Edge service invocation method and system

Abstract

This invention relates to the field of edge service invocation technology, and particularly to an edge service invocation method and system. The method includes: defining the access range of the edge service, marking edge nodes, and calculating the distance between edge nodes; sorting the edge nodes in ascending order of distance to obtain a distance list, wherein each edge node corresponds to a distance list; selecting a preset number of edge nodes from the top of the distance list to obtain collaborative nodes; upon receiving an invocation request uploaded by a user terminal, finding the edge node closest to the user terminal and defining it as the target node; and calculating the load value of the target node. This invention, by establishing a data channel, enables direct processing of invocation requests, reducing network overhead and resource consumption, while meeting the requirements of low latency and resource saving, thus greatly improving the efficiency of edge service invocation and user experience.

Description

Technical Field

[0001] This invention relates to the field of edge service invocation technology, and in particular to an edge service invocation method and system. Background Technology

[0002] In edge computing environments, edge nodes typically exhibit characteristics such as wide geographical distribution, strong network connectivity heterogeneity, and limited resources, leading to frequent changes in edge services, such as going online / offline, migration, and load fluctuations. However, traditional edge service processing mostly relies on a centralized service architecture, where access requests are written to a single node and then distributed from there. While this approach can effectively balance loads, it is prone to instability in edge environments. When high concurrency demands arise, data congestion can easily occur, significantly reducing the efficiency of access request processing and impacting overall service quality and response performance.

[0003] Therefore, "how to prioritize writing user requests to the nearest edge node" is the technical problem that this invention needs to solve. Summary of the Invention

[0004] The purpose of this invention is to provide an edge service invocation method and system to solve the problem of "how to prioritize writing user requests to the nearest edge node" mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] An edge service invocation method, the method comprising:

[0007] The access range of the edge service is defined, the edge nodes are marked, and the distance between the edge nodes is calculated. The edge nodes are sorted in ascending order of the distance to obtain a distance list, where each edge node corresponds to a distance list.

[0008] From the top of the distance list, a preset number of edge nodes are selected to obtain collaborative nodes. When a call request is received from the user terminal, the edge node closest to the user terminal is found and defined as the target node. The load value of the target node is calculated, and it is determined whether the load value is greater than a threshold. If so, the probe request pre-stored in the target node is triggered and distributed to the collaborative nodes in parallel.

[0009] The system receives feedback information from the cooperating node, defines the cooperating node corresponding to the first received feedback information as the optimal node, and forwards the call request to the optimal node through the target node, thereby establishing a communication link between the optimal node and the user terminal.

[0010] Collect the unique identifier of the optimal node and cache it in the target node. When the target node receives the call request again, it traces back the source of the request and sends the unique identifier to the source of the request, thus establishing a data channel between the source of the request and the optimal node.

[0011] Furthermore, the step of marking edge nodes includes:

[0012] Select several metrics for the edge nodes and assign weights to them;

[0013] Real-time values ​​of metrics in edge nodes are collected, and the load value of each edge node is calculated based on the weights and real-time values.

[0014] Furthermore, the method also includes:

[0015] Construct a time window and update the load value after the time window arrives;

[0016] The editor adjusts the strategy by using the load value to adjust the sorting in the distance list.

[0017] Furthermore, the step of calculating the load value of the target node includes:

[0018] Plot a load fluctuation trend graph with time on the horizontal axis and load value on the vertical axis.

[0019] Collect attribute information of edge nodes, generate labels, and insert the labels into the load fluctuation trend graph, wherein the attribute information includes at least: deployment location and unique identifier.

[0020] Furthermore, the step of forwarding the call request to the optimal node via the target node includes:

[0021] Find the two closest edge nodes, form a cooperative pair, and select a set value from the load values;

[0022] When the load value of the edge node exceeds the set value, the call request is reconnected using the mutual assistance pair.

[0023] Furthermore, the method also includes:

[0024] Create a layered architecture consisting of a top layer, a parent layer, and child layers, and divide the access scope into several blocks;

[0025] Establish a correspondence between the hierarchical architecture and the blocks. Based on the correspondence, mount the edge nodes in the blocks to the hierarchical architecture to generate an edge service graph.

[0026] Furthermore, the step of establishing a data channel between the request source and the optimal node includes:

[0027] The processing task sent by the user terminal is received via the data channel and divided into several task segments;

[0028] Based on the edge node corresponding to the unique identifier, the task fragment is distributed to the mutual aid pair and the target node.

