Scalable Network Model Management Method and Apparatus Based on Digital Twin

Abstract

This application provides a method and apparatus for managing scalable network models based on digital twins. The method involves determining the network attributes corresponding to each meta-network model in a preset network model library and receiving a target network construction instruction. When no scalable network model of the same category exists in the network model registry, a category is created and associated with domain identifier information, and a cross-domain attribute mapping rule set is generated. The corresponding meta-network model is determined according to the target network construction instruction. In response to an instantiation request, incremental instantiation parameters are obtained by comparing the target sub-attributes with the meta-network model sub-attributes. The target network is then constructed based on the incremental parameters and network attributes. Finally, the target network is associated with physical network resources, and protocol adaptation is performed using the cross-domain attribute mapping rule set. This method effectively solves the shortcomings of low efficiency, low flexibility, and low stability in constructing target networks, significantly improving the flexibility and stability of the target network.

Description

Technical Field

[0001] This application relates to the field of network security technology, specifically to a scalable network model management method and apparatus based on digital twins. Background Technology

[0002] Digital twin technology is a current research hotspot in information networks. It can use software simulation and network simulation technology to achieve a highly realistic reproduction of nodes, links and topology relationships in a target network, making the reproduction results flexible, scalable and highly realistic.

[0003] However, cyberspace itself is a complex virtual system. In digital twin network applications, the target network has uncertainty, and the components of the target network also have uncertainty. This makes it difficult for digital twin network models to adapt to different types of target network characteristics and to cope with the ever-increasing types of network elements, links, and services. As a result, the efficiency of building the target network is low, and the resulting target network has reduced flexibility and stability to some extent. Summary of the Invention

[0004] To address the problems in the existing technology, this application provides a scalable network model management method and apparatus based on digital twins, which can effectively solve the shortcomings of traditional technologies in terms of low efficiency in constructing target networks and the resulting target networks having reduced flexibility and stability to some extent. It significantly improves the construction efficiency of target networks and enhances the flexibility and stability of target networks.

[0005] To solve at least one of the above problems, this application provides the following technical solution:

[0006] Firstly, this application provides a scalable network model management method based on digital twins, comprising:

[0007] Determine the network attributes corresponding to each meta-network model included in the preset network model library, and receive the target network construction instruction. The network attributes include static attributes, components, model rules and dynamic features. The target network construction instruction includes network model classification, target sub-attributes and domain identification information.

[0008] When there is no extensible network model with the same network model classification in the network model registry, a corresponding classification is created based on the network model classification and associated with domain identification information. A cross-domain attribute mapping rule set is generated based on the domain identification information, and the meta-network model of the current network model classification is determined in the preset network model library based on the target network construction instructions. The cross-domain attribute mapping rule set includes conversion rules for attributes between heterogeneous domains.

[0009] In response to the instantiation request, the target sub-attributes are compared with the corresponding sub-attributes of the meta-network model to obtain incremental instantiation parameters. Based on the meta-network model, the target network is constructed using the incremental instantiation parameters and network attributes.

[0010] The target network is associated with physical network resources, and protocol adaptation is performed between the target network and physical network resources through cross-domain attribute mapping rule sets, so that the parameters of the multi-domain model of the target network based on federated learning are updated collaboratively.

[0011] Furthermore, after determining the network attributes corresponding to each meta-network model included in the preset network model library and receiving the target network construction instruction, the process also includes:

[0012] When an extensible network model with the same network model category exists in the network model registry, the cross-domain attribute mapping rule set corresponding to the extensible network model is updated based on the domain identifier information.

[0013] In response to the instantiation request, the target sub-attributes and the corresponding sub-attributes of the scalable network model are compared to obtain the incremental instantiation parameters. Based on the scalable network model, the target network is constructed using the incremental instantiation parameters and network attributes.

[0014] Furthermore, after constructing the target network based on incrementally instantiated parameters and network properties on the basis of the meta-network model, it also includes:

[0015] Receive event triggering instructions, which include time-triggered instructions, status-triggered instructions, and external triggering instructions;

[0016] For the executable script corresponding to the current dynamic feature-related event triggering instruction, historical traffic data is injected into the simulation environment, and the execution result of the executable script is obtained based on the historical traffic data;

[0017] If the execution result contains rule conflicts or resource limits, generate optimization prompts based on the execution result and mark high-risk operations.

[0018] Furthermore, it also includes: parsing the protocol type and security level included in the domain identifier information, and selecting a conversion template that matches the protocol type and security level from the preset rule template library;

[0019] Based on the conversion template, the protocol type and security level are mapped to the network model registry to obtain a cross-domain attribute mapping rule set.

[0020] Furthermore, it also includes: performing syntax tree comparison on the target sub-attribute and the corresponding sub-attribute of the meta-network model, determining the update node based on the syntax tree comparison, and determining the update node as the incremental instantiation parameter;

[0021] After obtaining the incremental instantiation parameters, the following is also included:

[0022] Based on the incremental instantiation parameters, the update operations to be performed on the sub-attributes corresponding to the meta-network model are determined, and the transaction logs corresponding to the update operations are recorded. The update operations include add operations, delete operations, and modify operations.

[0023] Furthermore, it also includes: deploying a hierarchical parameter aggregator in the federated learning framework, which includes local gradient encryption units at the edge node level and differential privacy processing units at the domain level;

[0024] The model contribution of each domain to the target network is determined by a hierarchical parameter aggregator. The sampling weights for the next training round are adjusted based on the model contribution, so as to collaboratively update the multi-domain model parameters in federated learning according to the sampling weights.

[0025] Furthermore, it also includes: static attributes, including basic static attributes, structured attribute labels, and mapping relationship tables; wherein, the basic static attributes are obtained based on the topology, device type, and fixed configuration parameters of the meta-network model in the preset network model library;

[0026] Structured attribute tags are obtained by performing semantic analysis on technical documents;

[0027] The mapping table includes physical device characteristics and protocol constraints.

[0028] Secondly, this application provides a scalable network model management device based on digital twins, comprising:

[0029] The first processing module is used to determine the network attributes corresponding to each meta-network model included in the preset network model library, and to receive the target network construction instruction. The network attributes include static attributes, constituent elements, model rules and dynamic features. The target network construction instruction includes network model classification, target sub-attributes and domain identification information.

[0030] The second processing module is used to create a corresponding category based on the network model category and associate domain identification information when there is no extensible network model with the same network model category in the network model registry. It generates a cross-domain attribute mapping rule set based on the domain identification information and determines the meta-network model of the current network model category in the preset network model library based on the target network construction instruction. The cross-domain attribute mapping rule set includes conversion rules for attributes between heterogeneous domains.

[0031] The third processing module is used to respond to the instantiation request, compare the target sub-attributes with the corresponding sub-attributes of the meta-network model, obtain the incremental instantiation parameters, and construct the target network based on the incremental instantiation parameters and network attributes on the basis of the meta-network model.

[0032] The fourth module is used to associate the target network with physical network resources. It adapts the target network to physical network resources through a cross-domain attribute mapping rule set, so that the parameters of the target network's multi-domain model based on federated learning can be updated collaboratively.

[0033] Thirdly, this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method described.

[0034] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the described scalable network model management method based on digital twins.

[0035] Fifthly, this application provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the described scalable network model management method based on digital twins.

