Edge computing power network system based on digital twinning and route distribution method thereof

Abstract

The invention discloses an edge computing power network system based on digital twinning. The edge computing power network system comprises an entity network layer, a sensing control layer and a network twinning layer, the invention also discloses a route distribution method which is realized based on the edge computing power network system based on digital twinning. The invention discloses an edge computing power network system based on digital twinning and a routing distribution method thereof, and the method comprises the steps: constructing a twinning network of an entity network layer, and finding an optimal routing strategy in the twinning network through a genetic algorithm, so as to process a computing service request; the method not only can dynamically adjust the routing strategy, optimize the data transmission path, reduce the network load and improve the data transmission efficiency according to the real-time data and the model prediction result, but also can reasonably allocate the computing power resource and effectively solve the problem of load imbalance, thereby improving the performance and efficiency of the system.

Description

Technical Field

[0001] The present invention relates to the field of communication networks and computing power technology, and in particular to an edge computing power network system based on digital twins and a routing allocation method thereof. Background Art

[0002] With the development of new business applications such as autonomous driving, smart cities, and the industrial Internet, the demand for network computing power has gradually increased, and the number of devices connected to the network has grown exponentially. In response to this trend, digital twin technology has emerged. By creating an accurate virtual mirror of a physical entity or system, it enables engineers and technicians to test, optimize, and perform predictive maintenance in a simulated environment. Digital twins can not only reflect the status of physical objects in real time, but also simulate different operating conditions to help optimize resource allocation and management, and improve decision-making efficiency and service quality.

[0003] Edge computing networks deploy computing resources at the edge of the network to provide low latency, high bandwidth, and real-time processing capabilities, allowing data to be processed close to the data source, thereby reducing the time and cost of transmission to the central cloud and reducing the possibility of network congestion. However, existing edge computing network systems face the problem of network coexistence, namely load imbalance - some edge service nodes are overloaded due to too many requests and cannot meet the processing needs of business applications, while the computing resources of other nodes are idle, resulting in resource waste and increased management complexity. To meet these challenges, current research mainly introduces technologies such as intelligent task routing allocation algorithms and adaptive computing resource management to improve the overall efficiency and service quality of edge computing networks.

[0004] Route allocation is to ensure that data can reach its destination efficiently by optimizing the data transmission path selection from terminal devices to edge nodes and then to the cloud. Route allocation needs to comprehensively consider the changes in network topology, the characteristics of different types of traffic, and the real-time and reliability requirements to achieve optimal resource utilization and maximize performance. However, the difficulties in this process include how to maximize the use of existing resources while ensuring the quality of service and cope with rapidly changing workload patterns, especially to implement ideal routing strategies in complex and changing real environments. At present, although some progress has been made, there are still many challenges, such as accurately predicting network status, dynamically adjusting paths to adapt to rapidly changing needs, and solving compatibility issues in cross-regional and heterogeneous network environments. These are important directions for future research and development.

[0005] In order to solve the above problems, routing allocation technology has become a research hotspot. This technology aims to optimize the selection of data transmission paths from terminal devices to edge nodes and then to the cloud to ensure that data can reach its destination efficiently. Combined with digital twin technology, the routing strategy can be dynamically adjusted by building a digital twin model of the edge network to adapt to the changing workload pattern. Its characteristics are that it needs to take into account multiple factors such as changes in network topology, different types of traffic characteristics, and real-time and reliability requirements. The difficulties include how to maximize the use of existing resources while ensuring the quality of service, and to achieve accurate predictive maintenance and troubleshooting. The current technical status is that although there have been some preliminary results, there are still many unresolved problems in realizing ideal routing strategies in complex and changing real environments.

[0006] Therefore, in order to efficiently utilize ubiquitous and heterogeneous computing resources and improve the system's adaptability, it is necessary to study the task routing strategy and computing resource allocation mechanism of the edge computing network. This will not only help solve the current problem of low resource utilization, but also provide theoretical support and technical guarantee for the design of future intelligent network architecture. Therefore, in order to efficiently utilize ubiquitous and heterogeneous computing resources and improve the system's adaptability, it is urgent to conduct in-depth research on the task routing strategy and computing resource allocation mechanism of the edge computing network. This will not only help solve the current problem of low resource utilization, but also provide theoretical support and technical guarantee for the design of future intelligent network architecture. In addition, the application of digital twin technology will further promote the development of the edge computing field and provide new ideas and methods for achieving more intelligent and automated network management and operation and maintenance. Summary of the invention

[0007] In order to solve the problem of load imbalance in the existing edge computing power network system, the present invention provides an edge computing power network system based on digital twins and a routing allocation method thereof.

