Service node discovery method and device under edge computing network, equipment and medium

Abstract

The application discloses a service node discovery method and device under an edge computing network, and relates to the technical field of communication. The method comprises the following steps: updating a service registration table through a service registration center based on service information of a plurality of service nodes; calculating the total time delay of data packets sent by each service node in the updated service registration table to a receiving service node, and constructing a service table; calculating the comprehensive computing power and load of each service node in the updated service registration table, and constructing an information table; performing heartbeat checking and service health state checking on each service node, and updating the service table based on the checking result; and calculating the comprehensive score of each service node, and determining a target service node. Through the application, the technical problems of low service discovery efficiency, slow service call response, poor service stability and unreasonable resource allocation of the conventional service discovery method in the prior art in a large-scale distributed system are solved.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a method, apparatus, device and medium for service node discovery in an edge computing network. Background Technology

[0002] Against the backdrop of the rapid development of IoT and 5G technologies, edge computing and computing networks have become key technologies in distributed systems. Edge computing is a distributed computing model that migrates data processing and computing tasks from centralized data centers to devices or network edges closer to data sources or users, thereby reducing latency, saving bandwidth, and meeting the real-time requirements of business operations. Computing networks, on the other hand, integrate and virtualize distributed computing resources to form a unified, scheduled, and on-demand computing power pool, connecting the cloud, edge, and terminal devices to achieve full-domain coverage and efficient collaboration. The combination of the two enables full-stack computing capability synergy from the edge to the cloud, meeting the real-time, computing density, and resource elasticity requirements of complex business scenarios.

[0003] In edge computing networks, time-sensitive services (such as real-time video streaming and online financial transactions) place extremely high demands on the efficiency of service discovery and the response speed of service calls. However, traditional service discovery methods suffer from the following technical problems in large-scale distributed systems: (1) Low service discovery efficiency: In the traditional service discovery process, the service information of nodes is not updated in time, resulting in a long service discovery response time. That is, the process of users or applications searching for and locating service nodes that can provide specific functions or services is long and cannot meet the needs of real-time applications. (2) Slow service call response: Due to network latency and node computing power limitations, traditional methods cannot respond to user requests quickly, affecting user experience; (3) Poor service stability: The dynamic network environment and node load changes make it difficult for traditional service discovery methods to monitor and adjust the service status in real time, resulting in service instability and unavailability; (4) Unreasonable resource allocation: The computing power and load status of nodes are not considered, resulting in waste or overload of resources and failure to effectively utilize resources in the edge computing network. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a service node discovery method, apparatus, device, and medium in an edge computing network.

[0005] To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, embodiments of the present invention provide a service node discovery method in an edge computing network, the method comprising: Based on the service information of several service nodes, the service registry is updated through the service registry center to obtain the updated service registry. Calculate the total latency for each service node in the updated service registry to send data packets to the receiving service node, and construct a service table based on each total latency; Calculate the overall computing power and load of each service node in the updated service registry, and construct an information table based on the overall computing power and load; Perform heartbeat checks and service health status checks on each service node, and update the service table based on the check results to obtain the updated service table; Based on the updated service registry, the updated service table, and the information table, a comprehensive score is calculated for each service node, and the service node with the highest comprehensive score is determined as the target service node.

[0006] Optionally, the service information includes: service name, service address and port, response time, load capacity, and computing power. The service information based on several service nodes is used to update the service registry through a service registry center to obtain an updated service registry, which includes: After obtaining the service information of several service nodes, the service information of each service node and the service node itself are verified through the service registry center; The service information of each verified service node is updated in the service registry to obtain the updated service registry.

[0007] Optionally, calculating the total latency from the sending of data packets to the receiving service node for each service node in the updated service registry includes: Get the number of tasks arriving at the sending service node per unit time, the amount of task data processed by the sending service node per unit time, the size of the data packets sent by the sending service node, the transmission rate between adjacent nodes, and the calculation time of each service node in the updated service registry. For service nodes , will service nodes Calculation time Arrival at service node per unit time Number of tasks Service nodes per unit time The amount of task data processed Service Nodes Size of the data packets sent and the transmission rate between adjacent nodes Substituting into the first preset formula, the service node is calculated. Send data packets to the forwarding node Total latency ; The first preset formula is as follows: ; Similarly, the total latency for each service node in the updated service registry to send data packets to the receiving service node is calculated. In the formula, j is the total number of hops between the service node and the forwarding node. Starting from service node s, it reaches the forwarding node. Each forwarding node.

[0008] Optionally, the step of constructing a service table based on each of the total latency includes: Sequentially, the total delay and the minimum delay duration are compared. Perform the comparison and obtain the comparison results; If the comparison result shows that the total delay is less than or equal to the minimum delay duration, then the service name of the service node, the next-hop node, the total delay, and the timestamp of the sent data packet are combined into an entry. A service table is obtained based on the multiple entries.

[0009] Optionally, the calculation of the overall computing power and load of each service node in the updated service registry includes: The comprehensive computing power of the i-th service node in the updated service registry is calculated using the second preset formula. The load of the i-th service node in the updated service registry is calculated using the third preset formula. The second preset formula is as follows:

[0010] In the formula, This represents the overall computing power of the i-th service node; This represents the logical operation capability of the i-th service node. , This represents the number of idle CPU cores in the i-th service node. This represents the clock frequency of the CPU in the i-th service node. This represents the number of instructions per cycle for the i-th service node; This represents the parallel computing capability of the i-th service node. , This represents the number of idle parallel computing units in the i-th service node. This represents the clock frequency of the graphics processing unit of the i-th service node. This represents the number of floating-point operations that can be performed by the i-th service node in each clock cycle; This represents the neural network computing power of the i-th service node. , This represents the number of idle neural network computation cores in the i-th service node. This represents the clock frequency of the neural network processing unit of the i-th service node. This represents the number of neural network-related operations that can be performed by the i-th service node in each clock cycle; The third preset formula is as follows: ) In the formula, This represents the load of the i-th service node. This represents the CPU and memory utilization of the i-th service node. This represents the GPU memory utilization of the i-th service node. This represents the TPU memory utilization rate of the i-th service node. This represents the memory utilization rate of the i-th service node. This represents the bandwidth utilization of the i-th service node. This represents the disk utilization rate of the i-th service node. This represents the maximum CPU and memory load capacity of the i-th service node. This represents the maximum load capacity of the GPU memory of the i-th service node. This represents the maximum load capacity of the TPU memory of the i-th service node. This represents the maximum memory load capacity of the i-th service node. This represents the maximum bandwidth load capacity of the i-th service node. This represents the maximum load capacity of the disk of the i-th service node.