[0036] As can be seen from the above technical solution, this application provides a scalable network model management method and apparatus based on digital twins. It innovatively determines the network attributes corresponding to each meta-network model included in a preset network model library and the received target network construction instructions. When no scalable network model of the same category is matched in the network model registry, the category is created and associated with domain identification information, generating a cross-domain attribute mapping rule set including heterogeneous domain attribute conversion rules. Simultaneously, a meta-network model corresponding to the target category is selected from the preset network model library according to the target network construction instructions. In response to the instantiation request, the target sub-attributes and the corresponding sub-attributes of the meta-network model are compared to obtain incremental instantiation parameters. The target network is constructed by combining the network attributes, and the target network is associated with physical network resources. Protocol adaptation is performed between the cross-domain attribute mapping rule set and physical network resources, enabling the target network to be synchronously updated according to the multi-domain model parameters of federated learning. This method effectively solves the shortcomings of traditional techniques, such as low efficiency in constructing target networks and a certain degree of reduction in the flexibility and stability of the resulting target networks. It significantly improves the construction efficiency of target networks and enhances the flexibility and stability of the target networks. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1This is a flowchart illustrating the scalable network model management method based on digital twins in an embodiment of this application.

[0039] Figure 2 This is a structural diagram of the scalable network model management device based on digital twins in the embodiments of this application;

[0040] Figure 3 This is a schematic diagram of the structure of the electronic device in the embodiments of this application.

[0041] Figure label:

[0042] Electronic device 9600, central processing unit 9100, memory 9140, communication module 9110, input unit 9120, audio processor 9130, display 9160, power supply 9170, buffer memory 9141, application / function storage unit 9142, data storage unit 9143, driver storage unit 9144, antenna 9111, speaker 9131, microphone 9132. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0044] The acquisition, storage, use, and processing of data in this application all comply with the relevant provisions of national laws and regulations.

[0045] In existing technologies, the uncertainty of the target network in digital twin network applications comes from many components such as device type, link type, routing selection, topology, application requirements, and user characteristics. This means that digital twin network models need to adapt to the network characteristics of different types of target networks, and need to cover the simulation of a series of elements from composition, topology, status to users. Furthermore, the target network needs to be able to adapt to the ever-increasing types of network element devices, link types, and service types.

[0046] In view of the problems existing in the prior art, this application provides a scalable network model management method and apparatus based on digital twins. By determining the network attributes corresponding to each meta-network model included in the preset network model library and the received target network construction instructions, when no scalable network model of the same category is matched in the network model registry, the category is created and associated with domain identification information, generating a cross-domain attribute mapping rule set including heterogeneous domain attribute conversion rules. At the same time, the meta-network model corresponding to the category is selected from the preset network model library according to the target network construction instructions. In response to the instantiation request, the target sub-attributes and the corresponding sub-attributes of the meta-network model are compared to obtain incremental instantiation parameters. The target network is constructed in combination with the network attributes, and the target network is associated with physical network resources. Protocol adaptation is performed with physical network resources through the cross-domain attribute mapping rule set, so that the target network is updated synchronously according to the multi-domain model parameters of federated learning.

[0047] To effectively address the shortcomings of traditional technologies, such as low efficiency in constructing target networks and reduced flexibility and stability, and to significantly improve the construction efficiency, flexibility, and stability of target networks, this application provides an embodiment of a scalable network model management method based on digital twins. See [link to embodiment]. Figure 1 The scalable network model management method based on digital twins specifically includes the following:

[0048] Step S101: Determine the network attributes corresponding to each meta-network model included in the preset network model library, and receive the target network construction instruction.

[0049] The network attributes include static attributes, constituent elements, model rules, and dynamic features, while the target network construction instructions include network model classification, target sub-attributes, and domain identification information.

[0050] Optionally, this embodiment determines the network attributes corresponding to each meta-network model included in the preset network model library. Multiple meta-network models stored in the library can be accessed through the preset network model library interface, and the network attributes of each meta-network model can be parsed one by one. The network attributes include static attributes, constituent elements, model rules, and dynamic features.

[0051] Static attributes are used to represent the inherent characteristic information of the meta-network model. They are shared by instances of the same type of meta-network model and the information of instances of the same type of target model will not change within a simulation cycle. Static attributes can be used to describe the meta-network model and its bound resource instances. Different components have corresponding inherent and specific static attributes.

[0052] The components include, but are not limited to, node elements, software elements, link elements, and subnet elements. Node attributes vary depending on the node type and mainly include basic attributes (such as identifier, type, manufacturer, model, etc.), computing attributes (such as CPU, memory, hard disk, temperature, processes, registry, etc.), and network configuration attributes (such as IP address, port, number of network cards, VLAN, routing table, etc.).

[0053] The attributes of software elements include, but are not limited to, basic information, software functions, software operating environment, software dependencies, and software connectivity relationships.

[0054] The attributes of link elements include, but are not limited to, basic link information, link physical medium, link bandwidth, network latency, protocol type, power, channel frequency, packet loss rate, and quality of service (QoS).

[0055] Subnet element attributes include, but are not limited to, basic network information, network size, topology, network purpose, and network hierarchy.

[0056] Among them, model rules refer to the constraints that need to be followed when using meta-network models, including connection rules between meta-network models, deployment rules for software on nodes, configuration rules for static attributes, etc. They play a constraining role in the process of constructing the target network based on the meta-network model, and also serve as the basis for network verification during the construction and configuration of the target network.

[0057] In addition, the system can receive target network construction instructions and construct the target network according to the instructions. The target network construction instructions include network model classification, target sub-attributes, and domain identification information.

[0058] This embodiment enhances the adaptability and flexibility of target model construction by controlling network attributes. It can improve the accuracy of response to the target network construction instruction based on the content included in the instruction, thereby improving the construction efficiency of the target model.

[0059] Step S102: When there is no extensible network model with the same network model classification in the network model registry, create a corresponding classification based on the network model classification and associate it with domain identification information, generate a cross-domain attribute mapping rule set based on the domain identification information, and determine the meta-network model of the current network model classification in the preset network model library based on the target network construction instruction.

[0060] The cross-domain attribute mapping rule set includes conversion rules for attributes between heterogeneous domains.

[0061] Optionally, in this embodiment, if it is determined that there is no scalable network model in the network model registry that is the same as the network model category included in the target network construction instruction, a corresponding category entry is created according to the network model category in the target network construction instruction. Here, association can be understood as establishing a mapping relationship, so as to be able to identify the characteristics of different domains under the network model category.

[0062] Among them, network model classification can be used to organize and manage the target models belonging to the network model classification.

[0063] In addition, the newly created network model classification is associated with the received domain identification information, and a cross-domain attribute mapping rule set is generated based on the domain identification information. The cross-domain attribute mapping rule set can be used to define and encapsulate the mutual conversion rules of attributes (such as configuration parameters, performance indicators, and status identifiers) between different heterogeneous network domains (such as source domain and target domain, or domains of different technology stacks).

[0064] For example, a cross-domain attribute mapping rule set may include logic for mapping specific QoS parameters in domain A to their equivalent parameters or configuration items in domain B. In other embodiments, the cross-domain attribute mapping rule set may also include calculation rules for mutual conversion when different domains have different units of expression and calculation methods for bandwidth.

