Business data sharing fusion system and method based on trusted data space

Abstract

The invention provides a business data sharing fusion system and method based on a trusted data space, and relates to the technical field of data sharing. According to the method, sharing and fusion of service data are realized by constructing a trusted data space infrastructure. The method comprises the following steps: firstly, establishing a node path planning model for priority evaluation and data matching; then forming a model collaborative deployment scheme to distribute tasks to edge nodes; then perception scoring and resource scheduling are carried out on the tasks, and a multi-task concurrent transmission scheme is generated; constructing a multi-domain network service level protocol evaluation model to perform service decomposition and business collaborative management; and finally carrying out heterogeneous processing and data fusion on the business data. According to the method, the problems of information islands, multi-source heterogeneous data fusion and cross-domain business collaboration among regional mechanisms are solved, and the system performance and the user experience are improved.

Description

Technical Field

[0001] This application relates to the technical field of data sharing, and particularly to a business data sharing and fusion system and method based on a trusted data space. Background Art

[0002] With the rapid development of the digital economy, data, as a key production factor, requires breaking information silos to fully explore its value and achieve cross-departmental and cross-domain data interconnection and fusion. Data sharing and business collaboration between various organizations have become increasingly important.

[0003] Currently, two main technical solutions are adopted for data sharing. One is a centralized architecture, which collects and manages data from all parties by establishing a unified data center; the other is a point-to-point data exchange channel, where separate data sharing interfaces are established between different organizations. Although these traditional methods can achieve basic data sharing functions, their limitations gradually emerge as the data scale and complexity increase.

[0004] In recent years, newer technical solutions have started to use distributed technologies such as consortium blockchains to build data sharing platforms, manage data access rights through smart contracts, and achieve controllable data sharing. Such solutions ensure data trustworthiness through blockchain technology and improve the security and reliability of data sharing to a certain extent.

[0005] However, the existing technologies still have the following problems: lack of a dynamic adaptive data matching and fusion mechanism, making it difficult to meet business scenarios with high timeliness requirements; low utilization efficiency of edge computing resources and high deployment costs of large models; inflexible allocation of communication resources in multi-task concurrent scenarios; and the service quality guarantee mechanism in cross-domain business collaboration needs to be improved. Therefore, these problems seriously restrict the improvement of business collaboration efficiency. Summary of the Invention

[0006] In view of this, this application provides a business data sharing and fusion system and method based on a trusted data space, which solves the problem of limited improvement of business collaboration efficiency in the existing technology, realizes a dynamic adaptive data matching and fusion mechanism, can effectively improve the utilization efficiency of edge computing resources, achieve flexible allocation of communication resources in multi-task concurrent scenarios, and realize a perfect service quality guarantee mechanism in cross-domain business collaboration.

[0007] An embodiment of the present application provides a method for sharing and fusing service data based on a trusted data space, including: obtaining service data of at least one task, where the service data includes task requirement data; constructing a trusted data space infrastructure, which is used to indicate deployed node devices, and the node devices include data access gateways and node security communication components; based on the trusted data space infrastructure and the task requirement data, establishing a node path planning model, and based on the node path planning model, performing priority evaluation and data matching on the service data to generate a data matching scheme; according to the data matching scheme and the service data, forming a model collaborative deployment scheme, which is used to indicate allocating the tasks to each edge node; according to the model collaborative deployment scheme, performing perception scoring on each task, and combining the perception scoring results to perform resource scheduling on the tasks to generate a multi-task concurrent transmission scheme; according to the multi-task concurrent transmission scheme and the service data, constructing a multi-domain network service level agreement evaluation model, and based on the multi-domain network service level agreement evaluation model, performing service decomposition and business collaborative management to generate a service quality guarantee scheme; performing heterogeneous processing on the service data, and based on the heterogeneous processing results, performing data fusion on the service data to generate a data fusion scheme; under the synergistic effect of the data matching scheme, the model collaborative deployment scheme, the multi-task concurrent transmission scheme, the service quality guarantee scheme, and the data fusion scheme, performing fusion sharing on the service data.

[0008] The step of, based on the trusted data space infrastructure and the task requirement data, establishing a node path planning model, and based on the node path planning model, performing priority evaluation and data matching on the service data to generate a data matching scheme, includes: based on the trusted data space infrastructure, combining a preset spatio-temporal map to analyze the node state data of mobile nodes, where the mobile nodes are used to represent node devices related to the task, and the node state data includes node position changes and network connection states; performing path search based on the node state data and the service data, constructing a node path planning model that meets the preset timeliness requirements, and evaluating the connection reliability between the mobile nodes to obtain a reliability evaluation result; according to the node path planning model and the reliability evaluation result, performing priority evaluation on the service data to determine the calculation task weights related to the task priority evaluation result; based on the trusted data space infrastructure and the calculation task weights, performing data matching on the service data to form a matching scheme; combining the node path planning model, the calculation task weights, and the matching scheme to form a data matching scheme, and updating the data matching scheme according to network state changes, task urgency, and a preset performance index system.

[0009] Performing data matching on the service data based on the trusted data space infrastructure and the computing task weights to form a matching scheme, including: constructing a data source feature vector and a task requirement vector based on the trusted data space infrastructure, where the data source feature vector includes data type, update frequency, and quality score, and the task requirement vector includes data specification, timeliness requirement, and accuracy requirement; calculating the similarity between the data source feature vector and the task requirement vector based on the service data, and establishing a mapping relationship between the data source and the task requirement of the service data based on the similarity; forming a matching scheme based on the mapping relationship, and updating the matching scheme according to changes in task priority and network status adjustment.

[0010] Forming a model collaborative deployment scheme according to the data matching scheme and the service data, including: collecting edge node resource monitoring data through the deployed lightweight resource monitoring agent based on the data matching scheme, where the edge node resource monitoring data includes CPU utilization, memory usage, GPU computing load, and network bandwidth data; constructing a multi-dimensional resource profile including computing power, storage capacity, and network performance according to the edge node resource monitoring data, and generating an edge node performance evaluation result; establishing a load prediction model based on the edge node performance evaluation result; calculating the resource usage trend based on the load prediction model.

[0011] Forming a model collaborative deployment scheme according to the data matching scheme and the service data further includes: dividing the task into multiple computing tasks based on the model layer dependency relationship indicated by the service data, and forming a model task allocation scheme acting within or between node devices according to the resource usage trend, where the model task allocation scheme is used to indicate allocating the computing task to the edge node; performing INT8 quantization compression on the model parameters of the computing model related to the task according to the model task allocation scheme, and performing frequent co-occurrence token sequence fusion to construct an optimized model component; distributing and deploying the computing model to each node device and / or edge node based on the optimized model component by using a preset edge node load balancing strategy to generate a model collaborative deployment scheme.

[0012] According to the model collaborative deployment plan, perform perception scoring on each task, and combine the perception scoring results to perform resource scheduling on the task, generating a multi-task concurrent transmission plan, including: based on the model collaborative deployment plan, extract the task type, data scale, and time attribute features of the task from the service data to construct task features; according to the task features, perform task understanding on the service data to obtain the service context analysis result; based on the service context analysis result, evaluate the business value and execution priority of the task to form the importance evaluation result of the task; based on the importance evaluation result and the preset task performance, calculate the task score of the task, and output the multi-task priority sorting result according to the task score, where the task performance includes data timeliness, business relevance, and resource consumption.

[0013] According to the model collaborative deployment plan, perform perception scoring on each task, and combine the perception scoring results to perform resource scheduling on the task, generating a multi-task concurrent transmission plan, including: based on the multi-task priority sorting result, collect channel status data through distributed detection nodes deployed on the edge nodes, where the channel status data includes bandwidth utilization rate, link delay, and packet loss rate data; according to the channel status data, use a preset time series analysis model to predict the channel quality change trend, and construct a channel availability evaluation index, and determine the channel resource allocation strategy according to the channel quality change trend and the channel availability evaluation index; according to the channel resource allocation strategy and the multi-task priority sorting result, determine the multi-task concurrent transmission plan of the task.

[0014] According to the multi-task concurrent transmission plan and the service data, construct a multi-domain network service level agreement evaluation model, and perform service decomposition and business collaborative management based on the multi-domain network service level agreement evaluation model, generating a service quality guarantee plan, including: based on the multi-task concurrent transmission plan, establish a service quality index system, where the service quality index system includes delay, availability, and throughput; based on the service quality index system, evaluate the risk data of service risk events related to the task, and perform hierarchical early warning analysis according to the risk data to form a risk assessment plan, where the risk data includes occurrence probability and impact range; according to the risk assessment plan, combined with the service quality requirements of a specified business scenario, generate a multi-domain network service level agreement evaluation model.