[0011] Optionally, the step of performing heartbeat checks and service health status checks on each service node, and updating the service table based on the check results to obtain an updated service table, includes: When a heartbeat request is received from a service node and the heartbeat request is verified, the service node's status is updated to online, and the duration of receiving the heartbeat request is recorded. If the receiving duration is longer than the first timeout duration, the service node is marked as the first preset marker and its status is updated to offline. Detect whether each service node responds to the health check request within the second timeout period after receiving the health check request; If all service nodes respond to the health check request within the second timeout period, the response time of each service node is recorded, and the service node corresponding to the comparison result where the response time is greater than the third timeout period is marked as the second preset mark. If there is a service node that fails to respond to the health check request within the second timeout period, the service node that fails to respond to the health check request within the second timeout period will be marked as a second preset marker. Remove the entries corresponding to service nodes that are offline or marked with the second preset tag from the service table to obtain the updated service table.

[0012] Optionally, the step of calculating a comprehensive score for each service node based on the updated service registry, the updated service table, and the information table includes: When the number of service nodes in the updated service table is greater than the preset number, the comprehensive computing power and load corresponding to each service node in the updated service table are obtained from the information table; Obtain the response time corresponding to each service node in the updated service table from the updated service registry; For the comprehensive score of the i-th service node, the comprehensive computing power, load and response time corresponding to the i-th service node are substituted into the fourth preset formula to calculate the comprehensive score of the i-th service node. The fourth preset formula is as follows:

[0013] In the formula, This represents the overall score of the i-th service node. This indicates the weight of the delay in the overall score. This indicates the weight of the load on the overall score. This indicates the weight of computing power in the overall score. Indicates the sensitivity to delay decay. This indicates the sensitivity to load effects. This represents the response time of the i-th service node. This represents the response time of the j-th service node. This represents the load of the i-th service node. This represents the overall computing power of the i-th service node. This represents the total computing power of the j-th service node. This represents the total number of service nodes in the updated service table.

[0014] Secondly, embodiments of the present invention provide a service node discovery device under an edge computing network, the device comprising: The information update module is configured to update the service registry based on service information from several service nodes through the service registry center, thereby obtaining the updated service registry. The first calculation module is configured to calculate the total latency of each service node in the updated service registry sending data packets to the receiving service node, and to construct a service table based on each total latency. The second computing module is configured to calculate the overall computing power and load of each service node in the updated service registry, and to build an information table based on the overall computing power and the load. The inspection module is configured to perform heartbeat checks and service health status checks on each service node, and update the service table based on the inspection results to obtain the updated service table. The service node discovery module is configured to calculate a comprehensive score for each service node based on the updated service registry, the updated service table, and the information table, and determine the service node with the highest comprehensive score as the target service node.

[0015] Optionally, the first computing module is specifically configured to: Get the number of tasks arriving at the sending service node per unit time, the amount of task data processed by the sending service node per unit time, the size of the data packets sent by the sending service node, the transmission rate between adjacent nodes, and the calculation time of each service node in the updated service registry. For service nodes , will service nodes Calculation time Arrival at service node per unit time Number of tasks Service nodes per unit time The amount of task data processed Service Nodes Size of the data packets sent and the transmission rate between adjacent nodes Substituting into the first preset formula, the service node is calculated. Send data packets to the forwarding node Total latency ; The first preset formula is as follows: ; Similarly, the total latency for each service node in the updated service registry to send data packets to the receiving service node is calculated. In the formula, j is the total number of hops between the service node and the forwarding node. Starting from service node s, it reaches the forwarding node. Each forwarding node.

[0016] Optionally, the inspection module is configured to: When a heartbeat request is received from a service node and the heartbeat request is verified, the service node's status is updated to online, and the duration of receiving the heartbeat request is recorded. If the receiving duration is longer than the first timeout duration, the service node is marked as the first preset marker and its status is updated to offline. Detect whether each service node responds to the health check request within the second timeout period after receiving the health check request; If all service nodes respond to the health check request within the second timeout period, the response time of each service node is recorded, and the service node corresponding to the comparison result where the response time is greater than the third timeout period is marked as the second preset mark. If there is a service node that fails to respond to the health check request within the second timeout period, the service node that fails to respond to the health check request within the second timeout period will be marked as a second preset marker. Remove the entries corresponding to service nodes that are offline or marked with the second preset tag from the service table to obtain the updated service table.

[0017] Thirdly, embodiments of the present invention also provide an electronic device, including: a memory and a processor; the processor is configured to read and execute a computer program stored in the memory to implement the steps of the aforementioned service node discovery method under an edge computing network.

[0018] Fourthly, embodiments of the present invention also provide a computer storage medium storing computer-executable instructions, which, when executed, implement the steps of the aforementioned service node discovery method under an edge computing network.