[0065] In addition, in the preset network model library, the meta-network model of the current network model category is determined according to the target network construction instructions. The meta-network model can be a basic, templated model structure that can define the general architecture, components and basic relationships of the current network model category.

[0066] This embodiment implements a meta-network model that determines the network model classification through a network model registry and target network construction instructions. It generates a cross-domain attribute mapping rule set based on the domain identification information in the target network construction instructions, thereby enhancing the dynamism and adaptability of scalable network model management. It can automatically respond to the needs of new network model classifications and improve the ability of heterogeneous network domains to work together.

[0067] Step S103: In response to the instantiation request, compare the target sub-attributes with the corresponding sub-attributes of the meta-network model to obtain the incremental instantiation parameters. Based on the meta-network model, construct the target network using the incremental instantiation parameters and network attributes.

[0068] Optionally, in response to the instantiation request, this embodiment compares the target sub-attribute with the sub-attribute corresponding to the previously determined meta-network model to obtain incremental instantiation parameters. These incremental instantiation parameters can be viewed as personalized parameters for the meta-network model, and can be determined through attribute difference analysis logic, capturing the necessary changes between the target sub-attribute and the sub-attribute.

[0069] In addition, based on the meta-network model, the target network is obtained by combining incremental instantiation parameters and determined network attributes to complete the instantiation request.

[0070] Furthermore, after obtaining the target network based on the instantiation request, a network model registry for the target network is obtained. The network model registry includes the network attributes of the target network. If new sub-attribute fields need to be added, simply register the attribute field name, identifier, type, length, value, and other information according to the original model rules. This enables the addition and instantiation of network models with different categories and attribute information, thereby achieving the scalability of the network model.

[0071] This embodiment enables the construction of a target network by incrementally instantiating parameters and network attributes, enhancing the flexibility and efficiency of target network instantiation, improving resource utilization and response speed, and simultaneously improving the adaptability and scalability of the target network model.

[0072] Step S104: Associate the target network with physical network resources, and adapt the protocol with the physical network resources through the cross-domain attribute mapping rule set so that the parameters of the target network's multi-domain model based on federated learning can be updated collaboratively.

[0073] Optionally, this embodiment associates the target network with physical network resources, including hardware devices such as routers and switches. The association process involves identifying resource attributes and performing initial mapping, and performing protocol adaptation through a cross-domain attribute mapping rule set. The cross-domain attribute mapping rule set includes the conversion relationship between virtual domain attributes and physical domain attributes. For example, it maps virtual network bandwidth requirements to physical device configuration parameters and handles compatibility issues between different network protocols through protocol conversion, so that data streams can be seamlessly connected during cross-domain transmission.

[0074] In addition, in the adapted environment, the collaborative update mechanism of the federated learning framework can be used, in which the target network can act as a node in the federated learning, and exchange model parameter update gradients with other domain nodes through a secure communication channel, so as to achieve distributed collaborative optimization of multi-domain model parameters without sharing the original data.

[0075] This embodiment realizes protocol adaptation between cross-domain attribute mapping rule sets and physical network resources, enabling the target network to collaboratively update the parameters of the multi-domain model learned from both sides. This enhances the cross-domain compatibility and protocol interoperability of network resources, improves the model training efficiency and factor protection level of federated learning in multi-domain environments, and improves performance and resource utilization.

[0076] Furthermore, after the target network is constructed, it is published to the network model management platform and stored in the preset network model library, where it serves as a meta-network model. This allows the current target network to be used as a meta-network model and extended based on it, thus improving the scalability of the target network.

[0077] This embodiment enables the creation of categories and association of domain identifier information when no corresponding network model category exists in the network model registry, enhancing the scalability of network models. By determining the meta-network model from a preset network model library through target network construction instructions, the construction efficiency of the target network is improved, thus fully meeting the requirements of complex target network construction in the target network construction instructions. It can accurately locate network attributes. Effective handling of network model classification and cross-domain attribute mapping rule sets improves the construction efficiency, network adaptability, and target network quality, reducing errors and resource waste during the target network construction process. Simultaneously, it allows the target network to adapt to various physical network environments, improving the flexibility and scalability of the target network.

[0078] In some embodiments, after determining the network attributes corresponding to each meta-network model included in the preset network model library and receiving the target network construction instruction, the method further includes:

[0079] When an extensible network model with the same network model category exists in the network model registry, the cross-domain attribute mapping rule set corresponding to the extensible network model is updated based on the domain identifier information.

[0080] In response to the instantiation request, the target sub-attributes and the corresponding sub-attributes of the scalable network model are compared to obtain the incremental instantiation parameters. Based on the scalable network model, the target network is constructed using the incremental instantiation parameters and network attributes.

[0081] Optionally, in this embodiment, a matching process is performed in the network model registry based on the network model classification in the target network construction instruction. If an extensible network model with the same network model classification exists in the network model registry, the cross-domain attribute mapping rule set corresponding to the extensible network model will be updated according to the domain identifier information in the target network construction instruction, so that it can better adapt to the characteristics of the domain involved in the current target network construction instruction.

[0082] The scalable network model can be a meta-network model or a target network previously obtained based on the target network construction instructions. For example, if the network model classification is a common classification, there may be a pre-configured meta-network model in the network model classification. The meta-network model is then identified as the scalable network model. Alternatively, the network model classification may contain a pre-configured meta-network model and a pre-configured target network. The target network model can then be directly identified as the scalable network model.

[0083] Among them, the corresponding meta-network model can be determined from the preset network model library by using basic information such as the type, identifier, name, and description of the network model.

[0084] For example, the bandwidth conversion rules in the existing cross-domain attribute mapping rule set are old standards for a specific domain, while the new domain identification information display needs to adopt a new standard, and the corresponding rules need to be updated accordingly.

[0085] In addition, in response to the instantiation request, the target sub-attributes in the target network construction instruction are compared and analyzed with the sub-attributes corresponding to the scalable network model to determine the differences between the sub-attributes and the target sub-attributes, thereby obtaining the incremental instantiation parameters.

[0086] The target sub-attributes include, but are not limited to, information such as identifier, name, type, and value range.

[0087] Furthermore, based on the scalable network model, the incremental instantiation parameters and the determined network properties will be combined to adjust and optimize the scalable network model, thereby constructing a target network that conforms to the target network construction instructions.

[0088] This embodiment enables the adaptation of existing scalable network models by updating cross-domain attribute mapping rule sets, or the rapid construction of new target networks based on scalable models. Both methods ensure that the target network can maintain good performance and functionality in changing environments. Furthermore, by utilizing scalable network models and precise parameter adjustments, target networks can be quickly built and expanded, reducing the time and economic costs of network construction and upgrades.

[0089] In some embodiments, after constructing the target network based on incrementally instantiated parameters and network attributes on the basis of the meta-network model, the method further includes:

[0090] Receive event triggering instructions, which include time-triggered instructions, status-triggered instructions, and external triggering instructions;

[0091] For the executable script corresponding to the current dynamic feature-related event triggering instruction, historical traffic data is injected into the simulation environment, and the execution result of the executable script is obtained based on the historical traffic data;

[0092] If the execution result contains rule conflicts or resource limits, generate optimization prompts based on the execution result and mark high-risk operations.

