Distributed storage-based teaching and training talent resource management system and security authentication method

Abstract

This invention discloses a distributed storage-based education and training talent resource management system and security authentication method, belonging to the field of talent resource management technology. The system includes: a data decomposition module, a hash mapping module, a distributed node cluster, a load balancing module, and a data migration and adjustment module. This invention solves the problem that existing static partitions cannot match the dynamic changes in talent capabilities by constructing a dynamic partitioned storage mechanism driven by the ability dimension of education and training talents. It combines three-dimensional data decomposition and structured identification, ability feature hash mapping, and mapping index table design. Through the layered deployment and redundant backup design of the distributed node cluster, combined with load balancing algorithms and automatic data migration and adjustment, it realizes the aggregated storage of talent data with the same ability dimension, improves retrieval efficiency, and avoids the overload of a single node. Through the deep coupling of the storage architecture with the entire lifecycle of education and training talent capability management, it adapts to the full-process data management needs of talents from onboarding to growth.

Description

Technical Field

[0001] This invention relates to the field of human resource management technology, specifically to a distributed storage-based education and training human resource management system and a security authentication method. Background Technology

[0002] Education and training is a social service that enhances individuals' knowledge reserves, professional skills, and comprehensive qualities through systematic teaching activities. Its core significance lies in meeting individual growth needs, supplementing the social talent supply, promoting the standardized development of the industry, and improving the overall quality of society. Against the backdrop of the popularization of lifelong learning and industry upgrading, the scale of the education and training industry continues to expand, and the teaching fields involved are becoming increasingly segmented. The demand for professional and high-quality education and training personnel is becoming increasingly urgent. Education and training talent resource management is the core support for the development of the education and training industry. Specifically, it refers to the process of comprehensively collecting, classifying, dynamically updating, and efficiently allocating core elements such as the basic information, professional qualifications, teaching abilities, and teaching and research resources of education and training practitioners. Education and training talent data has significant multi-dimensional and dynamic characteristics. It not only includes relatively fixed basic qualification information such as academic qualifications and teacher certification, but also covers dynamic ability data that changes in real time, such as teaching evaluations and skills updates, as well as related resource data such as courseware and teaching and research achievements. Its management quality is not only related to the teaching quality and market competitiveness of education and training institutions, but also crucial to the optimal allocation of talent and sustainable development of the industry.

[0003] Security authentication, as a crucial safeguard for the education and training talent resource management system, is a core component in preventing the leakage, tampering, and forgery of talent data, protecting the legitimate rights and interests of education and training institutions and practitioners, and ensuring the reliable and stable operation of the system. It directly affects the compliance and credibility of management processes. With the widespread application of distributed storage technology in the field of data management, it has gradually been introduced into education and training talent resource management scenarios due to its advantages of high scalability and high availability, becoming an important technical means to solve the problem of large-scale talent data storage.

[0004] However, existing educational and training talent resource management systems still have certain shortcomings. Most existing distributed storage solutions adopt a static partitioning model based on data type, which cannot adapt to the dynamic changes in the teaching fields, skill levels, and other ability dimensions of educational and training talents. The scattered storage of talent data in the same ability dimension leads to low retrieval efficiency, and a single node is prone to overload problems due to high-frequency updated ability data. At the same time, existing storage solutions only focus on data encryption or traceability functions and have not achieved deep coupling with the entire life cycle of educational and training talent ability management, resulting in insufficient adaptability and practicality. Therefore, it is of great significance to develop an educational and training talent resource management system and security authentication method based on distributed storage. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a distributed storage-based education and training talent resource management system and security authentication method. It can solve the problem that existing static partitions cannot match the dynamic changes in talent capabilities by constructing a dynamic partitioned storage mechanism driven by the ability dimension of education and training talents, combined with three-dimensional data decomposition and structured identification, ability feature hash mapping and mapping index table design. Through the layered deployment and redundant backup design of distributed node clusters, combined with load balancing algorithms and automatic data migration and adjustment, it can achieve aggregated storage of talent data with the same ability dimension. Through the deep coupling of storage architecture with the entire life cycle of education and training talent capability management, it can adapt to the full-process data management needs of talents from onboarding to growth.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a training talent resource management system based on distributed storage, the system comprising: a data decomposition module, a hash mapping module, a distributed node cluster, a load balancing module, and a data migration and adjustment module;

[0007] The data decomposition module performs three-dimensional decomposition on the full data of education and training personnel, and the decomposed data of each dimension is transmitted to the hash mapping module.

[0008] The hash mapping module extracts the core capability features of data from each dimension, generates unique feature identifiers through a hash algorithm, establishes a mapping relationship with the distributed node cluster, and directs the corresponding data stream to the matching node in the distributed node cluster.