[0019] The beneficial effects of the technical solutions provided by the embodiments of the present invention include: Based on the service information of several service nodes, the service registry is updated through the service registry center to obtain the updated service registry; the total latency of each service node sending data packets to the receiving service node in the updated service registry is calculated, and a service table is constructed based on each total latency; by calculating the computation latency and transmission latency of the service nodes, the service table of the service nodes is constructed to ensure that the service node with the shortest response time can be quickly located during the service discovery process, so as to meet the needs of time-sensitive services. Calculate the overall computing power and load of each service node in the updated service registry, and construct an information table based on the overall computing power and load; by periodically evaluating the overall computing power (including logical operation capability, parallel computing capability and neural network computing capability) of each service node, and monitoring the load status of the nodes (such as CPU, memory, bandwidth, etc.), ensure that the selected service nodes have sufficient computing power and can bear the load. Each service node undergoes a heartbeat check and a service health status check. The service table is updated based on the check results to obtain the updated service table. By periodically sending health check requests to service nodes, the health status of the nodes is monitored in real time, avoiding the allocation of user requests to unhealthy service nodes, thereby improving the stability and reliability of the system. Based on the updated service registry, the updated service table, and the information table, a comprehensive score for each service node is calculated, and the service node with the highest comprehensive score is determined as the target service node. According to the latency, load status, and comprehensive computing power of the service node, a comprehensive score method is used to select the best service node. This scoring mechanism takes into account multiple dimensions, can accurately match user needs, and improve the response speed and accuracy of services. This embodiment improves the discovery efficiency and response speed of time-sensitive services in edge computing networks, while ensuring service stability and efficient resource utilization. It solves the technical problems of low service discovery efficiency, slow service call response, poor service stability, and unreasonable resource allocation in traditional service discovery methods in large-scale distributed systems. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating an embodiment of the service node discovery method under the edge computing network of the present invention. Figure 2 This is a schematic diagram of the network topology of the service nodes in this invention; Figure 3 This is a schematic diagram illustrating the allocation of service nodes for this invention. Figure 4 for Figure 1 Detailed flowchart of step S40; Figure 5 This is a schematic diagram of the functional modules of a service node discovery device under the edge computing network of the present invention. Figure 6 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0024] In a first aspect, embodiments of the present invention provide a service node discovery method in an edge computing network.

[0025] In one embodiment, reference is made to Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the service node discovery method in an edge computing network according to the present invention. Figure 1 As shown, the service node discovery methods under the edge computing network include: Step S10: Based on the service information of several service nodes, update the service registry through the service registry center to obtain the updated service registry. In some specific embodiments, the service information includes: service name, service address and port, response time, load capacity, and computing power. The updated service registry, based on service information from several service nodes, is updated through a service registry center, resulting in an updated service registry that includes: After obtaining the service information of several service nodes, the service information of each service node and the service node itself are verified through the service registry center; The service information of each verified service node is updated in the service registry to obtain the updated service registry.

[0026] In this embodiment, service node registration is a fundamental step in the entire time-sensitive service discovery method. The purpose of service node registration is to upload all necessary information about the service nodes to the service registry, ensuring that the service registry can effectively manage and schedule all service nodes. The service registry stores the service node information and evaluates the service nodes based on service quality indicators, thereby providing a basis for subsequent service selection and scheduling.

[0027] Upon startup, each service node first initializes itself, preparing to provide time-sensitive services such as real-time video streaming and online financial transactions. Node initialization includes loading necessary configuration files and initializing network connections to ensure the node can function properly and provide services.

[0028] The service registration information is shown in Table 1.

[0029] Table 1

[0030] During startup, the service node collects and calculates the following service information, which will be used for service registration and subsequent scheduling: Service Name: Each service node must have a unique service name, such as "Live Video Service" or "Online Payment Service," so that the service registry can identify the service.

[0031] Service address and port: The network address and service port of the service node. Clients will connect to the service node through this service address. The service node must ensure that the address and port are configured correctly so that it can receive and process client requests normally.

[0032] Response time: The response time of a service node, which is usually the result of performance testing after the service starts.

[0033] Service node network bandwidth: This indicates the maximum network transmission rate that the node can provide. Bandwidth information helps in selecting a suitable node for services that require large amounts of data transmission. For example, real-time video services require high bandwidth to ensure smooth transmission of video data.

[0034] Load: The current load status of the service node, including the current number of connections, CPU utilization, memory usage, etc. This information helps the service registry center to determine the node's carrying capacity. The service node can monitor the system resource usage in real time and collect load information.

[0035] Overall computing power: The current computing power of the service node, including logical operation capabilities, parallel computing capabilities, and neural network computing capabilities. The service registry center uses this information to determine the overall computing power of the node.

[0036] After receiving a registration request from a service node, the service registry first verifies the service information and the service node itself to ensure that the information in the registration request is complete and meets the requirements. The verification includes: Check if the service address and port are valid, i.e., whether the address can be resolved correctly and whether the port is open and not occupied.

[0037] Whether metrics such as response time and load are within reasonable ranges is important. For example, does the response time meet the requirements of time-sensitive services, and is the load within the node's capacity? It's worth noting that while real-time load information may not be available during service node registration, the service registry can verify basic information such as service address, port, response time, and health check interface to ensure the correctness of the service node's basic functions and configuration. Load information, on the other hand, is dynamically collected and updated during service node operation to support subsequent service scheduling and management decisions.

[0038] Does the health check interface exist and can return a valid response? In other words, can the interface be accessed normally, and is the returned health status information accurate and reliable?

[0039] When the service registry receives a registration request from a service node, the verification process primarily relies on the validity of the service address and port, whether the response time is within the required range, and the validity of the health check interface to determine whether the verification passes. Specifically, the service registry checks whether the service address can be correctly resolved to an IP address, whether the port is open and not occupied; verifies whether the response time meets the requirements for time-sensitive services; and confirms that the health check interface exists and can return a valid response. If all the above verification conditions are met, the service registry considers the verification successful, stores the service node's information in the database, and updates the service registry, making the service node's information available for client queries.

[0040] Once the service information and verification of each service node are successful, the service registry stores the service node's registration information in the database and updates the service registry. At this point, the service node's service information becomes a usable resource that clients can query. The service registry must ensure the stability of the database and the consistency of the data to accurately store and manage service node information.

[0041] Step S20: Calculate the total latency for each service node in the updated service registry to send data packets to the receiving service node, and construct a service table based on each total latency; In some specific embodiments, calculating the total latency for each service node in the updated service registry to send data packets to the receiving service node includes: Get the number of tasks arriving at the sending service node per unit time, the amount of task data processed by the sending service node per unit time, the size of the data packets sent by the sending service node, the transmission rate between adjacent nodes, and the calculation time of each service node in the updated service registry. For service nodes , will service nodes Calculation time Arrival at service node per unit time Number of tasks Service nodes per unit time The amount of task data processed Service Nodes Size of the data packets sent and the transmission rate between adjacent nodes Substituting into the first preset formula, the service node is calculated. Send data packets to the forwarding node Total latency ; The first preset formula is as follows: ; Similarly, the total latency for each service node in the updated service registry to send data packets to the receiving service node is calculated. In the formula, j is the total number of hops between the service node and the forwarding node. Starting from service node s, it reaches the forwarding node. Each forwarding node.