[0093] Optionally, this embodiment receives an event triggering instruction, wherein the event triggering instruction includes a time triggering instruction, a state triggering instruction, and an external triggering instruction, wherein:

[0094] Time-triggered commands: These are used to trigger corresponding operations based on time conditions. For example, a performance testing task for the target network can be triggered during a specific time period each day (e.g., from 2 a.m. to 4 a.m. when network traffic is low), or network billing and settlement related operations can be triggered on a specific date each month (e.g., the 1st of each month).

[0095] Status trigger command: Used to initiate corresponding processes based on specific status changes of the target network. For example, when the bandwidth utilization of the target network exceeds a preset threshold (e.g., 80%), it triggers automatic adjustment of the target network traffic, or when a network device malfunctions, it triggers fault alarms and backup device switching operations.

[0096] External trigger command: An event originating from outside the target network, such as receiving an operation request manually initiated by the user. The operation request can be by clicking an operation button through the network management interface, or by an interaction signal from other systems, such as the power system sending a power outage maintenance notification to the target network, triggering the network's energy-saving mode operation.

[0097] In addition, the dynamic characteristics of the current target network are associated with the executable scripts corresponding to the event triggering instructions. That is, the real-time dynamic situation of the target network is linked with a series of preset operation scripts, so that the target network can automatically execute the corresponding script operations according to its own dynamic characteristics.

[0098] For example, the dynamic characteristics of the target network include real-time fluctuations in network traffic. When a time-triggered command based on traffic monitoring is received, the corresponding traffic monitoring and analysis scripts are associated.

[0099] Furthermore, historical traffic data is injected into the simulation environment. This historical traffic data includes, but is not limited to, information such as traffic volume and type under different time periods and business scenarios. By using historical traffic data in the simulation environment, scenarios that closely resemble real network operation can be simulated.

[0100] Execute an executable script associated with an event-triggered command based on historical traffic data, and obtain the execution result. For example, execute a network congestion prediction script based on historical traffic data to obtain a prediction of network congestion that may occur at a future time.

[0101] Furthermore, when execution results in rule conflicts or resource overruns, optimization prompts are generated based on the execution results, and high-risk operations are marked. Rule conflicts may occur between multiple executable script operations. For example, one executable script may require limiting video traffic when network bandwidth is limited, while another script may require prioritizing video conferencing traffic during a specific time period. A rule conflict may occur when both conditions are met simultaneously.

[0102] Resource exceeding limits may occur because executing an executable script requires excessive network resources (such as computing and storage resources), exceeding the capacity of existing resources. In this case, the generated optimization prompts may suggest adjusting the execution priority of the executable script or modifying resource allocation strategies. Simultaneously, high-risk operations that could lead to serious consequences (such as operations that could paralyze the network) are marked so that relevant operators are notified through optimization prompts.

[0103] Furthermore, formal rule conflict detection algorithms can be embedded to comprehensively and accurately detect rule overlaps between executable scripts. For example, by establishing a mathematical model of the executable script operations, model detection technology can be used to find potential conflict points.

[0104] This embodiment achieves intelligent network management by receiving event-triggered commands, generating optimization prompts, and marking high-risk operations. The target network can automatically execute corresponding operations based on various event-triggered conditions, and provide early warnings and optimization suggestions for potential problems, improving the efficiency and quality of scalable network management. Furthermore, through the execution of executable scripts and simulation analysis, potential rule conflicts and resource overruns can be identified and resolved in advance.

[0105] In some embodiments, a cross-domain attribute mapping rule set is generated based on domain identification information, including:

[0106] The protocol type and security level included in the parsed domain identifier information are selected from the preset rule template library to match the protocol type and security level;

[0107] Based on the conversion template, the protocol type and security level are mapped to the network model registry to obtain a cross-domain attribute mapping rule set.

[0108] Optionally, in the process of generating the cross-domain attribute mapping rule set, this embodiment parses the domain identification information. This domain identification information includes, but is not limited to, protocol type and security level. Protocol type covers various network communication protocols, such as TCP / IP for reliable data transmission and HTTP for web browsing. Security level reflects the network domain's requirements for data protection, access control, etc. For example, a financial network domain may require a high security level, including encrypted transmission and strict authentication, while a public wireless network domain may have a relatively lower security level.

[0109] In addition, after parsing out the protocol type and security level, the system searches for a matching conversion template from the preset rule template library. The preset rule template library contains conversion templates corresponding to various combinations of different protocol types and security levels. These conversion templates are pre-designed by network experts and engineers based on long-term practical experience and industry standards.

[0110] For example, for a domain identifier that resolves to a TCP / IP protocol type and a high security level (e.g., financial network requirements), a conversion template specifically for this combination can be found in the template library. The conversion template includes a conversion method to transform the protocol and security attributes of the domain into a form compatible with other domains.

[0111] In addition, after finding a suitable conversion template, the protocol type and security level are mapped to the network model registry based on the conversion template. The network model registry records various network models and their attribute information.

[0112] Through mapping operations, new cross-domain attribute information (conversion forms of protocol type and security level) is added to the network model registry, thereby obtaining a complete set of cross-domain attribute mapping rules.

[0113] Furthermore, a template intelligent recommendation algorithm can be constructed to automatically recommend the most suitable transformation template based on the characteristics of domain identifier information. This algorithm can learn from historical template usage and the attribute characteristics of network domains, improving the efficiency and accuracy of template selection. For example, when inputting the identifier information of a new financial network domain, it can quickly recommend matching transformation templates based on existing financial network cases.

[0114] Simultaneously, an automatic template generation and update mechanism is being developed, capable of automatically expanding the rule template library based on the development of network technology and the emergence of new security protocols. For example, when a new network protocol emerges, a preliminary conversion template can be automatically generated based on existing template generation rules and the general characteristics of the protocol, and continuously updated and improved with practical applications.

[0115] This embodiment achieves efficient and reliable cross-domain network communication through precise parsing of domain identifier information, matching and conversion templates, and unified mapping rule management. Different network domains can seamlessly exchange data and collaborate according to the mapping rule set, improving network communication efficiency and business collaboration capabilities. At the same time, through reasonable security level mapping, security vulnerabilities can be prevented, protecting the confidentiality and integrity of data in the target network.

[0116] In some embodiments, the incremental instantiation parameters are obtained by comparing the target sub-attribute with the corresponding sub-attribute of the meta-network model, including:

[0117] Perform syntax tree comparison on the target sub-attribute and the corresponding sub-attribute of the meta-network model, determine the update node based on the syntax tree comparison, and determine the update node as the incremental instantiation parameter;

[0118] After obtaining the incremental instantiation parameters, the following is also included:

[0119] Based on the incremental instantiation parameters, the update operations to be performed on the sub-attributes corresponding to the meta-network model are determined, and the transaction logs corresponding to the update operations are recorded. The update operations include add operations, delete operations, and modify operations.

[0120] Optionally, in the process of constructing the target network, this embodiment parses the target sub-attributes and the corresponding sub-attributes of the meta-network model into syntax tree structures. A syntax tree is an intuitive and efficient representation that clearly displays the hierarchical relationships, data types, and logical structures of attributes. For example, for a network bandwidth attribute, its syntax tree may contain nodes such as bandwidth value, unit, and possible fluctuation range.