[0015] Construct a multi-domain network service level agreement evaluation model according to the multi-task concurrent transmission scheme and the service data, and perform service decomposition and business collaboration management based on the multi-domain network service level agreement evaluation model to generate a service quality assurance scheme. It further includes: constructing a business process dependency graph of the task based on the multi-domain network service level agreement evaluation model, and constructing an output service topology structure according to the business process dependency graph. The service topology structure is used to indicate the call relationship and resource dependency between service components in the task; determine a search strategy according to the service topology structure. The search strategy is used to indicate iteratively optimizing the service decomposition scheme and evaluating the scheme metrics of the service decomposition scheme. When the scheme metrics reach the preset convergence condition, output the service decomposition result. The scheme metrics include the feasibility and optimization space of the service decomposition scheme. The service decomposition scheme is used to indicate adjusting the service component configuration; perform cross-domain collaboration management based on the service decomposition result to generate a service quality assurance scheme; then perform heterogeneous processing on the service data, and perform data fusion on the service data based on the heterogeneous processing result to generate a data fusion scheme, including: using a preset semantic knowledge graph to extract the semantic features of different data sources indicated by the service data; using preset quality evaluation metrics to perform data quality evaluation on the service data, and cleaning the service data according to the data quality evaluation result to obtain the cleaned service data. The quality evaluation metrics include the integrity, accuracy, consistency, and timeliness of the service data; establish a data feature model based on the semantic features, and analyze the data feature portrait using the data feature model. The data feature portrait includes the structural features and data distribution rules of the data source; generate a data fusion scheme based on the data feature portrait and the cleaned service data. The data fusion scheme is used to indicate fusing multi-source heterogeneous data in the service data.

[0016] The embodiment of the present application further provides a business data sharing and fusion system based on a trusted data space, including: an acquisition module, configured to acquire business data of at least one task, where the business data includes task requirement data; a construction module, configured to construct a trusted data space infrastructure, where the trusted data space infrastructure is used to indicate deployed node devices, and the node devices include data access gateways and node security communication components; a data matching module, configured to establish a node path planning model based on the trusted data space infrastructure and the task requirement data, and perform priority evaluation and data matching on the business data based on the node path planning model to generate a data matching scheme; a deployment module, configured to form a model collaborative deployment scheme according to the data matching scheme and the business data, where the model collaborative deployment scheme is used to indicate allocating the tasks to each edge node; a perception scoring module, configured to perform perception scoring on each task according to the model collaborative deployment scheme, and perform resource scheduling on the tasks in combination with the perception scoring results to generate a multi-task concurrent transmission scheme; a collaborative management module, configured to construct a multi-domain network service level agreement evaluation model according to the multi-task concurrent transmission scheme and the business data, and perform service decomposition and business collaborative management based on the multi-domain network service level agreement evaluation model to generate a service quality guarantee scheme; a data fusion module, configured to perform heterogeneous processing on the business data, and perform data fusion on the business data based on the heterogeneous processing results to generate a data fusion scheme; a sharing module, configured to perform fusion sharing on the business data under the synergistic effect of the data matching scheme, the model collaborative deployment scheme, the multi-task concurrent transmission scheme, the service quality guarantee scheme, and the data fusion scheme.

[0017] The embodiment of the present application further provides a computer device, where the computer device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned business data sharing and fusion method based on a trusted data space.

[0018] The embodiment of the present application further provides a computer-readable storage medium, which stores computer instructions for causing a computer to execute the above-mentioned business data sharing and fusion method based on a trusted data space.

[0019] The embodiment of the present application further provides a computer program product, including computer instructions that, when executed by a processor, implement the steps of the above-mentioned business data sharing and fusion method based on a trusted data space.

[0020] The present application has the following technical effects: By constructing a trusted data space infrastructure, data interconnection and interoperability among multiple regional institutions are achieved, effectively solving the problem of information silos among regional institutions; based on the priority evaluation and data matching mechanism of the node path planning model, dynamic adaptive data matching is realized, meeting the requirements of business scenarios with high timeliness requirements; by adopting the edge LLM collaboration framework, through model segmentation, collaborative deployment and compression optimization, the utilization efficiency of edge computing resources is improved, and the deployment cost of large models is reduced; the multi-task communication mechanism realizes flexible communication resource allocation and improves the system performance in multi-task concurrent scenarios; the risk-aware SLA decomposition and business management mechanism improves the service quality guarantee system for cross-domain business collaboration and ensures the stable operation of the business. Therefore, this application solves the problem that the improvement of business collaboration efficiency in the prior art is limited, realizes a dynamic adaptive data matching and fusion mechanism, can effectively improve the utilization efficiency of edge computing resources, realize flexible allocation of communication resources in multi-task concurrent scenarios, and realize a perfect service quality guarantee mechanism in cross-domain business collaboration. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings required to be used in the embodiments will be briefly introduced below. The drawings herein are incorporated into the specification and constitute a part of this specification. These drawings show the embodiments consistent with the present disclosure and are used together with the specification to illustrate the technical solutions of the present disclosure. It should be understood that the following drawings only show some embodiments of the present disclosure and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0022] Figure 1 is the flowchart of the business data sharing and fusion method based on a trusted data space in the embodiments of this application; Figure 2 is the flowchart of the node path planning model establishment and data matching method in the embodiments of this application; Figure 3 is the flowchart of the method for forming a model collaborative deployment plan in the embodiments of this application; Figure 4 is the flowchart of the task-aware scoring and resource scheduling method in the embodiments of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present disclosure clearer, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present disclosure. Apparently, the described embodiments are only a part rather than all of the embodiments of the present disclosure. The components of the embodiments of the present disclosure described and illustrated in the accompanying drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present disclosure provided in the accompanying drawings is not intended to limit the scope of the claimed present disclosure, but merely represents selected embodiments of the present disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of the present disclosure without creative efforts shall fall within the scope of protection of the present disclosure.

[0024] It should be noted that like reference numerals and letters denote like items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0025] As used herein, the term "and / or" merely describes an association relationship and indicates that three relationships may exist. For example, A and / or B may represent three cases: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the term "at least one" as used herein means any one of a plurality or any combination of at least two of a plurality. For example, including at least one of A, B, and C may represent any one or more elements selected from the set composed of A, B, and C.

[0026] As Figure 1 shown, an embodiment of the present application provides a method for sharing and integrating service data based on a trusted data space, including steps S1 to S8, namely, obtaining service data, constructing a trusted data space infrastructure, establishing a node path planning model, forming a model collaborative deployment plan, performing task awareness scoring and resource scheduling, constructing a service level agreement evaluation model, performing data heterogeneity processing and integration, and realizing the integration and sharing of service data. The following is a detailed description of these steps.

[0027] S1: Obtain service data of at least one task, where the service data includes task requirement data. It should be noted that these service data may come from business systems in different departments and different fields and include various types of information such as task requirement data, business process data, and resource status data. By obtaining these service data, the specific content and priority of business requirements can be understood, providing a basis for subsequent data processing and task allocation.

[0028] In step S1, various types of business data related to the task are collected through multiple channels. These data come from different business systems and departments and may have diverse formats and structures. This application needs to perform preliminary processing on this data to ensure its integrity and usability. Obtaining business data is the starting point of the entire method and provides the basic materials for subsequent data processing and analysis. By comprehensively and accurately obtaining business data, a clearer understanding of the task requirements can be achieved, laying a solid foundation for subsequent data sharing and integration work.

[0029] S2: Construct a trusted data space infrastructure, which is used to indicate the deployed node devices. The node devices include data access gateways and node security communication components. In this application, the trusted data space provides a secure and reliable environment for data sharing among different departments and different fields. During the construction process, this application needs to deploy data access gateways and node security communication components, establish unified data standards and exchange specifications, and achieve trusted access and management of data resources. Through a distributed architecture design, the trusted data space avoids the single-point failure risk of the traditional centralized mode, improving the reliability and scalability of the system. At the same time, by establishing unified data classification and grading standards, metadata specifications, and data quality standards, the standardization and consistency of data during the circulation process are ensured. The deployment of data security control components guarantees the security of data during the sharing process and prevents the leakage and abuse of sensitive data.

[0030] A trusted data space refers to a data circulation and utilization infrastructure that, based on consensus rules, connects multiple parties to achieve the shared use of data resources. It is an application ecosystem for the co-creation of the value of data elements and an important carrier for supporting the construction of a national integrated data market. A trusted data space has three core capabilities: data trusted control, resource interaction, and value co-creation. Among them, the data trusted control capability ensures the security and compliance of data during the circulation and use processes; the resource interaction capability supports data exchange and service invocation among multiple parties; the value co-creation capability promotes the deep integration and value mining of data resources. Through the organic combination of these three core capabilities, the trusted data space provides a reliable basic support for cross-departmental and cross-field data sharing and business collaboration. It should be noted that in this application, the three terms "trusted data space", "data space", and "space" have the same meaning and refer to the aforementioned data circulation and utilization infrastructure.

[0031] In addition, the trusted data space infrastructure includes components such as a unified data standard system, a distributed node network, data resource directory management, and data security control. Specifically, the unified data standard system includes data classification and grading standards, metadata specifications, and data quality standards, etc.; the distributed node network realizes secure communication between nodes through the deployment of data access gateways; the data resource directory management establishes a unified resource registration, discovery, and access mechanism; the data security control component realizes functions such as data desensitization, access control, and audit traceability. Through the collaborative work of these components, the trusted data space infrastructure provides a secure and reliable environment for the sharing and integration of business data.

[0032] S3: Based on the trusted data space infrastructure and the task requirement data, establish a node path planning model, and based on the node path planning model, conduct priority evaluation and data matching for the business data to generate a data matching scheme. In this step, establishing a node path planning model and conducting data matching means analyzing the node status data in combination with the spatio-temporal map, constructing a path planning model that meets the timeliness requirements, and conducting priority evaluation and data matching for the business data. The node path planning model takes into account the mobility and network connection status of the nodes and can provide a reliable path selection for data transmission. The priority evaluation mechanism assigns reasonable priorities to different business data according to factors such as the urgency and importance of the tasks to ensure that critical businesses are processed first. The data matching process establishes a mapping relationship between them by analyzing the data source characteristics and task requirements, realizing the precise matching of data and tasks. This dynamic and adaptive data matching mechanism can effectively cope with complex and changing business environments and meet the requirements of high-timeliness business scenarios.

