0216] The local cluster of each electronic device is assigned to a global cluster, and its corresponding global cluster center v is recorded. j .

[0217] For each global cluster, a corresponding global cluster index node is designed to complete the construction of the global cluster index layer; the global cluster index node includes: global cluster center, related device list, and global priority;

[0218] For example, the global index is divided into two layers: the first layer is the global clustering index layer, and the second layer is the electronic device index layer.

[0219] For each global cluster, a global cluster index node is created. The global cluster index node records the global cluster center v. j Related Equipment List D j And global priority, the formula for calculating global priority is as follows:

[0220]

[0221] Among them, P g This is the global priority, or global query priority, and represents the importance score of the electronic device during the query process. (F) j For the characteristics of the j-th electronic device, dist(X) j ,v j ) represents the multi-dimensional feature vector of the j-th electronic device and the global cluster center v. j The distance is N, where N is the number of features of the electronic device.

[0222] The specific construction process of the global clustering index layer is as follows: Centered on global clustering, it records all global cluster centers and their related attributes. Each global clustering index node corresponds to a global cluster, and the index information includes: global cluster center, related device list, and global priority. It is used to quickly locate the global cluster center containing the target data. Based on the global priority, it prioritizes searching high-priority local clusters, reducing the search scope.

[0223] For each electronic device in a global cluster, design a corresponding device index node to complete the construction of the power device index layer;

[0224] For example, the power equipment index layer is sorted by global priority, with higher priority data being retrieved first.

[0225] The construction process of the power equipment index layer involves: based on the relevant equipment list provided by the global clustering index layer, sorting and indexing optimization are performed on specific electronic devices under each local cluster. Index information includes: the device's unique ID, the device's data range, access frequency, and the distance between the electronic device and the local cluster center. The priority of a specific device within a cluster is determined, prioritizing access to high-priority devices to further quickly locate data within the target device.

[0226] The device index nodes include: the unique ID of the electronic device, the range of electronic data, the access frequency, and the distance between the electronic device and the local cluster center. After establishing local and cross-device indexes at the device index layer, this invention needs to establish a global index at the central control center to globally cluster the clustering information on these different electronic devices, generating a higher-level global clustering structure. The central control center can uniformly manage these local clusters, thereby avoiding repeated device searches and reducing the overall search scope. The central control center integrates all local indexes and creates global clusters, optimizing the search path through recursive aggregation algorithms, reducing redundant device access, and thus achieving fast access and centralized management of cross-device data. By assigning global labels to local clusters, the central control center can quickly find the devices or device groups storing relevant data through global labels, and use recursive aggregation algorithms to optimize the search path and accelerate data access. This invention establishes a global index, integrating dynamic local indexes and cross-device indexes into a unified global index management mode, significantly improving the efficiency of cross-device data retrieval. The introduction of multi-level recursive aggregation search algorithms can effectively optimize the search path, avoid repeated device access, and further accelerate data acquisition.

[0227] In one implementation, the data reorganization rule includes a cross-block association verification mechanism and a dynamic reorganization optimization algorithm; the step of using an improved block-based homomorphic encryption method combined with the data reorganization rule to feed the search data back to the user includes:

[0228] The search data is segmented to generate multiple data blocks, and each data block is marked with a segmentation point.

[0229] For example, the size B of each data block is dynamically determined based on the stability parameters of the transmission network (such as network latency and packet loss rate) and the size of the data block. i The calculation formula is as follows:

[0230]

[0231] Among them, B i Let L be the size of the i-th data block. D Where α1 is the length of the search data, T is the network settling time, and α1 and β2 are dynamic adjustment coefficients.

[0232] The search data is divided into blocks according to the size determined above, and the position P of each cut point is marked. i Record the relationships between data blocks.

[0233] The tangent marker is used to ensure that the order of data blocks is not lost during transmission; each data block S i After being cut, a unique cut point marker is generated for it to ensure the integrity and order of the data.

[0234] Generate the hash value H(S) for each data block i ), used for subsequent verification; combining the order information of the data blocks and the hash value to generate the cutoff point marker M. i :

[0235]

[0236] Among them, M i Let σ be the cutoff point marker, and σ be an activation function used to apply the hash value H(S) to the hash value. i Mapping H(S) i ) is data block S i Hash values ​​are used to ensure the integrity of data blocks. The XOR operation is used to mix the hash value of the data block with other information, and γ(s,T) is a correction factor generated based on the data block number S and the network stability parameter T.