[0009] The distributed node cluster stores talent data in various dimensions, enabling layered data deployment and redundant backup.

[0010] The load balancing module collects load data from each node in the distributed node cluster in real time, executes a load balancing algorithm based on multi-dimensional evaluation results, and outputs resource allocation adjustment instructions to the distributed node cluster.

[0011] The data migration and adjustment module monitors the update status of the teaching and training personnel's ability dimension. After receiving the update information, it links with the hash mapping module to update the mapping relationship and triggers the cross-node migration instruction for the corresponding data shards to the distributed node cluster.

[0012] Furthermore, the data decomposition module performs the following operations when performing three-dimensional decomposition of the full dataset of education and training personnel:

[0013] Receive full data on education and training personnel entered by education and training institutions. The full data includes basic qualification-related data, dynamic ability-related data, and related resource-related data.

[0014] The received full data is classified and identified to distinguish the specific data content corresponding to basic qualifications, dynamic capabilities, and related resources;

[0015] Each dimension of data is structured and labeled to clarify the attribute information of each type of data, record the relationships between different dimensions of data, and form a relational index of data in each dimension.

[0016] Furthermore, the hash mapping module performs the following operations when extracting core capability features and establishing mapping relationships:

[0017] From the data processed by the data decomposition module, we extract the core competency features corresponding to the teaching field, qualification level, and skill tags.

[0018] The weights of each core capability feature are determined, and a pre-defined hash algorithm is used to calculate the extracted core capability features in combination with the weights to generate a unique feature identifier. The calculation formula is as follows: ,in For feature identification, For the first The weights of each core capability feature By collecting recruitment and assessment standards from institutions of different sizes in the education and training industry, and statistically analyzing the proportion of each competency in talent matching decisions, the final determination was made after summarizing and analyzing industry data. These correspond to teaching areas, qualification levels, and skill tag characteristics, respectively. The basic hash calculation function, This is an XOR operation;

[0019] Pre-construct a mapping index table between capability characteristics and each node group in the distributed node cluster to clarify the target storage node group corresponding to different combinations of capability characteristics;

[0020] Based on the generated feature identifiers, query the mapping index table to determine the corresponding target storage node group, and then direct the data streams of each dimension to that node group to complete the storage deployment.

[0021] Furthermore, the load balancing module performs the following operations when executing load balancing adjustments:

[0022] Real-time collection of storage utilization, data read / write frequency, network bandwidth usage, and response latency of each node in the distributed node cluster;

[0023] The collected multi-dimensional load data is normalized to obtain normalized values ​​for each load metric. ,in A load assessment model is established, and the comprehensive load value of each node is obtained through weighted calculation. The calculation formula is as follows: ,in This represents the overall load value of the node. The weighting coefficients for storage utilization, data read / write frequency, network bandwidth utilization, and response latency are respectively. The correlation between various load indicators and node failure rates in the system's historical operation logs was analyzed, and the results were determined after training with a linear regression model.

[0024] Set a load threshold range, compare the overall load value of each node with the threshold range, and identify nodes with excessive load and nodes with idle load.

[0025] Based on the comparison results, a load balancing algorithm is executed to migrate some low-frequency access data from overloaded nodes to idle nodes, while adjusting the data access route.

[0026] Furthermore, the distributed node cluster consists of basic storage nodes, capability classification nodes, and resource backup nodes. The basic storage nodes store the basic qualification data of the training personnel, as well as the dynamic capability data and associated resource data that are accessed infrequently. The capability classification nodes are divided into multiple node groups according to the core capability characteristics of the training personnel. Each node group specifically stores the high-frequency access data of the corresponding capability characteristic combination. The resource backup nodes perform real-time redundant backup of the data in each dimension. When the basic storage node or capability classification node fails, the backup data is activated to support the operation of the system.

[0027] Furthermore, the data migration and adjustment module monitors the update status of the teaching and training personnel data update interface and the system-entered data in real time. When it detects changes in the capability dimension such as new teaching and research achievements, skills certification upgrades, and teaching evaluation updates, it first identifies the specific dimension to which the updated data belongs and the corresponding original storage node. Then, it regenerates feature identifiers based on the updated capability characteristics and queries the mapping index table. Finally, it determines the new target storage node through node adaptability calculation. The adaptability calculation formula is as follows: ,in For target node adaptation, The weight is the matching degree weight. Based on actual operational data on talent data retrieval priorities and node resource utilization in the education and training industry, and combined with industry experts' assessments of storage suitability, the following criteria were determined. To determine the matching degree between the updated capability features and the node group, The remaining resource percentage of the node is then used to initiate the data sharding migration process, which shards the updated data and related data to the new target node. After the migration is completed, the mapping index table and the relationship records of each node are updated immediately.