[0029] Furthermore, the system includes:

[0030] The module is used to define the access range of edge services, mark edge nodes, calculate the distance between edge nodes, and sort the edge nodes in ascending order of distance to obtain a distance list, where each edge node corresponds to a distance list.

[0031] The parallel module is used to select a preset number of edge nodes from the top of the distance list to obtain collaborative nodes. When a call request is received from the user terminal, the edge node closest to the user terminal is found and defined as the target node. The load value of the target node is calculated and it is determined whether the load value is greater than a threshold. If so, the probe request pre-stored in the target node is triggered and distributed to the collaborative nodes in parallel.

[0032] The module is used to receive feedback information sent by the collaborative node, define the collaborative node corresponding to the first received feedback information as the optimal node, and forward the call request to the optimal node through the target node, thereby establishing a communication link between the optimal node and the user terminal.

[0033] A module is established to collect the unique identifier of the optimal node and cache it in the target node. When the target node receives a call request again, it traces back the source of the request and sends the unique identifier to the source of the request, thus establishing a data channel between the source of the request and the optimal node.

[0034] Furthermore, the obtaining module includes:

[0035] The setting unit is used to select several indicators of the edge nodes and set their weights;

[0036] The calculation unit is used to collect real-time values ​​of indicators in the edge nodes and calculate the load value of each edge node based on the weights and real-time values.

[0037] Furthermore, the parallel module includes:

[0038] The plotting unit is used to plot a load fluctuation trend chart with time as the horizontal axis and load value as the vertical axis.

[0039] An insertion unit is used to collect attribute information of edge nodes, generate tags, and insert the tags into the load fluctuation trend graph, wherein the attribute information includes at least: deployment location and unique identifier.

[0040] Compared with the prior art, the beneficial effects of the present invention are:

[0041] By connecting call requests to the edge node closest to the user terminal, transmission latency can be significantly reduced, user request response speed can be improved, and user experience can be optimized. By identifying the optimal node, a "first response, takeover" approach can be used to quickly find nodes with available resources, avoiding request queuing, delays, or even loss due to node overload. At the same time, resource utilization is improved by minimizing forwarding paths, avoiding resource idleness. By establishing data channels, call requests can be processed directly, reducing network overhead and resource consumption, and taking into account the needs of low latency and resource conservation, greatly improving the call efficiency of edge services and user experience. Attached Figure Description

[0042] Figure 1 This is a flowchart of the edge service caller provided in an embodiment of the present invention;

[0043] Figure 2 This is a first sub-flow diagram of the edge service invocation method provided in an embodiment of the present invention;

[0044] Figure 3 This is a second sub-flow diagram of the edge service invocation method provided in an embodiment of the present invention;

[0045] Figure 4 This is a third sub-flow diagram of the edge service invocation method provided in an embodiment of the present invention;

[0046] Figure 5 This is a fourth sub-flow flowchart of the edge service invocation method provided in an embodiment of the present invention;

[0047] Figure 6 This is a block diagram of the edge service invocation system provided in an embodiment of the present invention;

[0048] Figure 7 A block diagram of the components of the edge service invocation system provided in this embodiment of the invention;

[0049] Figure 8 A block diagram illustrating the composition of a parallel module in an edge service invocation system provided in an embodiment of the present invention;

[0050] Figure 9 A block diagram illustrating the components of the construction module in the edge service invocation system provided in this embodiment of the invention;

[0051] Figure 10 This is a block diagram of the components of the establishment module in the edge service invocation system provided in an embodiment of the present invention. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0053] In Example 1, Figure 1 The implementation flow of the edge service invocation method provided in this embodiment of the invention is illustrated below, and is described in detail below:

[0054] S100: Define the access range of the edge service, mark the edge nodes, calculate the distance between the edge nodes, and sort the edge nodes in ascending order of the distance to obtain a distance list, wherein each edge node corresponds to a distance list.

[0055] Based on geographical location, user distribution density, and network topology information, the access range of edge services is defined, which is also the coverage area of ​​edge services. Within the access range, devices with computing, storage, and network capabilities are selected to obtain edge nodes. The specific location of the edge nodes within the access range is located, the latitude and longitude coordinates of each node are determined, and the distance between edge nodes is calculated. The nodes are then sorted according to this distance order to obtain a distance list.