[0237] In this way, the information carried by the data block includes not only the data itself, but also its location and network characteristics during transmission.

[0238] Each data block and the cut-off point marker are homomorphically encrypted, and the homomorphically encrypted data block is embedded with the data block number to generate an encrypted encoded data block.

[0239] For example, during data transmission, an innovative block-coded homomorphic encryption method is used to ensure that the data is encrypted during transmission, and that each data block carries its sequence information.

[0240] For each data block S i Perform homomorphic encryption and embed the data block number:

[0241] E(S i )=(S i +μ·s) e mod q

[0242] Among them, E(S) i S is a homomorphically encrypted data block. i For data blocks, μ is a weighting factor used to distinguish data block numbers, e and q are encryption parameters, mod is the modulo operator, and s is the data block number;

[0243] The encrypted data blocks not only encrypt the content but also contain block number information, which is used for order recovery during cloud reassembly.

[0244] The cutoff point marker for each data block also needs to be encrypted to ensure that the cutoff point location is not tampered with or lost. The cutoff point marker is encrypted using the data block number, as follows:

[0245]

[0246] Among them, C(S) i ) represents the encrypted pointcut marker, Extract(S i The last 8) is from data block S i Extract the last 8 bits of information, where H represents the hash function. This indicates an XOR operation, where μ is a weighting factor used to distinguish data block numbers, and s is the data block number.

[0247] This encrypted calculation ensures the security of the cutpoint markers, preventing them from being tampered with during transmission.

[0248] A cross-block association verification mechanism combined with a dynamic reorganization optimization algorithm is used to reorganize the encrypted data block according to the embedded data block number to generate a reorganized data block;

[0249] For example, at the central control center, after receiving a data block, it is necessary to decrypt the encrypted data block and reassemble the data according to the order information embedded during encryption.

[0250] The data block is decrypted according to the formula used during encryption. The decryption expression is as follows:

[0251] D i =(E(S)i )) d mod q-μ·s

[0252] Among them, D i For the decrypted data block S i E(S) i ) represents the encrypted data block, d represents the decryption exponent during the decryption process, which is usually the inverse of the encryption exponent; μ is the weighting factor used to distinguish the data block number, s is the data block number, μ·s is used to restore the original data during decryption, mod represents the modulo operation to ensure that the decrypted data conforms to the range of the original data, and q is the encryption parameter;

[0253] After decryption, the original data block is recovered by subtracting the block number information μ·s.

[0254] To ensure the correctness of data block order and cutoff point location, a cross-block association verification mechanism is introduced. This mechanism verifies the data by performing an XOR operation on the hash values ​​of adjacent data blocks. The expression for the cross-block verification mechanism is as follows:

[0255]

[0256] Among them, Check(S) i H(S) is an expression used to verify whether the order of data blocks and the position of the cutoff point are correct. i ) is data block S i The hash value, H(S) i-1 ) is data block S i-1 The hash value, H(S) i+1 ) is data block S i+1 hash value, This is an XOR operation used to mix the hash values ​​of consecutive data blocks, ensuring the correct order. This is another cryptographic operation used to further mix the hash values ​​of the current data block with those of the preceding and following data blocks;

[0257] This cross-block verification mechanism can effectively detect lost or erroneous blocks, ensuring the accuracy of the data reconstruction process.

[0258] To further improve the efficiency of data reassembly in the central control center, a dynamic reassembly optimization algorithm is proposed. This algorithm dynamically adjusts the reassembly order and strategy based on the packet loss rate and the difficulty of data block reassembly during transmission.

[0259] Based on the packet loss rate reported during transmission, adjust the priority of reassembling data blocks:

[0260]

[0261] Wherein, P(S) i ) is data block S iThe priority is δ, which is the adjustment coefficient, and Loss(i) is the estimated packet loss rate of the i-th reassembled data block.

[0262] If errors or delays occur during the transmission of certain data blocks, the central control center can infer or correct them based on previous data blocks. The specific correction formula is as follows:

[0263]

[0264] in, For the updated data block S i H(S) i-1 ) is data block S i-1 The hash value, H(S) i+1 ) is data block S i+1 hash value, This is an XOR operation used to combine the hash values ​​of preceding and following data blocks.