[0008] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0009] An edge computing network system based on digital twins includes a physical network layer, a perception control layer, and a network twin layer; wherein:

[0010] The physical network layer is used to initiate computing service requests and process computing service requests according to the optimal routing strategy issued by the network twin layer;

[0011] The perception control layer is used to collect data from the physical network layer in real time and upload the collected data to the network twin layer;

[0012] The network twin layer is used to perform virtual mapping according to the physical network layer to build a twin network in a digital way; simulate and process computing service requests in the twin network, use a genetic algorithm to find the optimal routing strategy, and send the optimal routing strategy to the physical network layer via the perception control layer.

[0013] In the above scheme, by constructing a twin network of the physical network layer and using genetic algorithms in the twin network to find the optimal routing strategy to process computing service requests, it is not only possible to dynamically adjust the routing strategy according to real-time data and model prediction results, optimize the data transmission path, reduce network load, and improve data transmission efficiency, but also to reasonably allocate computing resources and effectively solve the problem of load imbalance, thereby improving the performance and efficiency of the system.

[0014] Preferably, the physical network layer includes a cloud computing center, an edge service node, an edge manager, a computing routing node and a user terminal; wherein,

[0015] The cloud computing center is connected to the edge service node through an edge-cloud fixed link to process computing tasks that the edge service node cannot process;

[0016] The edge service node is connected to the computing power routing node through a fixed link and is used to process computing tasks according to computing service requests;

[0017] The edge manager is connected to the edge service node and the computing power routing node through fixed links, and is used to obtain the network topology relationship, the computing power status information of the edge service node and the network status information of the computing power routing node;

[0018] The computing power routing node is used for routing control and forwarding data; wherein each computing power routing node is connected via a dynamic link;

[0019] The user end is connected to the computing power routing node via a wireless link to initiate a computing service request.

[0020] Preferably, the types of computing power routing nodes include computing power routing entry nodes, computing power intermediate routing nodes and computing power routing exit nodes; wherein,

[0021] At least one end of the computing power routing entry node is connected to the user end via a wireless link;

[0022] The computing power intermediate routing node is connected to other computing power routing nodes through a dynamic link;

[0023] At least one end of the computing power routing exit node is connected to the edge service node via a fixed link.

[0024] Preferably, the data collected by the perception control layer includes network topology relationships, computing power status information, network status information and computing service requests.

[0025] Preferably, the computing power status information includes service ID, CPU usage and memory usage.

[0026] Preferably, the network status information includes time delay, delay jitter and packet loss rate.

[0027] Preferably, the network twin layer includes a network twin, a service mapping model and a data warehouse;

[0028] The network twin is used to perform virtual mapping according to the physical network layer and construct a twin network in a digital way;

[0029] The service mapping model is used to simulate and process computing service requests in the twin network and find the optimal routing strategy using a genetic algorithm;

[0030] The data warehouse is used to store and manage the data uploaded by the perception control layer and the data generated by the service mapping model simulation process.

[0031] Preferably, the specific steps of using a genetic algorithm to find the optimal routing strategy are:

[0032] A1: Initialize the strategy population: randomly generate M strategy individuals as the initial strategy population;

[0033] A2: Individual evaluation: Calculate the fitness of each strategy individual in the strategy group;

[0034] A3: Determine whether the current number of iterations reaches the preset maximum number of iterations;

[0035] If yes, go to step A7;

[0036] If not, proceed to step A4;

[0037] A4: Selection operation: Sort the fitness of each individual from small to large, retain the individuals with the highest ranking according to the preset ratio, and discard the individuals with the lowest ranking;

[0038] A5: Crossover: Cross the retained individuals with each other to generate new strategic individuals;

[0039] A6: Mutation: Randomly mutate the strategy arrangement of some new strategy individuals to obtain a new generation of strategy groups, and return to step A2 for the next round of iteration;

[0040] A7: Reach the optimal routing strategy: Take the strategy individual with the smallest fitness as the optimal routing strategy.