[0028] Furthermore, the hash mapping module uses an asymmetric hash algorithm, which dynamically adjusts the hash calculation dimension according to the complexity of the educational and training personnel's ability characteristics. The mapping index table supports dynamic updates. When new educational and training personnel ability characteristic types are added or the distributed node cluster is expanded, the mapping relationship in the index table is automatically updated through the system backend configuration.

[0029] Furthermore, the data decomposition module also has a data preprocessing function. Before performing three-dimensional decomposition, it performs format standardization processing on all received training personnel data, removes invalid data, duplicate data and data with incorrect format, and provides prompts to complete incomplete data.

[0030] The security authentication method for education and training human resource management based on distributed storage is applicable to the aforementioned education and training human resource management system based on distributed storage. The method includes the following steps:

[0031] S1. After standardizing the format of all data on education and training personnel, the data is broken down into three dimensions to clarify the attributes and relationships of each dimension.

[0032] S2. Extract the core capability features of data from each dimension, generate unique feature identifiers, verify the storage capacity and load status of the target node, and then map the data stream to the target node according to the mapping index table.

[0033] S3. Collect load data from distributed nodes, perform multi-dimensional evaluation, and execute a load balancing algorithm to dynamically adjust node resource allocation;

[0034] S4. Monitor the update status of the competency dimension of education and training personnel. When changes occur, trigger data sharding and encrypted migration. After the data consistency is confirmed by node consensus, update the mapping relationship and node association records.

[0035] Furthermore, the format standardization process in step S1 includes removing invalid, duplicate, and format-incorrect data; the feature identifier in step S2 is generated by a hash algorithm; when the target node load is close to the threshold, load adjustment is performed first; the load data in step S3 includes storage occupancy rate, data read / write frequency, network bandwidth usage, and response latency; and the integrity check is performed before data shard migration in step S4, and the mapping index table is updated synchronously after the migration is completed.

[0036] Compared with existing technologies, this distributed storage-based education and training talent resource management system and security authentication method have the following advantages:

[0037] This invention addresses the problem of existing static partitioning failing to match the dynamic changes in talent capabilities by constructing a dynamic partitioned storage mechanism driven by the capability dimension of education and training personnel. It combines three-dimensional data decomposition and structured identification, capability feature hash mapping, and mapping index table design. Through layered deployment and redundant backup design of distributed node clusters, coupled with load balancing algorithms and automatic data migration and adjustment, it achieves aggregated storage of talent data with the same capability dimension, improving retrieval efficiency while avoiding overload of a single node. By deeply coupling the storage architecture with the entire lifecycle of education and training talent capability management, it adapts to the full-process data management needs of talent from onboarding to growth, enhancing the system's adaptability to education and training industry scenarios, ensuring the real-time, efficient, and stable nature of talent data management, and providing reliable technical support for education and training talent resource management.

[0038] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0040] Figure 1 This is a schematic diagram of the structure of a distributed storage-based education and training human resource management system.

[0041] Figure 2 A flowchart illustrating a security authentication method for education and training human resource management based on distributed storage;

[0042] Figure 3 This is a flowchart of a security authentication method for education and training human resource management based on distributed storage. Detailed Implementation

[0043] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0044] The present invention provides a distributed storage-based education and training talent resource management system and security authentication method, which aims to solve the problems of existing static partitioned storage being unable to adapt to dynamic changes in talent capabilities, low retrieval efficiency, and easy node overload, and to achieve efficient management of education and training talent data throughout the entire life cycle.

[0045] See Figure 1 The system's core consists of five modules: The data decomposition module receives all training personnel data, preprocesses it according to standardized format, and then decomposes it into three dimensions: basic qualifications, dynamic capabilities, and related resources. It then performs structured identification and records the relationships between these dimensions. The hash mapping module extracts core capability features such as teaching areas, qualification levels, and skill tags, generates unique identifiers using an asymmetric hash algorithm, and directs the data stream to matching nodes based on a pre-set mapping index table. The distributed node cluster includes three types of nodes: basic storage, capability classification, and resource backup. These nodes store basic data, high-frequency capability data, and redundant backups, ensuring data security. The load balancing module collects multi-dimensional load data such as node storage occupancy and read / write frequency in real time. After evaluation, it executes an algorithm to migrate low-frequency data from overloaded nodes to idle nodes, preventing single-node overload. The data migration and adjustment module monitors updates to the personnel capability dimensions, identifies changes, and updates the mapping relationships accordingly. It determines new target nodes through node adaptability calculations, triggers data shard migration, and synchronously updates the index.