[0121] During the syntax tree comparison process, starting from the root node, each node is compared according to a certain traversal strategy (such as depth-first traversal or breadth-first traversal). The comparison includes, but is not limited to, attributes such as node name, data type, value range, and whether it is optional. When a difference is found between the syntax tree node of the target sub-attribute and the corresponding node in the meta-network model, such as different node values ​​or different node structures, these differing nodes are identified as update nodes, and the set of update nodes is used as incremental instantiation parameters.

[0122] In addition, after obtaining the incremental instantiation parameters, update operations are performed on the sub-attributes corresponding to the meta-network model based on the incremental instantiation parameters. The update operations include addition, deletion and modification operations.

[0123] The add operation is used to add a corresponding node when the target sub-attribute contains a new node that does not exist in the meta-network model. For example, if the target network requires the addition of a new network device type that is not defined in the meta-network model, then the attribute of the new network device type needs to be added to the meta-network model.

[0124] The delete operation is used to remove existing nodes from the meta-network model in the target sub-attribute. For example, if the target network no longer needs a specific network service module, the attribute node related to that service can be deleted from the meta-network model.

[0125] Modification operations are used when the node values ​​or structure in the target sub-attribute differ from the meta-network model. For example, the target network may modify the bandwidth value of a node from 100Mbps to 1000Mbps, or change the logical relationship of a node.

[0126] In addition, during the execution of update operations, the transaction log corresponding to the executed update operations is recorded. The transaction log records the type of each update operation, the object of the operation, the timestamp of the operation, and the execution status of the operation (success or failure).

[0127] Furthermore, an operational impact analysis mechanism can be constructed, which analyzes the potential impact of each update operation on the meta-network model before execution. For example, deleting an attribute node of a network device type may affect the associated network connection configuration; this potential risk can be analyzed and highlighted in advance.

[0128] Furthermore, a fine-grained network model version control mechanism can be established to manage the version of the target after each update operation. This means that by recording the incremental instantiation parameters and transaction logs for each version, switching and comparison between different versions can be performed. For example, if an update operation causes network problems, it can be quickly rolled back to a previous stable version.

[0129] This embodiment enables precise customization of the network model by identifying incremental instantiation parameters between the target sub-attribute and the sub-attribute of the meta-network model, and determining the corresponding update operation. This satisfies the personalized needs of the target network and improves the flexibility of the target network.

[0130] In some embodiments, the target network is collaboratively updated based on multi-domain model parameters using federated learning, including:

[0131] Deploy a hierarchical parameter aggregator in the federated learning framework. The hierarchical parameter aggregator includes a local gradient encryption unit at the edge node level and a differential privacy processing unit at the domain level.

[0132] The model contribution of each domain to the target network is determined by a hierarchical parameter aggregator. The sampling weights for the next training round are adjusted based on the model contribution, so as to collaboratively update the multi-domain model parameters in federated learning according to the sampling weights.

[0133] Optionally, this embodiment deploys a hierarchical parameter aggregator in a federated learning framework, wherein the hierarchical parameter aggregator includes a local gradient encryption unit at the edge node level and a differential privacy processing unit at the domain level.

[0134] Among them, the local gradient encryption unit at the edge node level is a local gradient encryption unit set on the edge node (such as various smart devices and terminal devices distributed at the edge of the network). The main function of the local gradient encryption unit is to encrypt the model gradients calculated locally by the edge node.

[0135] For example, when an IoT device (i.e., an edge node) calculates the gradient of model parameters based on local data, the local gradient encryption unit uses an advanced encryption algorithm to encrypt the gradient information so that the gradient information is not leaked during transmission, thus protecting the data privacy of the edge node.

[0136] Domain-level differential privacy processing units are deployed at the domain level. These units further process the encrypted gradients from different edge nodes. By adding appropriate noise, the differential privacy processing units can mask the influence of individual data, thereby ensuring the accuracy of the overall model parameter updates while preventing malicious attackers from inferring the specific data of a certain edge node through the aggregated gradient information.

[0137] For example, in a medical network domain, the gradients of medical diagnostic models submitted by different hospitals (i.e., edge nodes) can be protected from leakage of sensitive patient information by undergoing differential privacy processing.

[0138] In addition, during the federated learning process, the model contribution of each domain to the target network can be determined based on the hierarchical parameter aggregator. The model contribution reflects the size of each domain's contribution to the overall model optimization during the federated learning process.

[0139] Specifically, the contribution of the model is comprehensively evaluated by analyzing multiple factors, such as the quality of the encrypted gradient information submitted by each domain (e.g., the accuracy and richness of the gradient), the amount of data, and the performance improvement after the model update.

[0140] For example, a domain with a large amount of high-quality data and whose submitted gradients can effectively improve the overall model performance will have a higher model contribution.

[0141] Furthermore, the sampling weights for the next training epoch are adjusted based on the model's contribution. These sampling weights determine the probability of selecting edge nodes from each domain for model training in the next training epoch. If a domain has a high model contribution, its sampling weight in the next training epoch will increase accordingly.

[0142] For example, in a federated learning scenario involving multiple domains, edge nodes from domains with high contribution rates are more likely to be selected, allowing for more effective utilization of the high-quality data resources in these domains and accelerating the model optimization process. By collaboratively updating the parameters of multi-domain models in federated learning based on sampling weights, the entire federated learning process becomes more efficient and accurate, while also ensuring data privacy and security.

[0143] This embodiment achieves a highly efficient and secure federated learning ecosystem, allowing each domain to participate in the federated learning process safely without worrying about data leakage. By making reasonable use of the data resources of each domain, it accelerates model optimization and convergence, thereby improving the overall efficiency and practicality of federated learning.

[0144] In some embodiments, static attributes include basic static attributes, structured attribute tags, and mapping relationship tables;

[0145] Among them, the basic static attributes are obtained based on the topology, device type and fixed configuration parameters of the meta-network model in the preset network model library;

[0146] Structured attribute tags are obtained by performing semantic analysis on technical documents;

[0147] The mapping table includes physical device characteristics and protocol constraints.

[0148] Optionally, the static attributes involved in this embodiment include basic static attributes, structured attribute labels, and mapping relationship tables. The basic morphological attributes are determined based on the meta-network model in a preset network model library, and include, but are not limited to, the following elements:

[0149] Topology describes the connections between nodes (such as servers, routers, and switches) in a network. For example, in a star topology, all nodes are connected to a central node, and data transmission must be forwarded through this central node. In a mesh topology, nodes are interconnected, providing multiple data transmission paths and enhancing network reliability and redundancy. Topology information is extracted from the meta-network model.

[0150] Device Type: This section covers the type information of various devices used in the meta-network model. For example, an enterprise network may contain multiple types of devices such as routers, switches, firewalls, and servers. Each device performs different functions in the network. For instance, routers are used for routing and packet forwarding, while switches are used for exchanging data frames within a local area network. Device type information can be obtained from the device list and functional descriptions in the meta-network model.

[0151] Fixed configuration parameters: These are parameters that are determined when the network device leaves the factory or during the initial network configuration and are usually not easily changed. Examples include the device's IP address (in certain static network configurations), port number, and hardware serial number.

[0152] In addition, structured attribute tags are obtained by performing semantic analysis on technical documents. Technical documents contain detailed information on various aspects such as network design, device configuration, and communication protocols.