[0033] In some embodiments, as Figure 2 shown, step S3 includes the following sub-steps S3.1 to S3.5. S3.1: Based on the trusted data space infrastructure, analyze the node status data of the mobile nodes in combination with a preset spatio-temporal map, where the mobile nodes are used to represent the node devices related to the task, and the node status data includes node position changes and network connection status.

[0034] In this step, mobile node devices related to the task are first identified, which may be mobile terminals, edge servers, or other dynamic nodes accessing the network. Then, the state data of these nodes is analyzed through spatio-temporal graph technology, including node location changes and network connection status. It should be noted that the spatio-temporal graph is a data model that describes the change rules of nodes in the time and space dimensions, and it can capture the movement trajectories of nodes and the dynamic characteristics of network connections. By analyzing this state data, the future positions and connection states of nodes can be predicted, providing an important basis for path planning. Especially in scenarios where nodes move frequently, this spatio-temporal graph-based analysis method can effectively improve the accuracy and reliability of path planning.

[0035] S3.2: Perform path search based on the node state data and the service data, construct a node path planning model that meets the preset timeliness requirements, and evaluate the connection reliability between the mobile nodes to obtain a reliability evaluation result. In this step, an improved A algorithm can be used for path search. This algorithm takes into account node mobility and task timeliness constraints. It should be noted that this improved A algorithm introduces a time window constraint on the basis of the traditional A* algorithm to ensure that the planning result meets the timeliness requirements of task execution. For example, for a data collection task that needs to be completed within 15 minutes, nodes with stable connections within this time window are preferentially selected. Subsequently, through path search, a node path planning model is constructed, which describes the optimal path and time arrangement for data transmission. At the same time, the connection reliability between mobile nodes is evaluated to generate a reliability evaluation result, providing an important reference for subsequent task priority evaluation. This path planning method that considers node mobility and network reliability can effectively cope with data transmission challenges in a dynamic network environment.

[0036] In addition, the application of the improved A algorithm in node path planning is mainly reflected in its consideration of time factors. The traditional A algorithm mainly focuses on the shortest path in terms of spatial distance, while in this embodiment, a time window constraint is introduced, enabling path planning to meet specific timeliness requirements. Exemplarily, taking a specific scenario as an example: Suppose in a cross-departmental collaborative case handling system, relevant departments need to obtain on-site investigation data of a certain case within 15 minutes, and this data is distributed on multiple mobile law enforcement terminals. The system needs to plan a path from the data request origin to each mobile terminal and then to the data integration point to ensure that all data collection is completed within 15 minutes.

[0037] Specifically, the improved A algorithm is used for path search. First, a spatio-temporal graph is constructed, where nodes represent different mobile terminals, and edges represent the connection relationships between nodes. Each edge has two weights: transmission time and connection reliability. During the search process, the algorithm not only considers the spatial distance of the path (the focus of traditional A), but also considers the reachability and connection stability of each node within a specific time window.

[0038] In terms of specific operations, a time dimension attribute is added to each node to record the estimated time of arrival at that node. During the search process, the algorithm checks whether it can reach the target node within the time window (here it is 15 minutes). If the estimated time of arrival exceeds the time window, this path will be pruned and no longer explored. In addition, nodes with stable connections within the estimated time window will be preferentially selected. For example, if two paths can both be completed within 15 minutes, but the nodes on one path have more stable connections during this period (possibly because these law enforcement officers work in fixed locations rather than moving), then this more stable path will be preferentially selected.

[0039] In actual operation, when a relevant department initiates a data request, path planning is immediately executed. Suppose there are 5 mobile terminals holding relevant data, and an optimal path is calculated, indicating to first obtain data from node A (possibly the nearest and most stable-connected terminal), then nodes B, C, D, E, and finally integrate all the data and send it to the relevant department. The entire process is planned to be completed within 14 minutes, meeting the timeliness requirement of 15 minutes. If it is found during the execution process that a certain node (such as node C) suddenly moves to a weak signal area and the connection becomes unstable, the path will be re-planned in real time, possibly adjusted to the order of A→B→E→D→C, or if the connection of C is too unstable and it is not critical data, C may even be temporarily skipped to first complete the collection of other data. Therefore, this embodiment can effectively address the path planning challenges in a dynamic network environment, ensure that data transmission meets the timeliness requirements, and is of great significance for time-sensitive tasks.

[0040] S3.3: According to the node path planning model and the reliability evaluation result, perform a priority evaluation on the service data to determine the calculation task weight related to the task priority evaluation result.

[0041] In this step, a distributed computing architecture is adopted, and evaluation agents are locally deployed on each node to perform multi-dimensional priority evaluation on business data. Multiple factors are considered in the evaluation process: the urgency of tasks (such as the highest priority for emergency response), the timeliness requirements of data (such as real-time data stream processing), and the importance of the business (such as core business processes), etc. Subsequently, a weighted scoring model is used to dynamically calculate the task weights, and the priority information is synchronized in the distributed network through a consensus mechanism. The evaluation results are output in the form of standardized scores, representing the task weights related to business data. This distributed task priority evaluation mechanism can quickly respond to local needs while considering the overall situation, improving the overall efficiency and response speed of the system.

[0042] S3.4: Based on the trusted data space infrastructure and the computing task weights, perform data matching on the business data to form a matching scheme. In this step, first, construct a data source feature vector and a task requirement vector. The data source feature vector includes dimensions such as data type, update frequency, and quality score, describing the basic characteristics and quality status of the data source; the task requirement vector includes dimensions such as data specification, timeliness requirement, and accuracy requirement, describing the specific requirements of the task for the data. Then, calculate the similarity between these two types of vectors, through cosine similarity or other similarity calculation methods, to quantify the matching degree between the data source and the task requirements. Subsequently, based on the similarity calculation results, establish a mapping relationship between the data source and the task requirements of the business data to form an initial matching scheme. This data matching method based on vector similarity can evaluate the adaptability between the data source and the task requirements from multiple dimensions, improving the accuracy and effectiveness of the matching.

[0043] In an optional embodiment, in step S3.4, the specific implementation manner of data matching includes: Based on the trusted data space infrastructure, construct a data source feature vector and a task requirement vector, where the data source feature vector includes data type, update frequency, and quality score, and the task requirement vector includes data specification, timeliness requirement, and accuracy requirement; Based on the business data, calculate the similarity between the data source feature vector and the task requirement vector, and based on the similarity, establish a mapping relationship between the data source and the task requirements of the business data; Based on the mapping relationship, form a matching scheme, and update the matching scheme according to changes in task priorities and network status adjustments.

[0044] Specifically, the implementation of data matching first involves the construction of data source feature vectors and task requirement vectors. The data source feature vector is a multi-dimensional description of the characteristics of the data source, including key attributes such as data type, update frequency, and quality score. For example, for a real-time video stream data source, its feature vector may include attributes such as data type being "video stream", update frequency being "30 frames per second", and quality score being "high definition". The task requirement vector describes the specific requirements of the task for the data, including dimensions such as data specification, timeliness requirement, and accuracy requirement. For example, for a video analysis task, its requirement vector may include attributes such as data specification being "H.264 encoded video", timeliness requirement being "real-time processing", and accuracy requirement being "high definition resolution".

[0045] After constructing these two types of vectors, similarity calculation methods are used to evaluate the degree of matching between them. Commonly used similarity calculation methods include cosine similarity, Euclidean distance, and weighted similarity, etc. Exemplarily, taking cosine similarity as an example, after normalizing the two vectors, the cosine value of the angle between them is calculated. The closer the cosine value is to 1, the more similar the two vectors are, that is, the higher the matching degree between the data source and the task requirement. For example, if the similarity calculated between a high-definition video stream data source and an AI analysis task that requires high-definition video input is 0.95, while the similarity with a simple monitoring task that only requires low-resolution data is 0.6, the system will preferentially match this data source to the AI analysis task.

[0046] Based on the similarity calculation results, a mapping relationship between the data source and the task requirement is established. This process uses a bipartite matching algorithm, which considers the global optimal matching between multiple data sources and multiple task requirements. Specifically, first, an initial matching is established according to the similarity level, and then the matching relationship is adjusted through iterative optimization until a stable state is reached. In actual operation, a similarity threshold may be set, and a matching relationship is only established when the similarity exceeds a specific threshold to ensure the quality of the matching.

[0047] After forming the initial matching plan, the matching plan can be dynamically updated according to changes in task priorities and network status. When the task priority changes, such as when an emergency task is inserted or the priority of an original task is increased, the system will re-evaluate the resource allocation, which may break the original matching relationship and re-allocate high-quality data sources to high-priority tasks. When the network status changes, such as when the bandwidth of some transmission paths decreases or the latency increases, the selection of data sources is adjusted accordingly, and data sources on paths with better transmission conditions are preferentially used. This dynamic adjustment mechanism ensures that the system can adapt to environmental changes and always maintain a high matching efficiency.