[0265] In this way, even if some data blocks are lost, the central control center can perform a certain degree of speculative recovery, ensuring the continuity of the reorganization.

[0266] Send the reconstructed data block to the user;

[0267] The cross-block association verification mechanism verifies the hash values ​​of adjacent data blocks by performing an XOR operation; the dynamic reassembly optimization algorithm dynamically adjusts the order and strategy of data reassembly based on the packet loss rate and reassembly difficulty of data blocks during transmission; this invention performs homomorphic encryption on the searched data information at the device layer. To improve transmission efficiency, the data needs to be transmitted in blocks. The data is cut into blocks at the device layer and reassembled at the central control center. The invention designs reassembly rules, innovatively combines them with encryption methods, and explains how to mark the cutting. This invention employs data block encryption transmission technology. Power equipment transmits data in blocks with encryption. The central control center decrypts and reassembles the data using segmentation markers and hash verification techniques, ensuring data integrity and correctness. A dynamic reassembly algorithm enhances the reliability of data recovery. This invention addresses different network conditions by dividing power equipment into blocks and using homomorphic encryption to ensure data security during transmission. The block-based data is transmitted to the central control center in encrypted form. Each block generates a segmentation marker to ensure data order during transmission, and hash verification ensures data integrity after transmission. Upon receiving the data blocks, the central control center decrypts and reassembles the data based on the segmentation markers and encryption numbers. A dynamic reassembly algorithm adjusts the reassembly priority in real time, ensuring data recovery even during network fluctuations. This invention uses homomorphic encryption during data transmission between power equipment, combined with segmentation markers, to ensure data transmission integrity and correctness. The dynamic reassembly algorithm adjusts priorities based on the data block transmission status, ensuring the integrity and reliability of data reassembly and adapting to the power data management needs in complex network environments.

[0268] This invention can be widely applied to cross-regional data management in large-scale power systems, especially in power systems with smart grid functionality. With the increasing number of power devices in various locations, devices within the system need to transmit and access large amounts of data in real time between different regions. Therefore, the implementation of this invention helps achieve efficient management and secure transmission of power data. First, at the device management layer, electronic data is clustered based on a weighted feature space. An adaptive dynamic clustering update method is used to dynamically adjust the data point attribution, and fuzzy boundary processing and cluster center convergence algorithms are used to improve the accuracy of clustering. Second, for power devices, this invention optimizes the device retrieval order and improves the response speed of data access by extracting and ranking multi-dimensional features, combined with a weighted feature model and a dynamic adaptive ranking algorithm. Subsequently, by establishing local and cross-device indexes on power equipment and combining global indexes with recursive aggregation algorithms, the data indexes are uniformly managed in the central control center, significantly improving the efficiency of cross-device data retrieval. Finally, data block encryption transmission technology is adopted, where power equipment transmits data in blocks with encryption, and the central control center decrypts and reassembles the data using point marking and hash verification technology to ensure data integrity and correctness. Furthermore, a dynamic reassembly algorithm enhances the reliability of data recovery. Overall, this invention effectively improves the security and efficiency of cross-regional and cross-domain data interaction, and is suitable for data management and sharing scenarios in complex distributed network environments.

[0269] This invention introduces a power microservice architecture to achieve secure and efficient access to power equipment data, effectively solving the problems of efficient retrieval and secure transmission of data across regions and devices in distributed power systems. Through adaptive dynamic clustering based on weighted feature space, data can be accurately classified and archived, reducing the scope of irrelevant data retrieval and thus accelerating data access speed. By establishing a global index in the central control center and integrating local indexes into a unified global index management model, the efficiency of cross-device data retrieval is significantly improved. The introduction of a recursive aggregation algorithm effectively optimizes the retrieval path, avoids repeated device access, and further accelerates data acquisition. Data block encryption transmission technology is adopted, performing homomorphic encryption processing when transmitting data between power devices, combined with pointcut marking and hash verification technology, to ensure the integrity and correctness of data transmission. Through a dynamic reassembly algorithm, priority is adjusted according to the data block transmission status to ensure the integrity and reliability of data reassembly, adapting to the power data management needs in complex network environments.