[0046] See Figure 2 and Figure 3 The security authentication process is as follows: First, standardize and decompose all data in three dimensions; then extract core features to generate identifiers, verify node status, and map and store them in a targeted manner; next, collect node load data and dynamically adjust resource allocation; finally, monitor capability updates, trigger data encryption migration, and update mapping records after confirming data consistency.

[0047] This solution uses capability-driven dynamic partitioned storage to aggregate and store data with similar capabilities, improving retrieval efficiency. It is also deeply integrated with the entire talent management lifecycle, adapting to the needs of the entire process from onboarding to growth, and ensuring the real-time performance, stability, and security of data management.

[0048] Example 1

[0049] This embodiment is applied to a large chain of educational institutions. The institution's business covers multiple teaching areas, including K-12 education, vocational skills training, and adult skills enhancement. It has over 2,000 contracted teachers, and the talent data includes basic qualification information such as academic qualifications and teacher certifications, dynamic ability data such as teaching evaluations and skills certifications, and related resource data such as courseware and teaching research achievements. With the expansion of the institution and the specialization of teaching areas, the original static, partitioned talent management system showed significant drawbacks. Distributed storage of teacher data within the same ability dimension led to low retrieval efficiency during course matching, and the frequently updated dynamic ability data caused frequent overload of some storage nodes. Furthermore, the system could not adapt to the full-process data management needs of teachers from onboarding to promotion. Therefore, this invention employs a distributed storage-based educational talent resource management system and security authentication method to achieve efficient management and secure storage of talent data.

[0050] See Figure 1 , Figure 2 and Figure 3 The specific implementation process of this embodiment is as follows:

[0051] After educational institutions input all teacher data into the system, the data first enters the data decomposition module. This module first performs format standardization preprocessing on the entire dataset, automatically identifying and removing invalid blank data, duplicate qualification information, and incorrectly formatted teaching evaluation records. For incomplete data with missing key fields, it sends a prompt to the institution administrator to complete the data. After preprocessing, the module decomposes the data according to three dimensions: relatively fixed information such as teachers' education, professional background, and teaching certificates are classified as basic qualification data; real-time changing information such as teaching scores, student satisfaction feedback, and skills update records are classified as dynamic ability data; and information such as teaching materials, teaching research papers, and training handouts are classified as related resource data.

[0052] Subsequently, the three types of data were structured and labeled to clarify the attribute information of each type of data. For example, the educational attribute in the basic qualification data was labeled as undergraduate or above associate degree. At the same time, the relationship between different dimensions of data was recorded. For example, a correlation index was established between the dynamic ability data of a teacher's excellent course award record and the corresponding award-winning courseware related resource data to form a complete dimensional data system.

[0053] After data decomposition, data from each dimension is transmitted to the hash mapping module. This module extracts core competency features from the processed data, specifically including teaching areas such as primary school mathematics, Java programming, and English speaking; qualification levels such as junior teacher, intermediate teacher, and senior teacher; and skill tags such as multimedia teaching, interactive teaching, and postgraduate entrance examination tutoring. In the specific implementation of this embodiment, it is necessary to determine the weight of each core competency feature. The weights are determined by collecting talent recruitment and assessment standards from institutions of different sizes in the education and training industry, statistically analyzing the proportion of each competency feature in talent matching decisions, and then summarizing and analyzing industry data.

[0054] A pre-defined asymmetric hash algorithm combined with weights is used to calculate the extracted core capability features, generating a unique feature identifier. The calculation formula is as follows: ,in For feature identification, For the first The weights of each core capability feature These correspond to the skill tag characteristics of qualification levels in the teaching field. The basic hash calculation function, The hash mapping module performs an XOR operation. It pre-constructs a mapping index table between capability characteristics and node groups in the distributed node cluster, clearly defining the target storage node groups corresponding to different combinations of capability characteristics. For example, the combination of primary school mathematics, intermediate teacher, and multimedia teaching corresponds to node group A, while the combination of Java programming, advanced teacher, and project practice corresponds to node group B. The hash mapping module queries this mapping index table based on the generated characteristic identifier to determine the corresponding target storage node group and directs the data streams of each dimension to that node group to complete the storage deployment. When an institution adds the capability characteristic type of artificial intelligence teaching, or when the distributed node cluster expands by adding new node groups, the mapping index table is automatically updated through system backend configuration.