[0048] S3.5: Compose the node path planning model, the computing task weights, and the matching scheme into a data matching scheme, and update the data matching scheme according to network state changes, task urgency, and a preset performance metric system.

[0049] In this step, first, integrate the node path planning model, computing task weights, and the mapping relationship between data sources and task requirements generated in the previous steps into a complete data matching scheme. In this embodiment, it describes which data sources should be matched with which task requirements, the paths for data transmission, as well as the priority and resource allocation of each task. Then, monitor changes in network state, task urgency, and performance metrics, and dynamically update the data matching scheme based on these changes. For example, when the network connection state changes, the path planning can be recalculated; when a new high-priority task is inserted, the resource allocation and priority sorting can be adjusted; when a certain performance metric degrades, the relevant matching strategy can be optimized. This feedback optimization mechanism enables the data matching scheme to adapt to the dynamic changes in the environment and maintain the best matching effect.

[0050] S4: Based on the data matching scheme and the service data, form a model collaborative deployment scheme, which is used to indicate the allocation of the tasks to each edge node.

[0051] In this step, first, collect the resource monitoring data of edge nodes, construct a resource profile, and predict the resource usage trend. Then, based on the inter-model layer dependencies and resource usage trends, divide the tasks into multiple computing tasks to form a model task allocation scheme. To improve the computing efficiency of edge nodes, the model parameters are also quantized, compressed, and optimized to reduce the computational complexity and storage requirements. Finally, adopt a load balancing strategy to distribute the optimized model components to each node device and edge node to form a complete model collaborative deployment scheme. This edge LLM collaboration framework makes full use of distributed computing resources, improves the efficiency of edge computing, reduces the cost of large model deployment, and enables complex computing models to run efficiently in resource-constrained edge environments.

[0052] In some embodiments, as Figure 3 shown, step S4 includes the following sub-steps S4.1 to S4.7. S4.1: Based on the data matching scheme, collect edge node resource monitoring data through the deployed lightweight resource monitoring agent. The edge node resource monitoring data includes CPU utilization, memory usage, GPU computing load, and network bandwidth data.

[0053] In step S4.1, the lightweight resource monitoring agent is a low-overhead software component deployed on edge nodes to collect resource usage. These agents adopt a distributed architecture and can collect the resource status of edge nodes in real time with minimal system load. The specific monitoring data includes CPU utilization (percentage of core usage, thread load distribution), memory usage (physical memory occupancy, virtual memory allocation, memory fragmentation), GPU computing load (computing core utilization rate, video memory occupancy, computing queue length), and network bandwidth data (up and down traffic, link congestion status, packet loss rate). These agents transmit the collected data back through asynchronous communication to minimize the impact of the monitoring process on business execution.

[0054] S4.2: Construct a multi-dimensional resource profile including computing power, storage capacity, and network performance based on the edge node resource monitoring data, and generate an edge node performance evaluation result. In step S4.2, a multi-dimensional resource profile is constructed based on the resource monitoring data. This profile is a multi-dimensional vector that precisely describes the computing power of the edge node (such as floating-point operation performance, parallel processing ability, instruction set support), storage capacity (including storage space size, read / write speed, I / O concurrency ability), and network performance (such as link stability, communication latency, bandwidth fluctuation range). Standardized processing is used to normalize the metrics of different dimensions, and a weighted scoring mechanism is introduced to assign different weights to different resource dimensions according to the task characteristics, thereby generating a comprehensive edge node performance evaluation result. This multi-dimensional resource profile enables the system to comprehensively understand the resource status of edge nodes and provides an accurate basis for subsequent task allocation.

[0055] S4.3: Establish a load prediction model based on the edge node performance evaluation result.

[0056] In step S4.3, a load prediction model is established based on the edge node performance evaluation result. This model combines time series analysis and machine learning methods, integrates historical load data and current state information, and predicts the resource load changes of each node in the future for a period of time. Specifically, long short-term memory networks (LSTM) are used to process the historical load sequence to capture the periodic and trend changes of the load; at the same time, a regression analysis method is combined, considering the current known task schedule and resource allocation situation to improve the accuracy of the prediction. The model also considers the handling mechanism for emergencies and abnormal loads, and can dynamically adjust the prediction result when detecting load anomalies to maintain the adaptability and robustness of the model.

[0057] S4.4: Calculate the resource usage trend based on the load prediction model. In step S4.4, based on the established load prediction model, calculate the resource usage trend within the future time window. This calculation process includes trend analysis at multiple time scales, covering short-term (minute-level), medium-term (hour-level), and long-term (day-level) resource usage predictions. The sliding window method is used to continuously update the prediction results to ensure the real-time nature of the trend calculation. The calculation results include not only the predicted values of various resource utilization rates but also resource saturation assessment, bottleneck resource identification, and resource elasticity space analysis. These trend data provide important references for subsequent task allocation, enabling the advance planning of resource allocation strategies and avoiding resource competition and overload situations.

[0058] S4.5: Based on the inter-model layer dependencies indicated by the service data, divide the task into multiple computing tasks, and form a model task allocation scheme that acts within or between the node devices according to the resource usage trend. The model task allocation scheme is used to indicate the allocation of the computing tasks to the edge nodes.

[0059] In step S4.5, analyze the inter-model layer dependencies in the service data, and decompose complex tasks into multiple computing subtasks with clear dependencies. This process uses a directed acyclic graph (DAG) to represent task dependencies, identifying task groups that can be executed in parallel and task chains that must be processed serially. Combining the previously calculated resource usage trend, the system constructs a mixed-integer programming model to solve the optimal task allocation scheme. This scheme not only considers the multi-core parallel processing capabilities within the node but also fully utilizes the collaborative computing potential between nodes to form a cross-node distributed execution strategy. At the same time, comprehensively weigh the real-time requirements of tasks, data transmission overhead, and computational complexity to ensure the practicality and efficiency of the allocation scheme.

[0060] S4.6: According to the model task allocation scheme, perform INT8 quantization compression on the model parameters of the computing model related to the task, and execute frequent co-occurrence token sequence fusion to construct an optimized model component.

[0061] In step S4.6, optimizing the computing model is a key link to improve edge computing efficiency. First, perform INT8 quantization compression on the model parameters, converting the original floating-point parameters to 8-bit integer representation, significantly reducing the parameter storage space (usually reducing by 75%) and computational complexity, while maintaining the model accuracy through calibration techniques. In addition, also execute frequent co-occurrence token sequence fusion, identify the frequently occurring token sequence patterns in the model, and merge them into a single operation unit to reduce computational redundancy. This optimization not only reduces the storage and computational requirements of the model but also improves the execution efficiency of the model in resource-constrained environments, creating conditions for model deployment on edge devices.

[0062] S4.7: Based on the optimized model components, deploy the computing model distributively to each node device and / or edge node by using a preset edge node load balancing strategy to generate a model collaborative deployment plan.

[0063] In step S4.7, based on the optimized model components and the preset load balancing strategy, the computing model is distributively deployed to each node device and edge node. This load balancing strategy comprehensively considers static load balancing (based on node hardware configuration and network topology) and dynamic load adjustment (based on real-time resource status and task execution). A hierarchical deployment architecture is adopted to allocate components at different levels of the model to suitable computing nodes and establish a secure communication channel between components. In addition, a real-time monitoring and exception handling mechanism for task execution is implemented, which can trigger an emergency handling process when detecting a performance bottleneck or node failure to ensure the reliable execution of the model collaborative deployment plan. This distributive deployment method makes full use of edge computing resources and realizes the efficient operation of complex models in resource-constrained environments.

[0064] S5: According to the model collaborative deployment plan, perform a perception score for each task, and combine the perception score results to perform resource scheduling for the task to generate a multi-task concurrent transmission plan. In this step, first, task features are extracted from the service data, task understanding and business value evaluation are carried out, task scores are calculated and the priority ranking results are output. Then, channel status data is collected, the change trend of channel quality is predicted, and the channel resource allocation strategy is determined. Finally, based on the resource allocation strategy and task priorities, a multi-task concurrent transmission plan is generated. This LLM-based multi-task perception scoring mechanism, combined with channel adaptive technology, realizes the flexible allocation of communication resources, can guarantee the communication requirements of key services in a multi-task concurrent scenario, and improves the overall performance. Especially in the case of tight or highly fluctuating network resources, it can dynamically adjust the resource allocation strategy to ensure the execution quality of important tasks.

[0065] In some embodiments, as Figure 4 shown, S5 includes the following sub-steps S5.1 to S5.7. S5.1: Based on the model collaborative deployment plan, extract the task type, data scale, and time attribute features of the task from the service data to construct task features. Specifically, a feature extraction algorithm is used to identify and extract task types (such as real-time monitoring, batch processing analysis, interactive queries, etc.), data scales (such as data volume size, processing complexity, data distribution characteristics, etc.), and time attribute features (such as deadlines, execution cycles, response time requirements, etc.) from structured and unstructured business data. During the extraction process, multi-modal feature fusion technology is adopted to integrate different forms of information such as text descriptions, structured parameters, and historical execution records into a unified feature representation. This representation is in vector form, with each dimension corresponding to a key characteristic of the task, forming a comprehensive task feature portrait, laying the foundation for subsequent task understanding and evaluation.