[0042] In this embodiment, refer to Figure 2 , Figure 2 This is a schematic diagram of the network topology of the service nodes in this invention. Figure 2 As shown, service nodes D, A, F, B, G, K, S, J, N, E, H, L, and M are interconnected. Taking the calculation of the total latency of service node S as an example, if service node J is the service node receiving data packets sent by service node S, the latency of data packets arriving at service node S per unit time is calculated. Number of tasks Service nodes per unit time The amount of task data processed Service Nodes Size of the data packets sent Transmission rate between adjacent nodes and the computation time of each service node. The computation time for each service node is specified. It can be derived from analysis of log files or predictive models exported from machine learning.

[0043] The log file structure includes: task ID, task type, task input size, task start time, task end time, and computation duration. Analysis methods include: statistical averaging and historical data querying.

[0044] Task ID: Uniquely identifies each task.

[0045] Task types: For example, image processing, video streaming, financial transactions, etc.

[0046] Task input size: For example, the size of the input data (MB, GB, etc.).

[0047] Task start time: The timestamp at which the task begins processing on the service node.

[0048] Task completion time: The timestamp when the task is completed on the service node.

[0049] Calculation duration: Task end time minus task start time.

[0050] Statistical average: For each task type, calculate its average computation time on the service node.

[0051] Historical data query: Based on the input size and type of the task, query the computation time of similar tasks from the historical logs.

[0052] For service nodes , will service nodes Calculation time Arrival at service node per unit time Number of tasks Service nodes per unit time The amount of task data processed Service Nodes Size of the data packets sent and the transmission rate between adjacent nodes Substituting into the first preset formula, the service node is calculated. Send data packets to the forwarding node Total latency ; The first preset formula is as follows:

[0053] In the formula, , This indicates the queuing time of data packets at the sending nodes of adjacent nodes. , This indicates the processing time of the data packet at the sending node of the adjacent node. , It is a computing node To service node Transmission time, It is a service node The calculation time, It's the data packet size. It is the transmission rate between adjacent nodes. It is the number of tasks arriving in the queue per unit of time. This refers to the number of tasks processed per unit of time. These parameters can be inferred by using network performance testing tools such as iPerf and Speedtest for real-time measurement.

[0054] By analogy, the total latency for each service node in the updated service registry to send data packets to the receiving service node can be calculated. In the formula, j is the total number of hops between the service node and the forwarding node. Starting from service node s, it reaches the forwarding node. Each forwarding node.

[0055] In some specific embodiments, constructing a service table based on each of the total latency includes: Sequentially, the total delay and the minimum delay duration are compared. Perform the comparison and obtain the comparison results; If the comparison result shows that the total delay is less than or equal to the minimum delay duration, then the service name of the service node, the next-hop node, the total delay, and the timestamp of the sent data packet are combined into an entry. A service table is obtained based on the multiple entries.

[0056] In this embodiment, after calculating the total latency of each service node sending data packets to the receiving service node in the updated service registry, each total latency is sequentially compared with the set minimum latency. Perform the comparison and obtain the comparison results; If the comparison result shows that the total delay of a service node sending a data packet to the receiving service node is greater than the minimum delay duration, or if the timestamp of the data packet received by the node is significantly earlier than the current time or appears outdated compared to the timestamp of the most recently received data packet, then the data packet will be discarded directly.

[0057] If the comparison result shows that the total delay from a service node sending a data packet to the receiving service node is less than or equal to the minimum delay, the service node's information and next-hop route are added to the node's service table. If the service node... If it already exists in the service table, then update. If the value is not specified, then the service name of the service node, the next-hop node, the total latency, and the timestamp of the sent data packet are combined into an entry, for example: Insert into the service table, obtain the service table based on the obtained multiple entries, and forward the update operation to its next adjacent service node.

[0058] The service table is shown in Table 2.

[0059] Table 2

[0060] As shown in Table 2, the service table contains four entries: service name, next-hop node providing the time-sensitive service, total service latency at the current node, and timestamp of the data packet. The total service latency mainly consists of two parts: computation latency and transmission latency. Computation latency can be analyzed from log files or prediction models derived from machine learning, while transmission latency can be calculated using a hop-by-hop accumulation method.

[0061] Step S30: Calculate the comprehensive computing power and load of each service node in the updated service registry, and construct an information table based on the comprehensive computing power and the load; In this embodiment, the comprehensive computing power of the i-th service node in the updated service registry is calculated using a second preset formula. The second preset formula is as follows:

[0062] In the formula, This represents the overall computing power of the i-th service node; This represents the logical operation capability of the i-th service node. , This represents the number of idle CPU cores in the i-th service node. This represents the clock frequency of the CPU in the i-th service node. This represents the number of instructions per cycle for the i-th service node; This represents the parallel computing capability of the i-th service node. , This represents the number of idle parallel computing units in the i-th service node. This represents the clock frequency of the graphics processing unit of the i-th service node. This represents the number of floating-point operations that can be performed by the i-th service node in each clock cycle; This represents the neural network computing power of the i-th service node. , This represents the number of idle neural network computation cores in the i-th service node. This represents the clock frequency of the neural network processing unit of the i-th service node. This represents the number of neural network-related operations that can be performed in each clock cycle of the i-th service node.

[0063] The load of the i-th service node in the updated service registry is calculated using the third preset formula. The third preset formula is as follows: ) In the formula, This represents the load of the i-th service node. This represents the CPU and memory utilization of the i-th service node. This represents the GPU memory utilization of the i-th service node. This represents the TPU memory utilization rate of the i-th service node. This represents the memory utilization rate of the i-th service node. This represents the bandwidth utilization of the i-th service node. This represents the disk utilization rate of the i-th service node. This represents the maximum CPU and memory load capacity of the i-th service node. This represents the maximum load capacity of the GPU memory of the i-th service node. This represents the maximum load capacity of the TPU memory of the i-th service node. This represents the maximum memory load capacity of the i-th service node. This represents the maximum bandwidth load capacity of the i-th service node. This represents the maximum load capacity of the disk of the i-th service node.

[0064] The service node's service name, load, and overall computing power are combined into an entry and inserted into the information table. The information table is then generated based on multiple entries. The information table is shown in Table 3.