[0056] For example, if there are five edge nodes A, B, C, D, and E within the access range, calculate the distances from B, C, D, and E to A, centered on node A, and sort them in ascending order of these distances. Define the sorted results as a distance list. Then, calculate the distances from A, C, D, and E to B, centered on B, to obtain the distance list corresponding to B. And so on, with each edge node corresponding to a distance list.

[0057] The distance list shows the distance between edge nodes. The smaller the distance, the lower the communication latency, which allows for the use of higher frequency and larger bandwidth links, reducing congestion and transmission bottlenecks.

[0058] S200: Select a preset number of edge nodes from the top of the distance list to obtain collaborative nodes. When a call request is received from the user terminal, find the edge node closest to the user terminal and define it as the target node. Calculate the load value of the target node and determine whether the load value is greater than a threshold. If so, trigger the pre-stored probe request in the target node and distribute it to the collaborative nodes in parallel.

[0059] From the top of the distance list, a predetermined number of edge nodes are selected and defined as collaborative nodes. This predetermined number is pre-determined by professionals. Upon receiving a call request from a user terminal (which is essentially a service request from the user terminal to the edge nodes, requesting the execution of a specific computational, processing, or data interaction task), the user's location information is traced using GPS, base station positioning, or Wi-Fi positioning to locate the nearest edge node, which is then defined as the target node. Several metrics reflecting the operational status of the edge nodes are selected, such as CPU utilization, memory usage, and bandwidth. For each metric, such as occupancy rate, storage space utilization rate, and current task queue length, a weighted value is assigned. A weighted calculation method is used to multiply the real-time value of each metric by its corresponding weight, and then the results are summarized to obtain the load value of each edge node. The load value of the target node is compared with a preset load threshold. If the load value is greater than the threshold, it indicates that the target node is at risk of overload, triggering a pre-stored probe request in the target node. The probe request usually contains information such as the current task status, resource requirements, and data scale. The probe request is mainly used to query the available resources and processing capabilities of other nodes. The probe request is distributed in parallel to the aforementioned determined collaborative nodes.

[0060] S300: Receive feedback information sent by the cooperating node, define the cooperating node corresponding to the first received feedback information as the optimal node, forward the call request to the optimal node through the target node, and establish a communication link between the optimal node and the user terminal.

[0061] The system receives feedback information from collaborating nodes, which typically includes the node's current available resource status, processing capacity, and estimated response time. The collaborating node that receives the feedback information first is considered to have the fastest response speed and the best network connectivity, and is defined as the optimal node. The target node acts as a relay, forwarding the call requests received by the target node to the optimal node, directly establishing a communication link between the optimal node and the user terminal, and directly uploading subsequent tasks to the optimal node for processing.

[0062] S400: Collects the unique identifier of the optimal node and caches it in the target node. When the target node receives the call request again, it traces back the source of the request and sends the unique identifier to the source of the request, thus establishing a data channel between the source of the request and the optimal node.

[0063] Collect the unique identifier of the optimal node. The unique identifier can be the node's IP address, node number, MAC address, or other marker information that can uniquely identify the optimal node. Cache the unique identifier of the optimal node in the target node. When the target node receives a call request from the same or other user terminals again, trace back to the user terminal corresponding to the call request and define it as the request source. Send the unique identifier of the optimal node to the request source, establish a data channel between the request source and the optimal node, and directly transfer the call request to the optimal node through this data channel.

[0064] In Example 2, Figure 2 The implementation flow of the edge service invocation method provided by an embodiment of the present invention is shown below. The steps of marking edge nodes are described in detail below:

[0065] S101: Select several metrics for the edge nodes and set their weights.

[0066] Several metrics for edge nodes are identified, such as CPU utilization, memory utilization, bandwidth utilization, storage space utilization, and current task queue length, and a corresponding weight is assigned to each metric.

[0067] S102: Collect real-time values ​​of indicators in edge nodes, and calculate the load value of each edge node based on the weights and real-time values.

[0068] The load value of each edge node is obtained by multiplying the real-time value of each indicator by its weight and then summing the calculation results of all indicators.

[0069] In Example 3, Figure 3 The implementation flow of the edge service invocation method provided by an embodiment of the present invention is illustrated. The steps for calculating the load value of the target node are described in detail below:

[0070] S201: Plot a load fluctuation trend graph with time on the horizontal axis and load value on the vertical axis.