[0066] S5.2: Based on the task characteristics, perform task understanding on the business data to obtain the business context analysis result. In this step, semantic analysis technology is adopted, combined with a pre-trained domain knowledge model, to understand the role and significance of the task in the business process. Through comprehensive analysis of the task description, relevant business rules, and historical execution situations, the business objectives, precondition dependencies, and subsequent impact scopes of the task can be identified. On this basis, a business context graph is constructed to depict the association relationships between the task and other business activities, including direct dependencies and indirect impacts. This in-depth task understanding enables the system to evaluate the importance of the task from the overall business perspective, rather than being limited to the technical parameter level, thereby obtaining a more comprehensive and accurate business context analysis result.

[0067] S5.3: Based on the business context analysis result, evaluate the business value and execution priority of the task to form the importance evaluation result of the task. In this step, based on the business context analysis result, multi-dimensional evaluation of the business value and execution priority of the task is carried out. The business value evaluation involves both quantitative and qualitative aspects: quantitative evaluation includes the quantitative calculation of the direct economic benefits, risk avoidance value, and compliance guarantee value of the task; qualitative evaluation considers the impact of the task on business continuity, user experience, and organizational reputation. The execution priority evaluation comprehensively considers the urgency of the task (such as whether there are strict time limits), importance (such as whether it is related to core business), and association degree (such as the impact scope on other tasks). The analytic hierarchy process (AHP) and fuzzy comprehensive evaluation method are used to integrate these factors into a comprehensive score to form the importance evaluation result of the task. This evaluation result not only reflects the absolute importance of the task but also considers the relative value of the task in the entire business ecosystem, providing a scientific basis for subsequent resource allocation.

[0068] S5.4: Based on the importance assessment result and the preset task performance, calculate the task score of the task, and output the multi-task priority sorting result according to the task score, where the task performance includes data timeliness, business relevance, and resource consumption. In this step, based on the importance assessment result, combined with the preset task performance indicators, calculate the comprehensive score of the task. The task performance indicators mainly include three aspects: data timeliness (real-time requirements and expiration date of data), business relevance (degree of closeness to the core business process), and resource consumption (computing, storage, and network resources required to complete the task). Using a weighted summation model, different weights are assigned to these indicators according to the characteristics of different business scenarios, and the comprehensive score of the task is calculated. In the case of multi-task concurrency, based on these score results, use an improved priority queue algorithm to generate the multi-task priority sorting result. This sorting takes into account the dependencies between tasks and resource competition, ensuring that high-priority tasks can obtain sufficient resources while avoiding long-term starvation of low-priority tasks, achieving fairness and efficiency in task scheduling.

[0069] S5.5: Based on the multi-task priority sorting result, collect channel status data through distributed detection nodes deployed on the edge nodes, where the channel status data includes bandwidth utilization, link delay, and packet loss rate data. In this step, based on the multi-task priority sorting result, collect channel status data through the distributed detection node network. These detection nodes are strategically deployed at key positions in the network to form a monitoring network covering the entire network, capable of collecting the status information of each communication channel in real time. The specific data collected includes bandwidth utilization (the ratio of the current transmission traffic of the channel to the maximum capacity), link delay (the transmission time of the data packet from the source to the destination), and packet loss rate data (the proportion of lost data packets during transmission). The collection process uses a method combining low-overhead active detection and passive listening to ensure that the monitoring itself does not have a significant impact on network performance. These detection nodes adopt a hierarchical reporting mechanism to transmit the collected raw data to the central analysis system after preliminary processing, providing comprehensive and accurate basic data for subsequent channel quality prediction.

[0070] S5.6: According to the channel status data, use a preset time series analysis model to predict the change trend of channel quality, and construct a channel availability evaluation index. Based on the channel quality change trend and the channel availability evaluation index, determine the channel resource allocation strategy. In step S5.6, based on the collected channel status data, a preset time series analysis model is used to predict the change trend of channel quality. This model combines the autoregressive integrated moving average (ARIMA) method and deep learning technology, and can capture the regularity and uncertainty of channel quality changes over time. The prediction results include short-term (from seconds to minutes) and medium-term (hourly) bandwidth changes, delay fluctuations, and packet loss rate trends. Based on these predictions, a comprehensive channel availability evaluation index system is constructed, including a stability index (measuring the fluctuation degree of channel quality), a reliability score (measuring the ability of the channel to meet service quality requirements), and an adaptability index (measuring the ability of the channel to cope with bursty traffic). According to these quality change trends and evaluation indicators, a multi-objective optimization algorithm is used to determine the channel resource allocation strategy, which takes into account both the rigid constraints of meeting the needs of high-priority tasks and the improvement of the overall network resource utilization efficiency.

[0071] S5.7: Determine the multi-task concurrent transmission scheme for the tasks according to the channel resource allocation strategy and the multi-task priority ranking result.

[0072] In step S5.7, based on the channel resource allocation strategy and the multi-task priority ranking result, a multi-task concurrent transmission scheme is finally determined. This scheme is a detailed scheduling plan that assigns specific transmission time windows, channel selections, and resource quotas to each task. When formulating the scheme, the system uses a heuristic scheduling algorithm, comprehensively considering task priorities, data dependencies, channel quality predictions, and the overall system load. For high-priority tasks, sufficient channel resources are reserved and quality of service guarantees are provided; for general-priority tasks, their resource allocations are dynamically adjusted according to the current network conditions; for low-priority tasks, an opportunistic scheduling strategy is adopted to utilize idle network periods for transmission. In addition, the scheme also includes an emergency handling mechanism that can quickly adjust the transmission strategy when a sudden deterioration of channel quality is detected to ensure the normal execution of critical tasks. This multi-level and adaptive transmission scheme effectively improves the network resource utilization efficiency while ensuring the reliability and timeliness of task execution.

[0073] In the process of multi-task perception scoring and resource scheduling, task features are first extracted from business data, and task understanding is performed based on these features to obtain the business context analysis results. Then, the business value and execution priority of the tasks are evaluated to form the importance evaluation results, and the task scores are calculated by combining task performance (such as data timeliness, business relevance, and resource consumption), and the multi-task priority sorting results are output. On this basis, the channel status data is collected through distributed detection nodes, the change trend of the channel quality is predicted, the channel availability evaluation index is constructed, and the channel resource allocation strategy is determined. Finally, according to the channel resource allocation strategy and the multi-task priority sorting results, the multi-task concurrent transmission scheme is determined. This design enables the communication requirements of important tasks to be preferentially guaranteed under limited resources, improving the overall efficiency.

[0074] S6: According to the multi-task concurrent transmission scheme and the business data, construct a multi-domain network service level agreement evaluation model, and perform service decomposition and business collaborative management based on the multi-domain network service level agreement evaluation model to generate a service quality guarantee scheme. In this step, first, a service quality index system is established, the service risks are evaluated to form a risk assessment scheme, and a multi-domain network SLA evaluation model is generated. Then, the system constructs a business process dependency graph and a service topology structure, determines the search strategy, and iteratively optimizes the service decomposition scheme. Finally, cross-domain collaborative management is executed to generate a service quality guarantee scheme. This risk-aware SLA decomposition and business management mechanism innovatively solves the problem of service quality guarantee in multi-domain networks and improves the service quality guarantee system for cross-domain business collaboration. Through the accurate identification and quantification of service risks and the dynamic optimization of the service decomposition scheme, it is possible to ensure the stable operation of the business in a complex multi-domain network environment and improve the reliability and continuity of the service.

[0075] In some embodiments, step S6 includes the following sub-steps S6.1 to S6.6. S6.1: Based on the multi-task concurrent transmission scheme, establish a service quality index system, and the service quality index system includes delay, availability, and throughput. In step S6.1, a quality of service (QoS) indicator system is established based on the multi-task concurrent transmission scheme, which is the foundation for constructing the multi-domain network service level agreement (SLA) evaluation model. The indicator system adopts a hierarchical design, including basic layer indicators and derivative layer indicators. The basic layer indicators include latency (network transmission delay, processing response time, end-to-end completion time), availability (service continuity guarantee rate, system recovery time objective, fault recovery point objective), and throughput (business volume processed per unit time, peak processing capacity, data transmission rate). The derivative layer indicators are derived from the basic indicators and include the business resilience index (measuring the system's ability to cope with load fluctuations), service coherence score (measuring the smoothness of business process execution), and user experience indicators (evaluating the service quality perception from the user's perspective). This set of indicator systems has the characteristics of being quantifiable, comparable, and traceable, providing a scientific basis for the evaluation and management of service quality, and taking into account the dual requirements of the technical dimension and the business dimension.

[0076] S6.2: Based on the quality of service indicator system, evaluate the risk data of service risk events related to the task, and perform hierarchical early warning analysis based on the risk data to form a risk assessment plan, where the risk data includes the occurrence probability and the impact scope.

[0077] In step S6.2, based on the quality of service indicator system, a comprehensive evaluation of service risk events related to the task is carried out. The risk data evaluation uses a structured method to quantitatively analyze each type of potential risk event from two dimensions: the occurrence probability and the impact scope. The occurrence probability evaluation is based on historical data statistics and expert knowledge models, considering various information such as the current state, environmental factors, and known threats; the impact scope evaluation analyzes factors such as the scope of business processes that may be affected by the risk event, the scale of the user group, and the potential economic losses. On this basis, hierarchical early warning analysis is performed, and the risk events are classified into different levels according to the severity (such as critical level, high level, medium level, and low level), and corresponding treatment strategies and resource reservation plans are formulated for each level. This risk perception-based analysis method can identify potential service quality threats in advance and conduct targeted risk prevention and control, forming a comprehensive and detailed risk assessment plan.