[0270] like Figure 2As shown, the method of this invention combines data clustering, device sorting, index building, multi-level cluster recursive search, and data block transmission to form a microservice architecture. In the data clustering process, electronic data is first collected and preprocessed. Then, multi-dimensional features are extracted from the electronic data. After adaptive local clustering initialization of the extracted multi-dimensional features, the electronic data point with the widest coverage is selected as the initial center. From the collected electronic data dataset, the electronic data point farthest from the initial center is selected as the initial local cluster center. The distance metric calculation formula in the clustering process is dynamically adjusted to update the local cluster centers. Boundary fuzzing is introduced to adjust the clustering affiliation of electronic data points at the boundaries, completing the local clustering of electronic data points. In the device sorting process… In this process, multidimensional features of electronic devices are extracted. Weighted feature calculations based on the correlation between electronic devices and the local clustering results of electronic data points are performed. Combined with the current status and historical performance of the electronic devices, the electronic devices are ranked. The index building process includes building dynamic local indexes, cross-device indexes, and global indexes. When building dynamic local indexes, the relationships between data points are ranked according to the weighted distance, placing data points with smaller distances at the top level of the index. A dynamic hierarchical index structure is introduced to complete the construction of the dynamic local index. When building cross-device indexes, a cross-device correlation calculation formula is first defined. Based on this formula, electronic device groups are formed, and a cross-device index is built for each group to determine the optimal cross-device index. The retrieval path completes the construction of the cross-device index. During the construction of the global index, the local clustering information of electronic devices is globally standardized at the central control center. Local cluster centers are then globally clustered, assigned global cluster labels, and global cluster index nodes are created, prioritizing the retrieval of high-priority data. In the multi-level clustering recursive search process, the keywords and features of the query request are mapped to the global index to initially determine the location of the target data. High-priority global clusters and related devices are prioritized for retrieval, and the search path is dynamically adjusted. Then, cross-device relevance calculations are used to optimize the retrieval order, sequentially accessing related electronic devices. Within the local index of each electronic device, the most relevant local cluster centers are located, and high-priority electronic devices are prioritized for retrieval. Data points are collected, and the search results are summarized and sorted according to query relevance. The index priority is merged and updated to complete multi-level clustering recursive search. During the data block transmission process, firstly, the search data is input, the data block size is calculated, the data blocks are cut and marked, the data block hash value and label are generated, the data is encrypted, and the data blocks are transmitted. The received encrypted data blocks are decrypted, cross-block verification, data reorganization and verification are performed, and dynamic reorganization optimization is carried out to complete the data block transmission. This invention aims to solve the problems of security, real-time performance and cross-device retrieval efficiency in power equipment data access, and provides a technical solution for achieving secure and fast data access in a distributed power microservice architecture, which is suitable for data management and sharing applications in complex power network environments.The introduction of microservice architecture provides a new solution for distributed data management in power systems, such as; Figure 3 As shown, a microservice architecture can decompose complex system functions into multiple independent services, processed at the user layer, central control center, and device management layer respectively. These services collaborate with each other through lightweight communication protocols, greatly improving the system's flexibility and scalability. In power systems, using a microservice architecture to manage and access data between power devices can effectively solve the performance bottlenecks of traditional centralized management, while also improving the efficiency of cross-device data access.

[0271] Example 2:

[0272] Based on the same inventive concept, this invention patent application also provides a data security and efficient access system, such as... Figure 4 As shown, it includes: an access request receiving module, a search data aggregation module, and a search data feedback module;

[0273] The access request receiving module is used to receive user access requests;

[0274] The search data aggregation module is used to perform data search by using a multi-level recursive aggregation search algorithm according to the access request, sequentially through a pre-established global index, a cross-device index, and a dynamic local index, to obtain aggregated search data.

[0275] The search data feedback module is used to feed the search data back to the user using an improved block-coded homomorphic encryption method combined with data recombination rules;

[0276] The global index is constructed based on the local clustering results of electronic data and the multi-dimensional features of electronic devices, combined with a global index structure; the cross-device index is constructed based on the correlation of multi-dimensional features of electronic devices, combined with a hierarchical index structure; and the dynamic local index is constructed based on power data, combined with a dynamic hierarchical index structure.