[0065] Table 3

[0066] As shown in Table 3, the information table contains three entries: service name, service load, and overall service computing power.

[0067] Each service node autonomously collects its own resource usage (such as CPU and load) and shares it with neighboring service nodes at fixed time intervals (such as 5 or 10 seconds). In this propagation method, each service node randomly selects several neighboring service nodes and sends them its latest resource status. The receiving neighboring service nodes further share these statuses with their neighboring service nodes, forming a gradual propagation process. Furthermore, each service node adds its load and overall computing power to its own information table.

[0068] Step S40: Perform a heartbeat check and a service health status check on each service node, and update the service table based on the check results to obtain the updated service table; In some specific embodiments, reference is made to Figure 4 Step S40 involves performing a heartbeat check and a service health status check on each service node, and updating the service table based on the check results to obtain an updated service table, including: Step S401: When a heartbeat request is received from a service node and the verification of the heartbeat request is successful, update the status of the service node to online and record the reception duration of the heartbeat request. Step S402: If the reception duration is greater than the first timeout duration, then mark the service node as the first preset mark and update its status to offline; Step S403: Detect whether each service node responds to the health check request within the second timeout period after receiving the health check request; Step S404: If all the health check requests are responded to within the second timeout period, the response duration of each service node is recorded, and the service node corresponding to the comparison result where the response duration is greater than the third timeout period is marked as the second preset mark. Step S405: If there is a service node that does not respond to the health check request within the second timeout period, then mark the service node that does not respond to the health check request within the second timeout period as a second preset mark. Step S406: Remove the entries corresponding to service nodes that are offline or marked with a second preset mark from the service table to obtain the updated service table.

[0069] In this embodiment, the configuration file of the service node sets the heartbeat request sending interval and the first timeout duration for the service registry to wait for a heartbeat response. The heartbeat request sending interval should be reasonably set according to the real-time nature and importance of the service node to ensure timely updates to the service status; the first timeout duration for the service registry to wait for a heartbeat response should be slightly longer than the normal heartbeat response duration to avoid misjudgment.

[0070] In the service node's code, implement logic to periodically send heartbeat requests, ensuring that heartbeat requests are sent at set intervals. The heartbeat request should contain basic information about the service node, such as its service name, address, and port, so that the service registry can accurately identify the service node.

[0071] When a service registry receives a heartbeat request, it performs a series of verifications to determine the request's validity. First, it checks the authentication information in the heartbeat request. For digital signatures, the service registry uses a pre-allocated public key to verify that the signature originates from a service node possessing the corresponding private key. For tokens, the service registry verifies the token's signature, decodes the token content, and checks the service node's identity information and expiration time to ensure the token has not been tampered with and is still valid. Next, it verifies that the source IP address and port of the heartbeat request match the registered service node information, ensuring the request originates from a legitimate service node. Furthermore, the service registry checks the data format of the heartbeat request to ensure it contains all necessary fields, such as service name, address, port, and heartbeat timestamp. It also verifies the consistency and reasonableness of the data content, such as checking if the heartbeat timestamp is within a reasonable timeframe and if the load information is within a normal range. Only when the heartbeat request passes all these verification steps—including confirming the legitimate source, data integrity, and correct format—will the service registry consider the verification successful. This ensures that the heartbeat request indeed originates from a legitimate service node and that the provided information is complete and reliable. After successful verification, the service registry updates the status of the corresponding service node to online and records the latest heartbeat request reception duration.

[0072] The service center periodically checks the heartbeat request reception time of each service node. If the heartbeat request reception time of a service node exceeds the first timeout period set by the service registry for waiting for a heartbeat response, the service center marks the service node with the first preset flag, for example, heartbeat failure, and immediately updates its status to offline. The service node's relevant information is also removed from the service directory to avoid providing unavailable services to users.

[0073] Furthermore, in the service registry or relevant configuration files, a second timeout duration, a health status check frequency, and a third timeout duration should be set for each service node. These parameters should be reasonably set according to the actual needs of the service nodes and the network environment to ensure the effectiveness and accuracy of health checks.

[0074] Based on the set health status check frequency, health check requests are periodically sent to each service node, and the response time of each service node to the health check request is recorded. The health check request should contain the necessary check instructions so that the service node can respond correctly.

[0075] Check whether each service node responds to the health check request within the second timeout period after receiving the health check request; If all nodes respond, the response time of each service node is recorded, and the service node whose response time is greater than the third timeout ...

[0076] If there is a service node that fails to respond to the health check request within the second timeout period, the service node that fails to respond to the health check request within the second timeout period will be marked with a second preset mark (e.g., unhealthy).

[0077] Based on the health check results, after updating the status information of the service nodes, the system removes the entries corresponding to service nodes that are offline, marked with a first preset flag (e.g., heartbeat failure), or marked with a second preset flag (e.g., unhealthy) from the service table, resulting in an updated service table. This avoids allocating traffic to service nodes that are offline or marked with a second preset flag (e.g., unhealthy), thereby ensuring high availability and stability of the service. At the same time, the system dynamically adjusts the traffic allocation for healthy service nodes to fully utilize available resources and improve the overall performance of the service.

[0078] Step S50: Based on the updated service registry, the updated service table, and the information table, calculate the comprehensive score for each service node, and determine the service node with the highest comprehensive score as the target service node.

[0079] In some specific embodiments, calculating the comprehensive score for each service node based on the updated service registry, the updated service table, and the information table includes: When the number of service nodes in the updated service table is greater than the preset number, the comprehensive computing power and load corresponding to each service node in the updated service table are obtained from the information table; Obtain the response time corresponding to each service node in the updated service table from the updated service registry; For the comprehensive score of the i-th service node, the comprehensive computing power, load and response time corresponding to the i-th service node are substituted into the fourth preset formula to calculate the comprehensive score of the i-th service node. The fourth preset formula is as follows:

[0080] In the formula, This represents the overall score of the i-th service node. This indicates the weight of the delay in the overall score. This indicates the weight of the load on the overall score. This indicates the weight of computing power in the overall score. Indicates the sensitivity to delay decay. This indicates the sensitivity to load effects. This represents the response time of the i-th service node. This represents the response time of the j-th service node. This represents the load of the i-th service node. This represents the overall computing power of the i-th service node. This represents the total computing power of the j-th service node. This represents the total number of service nodes in the updated service table.