[0055] The distributed node cluster consists of basic storage nodes, capability classification nodes, and resource backup nodes. Basic storage nodes specifically store teachers' basic qualification data, as well as infrequently accessed dynamic capability data and related resource data, such as teachers' educational background, teaching evaluation records from three years ago, and outdated courseware. Capability classification nodes are divided into multiple node groups based on core capability characteristics. Each node group specifically stores frequently accessed data for the corresponding combination of capability characteristics. For example, node group A stores recent teaching evaluations and frequently used courseware for teachers specializing in elementary mathematics, intermediate, and multimedia teaching. Resource backup nodes provide real-time redundant backups of data across all dimensions. When a hardware failure or network interruption occurs in a basic storage node or capability classification node, the system automatically activates the backup data to ensure that course scheduling, teacher evaluation, and other business operations are not affected.

[0056] The load balancing module collects four types of load data in real time from each node in the distributed node cluster: storage utilization, data read / write frequency, network bandwidth utilization, and response latency. In the specific implementation of this embodiment, the collected multi-dimensional load data is normalized to obtain normalized values ​​for each load indicator. ,in These correspond to four types of load data. A load assessment model is established based on normalized data, and the comprehensive load value of each node is obtained through weighted calculation. The calculation formula is as follows: ,in This represents the overall load value of the node. to These are the weighting coefficients for storage utilization, data read / write frequency, network bandwidth utilization, and response latency, respectively. These weighting coefficients are determined by training a linear regression model after analyzing the correlation between various load indicators and node failure rates in the system's historical operation logs.

[0057] The system sets a fixed load threshold range and compares the overall load value of each node with the threshold range to identify nodes with excessive load and nodes with idle load. For example, node group A may exceed the threshold due to a surge in data read / write frequency caused by the recent surge in enrollment for elementary school math courses, while node group C is idle because its corresponding adult calligraphy training courses are in the off-season and its overall load value is below the threshold. The load balancing module then executes a load balancing algorithm to migrate some of the less frequently accessed older courseware data from node group A to node group C. At the same time, it adjusts the data access route so that subsequent access to this data is automatically directed to node group C, preventing node group A from becoming continuously overloaded.

[0058] The data migration and adjustment module monitors the update status of teachers' competency dimensions by listening to the data update interface for education and training personnel and the data entered into the system in real time. When it detects changes in competency dimensions such as new teaching and research achievements, skills certification upgrades, and updated teaching evaluations, for example, when a teacher passes the senior teacher qualification certification or wins a provincial-level excellent lesson award, the module first identifies the specific dimension to which the updated data belongs and the corresponding original storage node.

[0059] Subsequently, feature identifiers are regenerated based on the updated capability characteristics, and the mapping index table is queried. A new target storage node is determined through node adaptability calculation. In the specific implementation of this embodiment, the adaptability calculation formula is: ,in For target node adaptation, The matching score weight is determined based on actual operational data regarding the priority of talent data retrieval in the education and training industry and the utilization rate of node resources, combined with the assessment of storage suitability by industry experts. To determine the matching degree between the updated capability features and the node group, This represents the percentage of remaining resources for the node. After calculation, the system initiates the data sharding migration procedure, sharding and migrating the teacher's updated data and related data to the new target node. Before migration, the data integrity is verified to ensure that no data is lost or damaged. After migration, the mapping index table and the relationship records of each node are updated immediately to ensure the accuracy of subsequent data retrieval and retrieval.

[0060] The security authentication process strictly follows preset steps: First, all teacher data is standardized in format, and invalid, duplicate, and incorrectly formatted data is removed. Then, the data is broken down into three dimensions to clarify the attributes and relationships of each dimension. Second, the core capability characteristics of each dimension are extracted, and a unique feature identifier is generated using a hash algorithm. The storage capacity and load status of the target node are verified. If the load of the target node is close to the threshold, load adjustment is performed first, and then the data flow is mapped to the target node according to the mapping index table. Third, the load data of distributed nodes is continuously collected. After multi-dimensional evaluation, a load balancing algorithm is executed to dynamically adjust the allocation of node resources. Fourth, the update status of teacher capability dimensions is monitored in real time. When changes occur, data sharding and encrypted migration are triggered. After the data consistency is confirmed by node consensus, the mapping relationship and node association records are updated synchronously.

[0061] In summary, this embodiment, through the complete implementation process described above, successfully addresses the shortcomings of the original system of this large-scale chain of educational institutions. The ability-driven dynamic partitioned storage mechanism enables aggregated storage of teacher data within the same ability dimension, improving retrieval efficiency during course matching compared to the original system. The layered deployment and redundant backup design of the distributed node cluster, combined with load balancing algorithms and automatic data migration and adjustment, completely solves the problem of single-node overload, significantly improving system stability. The deep coupling of the storage architecture with the entire lifecycle of teacher ability management perfectly adapts to the full-process data management needs of teachers, from basic information entry upon joining the company to ability updates during their professional development, and finally to qualification changes after promotion.