[0153] In addition, the mapping table establishes a mapping relationship between physical device characteristics and protocol constraints. Physical device characteristics refer to the physical features of network devices, such as the hardware type of the device (e.g., routers with fiber optic interfaces, wireless access points, etc.), the performance parameters of the device (e.g., processing power, memory size, etc.), and the physical location of the device (e.g., rack location in a data center, installation location within a building, etc.).

[0154] Protocol constraints refer to the requirements and limitations that network protocols impose on devices during operation. For example, some protocols may require devices to have specific hardware characteristics (such as high-speed processing capabilities to support high-bandwidth data transmission), specific software configurations (such as installing a specific version of the protocol stack), or specific network environments (such as low latency, high reliability, etc.).

[0155] The mapping table associates physical device characteristics with protocol constraints, which can clarify which protocols are suitable for each network device to run, or which protocols can achieve optimal performance on specific network devices.

[0156] For example, for server equipment with high-bandwidth fiber optic interfaces, the mapping table can indicate which protocols are suitable for running (such as 4K video streaming protocols).

[0157] This embodiment enables the efficient construction and management of network models through detailed basic static attributes, structured attribute labels, and mapping relationship tables. This makes the network model more accurate, complete, and easier to manage. At the same time, when performing network upgrades, fault diagnosis, resource allocation, and other operations, it can quickly and accurately obtain the required information, make reasonable decisions, and improve the network's operating efficiency and reliability.

[0158] To effectively address the shortcomings of traditional technologies, such as low efficiency in constructing target networks and reduced flexibility and stability, and to significantly improve the construction efficiency, flexibility, and stability of target networks, this application provides an embodiment of a digital twin-based scalable network model management device for implementing all or part of the aforementioned digital twin-based scalable network model management. See [link to embodiment]. Figure 2 The scalable network model management device based on digital twins specifically includes the following components:

[0159] The first processing module 10 is used to determine the network attributes corresponding to each meta-network model included in the preset network model library, and to receive the target network construction instruction. The network attributes include static attributes, constituent elements, model rules and dynamic features. The target network construction instruction includes network model classification, target sub-attributes and domain identification information.

[0160] The second processing module 20 is used to create a corresponding category based on the network model category and associate domain identification information when there is no extensible network model with the same network model category in the network model registry, generate a cross-domain attribute mapping rule set based on the domain identification information, and determine the meta-network model of the current network model category in the preset network model library based on the target network construction instruction. The cross-domain attribute mapping rule set includes conversion rules for attributes between heterogeneous domains.

[0161] The third processing module 30 is used to respond to the instantiation request, compare the target sub-attributes with the corresponding sub-attributes of the meta-network model, obtain the incremental instantiation parameters, and construct the target network based on the incremental instantiation parameters and network attributes on the basis of the meta-network model.

[0162] The fourth module 40 is used to associate the target network with physical network resources and adapt the protocol to the physical network resources through a cross-domain attribute mapping rule set, so as to enable the target network to collaboratively update the parameters of the multi-domain model based on federated learning.

[0163] As described above, the scalable network model management device based on digital twins provided in this application can innovatively determine the network attributes corresponding to each meta-network model included in the preset network model library and the received target network construction instructions. When no scalable network model of the same category is matched in the network model registry, it creates the category and associates it with domain identification information, generating a cross-domain attribute mapping rule set including heterogeneous domain attribute conversion rules. Simultaneously, it selects the corresponding category of meta-network model from the preset network model library according to the target network construction instructions. In response to the instantiation request, it compares the target sub-attributes with the corresponding sub-attributes of the meta-network model to obtain incremental instantiation parameters. It then constructs the target network by combining the network attributes and associates the target network with physical network resources. Through the cross-domain attribute mapping rule set, it performs protocol adaptation with the physical network resources, enabling the target network to be synchronously updated according to the multi-domain model parameters of federated learning. This method effectively solves the shortcomings of traditional technologies, such as low efficiency in constructing target networks and a certain degree of reduced flexibility and stability of the resulting target networks. It significantly improves the construction efficiency of target networks and enhances their flexibility and stability.

[0164] From a hardware perspective, in order to effectively address the shortcomings of traditional technologies in constructing target networks, such as low efficiency and reduced flexibility and stability, and to significantly improve the construction efficiency, flexibility, and stability of target networks, this application provides an embodiment of an electronic device for implementing all or part of the aforementioned scalable network model management method based on digital twins. The electronic device specifically includes the following components:

[0165] The system comprises a processor, memory, a communications interface, and a bus; wherein the processor, memory, and communications interface communicate with each other via the bus; the communications interface is used to realize information transmission between the digital twin-based scalable network model management device and core business systems, user terminals, and related databases and other related devices; the logic controller can be a desktop computer, tablet computer, or mobile terminal, etc., and this embodiment is not limited to these. In this embodiment, the logic controller can be implemented with reference to the embodiments of the digital twin-based scalable network model management method and the digital twin-based scalable network model management device described in the embodiments, the content of which is incorporated herein, and repeated details will not be described again.

[0166] It is understood that the user terminal may include smartphones, tablet computers, network set-top boxes, portable computers, desktop computers, personal digital assistants (PDAs), in-vehicle devices, smart wearable devices, etc. Among these, the smart wearable devices may include smart glasses, smartwatches, smart bracelets, etc.

[0167] In practical applications, parts of the scalable network model management method based on digital twins can be executed on the electronic device side as described above, or all operations can be completed in the client device. The choice can be made based on the processing power of the client device and the limitations of the user's usage scenario. This application does not impose any limitations on this. If all operations are completed in the client device, the client device may further include a processor.

[0168] The aforementioned client device may have a communication module (i.e., a communication unit) that can communicate with a remote server to achieve data transmission. The server may include a server on the task scheduling center side; in other implementation scenarios, it may also include a server on an intermediate platform, such as a server on a third-party server platform that has a communication link with the task scheduling center server. The server may include a single computer device, a server cluster consisting of multiple servers, or a distributed server structure.

[0169] Figure 3 This is a schematic block diagram illustrating the system configuration of the electronic device 9600 according to an embodiment of this application. Figure 3 As shown, the electronic device 9600 may include a central processing unit 9100 and a memory 9140; the memory 9140 is coupled to the central processing unit 9100. It is worth noting that... Figure 3 This is an example; other types of structures can also be used to supplement or replace this structure to achieve telecommunications functions or other functions.

[0170] In one embodiment, the scalable network model management method based on digital twins can be integrated into the central processing unit 9100. The central processing unit 9100 can be configured to perform the following control:

[0171] Step S101: Determine the network attributes corresponding to each meta-network model included in the preset network model library, and receive the target network construction instruction. The network attributes include static attributes, constituent elements, model rules and dynamic features. The target network construction instruction includes network model classification, target sub-attributes and domain identification information.

[0172] Step S102: When there is no extensible network model in the network model registry that is the same as the network model category, create a corresponding category based on the network model category and associate it with the domain identification information, generate a cross-domain attribute mapping rule set based on the domain identification information, and determine the meta-network model of the current network model category in the preset network model library based on the target network construction instruction. The cross-domain attribute mapping rule set includes conversion rules for attributes between heterogeneous domains.