[0277] Preferably, the search data aggregation module is specifically used for:

[0278] Extract key query parameters from the access request;

[0279] The query key parameters are mapped to the global cluster labels of the global index to initially determine the global cluster center, global cluster, and related device list containing the target data;

[0280] Based on the global cluster center, global cluster, and related device list containing the target data, the most relevant global cluster is selected on the global cluster index layer of the global index. After searching for the target data in the most relevant global cluster, the most relevant power equipment is selected on the power equipment index layer of the global index according to the related device list and the calculated global priority of electronic equipment.

[0281] Using the cross-device index, based on the calculated multi-dimensional feature correlation between the most relevant power equipment and each electronic device group, electronic devices that exceed the multi-dimensional feature correlation threshold are filtered out, and target data is searched by accessing the electronic devices that exceed the multi-dimensional feature correlation threshold;

[0282] Using the local dynamic index of the electronic device that exceeds the multi-dimensional feature relevance threshold, and based on the query key parameters, after locating the most relevant local cluster center, the target data is searched according to the index priority;

[0283] All target data is aggregated, and the aggregated target data is processed to serve as the final search data.

[0284] The key parameters for querying include one or more of the following: query keywords, query time range, and required data type; the data processing includes: target data sorting, target data merging, target data deduplication, and target data filtering.

[0285] Preferably, the dynamic local index structure includes a clustering center layer and a data point ranking layer; the system further includes a dynamic local index construction module, which includes:

[0286] The local clustering submodule is used to perform local clustering of power data on all power equipment to obtain local clustering results; the local clustering results include: multiple local clusters and multiple local cluster centers;

[0287] The power data point sequence generation submodule is used to calculate the multidimensional weighted distance based on the power data points after local clustering and the local cluster centers obtained from the local clustering results, and to sort the power data points corresponding to the multidimensional weighted distance to generate a power data point sequence.

[0288] The cluster center layer construction submodule is used to design a corresponding local clustering index for each of the local clusters and complete the construction of the cluster center layer;

[0289] The data point sorting layer construction submodule is used to place the power data points into the corresponding local clustering index. Based on the power data point sequence and according to the index priority calculation formula, the index position of the power data point in each local clustering index is determined, thus completing the construction of the data point sorting layer.

[0290] Preferably, the local clustering submodule includes:

[0291] A multidimensional data feature extraction unit is used to extract multidimensional data features of multiple power data from all power equipment.

[0292] The feature mapping unit is used to map the multidimensional data features to a weighted feature space, and then perform feature weighting on the multidimensional data features to obtain power data points.

[0293] The local clustering unit is used to dynamically update the local cluster centers and optimize the boundary data of the power data points by using an adaptive dynamic clustering update method combined with a boundary fuzzing processing mechanism, and to generate local clustering results.

[0294] The power data clustering result generation unit is used to terminate the adaptive dynamic clustering update method when the local clustering result meets the clustering termination condition, and to use the local clustering result as the power data clustering result;

[0295] The multidimensional data features include one or more of the following: data type, data size, access frequency, data relevance, timeliness, and access pattern; the clustering termination condition is that the movement of the local cluster centers meets a preset movement threshold, and the change in the affiliation of the power data points meets a preset change threshold.

[0296] Preferably, the local data clustering unit is specifically used for:

[0297] From the power data points, select the power data point with the widest coverage as the initial center;

[0298] Select the power data point that is farthest from the initial center from the power data points, and use it as the initial local cluster center;

[0299] Calculate the distance metric in each clustering process, and perform local clustering on the power data points based on the initial local cluster centers and the distance metric to obtain multiple local clusters;

[0300] A local cluster center update calculation formula is used to dynamically adjust the local cluster centers of each local cluster to obtain dynamically adjusted local cluster centers.

[0301] By employing a fuzzy attribution metric, the multidimensional data features at the fuzzy boundaries of the local clusters are adjusted for multi-cluster attribution to obtain the local clustering results.

[0302] Preferably, the system further includes: a cross-device index building module; the cross-device index building module is used for:

[0303] The multi-dimensional feature correlation of all electronic devices is calculated using the electronic device feature correlation calculation formula.

[0304] Based on the correlation of the multi-dimensional features, the electronic devices are combined to generate multiple electronic device groups;

[0305] A hierarchical index structure is adopted to hierarchically classify electronic data and electronic devices in each electronic device group, generating a provisional cross-device index hierarchy.