[0081] In this embodiment, refer to Figure 3 Taking a preset quantity of 1 as an example, when a user requests a service through the access point, it checks whether the number of service nodes in the updated service table is greater than 1. If there is only one service node in the updated service table, the request is forwarded to the next-hop route of that service node according to the information in the updated service table.

[0082] If the updated service table contains two or more service nodes, then retrieve the overall computing power and load corresponding to each service node in the updated service table from the information table. Finally, retrieve the response time corresponding to each service node in the updated service table from the updated service registry.

[0083] For the comprehensive score of the i-th service node, the comprehensive computing power, load and response time corresponding to the i-th service node are substituted into the fourth preset formula to calculate the comprehensive score of the i-th service node. The fourth preset formula is as follows:

[0084] In the formula, This represents the overall score of the i-th service node. This indicates the weight of the delay in the overall score. This indicates the weight of the load on the overall score. This indicates the weight of computing power in the overall score. Indicates the sensitivity to delay decay. This indicates the sensitivity to load effects. This represents the response time of the i-th service node. This represents the response time of the j-th service node. This represents the load of the i-th service node. This represents the overall computing power of the i-th service node. This represents the total computing power of the j-th service node. This represents the total number of service nodes in the updated service table.

[0085] This process is repeated to obtain the comprehensive score for each service node in the updated service table. Based on the comprehensive score of each service node, the service node with the highest comprehensive score is determined as the target service node, and the request is sent to the target service node. The routing node of the target service node serves as the next-hop route for the instance to ensure that the selected service meets the user's needs.

[0086] Furthermore, the system continuously monitors the health status of service nodes and performs heartbeat and service health status checks again at the end of each check cycle. Simultaneously, the system records the results of each heartbeat and service health status check, including response time, health status changes, failures, and timeouts, for subsequent analysis, latency calculation, troubleshooting, and auditing.

[0087] In this embodiment, based on the service information of several service nodes, the service registry is updated through the service registry center to obtain the updated service registry; the total latency of each service node sending data packets to the receiving service node in the updated service registry is calculated, and a service table is constructed based on each total latency; by calculating the computation latency and transmission latency of the service nodes, the service table of the service nodes is constructed to ensure that the service node with the shortest response time can be quickly located during the service discovery process, thus meeting the needs of time-sensitive services; Calculate the overall computing power and load of each service node in the updated service registry, and construct an information table based on the overall computing power and load; by periodically evaluating the overall computing power (including logical operation capability, parallel computing capability and neural network computing capability) of each service node, and monitoring the load status of the nodes (such as CPU, memory, bandwidth, etc.), ensure that the selected service nodes have sufficient computing power and can bear the load. Each service node undergoes a heartbeat check and a service health status check. The service table is updated based on the check results to obtain the updated service table. By periodically sending health check requests to service nodes, the health status of the nodes is monitored in real time, avoiding the allocation of user requests to unhealthy service nodes, thereby improving the stability and reliability of the system. Based on the updated service registry, the updated service table, and the information table, a comprehensive score for each service node is calculated, and the service node with the highest comprehensive score is determined as the target service node. According to the latency, load status, and comprehensive computing power of the service node, a comprehensive score method is used to select the best service node. This scoring mechanism takes into account multiple dimensions, can accurately match user needs, and improve the response speed and accuracy of services. This embodiment improves the discovery efficiency and response speed of time-sensitive services in edge computing networks, while ensuring service stability and efficient resource utilization. It solves the technical problems of low service discovery efficiency, slow service call response, poor service stability, and unreasonable resource allocation in traditional service discovery methods in large-scale distributed systems.

[0088] Secondly, embodiments of the present invention also provide a service node discovery device under an edge computing network.

[0089] In one embodiment, reference is made to Figure 5 , Figure 5 This is a functional module diagram of an embodiment of the service node discovery device in the edge computing network of the present invention. Figure 5 As shown, the service node discovery device under the edge computing network includes: The information update module 10 is configured to update the service registry through the service registry center based on the service information of several service nodes, so as to obtain the updated service registry. The first calculation module 20 is configured to calculate the total latency of each service node in the updated service registry sending data packets to the receiving service node, and to construct a service table based on each total latency. The second computing module 30 is configured to calculate the comprehensive computing power and load of each service node in the updated service registry, and to build an information table based on the comprehensive computing power and the load. Inspection module 40 is configured to perform heartbeat checks and service health status checks on each service node, and update the service table based on the inspection results to obtain an updated service table; The service node discovery module 50 is configured to calculate a comprehensive score for each service node based on the updated service registry, the updated service table, and the information table, and determine the service node with the highest comprehensive score as the target service node.

[0090] Optionally, in one embodiment, the service information includes: service name, service address and port, response time, load capacity, and computing power, and the information update module 10 is configured to: After obtaining the service information of several service nodes, the service information of each service node and the service node itself are verified through the service registry center; The service information of each verified service node is updated in the service registry to obtain the updated service registry.

[0091] Optionally, in one embodiment, the first computing module 20 is configured to: Get the number of tasks arriving at the sending service node per unit time, the amount of task data processed by the sending service node per unit time, the size of the data packets sent by the sending service node, the transmission rate between adjacent nodes, and the calculation time of each service node in the updated service registry. For service nodes , will service nodes Calculation time Arrival at service node per unit time Number of tasks Service nodes per unit time The amount of task data processed Service Nodes Size of the data packets sent and the transmission rate between adjacent nodes Substituting into the first preset formula, the service node is calculated. Send data packets to the forwarding node Total latency ; The first preset formula is as follows: ; Similarly, the total latency for each service node in the updated service registry to send data packets to the receiving service node is calculated. In the formula, j is the total number of hops between the service node and the forwarding node. Starting from service node s, it reaches the forwarding node. Each forwarding node.

[0092] Optionally, in one embodiment, the first computing module 20 is further configured to: Sequentially, the total delay and the minimum delay duration are compared. Perform the comparison and obtain the comparison results; If the comparison result shows that the total delay is less than or equal to the minimum delay duration, then the service name of the service node, the next-hop node, the total delay, and the timestamp of the sent data packet are combined into an entry. A service table is obtained based on the multiple entries.