[0062] Example 2

[0063] This embodiment applies to a regional education and training alliance, which consists of more than ten primary and secondary school education and training institutions of different types, including after-school tutoring, art training, and vocational skills training. The alliance shares talent resources and teaching and research results, and the talent data involved covers the basic qualifications, dynamic abilities, and related resource data of teachers from each institution. Due to the inconsistent data formats among the institutions within the alliance, frequent cross-institutional talent movement leads to frequent changes in ability dimensions. The existing storage system suffers from poor data interoperability and compatibility, low efficiency in cross-institutional talent retrieval, and overload of some nodes due to centralized access to shared resources. Furthermore, it cannot adapt to the full-process talent management needs under alliance-based operation. Therefore, based on the aforementioned embodiment, this invention employs a distributed storage-based education and training talent resource management system and security authentication method to achieve unified management, secure interoperability, and efficient allocation of talent data within the alliance.

[0064] See Figure 1 , Figure 2 and Figure 3 The specific implementation process of this embodiment is as follows:

[0065] Each institution within the alliance uploads its full dataset of training personnel data through a dedicated system entry port. Upon receiving the data, the data decomposition module first performs format standardization preprocessing. In addition to removing invalid, duplicate, and incorrectly formatted data and prompting for incomplete data, a new cross-institutional data format compatibility processing step is added. This standardizes field naming conventions and data format standards across different institutions. For example, teaching qualifications marked by some institutions are standardized to qualification levels, and date data in different formats is converted to a standard format. After preprocessing, the data is decomposed according to the three dimensions of basic qualifications, dynamic capabilities, and associated resources. When structurally identifying the data in each dimension, an additional cross-institutional association index is added to record the personnel's work experience in different institutions within the alliance, cross-institutional teaching records, and other related information, forming a dimensional data system that balances institutional independence with alliance unity.

[0066] After data decomposition, data from each dimension is transmitted to the hash mapping module. This module, based on the aforementioned extraction of core competency characteristics of skill tags for qualification levels in the teaching field, and combined with the alliance's operational needs, unifies the classification standards for core competency characteristics within the alliance. For example, the arts training field is subdivided into subcategories such as fine arts, dance, and music, and qualification levels are divided into beginner, intermediate, advanced, and special levels according to the alliance's unified assessment standards. In the specific implementation of this embodiment, an asymmetric hash algorithm combined with weight calculation is used to generate a unique feature identifier. The calculation formula is: .

[0067] The pre-built mapping index table, in addition to containing the mapping relationship between single ability feature combinations and node groups, also includes a new mapping configuration for cross-institutional shared node groups. This maps high-quality teacher data and general teaching and research resource data that need to be shared across institutions within the alliance to cross-institutional shared node groups. When the alliance adds new ability feature types for special teaching areas, or when new node groups are added due to institutional expansion, the mapping index table automatically updates the mapping relationship through the system backend configuration, ensuring the alliance's business expansion needs are met.

[0068] The distributed node cluster, based on the aforementioned basic storage nodes, capability classification nodes, and resource backup nodes, adds cross-institutional shared node groups to specifically store high-frequency access data that is interconnected across institutions within the alliance, such as high-quality open course materials and dynamic capability data of cross-institutional instructors. The basic storage nodes are divided into dedicated storage areas according to the institutional dimension to store basic qualification data and low-frequency access data within each institution. The capability classification nodes are still divided into node groups according to core capability characteristics to store high-frequency data of talents with similar capabilities within the alliance. The resource backup nodes adopt a cross-regional redundant backup mode to back up the key data of each institution to backup nodes in different regions within the alliance to prevent data loss caused by regional failures.

[0069] The load balancing module collects real-time data on storage occupancy, read / write frequency, network bandwidth usage, and response latency of each node. It also adds cross-organizational data access concurrency as a load monitoring indicator. In this specific implementation, the collected multi-dimensional load data is normalized, and then a weighted calculation is performed to obtain the comprehensive load value of each node. The calculation formula is as follows: .

[0070] The system sets a load threshold range and compares the comprehensive load values ​​of each node. In addition to performing routine low-frequency data migration, for the load adjustment of cross-organization shared node groups, a cross-regional routing optimization strategy is adopted to divert the load of nodes far from the data access source organization to idle nodes in nearby regions, thereby reducing network latency for cross-regional data transmission and preventing the shared node group from becoming overloaded due to concentrated access.