[0173] Step S103: In response to the instantiation request, compare the target sub-attribute with the sub-attribute corresponding to the meta-network model to obtain incremental instantiation parameters, and construct the target network based on the incremental instantiation parameters and the network attribute on the basis of the meta-network model;

[0174] Step S104: Associate the target network with physical network resources, and adapt the cross-domain attribute mapping rule set with the physical network resources through protocol adaptation, so that the multi-domain model parameters of the target network based on federated learning are updated collaboratively.

[0175] As described above, the electronic device provided in this application innovatively determines the network attributes corresponding to each meta-network model included in a preset network model library and the received target network construction instructions. When no scalable network model of the same category is matched in the network model registry, it creates the category and associates it with domain identifier information, generating a cross-domain attribute mapping rule set including heterogeneous domain attribute conversion rules. Simultaneously, it selects the corresponding category of the meta-network model from the preset network model library according to the target network construction instructions. In response to the instantiation request, it compares the target sub-attributes with the corresponding sub-attributes of the meta-network model to obtain incremental instantiation parameters. It then constructs the target network by combining the network attributes and associates the target network with physical network resources. Through the cross-domain attribute mapping rule set, it performs protocol adaptation with the physical network resources, enabling the target network to be synchronously updated according to the multi-domain model parameters of federated learning. This method effectively solves the shortcomings of traditional techniques, such as low efficiency in constructing target networks and a certain degree of reduced flexibility and stability of the resulting target networks. It significantly improves the construction efficiency of target networks and enhances their flexibility and stability.

[0176] In another embodiment, the digital twin-based scalable network model management device can be configured separately from the central processing unit 9100. For example, the digital twin-based scalable network model management device can be configured as a chip connected to the central processing unit 9100, and the functions of the digital twin-based scalable network model management method can be implemented through the control of the central processing unit.

[0177] like Figure 3As shown, the electronic device 9600 may further include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It is worth noting that the electronic device 9600 does not necessarily need to include these components. Figure 3 All components shown; in addition, the electronic device 9600 may also include Figure 3 For components not shown, please refer to existing technologies.

[0178] like Figure 3 As shown, the central processing unit 9100, sometimes also referred to as a controller or operating control, may include a microprocessor or other processor device and / or logic device, which receives inputs and controls the operation of various components of the electronic device 9600.

[0179] The memory 9140 may be, for example, one or more of a cache, flash memory, hard drive, removable media, volatile memory, non-volatile memory, or other suitable devices. It may store the aforementioned failure-related information, and also store a program for executing that information. The central processing unit 9100 may execute the program stored in the memory 9140 to perform information storage or processing, etc.

[0180] Input unit 9120 provides input to central processing unit 9100. Input unit 9120 may be, for example, a keypad or touch input device. Power supply 9170 provides power to electronic device 9600. Display 9160 displays images and text. Display may be, for example, an LCD display, but is not limited thereto.

[0181] The memory 9140 can be a solid-state memory, such as a read-only memory (ROM), random access memory (RAM), a SIM card, etc. It can also be a memory that retains information even when power is off, can be selectively erased, and contains more data; examples of this type of memory are sometimes referred to as EPROMs. The memory 9140 can also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142 for storing application programs and function programs or processes for executing the operation of the electronic device 9600 via the central processing unit 9100.

[0182] The memory 9140 may also include a data storage unit 9143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various drivers for the electronic device's communication functions and / or for performing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0183] The communication module 9110 is a transmitter / receiver that sends and receives signals via the antenna 9111. The communication module 9110 (transmitter / receiver) is coupled to the central processing unit 9100 to provide input signals and receive output signals, which is the same as in a conventional mobile communication terminal.

[0184] Based on different communication technologies, multiple communication modules 9110 can be configured in the same electronic device, such as cellular network modules, Bluetooth modules, and / or wireless LAN modules. The communication module 9110 (transmitter / receiver) is also coupled to a speaker 9131 and a microphone 9132 via an audio processor 9130 to provide audio output via the speaker 9131 and receive audio input from the microphone 9132, thereby realizing typical telecommunications functions. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. Additionally, the audio processor 9130 is coupled to a central processing unit 9100, enabling on-device recording via the microphone 9132 and on-device playback of stored audio via the speaker 9131.

[0185] Embodiments of this application also provide a computer-readable storage medium capable of implementing all steps of the scalable network model management method based on digital twins, where the execution subject is a server or client, as described in the above embodiments. The computer-readable storage medium stores a computer program that, when executed by a processor, implements all steps of the scalable network model management method based on digital twins, where the execution subject is a server or client, as described in the above embodiments. For example, when the processor executes the computer program, it implements the following steps:

[0186] Step S101: Determine the network attributes corresponding to each meta-network model included in the preset network model library, and receive the target network construction instruction. The network attributes include static attributes, constituent elements, model rules and dynamic features. The target network construction instruction includes network model classification, target sub-attributes and domain identification information.

[0187] Step S102: When there is no extensible network model in the network model registry that is the same as the network model category, create a corresponding category based on the network model category and associate it with the domain identification information, generate a cross-domain attribute mapping rule set based on the domain identification information, and determine the meta-network model of the current network model category in the preset network model library based on the target network construction instruction. The cross-domain attribute mapping rule set includes conversion rules for attributes between heterogeneous domains.

[0188] Step S103: In response to the instantiation request, compare the target sub-attribute with the sub-attribute corresponding to the meta-network model to obtain incremental instantiation parameters, and construct the target network based on the incremental instantiation parameters and the network attribute on the basis of the meta-network model;

[0189] Step S104: Associate the target network with physical network resources, and adapt the cross-domain attribute mapping rule set with the physical network resources through protocol adaptation, so that the multi-domain model parameters of the target network based on federated learning are updated collaboratively.

[0190] As described above, the computer-readable storage medium provided in this application innovatively determines the network attributes corresponding to each meta-network model included in a preset network model library and the received target network construction instructions. When no matching scalable network model of the same category is found in the network model registry, it creates the category and associates it with domain identifier information, generating a cross-domain attribute mapping rule set including heterogeneous domain attribute conversion rules. Simultaneously, it selects the corresponding category of the meta-network model from the preset network model library according to the target network construction instructions. In response to an instantiation request, it compares the target sub-attributes with the corresponding sub-attributes of the meta-network model to obtain incremental instantiation parameters. It then constructs the target network by combining the network attributes and associates the target network with physical network resources. Through the cross-domain attribute mapping rule set and protocol adaptation with physical network resources, the target network is synchronously updated according to the multi-domain model parameters of federated learning. This method effectively solves the shortcomings of traditional techniques, such as low efficiency in constructing target networks and a certain degree of reduced flexibility and stability of the resulting target networks. It significantly improves the construction efficiency of target networks and enhances their flexibility and stability.

[0191] Embodiments of this application also provide a computer program product capable of implementing all steps of the scalable network model management method based on digital twins, where the execution subject is a server or client, as described in the above embodiments. When executed by a processor, this computer program / instruction implements the steps of the scalable network model management method based on digital twins. For example, the computer program / instruction implements the following steps:

[0192] Step S101: Determine the network attributes corresponding to each meta-network model included in the preset network model library, and receive the target network construction instruction. The network attributes include static attributes, constituent elements, model rules and dynamic features. The target network construction instruction includes network model classification, target sub-attributes and domain identification information.