[0078] S6.3: Based on the risk assessment plan, combined with the service quality requirements of the specified business scenario, generate a multi-domain network service level agreement evaluation model.

[0079] In step S6.3, according to the risk assessment plan and combined with the service quality requirements of the specified business scenario, a multi-domain network service level agreement evaluation model is generated. This model adopts a hierarchical design and includes two dimensions: the business layer SLA and the technical layer SLA. The business layer SLA defines the expectations for service quality from the user's perspective, such as the business processing time limit, service available time window, and business success rate, etc.; the technical layer SLA decomposes these business requirements into specific technical indicators, such as network response time, throughput capacity, and data processing accuracy, etc. The model realizes the two-way traceability from business requirements to technical implementation by establishing the mapping relationship between business indicators and technical indicators. In addition, the model also considers the particularity of the multi-domain network environment, introduces inter-domain cooperation indicators and boundary responsibility definitions, clarifies the service quality responsibilities and cooperation mechanisms between different network domains, and ensures the consistency and coherence of service quality for cross-domain services.

[0080] S6.4: Based on the multi-domain network service level agreement evaluation model, construct a business process dependency graph for the task, and construct an output service topology structure according to the business process dependency graph. The service topology structure is used to indicate the call relationship and resource dependency between service components in the task. In step S6.4, based on the multi-domain network service level agreement evaluation model, a business process dependency graph for the task is constructed. This dependency graph adopts a directed graph structure, where nodes represent business activities or service components, and edges represent the dependency relationships and interaction methods between them. During the construction process, the system analyzes business process documents, historical execution logs, and service call relationships, and identifies key business nodes and dependency paths. The dependency graph not only describes explicit call relationships but also considers implicit data dependencies and state dependencies, comprehensively reflecting the complex interactions in the actual operation of the business. Based on this dependency graph, a service topology structure is further constructed. This structure describes the call relationship and resource dependency between service components in the task from the perspective of technical implementation. The service topology structure includes detailed information such as the deployment location, interface definition, communication method, and resource requirements of service components, providing a technical basis for subsequent service decomposition. This mapping from business processes to service topologies realizes the seamless connection between the business perspective and the technical perspective, and helps to ensure the consistency between technical implementation and business requirements.

[0081] S6.5: According to the service topology structure, determine a search strategy. The search strategy is used to indicate the iterative optimization of the service decomposition plan and evaluate the plan indicators of the service decomposition plan. When the plan indicators reach the preset convergence conditions, the service decomposition result is output. The plan indicators include the feasibility and optimization space of the service decomposition plan, and the service decomposition plan is used to indicate the adjustment of service component configuration. In step S6.5, according to the service topology structure, determine the search strategy for service decomposition optimization. This strategy adopts an iterative optimization method and searches for the optimal service decomposition scheme through multiple rounds of search. Specifically, first generate an initial set of decomposition schemes based on the service topology structure, and then evaluate and improve the candidate schemes in the set in each round of iteration. The search process uses an improved genetic algorithm to continuously optimize the solution space through crossover, mutation, and selection operations. In each round of evaluation, calculate the scheme metrics, including the feasibility of the scheme (whether it meets all technical constraints and business requirements) and the optimization space (the resource efficiency and performance improvement potential of the scheme). When the scheme metrics reach the preset convergence condition, stop the iteration and output the final service decomposition result. During this iterative optimization process, continuously adjust the configuration parameters, deployment locations, and interaction methods of service components to find the optimal decomposition scheme that not only meets business requirements but also maximizes resource utilization efficiency.

[0082] S6.6: Based on the service decomposition result, perform cross-domain collaborative management to generate a service quality assurance scheme.

[0083] In step S6.6, based on the service decomposition result, perform cross-domain collaborative management to generate the final service quality assurance scheme. Cross-domain collaborative management is a key link to ensure service quality consistency in a multi-domain network environment. First, establish an inter-domain service coordination mechanism, define the collaboration rules, data exchange standards, and service interoperability protocols between different network domains. Then, based on the pre-defined multi-domain network SLA evaluation model, implement distributed service monitoring, collect the service operation status in each domain in real time, and identify potential service quality problems through collaborative analysis. When a service quality anomaly is detected, trigger the preset collaborative processing process, coordinate the resource scheduling and service adjustment within the relevant domains, and ensure end-to-end service quality. The finally formed service quality assurance scheme includes the service quality maintenance strategy during normal operation, the emergency handling mechanism in case of anomalies, and the methodology for continuous service quality optimization, constituting a complete service quality life cycle management system. This risk-aware cross-domain collaborative management mechanism effectively solves the problem of service quality assurance in a multi-domain network environment and provides a reliable guarantee for the stable operation of cross-domain services.

[0084] During the service level agreement evaluation and business collaborative management process, first establish a service quality index system, evaluate the risk data of service risk events, form a risk assessment scheme, and generate a multi-domain network service level agreement evaluation model. Then, construct a business process dependency graph and a service topology structure, determine the search strategy, iteratively optimize the service decomposition scheme, and output the service decomposition result when the scheme metrics reach the convergence condition. Finally, based on the service decomposition result, perform cross-domain collaborative management to generate a service quality assurance scheme. This design enables the system to ensure the service quality of various services in a complex multi-domain network environment and improve the reliability and stability of the system.

[0085] S7: Perform heterogeneous processing on the service data, and perform data fusion on the service data based on the heterogeneous processing result to generate a data fusion scheme. In this step, use a semantic knowledge graph to extract the semantic features of the data sources, perform data quality assessment and cleaning, establish a data feature model, analyze the data feature portrait, and finally generate a data fusion scheme. The fusion and sharing mechanism for the multi-source heterogeneous characteristics of service data solves the integration problem of data from different sources and in different formats, and improves the efficiency and value of data utilization. Through semantic understanding and mapping, the semantic consistency between different data sources can be ensured; through data quality assessment and cleaning, the usability and reliability of the data can be improved; through data feature analysis and fusion, deeper information and value can be mined from multi-source data to support more complex business requirements and decision-making analysis.

[0086] In some embodiments, step S7 includes the following sub-steps S7.1 to S7.4. S7.1: Use a preset semantic knowledge graph to extract the semantic features of different data sources indicated by the service data. In step S7.1, use a preset semantic knowledge graph to extract the semantic features of different data sources in the service data. This semantic knowledge graph is a multi-level knowledge representation structure, including three levels: domain concepts, attribute relationships, and entity instances. In the extraction process, first identify the key entities and concepts in the service data and map them to the corresponding nodes in the knowledge graph; then analyze the relationships between the entities and identify the implicit semantic connections in the data; finally, extract the structural features and content features of the data sources to form a complete semantic feature representation. Specifically, adopt deep semantic analysis technology, combine natural language processing and ontology reasoning methods to extract unified semantic representations from structured data, semi-structured data, and unstructured data. This semantic feature extraction can not only understand the surface meaning of the data, but also deeply mine the context correlation and implicit information of the data, providing a semantic-level foundation for subsequent data fusion and ensuring the semantic consistency and compatibility of data from different sources.

[0087] S7.2: Use preset quality assessment indicators to perform data quality assessment on the service data, and clean the service data according to the data quality assessment result to obtain the cleaned service data, where the quality assessment indicators include the integrity, accuracy, consistency, and timeliness of the service data. In step S7.2, a comprehensive quality assessment and cleaning of the business data is carried out using preset quality assessment metrics. The quality assessment metric system covers four core dimensions of data quality: integrity (the missing situation of data records and fields), accuracy (the degree of conformity of data values with the actual situation), consistency (the degree of consistency of data across different systems and time points), and timeliness (the update frequency and expiration date of data). A method combining a rule engine and a statistical model is adopted to perform multi-dimensional quality scoring on each data record and generate a detailed quality assessment report. Based on the assessment results, a data cleaning process is executed, including missing value handling (such as imputation, deletion, or marking), outlier detection and correction (based on statistical models and domain rules), redundant data removal (using similarity calculation and entity matching), and format standardization (unifying data formats and unit systems). The cleaning process adopts an incremental design, generating a cleaned data version while retaining the original data and recording all cleaning operations to ensure the traceability and reversibility of data processing. The execution of this step significantly improves the quality and reliability of subsequent data fusion and lays a solid foundation for the full excavation of data value.

[0088] S7.3: Based on the semantic features, a data feature model is established, and the data feature portrait is analyzed using the data feature model, where the data feature portrait includes the structural features and data distribution rules of the data source. In step S7.3, a feature model of multi-source data is established based on the extracted semantic features. This model is represented in a graph structure, where nodes represent data entities or attributes, and edges represent the semantic relationships or data dependencies between them. During the model construction process, machine learning methods are used to learn the correlations and evolution rules between data from historical data, and explicit semantic constraints are supplemented in combination with domain expert knowledge to form a hybrid feature model with both data-driven and knowledge-driven characteristics. Using this feature model, in-depth analysis of the data source is carried out to generate a data feature portrait. This portrait not only includes the structural features of the data source (such as data schema, type distribution, association relationships) and data distribution rules (such as numerical distribution characteristics, time variation trends, spatial distribution patterns), but also higher-level semantic features and quality features. This all-round feature analysis enables the system to deeply understand the characteristics and values of each data source, provides a decision-making basis for the selection of intelligent data fusion strategies, and effectively improves the accuracy and efficiency of data fusion.