[0071] In addition to monitoring changes in competency dimensions such as new teaching and research achievements, skills certification upgrades, and teaching evaluation updates, the data migration and adjustment module focuses on monitoring changes in competency dimensions brought about by cross-institutional talent mobility. For example, this includes the review and update of teaching qualification levels after a teacher moves from Institution A to Institution B within the alliance. When such changes are detected, the module first identifies the dimension to which the updated data belongs and the original storage node. Based on the updated competency characteristics, it regenerates feature identifiers and queries the mapping index table. Finally, it determines the new target storage node through node adaptability calculation. In this specific implementation, the adaptability calculation formula is: .

[0072] Before data sharding and migration, in addition to integrity verification, cross-institutional data permission synchronization is added. Based on the access permission configuration of the talent's new employing institution, the access permission identifier of the data is updated synchronously. After migration, the mapping index table and the relationship records of each node are updated immediately, and simultaneously synchronized to the system ports of relevant institutions within the alliance, ensuring that each institution can obtain the latest storage location and data information of the talent in real time.

[0073] The security authentication process is further optimized based on the above. The first step is to add cross-organizational data encryption identifiers in the data format standardization process to ensure the privacy of the original data. The second step is to add a cross-organizational data access permission verification step after extracting core capability features and generating feature identifiers. Only the organizational nodes with the corresponding permissions can obtain the mapping access permissions. The third step is to perform real-time encryption processing on the data transmitted across organizations during the load adjustment process. The fourth step is to adopt a cross-organizational node consensus mechanism when the capability dimension update triggers data migration. The data consistency must be confirmed by the original storage organization node, the target storage organization node, and the alliance center node before the mapping relationship and node association records are updated to ensure the security and accuracy of cross-organizational data migration.

[0074] In summary, this embodiment effectively solves the talent data management challenges of regional education and training alliances through targeted optimization and expansion. Cross-institutional data format compatibility processing and a unified feature classification standard break down data barriers between institutions, enabling seamless data exchange within the alliance. The newly added cross-institutional shared node groups and routing optimization strategies improve the access efficiency of shared resources and avoid node overload caused by centralized access. The data migration and adjustment module's adaptability to cross-institutional talent mobility ensures the continuity of talent lifecycle data management. The enhanced security authentication process balances data privacy with the security of cross-institutional access.

[0075] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A distributed storage-based teaching and training talent resource management system, characterized in that, The system includes: a data decomposition module, a hash mapping module, a distributed node cluster, a load balancing module, and a data migration and adjustment module; The data decomposition module performs three-dimensional decomposition on the full data of education and training personnel, and the decomposed data of each dimension is transmitted to the hash mapping module. The hash mapping module extracts the core capability features of data from each dimension, generates unique feature identifiers through a hash algorithm, establishes a mapping relationship with the distributed node cluster, and directs the corresponding data stream to the matching node in the distributed node cluster. The distributed node cluster stores talent data in various dimensions, enabling layered data deployment and redundant backup. The load balancing module collects load data from each node in the distributed node cluster in real time, executes a load balancing algorithm based on multi-dimensional evaluation results, and outputs resource allocation adjustment instructions to the distributed node cluster. The data migration and adjustment module monitors the update status of the teaching and training personnel's ability dimension. After receiving the update information, it links with the hash mapping module to update the mapping relationship and triggers the cross-node migration instruction for the corresponding data shards to the distributed node cluster.

2. The education and training human resource management system based on distributed storage according to claim 1, characterized in that, The data decomposition module performs the following operations when performing three-dimensional decomposition on the full dataset of education and training personnel: Receive full data on education and training personnel entered by education and training institutions. The full data includes basic qualification-related data, dynamic ability-related data, and related resource-related data. The received full data is classified and identified to distinguish the specific data content corresponding to basic qualifications, dynamic capabilities, and related resources; Each dimension of data is structured and labeled to clarify the attribute information of each type of data, record the relationships between different dimensions of data, and form a relational index of data in each dimension.

3. The education and training human resource management system based on distributed storage according to claim 1, characterized in that, The hash mapping module performs the following operations when extracting core capability features and establishing mapping relationships: From the data processed by the data decomposition module, we extract the core competency features corresponding to the teaching field, qualification level, and skill tags. The weights of each core capability feature are determined, and a pre-defined hash algorithm is used to calculate the extracted core capability features in combination with the weights to generate a unique feature identifier. The calculation formula is as follows: ,in For feature identification, For the first The weights of each core capability feature These correspond to teaching areas, qualification levels, and skill tag characteristics, respectively. The basic hash calculation function, This is an XOR operation; Pre-construct a mapping index table between capability characteristics and each node group in the distributed node cluster to clarify the target storage node group corresponding to different combinations of capability characteristics; Based on the generated feature identifiers, query the mapping index table to determine the corresponding target storage node group, and then direct the data streams of each dimension to that node group to complete the storage deployment.