[0071] Plot a load fluctuation trend chart with time on the horizontal axis and the load value at the corresponding time on the vertical axis; since each edge node corresponds to a load value, each edge node also corresponds to a load fluctuation trend chart.

[0072] S202: Collect attribute information of edge nodes, generate labels, and insert the labels into the load fluctuation trend graph, wherein the attribute information includes at least: deployment location and unique identifier.

[0073] Collect the attribute information of the edge nodes, including the deployment location of the edge nodes and the unique identifier of the edge nodes. Write the attribute information into the corresponding load fluctuation trend map in the form of tags. For example, write "Node A (Shanghai-ID:1003)" in the blank space of the load fluctuation trend map corresponding to edge node A, where "Shanghai" is the deployment location, which should be accurate to the specific latitude and longitude in actual operation, and ID is the unique identifier.

[0074] In Example 4, Figure 4 The implementation flow of the edge service invocation method provided by an embodiment of the present invention is illustrated. The following details the step of forwarding the invocation request to the optimal node via the target node:

[0075] S301: Find the two closest edge nodes, form a cooperative pair, and select a set value from the load values.

[0076] Identify the two closest edge nodes and form a cooperative pair; for example, if B is closest to A among all edge nodes, then use A and B to form a cooperative pair, and form cooperative pairs for all edge nodes in a certain order (such as numbering order).

[0077] S302: When the load value of the edge node is greater than the set value, the call request is reconnected using the mutual assistance pair.

[0078] Select a set value from the load values. When the load value of an edge node is greater than the set value, after receiving the call request, the edge node will transfer the call request to another edge node of the mutual assistance pair.

[0079] In Example 5, Figure 5 The implementation flow of the edge service invocation method provided by an embodiment of the present invention is illustrated below. The steps for establishing a data channel between the request source and the optimal node are described in detail below:

[0080] S401: Receive the processing task sent by the user terminal via the data channel and divide it into several task segments.

[0081] The processing task is analyzed to extract its internal structure, resource requirements and processing flow. Based on the task size and logical dependencies, the processing task is divided into several relatively independent and parallel-executable task segments. Each task segment can be processed through specific data blocks or computing units.

[0082] S402: Based on the edge node corresponding to the unique identifier, the task fragment is distributed to the mutual aid pair and the target node.

[0083] Task segments are distributed to the mutual aid pairs and edge nodes in the target node for parallel processing.

[0084] In Example 6, unlike Example 1, the method further includes:

[0085] Construct a time window and update the load value after the time window arrives;

[0086] The editor adjusts the strategy by using the load value to adjust the sorting in the distance list.

[0087] Select two adjacent time points and construct a time window. Within each time window, update the load value and adjust the sorting of the distance list using the load value. Assume that several time points are selected at 5-second intervals and time windows are constructed accordingly. In other words, update the load value every 5 seconds.

[0088] For example, if the distances from edge nodes B and C to A are roughly the same (within 5 meters), then the load values ​​of B and C are compared. If the load value of B is smaller, then B will be ranked higher. It should be noted that the larger the load value, the greater the load of the corresponding edge node, and the more likely data congestion will occur.

[0089] In Example 7, unlike Example 1, the method further includes:

[0090] Create a layered architecture consisting of a top layer, a parent layer, and child layers, and divide the access scope into several blocks;

[0091] Establish a correspondence between the hierarchical architecture and the blocks. Based on the correspondence, mount the edge nodes in the blocks to the hierarchical architecture to generate an edge service graph.

[0092] A hierarchical architecture similar to a binary tree in existing technologies is created, with the top layer as the top level, the parent layer as the middle layer, and the child layer as the bottom level. Furthermore, based on the attribute information and metrics of each edge node, it is divided into three levels according to performance: optimal, normal, and poor. Edge nodes corresponding to the optimal level are moved to the top layer, those corresponding to the normal level to the parent layer, and those corresponding to the poor level to the child layer, generating an edge service graph. This edge service graph can intuitively display the performance of each edge node. When a user terminal initiates a call request, it can query the edge service graph and prioritize uploading the call request to the edge node with the best performance. Based on the number of call requests, administrative divisions, etc., the access range is divided into multiple blocks, each block corresponding to an edge service graph.