[0093] Optionally, in one embodiment, the second computing module 30 is configured to: The comprehensive computing power of the i-th service node in the updated service registry is calculated using the second preset formula. The load of the i-th service node in the updated service registry is calculated using the third preset formula. The second preset formula is as follows:

[0094] In the formula, This represents the overall computing power of the i-th service node; This represents the logical operation capability of the i-th service node. , This represents the number of idle CPU cores in the i-th service node. This represents the clock frequency of the CPU in the i-th service node. This represents the number of instructions per cycle for the i-th service node; This represents the parallel computing capability of the i-th service node. , This represents the number of idle parallel computing units in the i-th service node. This represents the clock frequency of the graphics processing unit of the i-th service node. This represents the number of floating-point operations that can be performed by the i-th service node in each clock cycle; This represents the neural network computing power of the i-th service node. , This represents the number of idle neural network computation cores in the i-th service node. This represents the clock frequency of the neural network processing unit of the i-th service node. This represents the number of neural network-related operations that can be performed by the i-th service node in each clock cycle; The third preset formula is as follows: ) In the formula, This represents the load of the i-th service node. This represents the CPU and memory utilization of the i-th service node. This represents the GPU memory utilization of the i-th service node. This represents the TPU memory utilization rate of the i-th service node. This represents the memory utilization rate of the i-th service node. This represents the bandwidth utilization of the i-th service node. This represents the disk utilization rate of the i-th service node. This represents the maximum CPU and memory load capacity of the i-th service node. This represents the maximum load capacity of the GPU memory of the i-th service node. This represents the maximum load capacity of the TPU memory of the i-th service node. This represents the maximum memory load capacity of the i-th service node. This represents the maximum bandwidth load capacity of the i-th service node. This represents the maximum load capacity of the disk of the i-th service node.

[0095] Optionally, in one embodiment, the inspection module 40 is configured to: When a heartbeat request is received from a service node and the heartbeat request is verified, the status of the service node is updated to online, and the time when the heartbeat request is received is recorded. If the receiving time is greater than the first timeout duration, the service node is marked as the first preset marker, and its status is updated to offline. Detect whether each service node responds to the health check request within the second timeout period after receiving the health check request; If all respond, the response time of each service node is recorded, and the service node whose response time is greater than the third timeout time is marked as the second preset marker; If there is a service node that does not respond to the health check request, then the service node that responds to the health check request is marked with a second preset mark; Remove the entries corresponding to service nodes that are offline or marked with the second preset tag from the service table to obtain the updated service table.

[0096] Optionally, in one embodiment, the service node discovery module 50 is configured to: When the number of service nodes in the updated service table is greater than the preset number, the comprehensive computing power and load corresponding to each service node in the updated service table are obtained from the information table; Obtain the response time corresponding to each service node in the updated service table from the updated service registry; For the comprehensive score of the i-th service node, the comprehensive computing power, load and response time corresponding to the i-th service node are substituted into the fourth preset formula to calculate the comprehensive score of the i-th service node. The fourth preset formula is as follows:

[0097] In the formula, This represents the overall score of the i-th service node. This indicates the weight of the delay in the overall score. This indicates the weight of the load on the overall score. This indicates the weight of computing power in the overall score. Indicates the sensitivity to delay decay. This indicates the sensitivity to load effects. This represents the response time of the i-th service node. This represents the response time of the j-th service node. This represents the load of the i-th service node. This represents the overall computing power of the i-th service node. This represents the total computing power of the j-th service node. This represents the total number of service nodes in the updated service table.

[0098] The functions of each module in the service node discovery device under the aforementioned edge computing network correspond to the steps in the embodiment of the service node discovery method under the aforementioned edge computing network, and their functions and implementation processes will not be described in detail here.

[0099] Thirdly, embodiments of the present invention also provide an electronic device, the structure of which is as follows: Figure 6 As shown, it includes: a memory and a processor, wherein the processor is used to read and execute the computer program stored in the memory to implement the aforementioned service node discovery method under an edge computing network.

[0100] Fourthly, embodiments of the present invention also provide a computer storage medium storing computer-executable instructions, which, when executed, implement the aforementioned service node discovery method under an edge computing network.

[0101] Finally, it should be noted that while some processes described in the embodiments of the present invention include multiple operations or steps that appear in a specific order, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of the present invention, or may be executed in parallel. The sequence number of the operation is only used to distinguish different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.

[0102] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A service node discovery method in an edge computing network, characterized in that, The method includes: Based on the service information of several service nodes, the service registry is updated through the service registry center to obtain the updated service registry. Calculate the total latency for each service node in the updated service registry to send data packets to the receiving service node, and construct a service table based on each total latency; Calculate the overall computing power and load of each service node in the updated service registry, and construct an information table based on the overall computing power and load; Perform a heartbeat check and a service health status check on each service node, and update the service table based on the check results to obtain the updated service table; Based on the updated service registry, the updated service table, and the information table, a comprehensive score is calculated for each service node, and the service node with the highest comprehensive score is determined as the target service node. The calculation of the total latency for each service node in the updated service registry to send data packets to the receiving service node includes: Get the number of tasks arriving at the sending service node per unit time, the amount of task data processed by the sending service node per unit time, the size of the data packets sent by the sending service node, the transmission rate between adjacent nodes, and the calculation time of each service node in the updated service registry. For service nodes , will service nodes Calculation time The number of tasks arriving at the service node per unit time Service nodes per unit time The amount of task data processed Service Nodes Size of the data packets sent and the transmission rate between adjacent nodes Substituting into the first preset formula, the service node is calculated. Send data packets to the forwarding node Total latency ; The first preset formula is as follows: ； Similarly, the total latency for each service node in the updated service registry to send data packets to the receiving service node is calculated. In the formula, j is the total number of hops between the service node and the forwarding node. Starting from service node s, it reaches the forwarding node. Each forwarding node; The calculation of the overall computing power and load of each service node in the updated service registry includes: The comprehensive computing power of the i-th service node in the updated service registry is calculated using the second preset formula. The load of the i-th service node in the updated service registry is calculated using the third preset formula. The second preset formula is as follows: In the formula, This represents the overall computing power of the i-th service node; This represents the logical operation capability of the i-th service node. , This represents the number of idle CPU cores in the i-th service node. This represents the clock frequency of the CPU in the i-th service node. This represents the number of instructions per cycle for the i-th service node; This represents the parallel computing capability of the i-th service node. , This represents the number of idle parallel computing units in the i-th service node. This represents the clock frequency of the graphics processing unit of the i-th service node. This represents the number of floating-point operations that can be performed by the i-th service node in each clock cycle; This represents the neural network computing power of the i-th service node. , This represents the number of idle neural network computation cores in the i-th service node. This represents the clock frequency of the neural network processing unit of the i-th service node. This represents the number of neural network-related operations that can be performed by the i-th service node in each clock cycle; The third preset formula is as follows: ) In the formula, This represents the load of the i-th service node. This represents the CPU and memory utilization of the i-th service node. This represents the GPU memory utilization of the i-th service node. This represents the TPU memory utilization rate of the i-th service node. This represents the memory utilization rate of the i-th service node. This represents the bandwidth utilization of the i-th service node. This represents the disk utilization rate of the i-th service node. This represents the maximum CPU and memory load capacity of the i-th service node. This represents the maximum load capacity of the GPU memory of the i-th service node. This represents the maximum load capacity of the TPU memory of the i-th service node. This represents the maximum memory load capacity of the i-th service node. This represents the maximum bandwidth load capacity of the i-th service node. This represents the maximum load capacity of the disk of the i-th service node.