[0306] The provisional cross-device index hierarchy is optimized using the pre-obtained power equipment score sequence to obtain the final cross-device index hierarchy;

[0307] The optimal cross-device retrieval path for the cross-device index hierarchy is determined by using a cross-device retrieval path calculation formula, thus completing the construction of the cross-device index.

[0308] Preferably, the system further includes: a power equipment score sequence generation module, used for:

[0309] Calculate the correlation between the multi-dimensional features of each power device and each local cluster center;

[0310] Based on the relevant parameters for power equipment ranking and the correlation degree, the ranking score of each power equipment is calculated using the power equipment ranking score calculation formula, the power equipment scores are ranked, and a provisional power equipment score sequence is generated.

[0311] A dynamic adaptive sorting model is adopted to dynamically adjust each electronic device in the provisional power equipment score sequence based on its current status and historical performance, thereby generating the current power equipment score sequence.

[0312] The multi-dimensional characteristics of the power equipment include one or more of the following: equipment computing power, equipment storage capacity, equipment location effect, and the matching degree between the power data stored on the equipment and the local cluster centers.

[0313] Preferably, the global index structure includes a global clustering index layer and a power equipment index layer; the system further includes a global index construction module; the global index construction module is used for:

[0314] Perform global standardization on local cluster centers and multidimensional data features;

[0315] The local cluster centers after global standardization are then subjected to global clustering to obtain global clustering results and global clustering labels; the global clustering results include: multiple global clusters and multiple global cluster centers;

[0316] For each global cluster, a corresponding global cluster index node is designed to complete the construction of the global cluster index layer; the global cluster index node includes: global cluster center, related device list, and local cluster global priority;

[0317] For each electronic device under a global cluster, a corresponding device index node is designed to complete the construction of the power equipment index layer; the device index node includes: the unique ID of the electronic device, the range of electronic data, the access frequency, and the distance between the electronic device and the local cluster center.

[0318] Preferably, the data reorganization rules include a cross-block association verification mechanism and a dynamic reorganization optimization algorithm; the search data feedback module is specifically used for:

[0319] The search data is segmented to generate multiple data blocks, and each data block is marked with a segmentation point.

[0320] Each data block and the cut-off point marker are homomorphically encrypted, and the homomorphically encrypted data block is embedded with the data block number to generate an encrypted encoded data block.

[0321] A cross-block association verification mechanism combined with a dynamic reorganization optimization algorithm is used to reorganize the encrypted data block according to the embedded data block number to generate a reorganized data block;

[0322] Send the reconstructed data block to the user;

[0323] The cross-block association verification mechanism verifies the hash values ​​of adjacent data blocks by performing an XOR operation; the dynamic reassembly optimization algorithm dynamically adjusts the order and strategy of data reassembly based on the packet loss rate and the difficulty of data block reassembly during transmission.

[0324] Example 3

[0325] like Figure 5 As shown, the present invention also provides an electronic device, which may be a computer device, a microcontroller device, a smart mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, processor, and transceiver component are connected via a bus; the memory can be used to store executable programs, and an exemplary executable program may include instructions; the processor is used to execute the instructions stored in the memory. The memory can also be used to store data, which can be accessed and / or modified when instructions are executed.

[0326] The processor may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and it is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions in the storage medium to implement the corresponding method flow or corresponding function, so as to realize the steps of a data security and efficient access method in the above embodiments.

[0327] Example 4

[0328] Based on the same inventive concept, this invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory). This readable storage medium is a memory device within an electronic device used to store programs and data. It is understood that the storage medium here can include both built-in storage media within the electronic device and extended storage media supported by the electronic device. The storage medium provides storage space, which stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more executable programs (including program code). It should be noted that the storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device. Loading and executing one or more instructions stored in the storage medium by the processor enables the implementation of the steps of the data security and efficient access method described in the above embodiments.

[0329] Those skilled in the art will understand that embodiments of this patent application can be provided as methods, systems, or computer program products. Therefore, this patent application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this patent application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0330] This patent application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the patent application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0331] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0332] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0333] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this patent application and not to limit its scope of protection. Although the patent application has been described in detail with reference to the above embodiments, those skilled in the art should understand that after reading this patent application, they can still make various changes, modifications or equivalent substitutions to the specific implementation of the application, but these changes, modifications or equivalent substitutions are all within the scope of protection of the claims pending approval.