[0041] Preferably, the fitness is the shortest transmission delay of each strategy individual, and the computing resources and storage resources available to all users routed to the edge server do not exceed the computing resources and storage resources that the edge server can provide as a constraint.

[0042] A routing allocation method is implemented based on the edge computing network system based on digital twins, comprising the following steps:

[0043] S1: Collect data from the physical network layer in real time through the perception control layer, and upload the collected data to the network twin layer;

[0044] S2: Virtually map the physical network layer through the network twin of the network twin layer to build a twin network in a digital way;

[0045] S3: Use the service mapping model of the network twin layer to simulate and process computing service requests in the twin network, and use genetic algorithms to find the optimal routing strategy;

[0046] S4: The service mapping model sends the optimal routing strategy to the physical network layer through the perception control layer for execution.

[0047] Beneficial technical effects of the present invention:

[0048] The present invention provides an edge computing network system based on digital twins and a routing allocation method thereof. By constructing a twin network of the physical network layer and using a genetic algorithm in the twin network to find the optimal routing strategy to process computing service requests, it is not only possible to dynamically adjust the routing strategy according to real-time data and model prediction results, optimize the data transmission path, reduce network load, and improve data transmission efficiency, but also to reasonably allocate computing resources and effectively solve the problem of load imbalance, thereby improving the performance and efficiency of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is the overall structural diagram of the present invention;

[0050] Figure 2 It is a schematic diagram of module connection of the present invention;

[0051] Figure 3 It is a schematic diagram of the structure of the physical network layer in the present invention;

[0052] Figure 4 A flowchart of the steps of using a genetic algorithm to find the optimal routing strategy in the present invention;

[0053] Figure 5 It is a flow chart of the implementation steps of the technical solution of the present invention. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments, but the scope of protection claimed in the present invention is not limited to the following specific embodiments.

[0055] Example 1

[0056] like Figure 1-2 As shown, an edge computing network system based on digital twins includes a physical network layer, a perception control layer and a network twin layer; wherein,

[0057] The physical network layer is used to initiate computing service requests and process computing service requests according to the optimal routing strategy issued by the network twin layer;

[0058] The perception control layer is used to collect data from the physical network layer in real time and upload the collected data to the network twin layer;

[0059] The network twin layer is used to perform virtual mapping according to the physical network layer to build a twin network in a digital way; simulate and process computing service requests in the twin network, use a genetic algorithm to find the optimal routing strategy, and send the optimal routing strategy to the physical network layer via the perception control layer.

[0060] In the specific implementation process, by constructing a twin network of the physical network layer and using genetic algorithms in the twin network to find the optimal routing strategy to process computing service requests, it can not only dynamically adjust the routing strategy according to real-time data and model prediction results, optimize the data transmission path, reduce network load, and improve data transmission efficiency, but also reasonably allocate computing resources and effectively solve the problem of load imbalance, thereby improving the performance and efficiency of the system.

[0061] Example 2

[0062] An edge computing network system based on digital twins includes a physical network layer, a perception control layer, and a network twin layer; wherein:

[0063] The physical network layer is used to initiate computing service requests and process computing service requests according to the optimal routing strategy issued by the network twin layer;

[0064] The computing service request includes information such as computing data packet, service type, priority, etc.

[0065] More specifically, the physical network layer includes a cloud computing center, an edge service node, an edge manager, a computing routing node, and a user terminal; wherein,

[0066] The cloud computing center is connected to the edge service node through an edge-cloud fixed link to process computing tasks that the edge service node cannot process;

[0067] In the specific implementation process, the cloud computing center has sufficient storage resources and computing resources, and is a large server cluster deployed at a location far away from users;

[0068] The edge service node (such as Figure 2 The middle edge service node 1, edge service node 2, and edge service node 3 are connected to the computing power routing node through a fixed link to process computing tasks according to computing service requests;

[0069] In the specific implementation process, the edge service node is an edge application platform with hardware infrastructure as virtualized resources, which is used to provide storage and computing resources, register or update the edge service node status information to the edge manager, and process the user's computing business;

[0070] The edge manager is connected to the edge service node and the computing power routing node through fixed links, and is used to obtain the network topology relationship, the computing power status information of the edge service node and the network status information of the computing power routing node;

[0071] During the specific implementation process, the edge manager is used to distribute and update the registration information of the edge service node computing service. When registering the computing service, a unique SID (service ID) will be assigned according to the service type; the edge manager is also used to perceive the computing power status and network status, form a computing power status information table and a network status information table, and upload them to the perception control layer.