[0089] S7.4: Based on the data feature portrait and the cleaned business data, a data fusion plan is generated, where the data fusion plan is used to indicate the fusion of multi-source heterogeneous data in the business data.

[0090] In this step, the solution adopts a multi-level fusion architecture, including three levels: entity-level fusion, attribute-level fusion, and relationship-level fusion. Entity-level fusion solves the problem of identifying and mapping the same entity in different data sources, and uses entity resolution technology to identify and merge different records referring to the same object; attribute-level fusion deals with the conflict and complementary relationship of entity attributes, and uses the truth discovery algorithm to select the most reliable attribute values from multiple sources; relationship-level fusion integrates the complex associations between entities and constructs a unified relationship network. During the fusion process, according to the data characteristics and business requirements, appropriate fusion strategies are dynamically selected. For example, strict matching rules are adopted for data with high consistency requirements, while more relaxed association strategies may be adopted for exploratory analysis. At the same time, a fusion quality monitoring mechanism is established to evaluate the fusion effect through a series of indicators (such as fusion coverage rate, conflict resolution rate, information gain), and the fusion strategy is continuously optimized based on the feedback. The finally generated data fusion solution not only guides the technical fusion implementation of multi-source heterogeneous data, but also ensures that the fusion result can effectively support business requirements and truly realizes the maximization of data value.

[0091] In the process of fusing and processing multi-source heterogeneous data, first, the preset semantic knowledge graph is used to extract the semantic features of different data sources to ensure the semantic consistency of the data. Then, quality assessment indicators are used to evaluate and clean the business data to improve the data availability. Next, a data feature model is established based on the semantic features to analyze the data feature portrait, including the structural features and distribution rules of the data sources. Finally, based on the data feature portrait and the cleaned business data, a data fusion solution is generated to achieve the effective fusion of multi-source heterogeneous data. This design enables the system to process data from different sources and in different formats, improving the quality and efficiency of data fusion.

[0092] S8: Under the synergistic effect of the data matching solution, the model co-deployment solution, the multi-task concurrent transmission solution, the service quality guarantee solution, and the data fusion solution, the business data is fused and shared.

[0093] In this application, this is the ultimate goal of the entire method. Through the organic combination of the data matching scheme, the model collaborative deployment scheme, the multi-task concurrent transmission scheme, the service quality guarantee scheme, and the data fusion scheme, the efficient sharing and deep integration of business data are achieved. In practical applications, this mechanism of multi-scheme collaboration can effectively handle various complex business scenarios and meet the data sharing needs of different departments and different fields. For example, in cross-departmental collaborative work, it can ensure the real-time transmission and processing of various business data and support the seamless connection of business processes; in data analysis and decision-making, it can integrate multi-source data and provide comprehensive and accurate information support; in emergency response scenarios, it can prioritize the execution of critical services and improve the response speed and efficiency. This all-round ability to share and integrate business data provides strong support for digital transformation and intelligent development.

[0094] Therefore, this application can dynamically adapt to changing network environments and task requirements, ensuring the accuracy and effectiveness of data matching. By constructing the data source feature vector and the task requirement vector, the matching degree between the data source and the task requirement can be evaluated from multiple dimensions, improving the accuracy of data matching. At the same time, the matching scheme can also be dynamically updated according to changes in task priorities and network status adjustments, ensuring the real-time and adaptability of data matching.

[0095] In the model collaborative deployment scheme, first, a lightweight resource monitoring agent is used to collect various resource data of edge nodes, construct a multi-dimensional resource profile, and based on this, establish a load prediction model to calculate the resource usage trend. This information provides a decision-making basis for subsequent task allocation. Then, according to the inter-layer dependence relationship of the model, the task is divided into multiple computing tasks, and a model task allocation scheme is formed. To improve the processing efficiency of edge nodes, the computing model is also optimized by INT8 quantization compression and token sequence fusion, significantly reducing the computational complexity and storage requirements of the model. Finally, a load balancing strategy is used to distribute the optimized model components to each node in a distributed manner, generating a complete model collaborative deployment scheme.

[0096] In addition, in one embodiment, step S8 is to fuse and share business data under the collaborative action of all the previous schemes. Specifically, based on the data matching scheme, the mapping relationship between the data source and the task requirement is determined, the computing model is efficiently deployed using the model collaborative deployment scheme, the efficiency and quality of data transmission are guaranteed through the multi-task concurrent transmission scheme, the service level of each service is ensured according to the service quality guarantee scheme, and the fusion processing of multi-source heterogeneous data is achieved according to the data fusion scheme. The collaborative action of these schemes enables the security, reliability, and efficiency of data sharing to be guaranteed while meeting various business needs.

[0097] Exemplarily, in an actual application scenario, the method of the present application can be applied to cross - department and cross - domain business data sharing and integration. For example, in the field work of grass - roots units, mobile terminals need to collect on - site data in real time and interact with the central system. Through pre - path planning, reliable data transmission can be ensured even in areas with unstable signals. The task priority evaluation mechanism gives priority to processing information with a high degree of urgency, such as information related to important cases. The adaptive matching algorithm ensures that the multimedia data collected on - site can be quickly associated and matched with the historical data in the archive system. Through resource status awareness, processing tasks are reasonably allocated to each computing node. This systematic design not only guarantees the efficiency of business handling but also ensures the security and standardization of information processing.

[0098] The embodiment of the present application also provides a business data sharing and integration system based on a trusted data space. The business data sharing and integration system 500 based on a trusted data space includes: An acquisition module 501, configured to acquire business data of at least one task, where the business data includes task requirement data; A construction module 502, configured to construct a trusted data space infrastructure, where the trusted data space infrastructure is used to indicate deployed node devices, and the node devices include data access gateways and node security communication components; A data matching module 503, configured to establish a node path planning model based on the trusted data space infrastructure and the task requirement data, and perform priority evaluation and data matching on the business data based on the node path planning model to generate a data matching scheme; A deployment module 504, configured to form a model collaborative deployment scheme according to the data matching scheme and the business data, where the model collaborative deployment scheme is used to indicate allocating the task to each edge node; A perception scoring module 505, configured to perform perception scoring on each task according to the model collaborative deployment scheme, and perform resource scheduling on the task in combination with the perception scoring result to generate a multi - task concurrent transmission scheme; A collaborative management module 506, configured to construct a multi - domain network service level agreement evaluation model according to the multi - task concurrent transmission scheme and the business data, and perform service decomposition and business collaborative management based on the multi - domain network service level agreement evaluation model to generate a service quality guarantee scheme; A data fusion module 507, configured to perform heterogeneous processing on the business data, and perform data fusion on the business data based on the heterogeneous processing result to generate a data fusion scheme; A shared module 508 is used to perform fusion and sharing of the service data under the collaborative effect of the data matching scheme, the model collaborative deployment scheme, the multi-task concurrent transmission scheme, the service quality guarantee scheme, and the data fusion scheme.

[0099] The functions of each module correspond one by one to the corresponding steps of the above method, and jointly implement the function of sharing and fusing service data. Through the above detailed description, it can be seen that the service data sharing and fusion system and method based on a trusted data space provided by this application have obvious technological innovation and practical value, can effectively solve the problems existing in the prior art, and improve the efficiency and quality of data sharing and service collaboration.

[0100] An embodiment of this application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is run by a processor, it executes the steps of the method for sharing and fusing service data based on a trusted data space described in the above method embodiment. Among them, the storage medium can be a volatile or non-volatile computer-readable storage medium.

[0101] In addition, an embodiment of this application also provides a computer program product, on which a computer program is stored. When the computer program is run by a processor, it executes the steps of the method for sharing and fusing service data based on a trusted data space provided in any of the above embodiments of the present disclosure. For details, reference can be made to the above method embodiments, and details are not described herein again.

[0102] Among them, the above computer program product can be specifically implemented in a manner of hardware, software, or a combination thereof. In an optional embodiment, the computer program product is specifically embodied as a computer storage medium, which can be a volatile or non-volatile computer-readable storage medium. In another optional embodiment, the computer program product is specifically embodied as a software product, such as a Software Development Kit (SDK), etc.

[0103] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described devices and apparatuses can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein. In several embodiments provided in the present disclosure, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For another example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the couplings or direct couplings or communication connections shown or discussed with each other can be through some communication interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0104] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0105] In addition, in each embodiment of the present disclosure, the functional units can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.

[0106] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium executable by a processor. Based on such an understanding, the technical solution of the present disclosure, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present disclosure. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0107] Finally, it should be noted that the above-described embodiments are only specific embodiments of the present disclosure, which are used to illustrate the technical solutions of the present disclosure, rather than limiting them. The protection scope of the present disclosure is not limited thereto. Although the present disclosure has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that any person skilled in the art within the technical scope disclosed by the present disclosure can still modify the technical solutions recorded in the foregoing embodiments, or can easily think of changes, or perform equivalent replacements on some of the technical features; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present disclosure, and should all be covered within the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be subject to the protection scope of the claims.