4. The education and training human resource management system based on distributed storage according to claim 1, characterized in that, The load balancing module performs the following operations when performing load balancing adjustments: Real-time collection of storage utilization, data read / write frequency, network bandwidth usage, and response latency of each node in the distributed node cluster; The collected multi-dimensional load data is normalized to obtain normalized values ​​for each load metric. A load assessment model is established, and the comprehensive load value of each node is obtained through weighted calculation. The calculation formula is as follows: ,in This represents the overall load value of the node. These are the weighting coefficients for storage utilization, data read / write frequency, network bandwidth usage, and response latency, respectively. Set a load threshold range, compare the overall load value of each node with the threshold range, and identify nodes with excessive load and idle load. Based on the comparison results, a load balancing algorithm is executed to migrate some low-frequency access data from overloaded nodes to idle nodes, while adjusting the data access route.

5. The education and training human resource management system based on distributed storage according to claim 1, characterized in that, The distributed node cluster consists of basic storage nodes, capability classification nodes, and resource backup nodes. The basic storage nodes store the basic qualification data of the training personnel, as well as the dynamic capability data and associated resource data that are accessed infrequently. The capability classification nodes are divided into multiple node groups according to the core capability characteristics of the training personnel. Each node group specifically stores the high-frequency access data of the corresponding capability characteristic combination. The resource backup nodes perform real-time redundant backup of the data in each dimension. When the basic storage node or capability classification node fails, the backup data is activated to support the operation of the system.

6. The education and training human resource management system based on distributed storage according to claim 1, characterized in that, The data migration and adjustment module monitors the update status of the teaching and training personnel data update interface and the system-entered data in real time. When it detects changes in the capability dimension such as new teaching and research achievements, skills certification upgrades, and teaching evaluation updates, it first identifies the specific dimension to which the updated data belongs and the corresponding original storage node. Then, it regenerates feature identifiers based on the updated capability characteristics and queries the mapping index table. The new target storage node is determined through node adaptability calculation, and the adaptability calculation formula is as follows: ,in For target node adaptation, For matching degree weight, To determine the matching degree between the updated capability features and the node group, The remaining resource percentage of the node is then used to initiate the data sharding migration process, which shards the updated data and related data to the new target node. After the migration is completed, the mapping index table and the relationship records of each node are updated immediately.

7. The education and training human resource management system based on distributed storage according to claim 1, characterized in that, The hash mapping module uses an asymmetric hash algorithm, which dynamically adjusts the hash calculation dimension according to the complexity of the educational and training personnel's ability characteristics. The mapping index table supports dynamic updates. When new educational and training personnel ability characteristic types are added or the distributed node cluster is expanded, the mapping relationship in the index table is automatically updated through the system backend configuration.

8. The education and training human resource management system based on distributed storage according to claim 1, characterized in that, The data decomposition module also has a data preprocessing function. Before performing three-dimensional decomposition, it performs format standardization processing on the received full amount of education and training personnel data, removes invalid data, duplicate data and data with incorrect format, and provides prompts to complete incomplete data.

9. A security authentication method for education and training human resource management based on distributed storage, applicable to the education and training human resource management system based on distributed storage as described in any one of claims 1-8, characterized in that, The method includes the following steps: S1. After standardizing the format of all data on education and training personnel, the data is broken down into three dimensions to clarify the attributes and relationships of each dimension. S2. Extract the core capability features of data from each dimension, generate unique feature identifiers, verify the storage capacity and load status of the target node, and then map the data stream to the target node according to the mapping index table. S3. Collect load data from distributed nodes, perform multi-dimensional evaluation, and execute a load balancing algorithm to dynamically adjust node resource allocation; S4. Monitor the update status of the competency dimension of education and training personnel. When changes occur, trigger data sharding and encrypted migration. After the data consistency is confirmed by node consensus, update the mapping relationship and node association records.

10. The security authentication method for education and training human resource management based on distributed storage according to claim 9, characterized in that, The format standardization process in step S1 includes removing invalid, duplicate, and formatted data. In step S2, the feature identifier is generated by a hash algorithm. When the target node load is close to the threshold, load adjustment is performed first. The load data in step S3 includes storage occupancy, data read / write frequency, network bandwidth usage, and response latency. In step S4, integrity verification is performed before data shard migration, and the mapping index table is updated synchronously after migration is completed.

Images