[0072] The computing power routing node is used for routing control and forwarding data, such as notification of computing power status and network status information, computing power routing addressing and forwarding, etc.;

[0073] Among them, each computing power routing node is connected through a dynamic link;

[0074] More specifically, the types of computing power routing nodes include computing power routing entry nodes, computing power intermediate routing nodes and computing power routing exit nodes; wherein,

[0075] The computing power routing entry node (such as Figure 3 At least one end of the computing power routing entry node 1, computing power routing entry node 2, and computing power routing entry node 3 is connected to the user end through a wireless link;

[0076] The computing power intermediate routing node (such as Figure 2 The middle computing power intermediate routing node 1, computing power intermediate routing node 2, computing power intermediate routing node 3) are connected to other computing power routing nodes through dynamic links;

[0077] The computing power routing exit node (such as Figure 2At least one end of the middle computing power routing exit node 1, computing power routing exit node 2, and computing power routing exit node 3) is connected to the edge service node through a fixed link.

[0078] The user terminal (such as Figure 2 Among them, user terminal 1, user terminal 2, and user terminal 3) are connected to the computing power routing node through a wireless link to initiate computing service requests.

[0079] In the specific implementation process, the user terminal uses the SID distributed by the edge manager to initiate a computing service request.

[0080] The perception control layer is used to collect data from the physical network layer in real time and upload the collected data to the network twin layer;

[0081] More specifically, the data collected by the perception control layer includes network topology relationships, computing power status information, network status information, and computing service requests.

[0082] More specifically, the computing power status information includes service ID, CPU usage and memory usage.

[0083] More specifically, the network status information includes time delay, delay jitter and packet loss rate.

[0084] During the specific implementation process, all edge service nodes of the physical network layer register the computing power status information to the edge manager, which generates a computing power status information table and uploads it to the perception control layer; at the same time, the registered computing power status information is sent to the computing power routing exit node, and the computing load information is updated to notify the computing power status.

[0085] The computing power routing exit node forwards the registration information it obtains to the computing power intermediate routing node connected to it, and the computing power intermediate routing node generates a local computing power status information table based on the registration information; the computing power intermediate routing nodes forward the obtained registration information between each other until all computing power routing nodes obtain the computing power status information and generate a local computing power status information table.

[0086] Each computing power routing entry node in the physical network layer sends a request to each computing power routing exit node to detect network status information. After each computing power routing entry node receives a reply from the corresponding computing power routing exit node, it generates a network status information table to record network time delay, delay jitter, packet loss rate and other information; within the specified time, if the computing power routing entry node does not receive a reply from the corresponding computing power routing exit node, it means that the computing power routing exit node is unreachable. At this time, the packet loss information is recorded and uploaded to the network status information table; if the computing power routing entry node receives a reply from the corresponding computing power routing exit node, it means that the computing power routing exit node is reachable. At this time, the computing power routing entry node calculates the delay based on the time of sending the request and receiving the reply, and updates the delay and packet loss information and uploads it to the network status information table.

[0087] Each computing power routing entry node uploads the local network status information table to the edge manager, the edge manager uploads the network status information table to the perception control layer, and the perception control layer then uploads the computing power status information table and the network status information table to the network twin layer.

[0088] The network twin layer is used to perform virtual mapping according to the physical network layer to build a twin network in a digital way; simulate and process computing service requests in the twin network, use a genetic algorithm to find the optimal routing strategy, and send the optimal routing strategy to the physical network layer via the perception control layer.

[0089] More specifically, the network twin layer includes a network twin, a service mapping model, and a data warehouse;

[0090] The network twin is used to perform virtual mapping according to the physical network layer and construct a twin network in a digital way;

[0091] The service mapping model is used to simulate and process computing service requests in the twin network and find the optimal routing strategy using a genetic algorithm;

[0092] The data warehouse is used to store and manage the data uploaded by the perception control layer and the data generated by the service mapping model simulation process.