2. The service node discovery method under the edge computing network according to claim 1, characterized in that, The service information includes: service name, service address and port, response time, load capacity, and computing power. Based on the service information from several service nodes, the service registry is updated through a service registry center to obtain the updated service registry, which includes: After obtaining the service information of several service nodes, the service information of each service node and the service node itself are verified through the service registry center; The service information of each verified service node is updated in the service registry to obtain the updated service registry.

3. The service node discovery method under the edge computing network according to claim 1, characterized in that, The process of constructing a service table based on each of the total latency includes: Sequentially, the total delay and the minimum delay duration are compared. The comparison is performed to obtain the comparison results; If the comparison result shows that the total delay is less than or equal to the minimum delay duration, then the service name of the service node, the next-hop node, the total delay, and the timestamp of the sent data packet are combined into an entry. A service table is obtained based on the multiple entries.

4. The service node discovery method under the edge computing network according to claim 1, characterized in that, The process of performing heartbeat checks and service health status checks on each service node, and updating the service table based on the check results, to obtain the updated service table includes: When a heartbeat request is received from a service node and the heartbeat request is verified, the service node's status is updated to online, and the duration of receiving the heartbeat request is recorded. If the receiving duration is longer than the first timeout duration, the service node is marked as the first preset marker and its status is updated to offline. Detect whether each service node responds to the health check request within the second timeout period after receiving the health check request; If all service nodes respond to the health check request within the second timeout period, the response time of each service node is recorded, and the service node corresponding to the comparison result where the response time is greater than the third timeout period is marked as the second preset mark. If there is a service node that fails to respond to the health check request within the second timeout period, the service node that fails to respond to the health check request within the second timeout period will be marked as a second preset marker. Remove the entries corresponding to service nodes that are offline or marked with the second preset tag from the service table to obtain the updated service table.

5. The service node discovery method under the edge computing network according to claim 1, characterized in that, The comprehensive score for each service node is calculated based on the updated service registry, the updated service table, and the information table, including: When the number of service nodes in the updated service table is greater than the preset number, the comprehensive computing power and load corresponding to each service node in the updated service table are obtained from the information table; Obtain the response time corresponding to each service node in the updated service table from the updated service registry; For the comprehensive score of the i-th service node, the comprehensive computing power, load and response time corresponding to the i-th service node are substituted into the fourth preset formula to calculate the comprehensive score of the i-th service node. The fourth preset formula is as follows: In the formula, This represents the overall score of the i-th service node. This indicates the weight of the delay in the overall score. This indicates the weight of the load on the overall score. This indicates the weight of computing power in the overall score. Indicates the sensitivity to delay decay. This indicates the sensitivity to load effects. This represents the response time of the i-th service node. This represents the response time of the j-th service node. This represents the load of the i-th service node. This represents the overall computing power of the i-th service node. This represents the overall computing power of the j-th service node. This represents the total number of service nodes in the updated service table.

6. A service node discovery device in an edge computing network, used to implement the method of claim 1, characterized in that, The device includes: The information update module is configured to update the service registry based on service information from several service nodes through the service registry center, thereby obtaining the updated service registry. The first calculation module is configured to calculate the total latency of each service node in the updated service registry sending data packets to the receiving service node, and to construct a service table based on each total latency. The second computing module is configured to calculate the overall computing power and load of each service node in the updated service registry, and to build an information table based on the overall computing power and the load. The inspection module is configured to perform heartbeat checks and service health status checks on each service node, and update the service table based on the inspection results to obtain the updated service table. The service node discovery module is configured to calculate a comprehensive score for each service node based on the updated service registry, the updated service table, and the information table, and determine the service node with the highest comprehensive score as the target service node.

7. The service node discovery device under the edge computing network according to claim 6, characterized in that, The inspection module is configured to: When a heartbeat request is received from a service node and the heartbeat request is verified, the service node's status is updated to online, and the duration of receiving the heartbeat request is recorded. If the receiving duration is longer than the first timeout duration, the service node is marked as the first preset marker and its status is updated to offline. Detect whether each service node responds to the health check request within the second timeout period after receiving the health check request; If all service nodes respond to the health check request within the second timeout period, the response time of each service node is recorded, and the service node corresponding to the comparison result where the response time is greater than the third timeout period is marked as the second preset mark. If there is a service node that fails to respond to the health check request within the second timeout period, the service node that fails to respond to the health check request within the second timeout period will be marked as a second preset marker. Remove the entries corresponding to service nodes that are offline or marked with the second preset tag from the service table to obtain the updated service table.

8. An electronic device, characterized in that, include: Memory and processor; The processor is configured to read and execute the computer program stored in the memory to implement the steps of the service node discovery method under the edge computing network as described in any one of claims 1-5.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed, implement the steps of the service node discovery method under the edge computing network as described in any one of claims 1-5.

Images