[0193] Step S102: When there is no extensible network model in the network model registry that is the same as the network model category, create a corresponding category based on the network model category and associate it with the domain identification information, generate a cross-domain attribute mapping rule set based on the domain identification information, and determine the meta-network model of the current network model category in the preset network model library based on the target network construction instruction. The cross-domain attribute mapping rule set includes conversion rules for attributes between heterogeneous domains.

[0194] Step S103: In response to the instantiation request, compare the target sub-attribute with the sub-attribute corresponding to the meta-network model to obtain incremental instantiation parameters, and construct the target network based on the incremental instantiation parameters and the network attribute on the basis of the meta-network model;

[0195] Step S104: Associate the target network with physical network resources, and adapt the cross-domain attribute mapping rule set with the physical network resources through protocol adaptation, so that the multi-domain model parameters of the target network based on federated learning are updated collaboratively.

[0196] As described above, the computer program product provided in this application innovatively determines the network attributes corresponding to each meta-network model included in a preset network model library and the received target network construction instructions. When no scalable network model of the same category is matched in the network model registry, it creates the category and associates it with domain identification information, generating a cross-domain attribute mapping rule set including heterogeneous domain attribute conversion rules. Simultaneously, it selects the corresponding category of the meta-network model from the preset network model library according to the target network construction instructions. In response to the instantiation request, it compares the target sub-attributes with the corresponding sub-attributes of the meta-network model to obtain incremental instantiation parameters. It then constructs the target network by combining the network attributes and associates the target network with physical network resources. Through the cross-domain attribute mapping rule set, it performs protocol adaptation with the physical network resources, enabling the target network to be synchronously updated according to the multi-domain model parameters of federated learning. This method effectively solves the shortcomings of traditional techniques, such as low efficiency in constructing target networks and a certain degree of reduced flexibility and stability of the resulting target networks. It significantly improves the construction efficiency of target networks and enhances their flexibility and stability.

[0197] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0198] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (devices), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0199] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0200] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0201] Specific embodiments have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A scalable network model management method based on digital twinning, characterized in that, The method comprises: determining the network attributes corresponding to each meta-network model included in the preset network model library, and receiving a target network construction instruction, wherein the network attributes comprise static attributes, component elements, model rules and dynamic characteristics, and the target network construction instruction comprises a network model classification, target sub-attributes and domain identification information; when there is no extensible network model with the same network model classification in the network model registry, creating a corresponding classification based on the network model classification and associating the domain identification information, generating a cross-domain attribute mapping rule set based on the domain identification information, and determining a meta-network model of the current network model classification in the preset network model library based on the target network construction instruction, wherein the cross-domain attribute mapping rule set comprises conversion rules for attributes between heterogeneous domains; in response to an instantiation request, comparing the target sub-attributes with the sub-attributes corresponding to the meta-network model to obtain incremental instantiation parameters, and constructing a target network based on the incremental instantiation parameters and the network attributes on the basis of the meta-network model; associating the target network with physical network resources, and adapting the protocol with the physical network resources through the cross-domain attribute mapping rule set, so that the target network is based on the collaborative update of multi-domain model parameters of federated learning.

2. The method of claim 1, wherein, After determining the network attributes corresponding to each meta-network model included in the preset network model library and receiving a target network construction instruction, the method further comprises: when there is an extensible network model with the same network model classification in the network model registry, updating the cross-domain attribute mapping rule set corresponding to the extensible network model based on the domain identification information; in response to an instantiation request, comparing the target sub-attributes with the sub-attributes corresponding to the extensible network model to obtain incremental instantiation parameters, and constructing a target network based on the incremental instantiation parameters and the network attributes on the basis of the extensible network model.

3. The method of claim 1, wherein, After constructing a target network based on the incremental instantiation parameters and the network attributes on the basis of the meta-network model, the method further comprises: receiving an event trigger instruction, wherein the event trigger instruction comprises a time trigger instruction, a state trigger instruction and an external trigger instruction; associating an executable script corresponding to the event trigger instruction with the current dynamic characteristics, injecting historical traffic data in a simulation environment, and obtaining an execution result of the executable script based on the historical traffic data; when the execution result has rule conflicts or resource overruns, generating an optimization prompt information based on the execution result and marking a high-risk operation.

4. The method of claim 1, wherein, The method of generating a cross-domain attribute mapping rule set based on the domain identification information comprises: parsing the protocol type and security level included in the domain identification information, and selecting a conversion template matching the protocol type and the security level from a preset rule template library; mapping the protocol type and the security level to the network model registry based on the conversion template to obtain a cross-domain attribute mapping rule set.

5. The method of claim 1, wherein, The method of comparing the target sub-attributes with the sub-attributes corresponding to the meta-network model to obtain incremental instantiation parameters comprises: Performing a syntax tree comparison on the target sub-attribute and the sub-attribute corresponding to the meta-network model, determining an update node based on the syntax tree comparison, and determining the update node as an incremental instantiation parameter; After obtaining the incremental instantiation parameter, further comprising: Based on the incremental instantiation parameter, determining the update operation performed on the sub-attribute corresponding to the meta-network model, and recording the transaction log corresponding to the update operation, the update operation including an addition operation, a deletion operation and a modification operation.

6. The method of claim 1, wherein, The target network based on the collaborative update of the multi-domain model parameter of the federated learning comprises: Deploying a hierarchical parameter aggregator in the federated learning framework, the hierarchical parameter aggregator comprising a local gradient encryption unit at the edge node level and a differential privacy processing unit at the domain level; Based on the hierarchical parameter aggregator, determine the model contribution degree of the target network of each domain, and adjust the sampling weight of the next training round based on the model contribution degree, to collaboratively update the multi-domain model parameter in the federated learning according to the sampling weight.

7. The method of claim 1, wherein, The static attribute includes a basic static attribute, a structured attribute label and a mapping relationship table; The basic static attribute is obtained based on the topology structure, device type and fixed configuration parameter of the meta-network model in the preset network model library; The structured attribute label is obtained by performing semantic analysis on the technical document; The mapping relationship table includes physical device features and protocol constraint conditions.

8. A scalable network model management apparatus based on digital twinning, characterized by, The device comprises: A first processing module for determining network attributes corresponding to each meta-network model included in a preset network model library, and receiving a target network construction instruction, the network attributes including static attributes, component elements, model rules and dynamic characteristics, the target network construction instruction including network model classification, target sub-attributes and domain identification information; A second processing module for creating a corresponding classification based on the network model classification and associating the domain identification information when there is no extensible network model with the same network model classification in the network model registry, generating a cross-domain attribute mapping rule set based on the domain identification information, and determining a meta-network model of the current network model classification in the preset network model library based on the target network construction instruction, the cross-domain attribute mapping rule set including conversion rules for attributes between heterogeneous domains; A third processing module for responding to an instantiation request, comparing the target sub-attributes with the sub-attributes corresponding to the meta-network model, obtaining an incremental instantiation parameter, and constructing a target network based on the incremental instantiation parameter and the network attributes on the basis of the meta-network model; A fourth processing module for associating the target network with physical network resources, and adapting the protocol between the cross-domain attribute mapping rule set and the physical network resources, so that the target network collaboratively updates the multi-domain model parameter based on the federated learning.

9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the steps of the extensible network model management method based on digital twinning of any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the extensible network model management method based on digital twinning of any one of claims 1 to 7.

Images