Claims

1. A secure and efficient data access method, characterized in that, include: Receive user access requests; A multi-level recursive aggregation search algorithm is adopted. Based on the access request, data is searched sequentially through a pre-established global index, a cross-device index, and a dynamic local index to obtain the summarized search data. An improved block-based homomorphic encryption method combined with data recombination rules is used to feed the search data back to the user; The global index is constructed based on the local clustering results of electronic data and the multi-dimensional features of electronic devices, combined with a global index structure; the cross-device index is constructed based on the correlation of multi-dimensional features of electronic devices, combined with a hierarchical index structure; and the dynamic local index is constructed based on power data, combined with a dynamic hierarchical index structure.

2. The method as described in claim 1, characterized in that, The algorithm employs a multi-level recursive aggregation search. Based on the access request, it sequentially searches for data using a pre-established global index, a cross-device index, and a dynamic local index to obtain summarized search data, including: Extract key query parameters from the access request; The query key parameters are mapped to the global cluster labels of the global index to initially determine the global cluster center, global cluster, and related device list containing the target data; Based on the global cluster center, global cluster, and related device list containing the target data, the most relevant global cluster is selected on the global cluster index layer of the global index. After searching for the target data in the most relevant global cluster, the most relevant power equipment is selected on the power equipment index layer of the global index according to the related device list and the calculated global priority of electronic equipment. Using the cross-device index, based on the calculated multi-dimensional feature correlation between the most relevant power equipment and each electronic device group, electronic devices that exceed the multi-dimensional feature correlation threshold are filtered out, and target data is searched by accessing the electronic devices that exceed the multi-dimensional feature correlation threshold; Using the local dynamic index of the electronic device that exceeds the multi-dimensional feature relevance threshold, and based on the query key parameters, after locating the most relevant local cluster center, the target data is searched according to the index priority; All target data is aggregated, and the aggregated target data is processed to serve as the final search data. The key parameters for querying include one or more of the following: query keywords, query time range, and required data type; the data processing includes: target data sorting, target data merging, target data deduplication, and target data filtering.

3. The method as described in claim 2, characterized in that, The dynamic local index structure includes a cluster center layer and a data point ranking layer; the construction of the dynamic local index includes: Local clustering is performed on the power data of all power equipment to obtain local clustering results; the local clustering results include: multiple local clusters and multiple local cluster centers; Based on the power data points after local clustering and the local cluster centers obtained from the local clustering results, a multidimensional weighted distance is calculated. Based on the multidimensional weighted distance, the power data points corresponding to the multidimensional weighted distance are sorted to generate a power data point sequence. For each local cluster, a corresponding local cluster index is designed to complete the construction of the cluster center layer; The power data points are placed into the corresponding local clustering index. Based on the power data point sequence and according to the index priority calculation formula, the index position of each power data point in the local clustering index is determined, thus completing the construction of the data point sorting layer.

4. The method as described in claim 3, characterized in that, The process of performing local clustering on the power data of all power equipment to obtain local clustering results includes: Extracting multidimensional data features from multiple power data from all power equipment; After mapping the multidimensional data features to a weighted feature space, the multidimensional data features are weighted to obtain power data points; An adaptive dynamic clustering update method combined with a boundary fuzzing processing mechanism is used to dynamically update the local cluster centers and optimize the boundary data of the power data points, generating local clustering results. When the local clustering result meets the clustering termination condition, the adaptive dynamic clustering update method ends, and the local clustering result is used as the power data clustering result. The multidimensional data features include one or more of the following: data type, data size, access frequency, data relevance, timeliness, and access pattern; the clustering termination condition is that the movement of the local cluster centers meets a preset movement threshold, and the change in the affiliation of the power data points meets a preset change threshold.

5. The method as described in claim 4, characterized in that, The method employs an adaptive dynamic clustering update approach combined with a boundary fuzzing mechanism to dynamically update the local cluster centers and optimize the boundary data of the power data points, generating local clustering results, including: From the power data points, select the power data point with the widest coverage as the initial center; Select the power data point that is farthest from the initial center from the power data points, and use it as the initial local cluster center; Calculate the distance metric in each clustering process, and perform local clustering on the power data points based on the initial local cluster centers and the distance metric to obtain multiple local clusters; A local cluster center update calculation formula is used to dynamically adjust the local cluster centers of each local cluster to obtain dynamically adjusted local cluster centers. By employing a fuzzy attribution metric, the multidimensional data features at the fuzzy boundaries of the local clusters are adjusted for multi-cluster attribution to obtain the local clustering results.