Claims

1. A business data sharing and fusion method based on a trusted data space, characterized in that, Including: Obtain the business data of at least one task, where the business data includes task requirement data; Construct a trusted data space infrastructure, which is used to indicate the deployment node devices, and the node devices include data access gateways and node security communication components; Based on the trusted data space infrastructure and the task requirement data, establish a node path planning model, and based on the node path planning model, perform priority evaluation and data matching on the business data to generate a data matching scheme; According to the data matching scheme and the business data, form a model collaborative deployment scheme, which is used to indicate allocating the tasks to each edge node; According to the model collaborative deployment scheme, perform perception scoring on each task, and combine the perception scoring results to perform resource scheduling on the tasks to generate a multi-task concurrent transmission scheme; According to the multi-task concurrent transmission scheme and the business data, construct a multi-domain network service level agreement evaluation model, and based on the multi-domain network service level agreement evaluation model, perform service decomposition and business collaborative management to generate a service quality guarantee scheme; Perform heterogeneous processing on the business data, and based on the heterogeneous processing results, perform data fusion on the business data to generate a data fusion scheme; Under the synergistic effect of the data matching scheme, the model collaborative deployment scheme, the multi-task concurrent transmission scheme, the service quality guarantee scheme and the data fusion scheme, perform fusion sharing on the business data.

2. The method according to claim 1, wherein The step of based on the trusted data space infrastructure and the task requirement data, establishing a node path planning model, and based on the node path planning model, performing priority evaluation and data matching on the business data to generate a data matching scheme includes: Based on the trusted data space infrastructure, combine the preset spatio-temporal map to analyze the node state data of the mobile nodes, where the mobile nodes are used to represent the node devices related to the tasks, and the node state data includes node position changes and network connection states; Perform path search based on the node state data and the business data, construct a node path planning model that meets the preset timeliness requirements, and evaluate the connection reliability between the mobile nodes to obtain a reliability evaluation result; According to the node path planning model and the reliability evaluation result, perform priority evaluation on the business data to determine the calculation task weights related to the task priority evaluation result; Based on the trusted data space infrastructure and the calculation task weights, perform data matching on the business data to form a matching scheme; Combine the node path planning model, the calculation task weights and the matching scheme to form a data matching scheme, and update the data matching scheme according to the network state changes, task urgency and the preset performance index system.

3. The method according to claim 2, wherein The step of based on the trusted data space infrastructure and the calculation task weights, performing data matching on the business data to form a matching scheme includes: Based on the trusted data space infrastructure, construct a data source feature vector and a task requirement vector, where the data source feature vector includes data type, update frequency, and quality score, and the task requirement vector includes data specification, timeliness requirement, and accuracy requirement; Based on the business data, calculate the similarity between the data source feature vector and the task requirement vector, and based on the similarity, establish a mapping relationship between the data source and the task requirement of the business data; Based on the mapping relationship, form a matching scheme, and update the matching scheme according to the change of task priority and network status adjustment.

4. The method according to claim 3, wherein The formation of the model collaborative deployment scheme according to the data matching scheme and the business data includes: Based on the data matching scheme, collect edge node resource monitoring data through the deployed lightweight resource monitoring agent, where the edge node resource monitoring data includes CPU utilization, memory usage, GPU computing load, and network bandwidth data; According to the edge node resource monitoring data, construct a multi-dimensional resource profile including computing power, storage capacity, and network performance, and generate an edge node performance evaluation result; Based on the edge node performance evaluation result, establish a load prediction model; Based on the load prediction model, calculate the resource usage trend.

5. The method according to claim 4, characterized in that, The formation of the model collaborative deployment scheme according to the data matching scheme and the business data further includes: Based on the model layer dependency relationship indicated by the business data, divide the task into multiple computing tasks, and based on the resource usage trend, form a model task allocation scheme for acting within or between node devices, where the model task allocation scheme is used to indicate the allocation of the computing tasks to the edge nodes; According to the model task allocation scheme, perform INT8 quantization compression on the model parameters of the computing model related to the task, and perform frequent co-occurrence token sequence fusion to construct an optimized model component; Based on the optimized model component, use a preset edge node load balancing strategy to distribute the computing model to each node device and / or edge node, and generate a model collaborative deployment scheme.

6. The method according to claim 5, characterized in that The generation of a multi-task concurrent transmission scheme by performing a perception score on each task according to the model collaborative deployment scheme and performing resource scheduling on the task in combination with the perception score result includes: Based on the model collaborative deployment scheme, extract the task type, data scale, and time attribute features of the task from the business data to construct a task feature; Based on the task feature, perform task understanding on the business data to obtain a business context analysis result; Based on the business context analysis result, evaluate the business value and execution priority of the task to form an importance evaluation result of the task; Based on the importance evaluation result and the preset task performance, calculate the task score of the task, and output a multi-task priority sorting result according to the task score, where the task performance includes data timeliness, business relevance, and resource consumption.

7. The method according to claim 6, characterized in that, Performing perception scoring on each task according to the model collaborative deployment plan, and combining the perception scoring results to perform resource scheduling on the task to generate a multi-task concurrent transmission plan, including: Based on the multi-task priority sorting result, collecting channel status data through distributed detection nodes deployed on the edge node, where the channel status data includes bandwidth utilization rate, link delay, and packet loss rate data; According to the channel status data, using a preset time series analysis model to predict the channel quality change trend, constructing a channel availability evaluation index, and determining a channel resource allocation strategy based on the channel quality change trend and the channel availability evaluation index; Based on the channel resource allocation strategy and the multi-task priority sorting result, determining the multi-task concurrent transmission plan of the task.

8. The method according to claim 1, wherein Constructing a multi-domain network service level agreement evaluation model according to the multi-task concurrent transmission plan and the service data, and performing service decomposition and service collaboration management based on the multi-domain network service level agreement evaluation model to generate a service quality guarantee plan, including: Based on the multi-task concurrent transmission plan, establishing a service quality index system, where the service quality index system includes delay, availability, and throughput; Based on the service quality index system, evaluating the risk data of service risk events related to the task, performing hierarchical early warning analysis based on the risk data to form a risk assessment plan, where the risk data includes occurrence probability and impact range; Based on the risk assessment plan, combining the service quality requirements of a specified business scenario to generate a multi-domain network service level agreement evaluation model.

9. The method according to claim 1, characterized in that Constructing a multi-domain network service level agreement evaluation model according to the multi-task concurrent transmission plan and the service data, and performing service decomposition and service collaboration management based on the multi-domain network service level agreement evaluation model to generate a service quality guarantee plan, further including: Based on the multi-domain network service level agreement evaluation model, constructing a business process dependency graph of the task, and constructing an output service topology structure based on the business process dependency graph, where the service topology structure is used to indicate the call relationship and resource dependency between service components in the task; According to the service topology structure, determining a search strategy, where the search strategy is used to indicate iteratively optimizing the service decomposition plan and evaluating the plan metrics of the service decomposition plan, and outputting the service decomposition result when the plan metrics reach preset convergence conditions, where the plan metrics include the feasibility and optimization space of the service decomposition plan, and the service decomposition plan is used to indicate adjusting the service component configuration; Based on the service decomposition result, performing cross-domain collaborative management to generate a service quality guarantee plan; Then, performing heterogeneous processing on the service data, and performing data fusion on the service data based on the heterogeneous processing result to generate a data fusion plan, including: Using a preset semantic knowledge graph to extract the semantic features of different data sources indicated by the service data; Using preset quality assessment indicators, perform data quality assessment on the business data, and clean the business data according to the data quality assessment results to obtain the cleaned business data, where the quality assessment indicators include the integrity, accuracy, consistency, and timeliness of the business data; Based on the semantic features, establish a data feature model, and use the data feature model to analyze the data feature portrait, where the data feature portrait includes the structural features of the data source and the data distribution law; Based on the data feature portrait and the cleaned business data, generate a data fusion plan, where the data fusion plan is used to indicate the fusion of multi-source heterogeneous data in the business data.

10. A business data sharing and fusion system based on a trusted data space, characterized in that, It includes: An acquisition module, configured to acquire business data of at least one task, where the business data includes task requirement data; A construction module, configured to construct a trusted data space infrastructure, where the trusted data space infrastructure is used to indicate the deployment of node devices, and the node devices include data access gateways and node security communication components; A data matching module, configured to establish a node path planning model based on the trusted data space infrastructure and the task requirement data, and perform priority assessment and data matching on the business data based on the node path planning model to generate a data matching plan; A deployment module, configured to form a model collaborative deployment plan according to the data matching plan and the business data, where the model collaborative deployment plan is used to indicate the allocation of the tasks to each edge node; A perception scoring module, configured to perform perception scoring on each task according to the model collaborative deployment plan, and perform resource scheduling on the tasks in combination with the perception scoring results to generate a multi-task concurrent transmission plan; A collaborative management module, configured to construct a multi-domain network service level agreement evaluation model according to the multi-task concurrent transmission plan and the business data, and perform service decomposition and business collaborative management based on the multi-domain network service level agreement evaluation model to generate a service quality guarantee plan; A data fusion module, configured to perform heterogeneous processing on the business data, and perform data fusion on the business data based on the heterogeneous processing results to generate a data fusion plan; A sharing module, configured to perform fusion sharing on the business data under the synergistic effect of the data matching plan, the model collaborative deployment plan, the multi-task concurrent transmission plan, the service quality guarantee plan, and the data fusion plan.

Images