[0093] In the specific implementation process, the network twin layer sends the optimal routing strategy to the perception control layer, which then sends it to the edge manager of the physical network layer; the edge manager sends the obtained optimal routing strategy to the computing power routing entry node that receives the computing service request initiated by the user end, and records the edge server and routing node used this time, updates the computing power status information table and the network status information table, and selects the routing node for the next computing service. The computing power routing entry node forwards the computing data packet to the computing power intermediate routing node passing through according to the received optimal routing strategy, until it is forwarded to the corresponding optimal computing power routing exit node; the optimal computing power routing exit node transmits the obtained computing data packet to the edge service node connected to it for computing and processing. After processing, the edge service node returns the calculation result to the user end that initiated the computing service request.

[0094] More specifically, if Figure 4 As shown in Figure 1, the specific steps of using genetic algorithm to find the optimal routing strategy are:

[0095] A1: Initialize the strategy population: randomly generate M strategy individuals as the initial strategy population;

[0096] A2: Individual evaluation: Calculate the fitness of each strategy individual in the strategy group;

[0097] More specifically, the fitness is the shortest transmission delay of each individual strategy, and the constraint condition is that the computing resources and storage resources available to all users routed to the edge server do not exceed the computing resources and storage resources available to the edge server.

[0098] A3: Determine whether the current number of iterations reaches the preset maximum number of iterations;

[0099] If yes, go to step A7;

[0100] If not, proceed to step A4;

[0101] A4: Selection operation: Sort the fitness of each individual from small to large, retain the individuals with the highest ranking according to the preset ratio, and discard the individuals with the lowest ranking;

[0102] A5: Crossover: Cross the retained individuals with each other to generate new strategic individuals;

[0103] A6: Mutation: Randomly mutate the strategy arrangement of some new strategy individuals (for example, 0 becomes 1 or 1 becomes 0) to obtain a new generation of strategy groups, and return to step A2 for the next round of iteration;

[0104] A7: Get the optimal routing strategy: take the strategy individual with the smallest fitness as the optimal routing strategy.

[0105] Example 3

[0106] like Figure 5 As shown, a routing allocation method is implemented based on the edge computing network system based on digital twins, comprising the following steps:

[0107] S1: Collect data from the physical network layer in real time through the perception control layer, and upload the collected data to the network twin layer;

[0108] The collected data includes network topology, computing power status information, network status information, and computing service requests;

[0109] S2: Virtually map the physical network layer through the network twin of the network twin layer to build a twin network in a digital way;

[0110] S3: Use the service mapping model of the network twin layer to simulate and process computing service requests in the twin network, and use genetic algorithms to find the optimal routing strategy;

[0111] More specifically, the specific steps to find the optimal routing strategy using a genetic algorithm are:

[0112] A1: Initialize the strategy population: randomly generate M strategy individuals as the initial strategy population;

[0113] A2: Individual evaluation: Calculate the fitness of each strategy individual in the strategy group;

[0114] More specifically, the fitness is the shortest transmission delay of each individual strategy, and the constraint condition is that the computing resources and storage resources available to all users routed to the edge server do not exceed the computing resources and storage resources available to the edge server.

[0115] A3: Determine whether the current number of iterations reaches the preset maximum number of iterations;

[0116] If yes, go to step A7;

[0117] If not, proceed to step A4;

[0118] A4: Selection operation: Sort the fitness of each individual from small to large, retain the individuals with the highest ranking according to the preset ratio, and discard the individuals with the lowest ranking;

[0119] A5: Crossover: Cross the retained individuals with each other to generate new strategic individuals;

[0120] A6: Mutation: Randomly mutate the strategy arrangement of some new strategy individuals (for example, 0 becomes 1 or 1 becomes 0) to obtain a new generation of strategy groups, and return to step A2 for the next round of iteration;

[0121] A7: Get the optimal routing strategy: take the strategy individual with the smallest fitness as the optimal routing strategy.

[0122] S4: The service mapping model sends the optimal routing strategy to the physical network layer through the perception control layer for execution.

[0123] According to the disclosure and teaching of the above description, those skilled in the art to which the present invention belongs may also make changes and modifications to the above embodiments. Therefore, the present invention is not limited to the specific embodiments disclosed and described above, and some modifications and changes to the invention should also fall within the scope of protection of the claims of the present invention. In addition, although some specific terms are used in this specification, these terms are only for the convenience of description and do not constitute any limitation to the present invention.