6. The method as described in claim 2, characterized in that, The construction of the cross-device index includes: The multi-dimensional feature correlation of all electronic devices is calculated using the electronic device feature correlation calculation formula. Based on the correlation of the multi-dimensional features, the electronic devices are combined to generate multiple electronic device groups; A hierarchical index structure is adopted to hierarchically classify electronic data and electronic devices in each electronic device group, generating a provisional cross-device index hierarchy. The provisional cross-device index hierarchy is optimized using the pre-obtained power equipment score sequence to obtain the final cross-device index hierarchy; The optimal cross-device retrieval path for the cross-device index hierarchy is determined by using a cross-device retrieval path calculation formula, thus completing the construction of the cross-device index.

7. The method as described in claim 6, characterized in that, The process of generating the power equipment score sequence includes: Calculate the correlation between the multi-dimensional features of each power device and each local cluster center; Based on the relevant parameters for power equipment ranking and the correlation degree, the ranking score of each power equipment is calculated using the power equipment ranking score calculation formula, the power equipment scores are ranked, and a provisional power equipment score sequence is generated. A dynamic adaptive sorting model is adopted to dynamically adjust each electronic device in the provisional power equipment score sequence based on its current status and historical performance, thereby generating the current power equipment score sequence. The multi-dimensional characteristics of the power equipment include one or more of the following: equipment computing power, equipment storage capacity, equipment location effect, and the matching degree between the power data stored on the equipment and the local cluster centers.

8. The method as described in claim 2, characterized in that, The global index structure includes a global clustering index layer and a power equipment index layer; the construction of the global index includes: Perform global standardization on local cluster centers and multidimensional data features; The local cluster centers after global standardization are then subjected to global clustering to obtain global clustering results and global clustering labels; the global clustering results include: multiple global clusters and multiple global cluster centers; For each global cluster, a corresponding global cluster index node is designed to complete the construction of the global cluster index layer; the global cluster index node includes: global cluster center, related device list, and global priority; For each electronic device under a global cluster, a corresponding device index node is designed to complete the construction of the power equipment index layer; the device index node includes: the unique ID of the electronic device, the range of electronic data, the access frequency, and the distance between the electronic device and the local cluster center.

9. The method as described in claim 1, characterized in that, The data recombination rules include a cross-block association verification mechanism and a dynamic recombination optimization algorithm; the improved block-based homomorphic encryption method combined with the data recombination rules, and the feedback of the search data to the user, includes: The search data is segmented to generate multiple data blocks, and each data block is marked with a segmentation point. Each data block and the cut-off point marker are homomorphically encrypted, and the homomorphically encrypted data block is embedded with the data block number to generate an encrypted encoded data block. A cross-block association verification mechanism combined with a dynamic reorganization optimization algorithm is used to reorganize the encrypted data block according to the embedded data block number to generate a reorganized data block; Send the reconstructed data block to the user; The cross-block association verification mechanism verifies the hash values ​​of adjacent data blocks by performing an XOR operation; the dynamic reassembly optimization algorithm dynamically adjusts the order and strategy of data reassembly based on the packet loss rate and the difficulty of data block reassembly during transmission.

10. A data security and efficiency access system, characterized in that, include: The module includes an access request receiving module, a search data aggregation module, and a search data feedback module. The access request receiving module is used to receive user access requests; The search data aggregation module is used to perform data search by using a multi-level recursive aggregation search algorithm according to the access request, sequentially through a pre-established global index, a cross-device index, and a dynamic local index, to obtain aggregated search data. The search data feedback module is used to feed back the search data to the user using an improved block-coded homomorphic encryption method combined with data recombination rules. The global index is constructed based on the local clustering results of electronic data and the multi-dimensional features of electronic devices, combined with a global index structure; the cross-device index is constructed based on the correlation of multi-dimensional features of electronic devices, combined with a hierarchical index structure; and the dynamic local index is constructed based on power data, combined with a dynamic hierarchical index structure.

Images