Claims

1. An edge computing network system based on digital twins, characterized in that: It includes the physical network layer, the perception control layer and the network twin layer; among them, The physical network layer is used to initiate computing service requests and process computing service requests according to the optimal routing strategy issued by the network twin layer; The perception control layer is used to collect data from the physical network layer in real time and upload the collected data to the network twin layer; The network twin layer is used to perform virtual mapping according to the physical network layer to build a twin network in a digital way; simulate and process computing service requests in the twin network, use a genetic algorithm to find the optimal routing strategy, and send the optimal routing strategy to the physical network layer via the perception control layer.

2. According to claim 1, the edge computing network system based on digital twins is characterized in that: The physical network layer includes a cloud computing center, an edge service node, an edge manager, a computing routing node and a user terminal; wherein, The cloud computing center is connected to the edge service node through an edge-cloud fixed link to process computing tasks that the edge service node cannot process; The edge service node is connected to the computing power routing node through a fixed link and is used to process computing tasks according to computing service requests; The edge manager is connected to the edge service node and the computing power routing node through fixed links, and is used to obtain the network topology relationship, the computing power status information of the edge service node and the network status information of the computing power routing node; The computing power routing node is used for routing control and forwarding data; wherein each computing power routing node is connected via a dynamic link; The user end is connected to the computing power routing node via a wireless link to initiate a computing service request.

3. According to claim 2, the edge computing network system based on digital twins is characterized in that: The types of computing power routing nodes include computing power routing entry nodes, computing power intermediate routing nodes and computing power routing exit nodes; wherein, At least one end of the computing power routing entry node is connected to the user end via a wireless link; The computing power intermediate routing node is connected to other computing power routing nodes through a dynamic link; At least one end of the computing power routing exit node is connected to the edge service node via a fixed link.

4. According to the edge computing network system based on digital twins according to claim 1, it is characterized in that: The data collected by the perception control layer includes network topology relationships, computing power status information, network status information and computing service requests.

5. According to claim 4, the edge computing network system based on digital twins is characterized in that: The computing power status information includes service ID, CPU usage and memory usage.

6. The edge computing network system based on digital twin according to claim 4 is characterized in that: The network status information includes time delay, delay jitter and packet loss rate.

7. The edge computing network system based on digital twin according to claim 1 is characterized in that: The network twin layer includes a network twin, a service mapping model and a data warehouse; The network twin is used to perform virtual mapping according to the physical network layer and construct a twin network in a digital way; The service mapping model is used to simulate and process computing service requests in the twin network and find the optimal routing strategy using a genetic algorithm; The data warehouse is used to store and manage the data uploaded by the perception control layer and the data generated by the service mapping model simulation process.

8. The edge computing network system based on digital twin according to claim 1, characterized in that: The specific steps to find the optimal routing strategy using genetic algorithms are: A1: Initialize the strategy population: randomly generate M strategy individuals as the initial strategy population; A2: Individual evaluation: Calculate the fitness of each strategy individual in the strategy group; A3: Determine whether the current number of iterations reaches the preset maximum number of iterations; If yes, go to step A7; If not, proceed to step A4; A4: Selection operation: Sort the fitness of each individual from small to large, retain the individuals with the highest ranking according to the preset ratio, and discard the individuals with the lowest ranking; A5: Crossover: Cross the retained individuals with each other to generate new strategic individuals; A6: Mutation: Randomly mutate the strategy arrangement of some new strategy individuals to obtain a new generation of strategy groups, and return to step A2 for the next round of iteration; A7: Get the optimal routing strategy: take the strategy individual with the smallest fitness as the optimal routing strategy.

9. The edge computing network system based on digital twin according to claim 8, characterized in that: The fitness is the shortest transmission delay of each strategy individual, and the computing resources and storage resources available to all users routed to the edge server do not exceed the computing resources and storage resources that the edge server can provide as a constraint.

10. A routing allocation method for an edge computing network system based on digital twins according to any one of claims 1 to 9, characterized in that: The following steps are involved: S1: Collect data from the physical network layer in real time through the perception control layer, and upload the collected data to the network twin layer; S2: Virtually map the physical network layer through the network twin of the network twin layer to build a twin network in a digital way; S3: Use the service mapping model of the network twin layer to simulate and process computing service requests in the twin network, and use genetic algorithms to find the optimal routing strategy; S4: The service mapping model sends the optimal routing strategy to the physical network layer through the perception control layer for execution.

Images