[0093] Figure 6This diagram illustrates the structural block diagram of an edge service invocation system provided in an embodiment of the present invention. The edge service invocation system 1 includes:

[0094] Module 11 is used to divide the access range of edge services, mark edge nodes, calculate the distance between edge nodes, and sort the edge nodes in ascending order of the distance to obtain a distance list, wherein each edge node corresponds to a distance list;

[0095] Parallel module 12 is used to select a preset number of edge nodes from the top of the distance list to obtain collaborative nodes. When a call request uploaded by a user terminal is received, the edge node closest to the user terminal is found and defined as the target node. The load value of the target node is calculated and it is determined whether the load value is greater than a threshold. If so, the probe request pre-stored in the target node is triggered and distributed to the collaborative nodes in parallel.

[0096] The module 13 is used to receive feedback information sent by the collaborative node, define the collaborative node corresponding to the first received feedback information as the optimal node, and forward the call request to the optimal node through the target node, thereby establishing a communication link between the optimal node and the user terminal.

[0097] Module 14 is established to collect the unique identifier of the optimal node and cache it in the target node. When the target node receives the call request again, it traces back the source of the request and sends the unique identifier to the source of the request, thus establishing a data channel between the source of the request and the optimal node.

[0098] Figure 7 This diagram illustrates the structural composition of an edge service invocation system provided in an embodiment of the present invention. The obtaining module 11 includes:

[0099] Setting unit 111 is used to select several indicators of edge nodes and set their weights;

[0100] The calculation unit 112 is used to collect real-time values ​​of indicators in the edge nodes and calculate the load value of each edge node based on the weight and the real-time values.

[0101] Figure 8 This diagram illustrates the structural composition of an edge service invocation system provided in an embodiment of the present invention. The parallel module 12 includes:

[0102] Drawing unit 121 is used to draw a load fluctuation trend chart with time as the horizontal axis and load value as the vertical axis.

[0103] Insertion unit 122 is used to collect attribute information of edge nodes, generate tags, and insert the tags into the load fluctuation trend map, wherein the attribute information includes at least: deployment location and unique identifier.

[0104] Figure 9 This diagram illustrates the structural composition of an edge service invocation system provided in an embodiment of the present invention. The construction module 13 includes:

[0105] The selection unit 131 is used to find the two closest edge nodes, form a cooperative pair, and select a set value from the load values;

[0106] Access unit 132 is used to re-access the call request using the mutual aid pair when the load value of the edge node is greater than a set value.

[0107] Figure 10 This diagram illustrates the structural composition of an edge service invocation system provided in an embodiment of the present invention. The establishment module 14 includes:

[0108] The segmentation unit 141 is used to receive the processing task sent by the user terminal via the data channel and segment it into several task segments.

[0109] The distribution unit 142 is used to distribute the task fragment to the mutual aid pair and the target node according to the edge node corresponding to the unique identifier.

[0110] The module 11 is mainly used to complete step S100, the parallel module 12 is mainly used to complete step S200, the construction module 13 is mainly used to complete step S300, and the establishment module 14 is mainly used to complete step S400.

[0111] The setting unit 111 is mainly used to complete step S101, and the calculation unit 112 is mainly used to complete step S102.

[0112] The drawing unit 121 is mainly used to complete step S201, and the insertion unit 122 is mainly used to complete step S202;

[0113] The selection unit 131 is mainly used to complete step S301, and the access unit 132 is mainly used to complete step S302;

[0114] The segmentation unit 141 is mainly used to complete step S401, and the distribution unit 142 is mainly used to complete step S402.

[0115] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0116] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

[0117] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An edge service invocation method, characterized by, The method comprises: dividing the access range of the edge service, marking the edge nodes, and calculating the distances between the edge nodes, sorting the edge nodes in the order of the distances from small to large to obtain a distance list, wherein each edge node corresponds to a distance list; wherein the access range at least includes five edge nodes A, B, C, D and E, the distances from B, C, D and E to A are calculated with A as the center, and the sorted results are defined as the distance list, the distances from A, C, D and E to B are calculated with B as the center to obtain the distance list corresponding to B, and so on, each edge node corresponds to a distance list; from the front row of the distance list, a preset number of edge nodes are selected to obtain cooperative nodes, when receiving the calling request uploaded by the user terminal, the edge node closest to the user terminal is found out and defined as the target node, the load value of the target node is calculated, and it is judged whether the load value is greater than a threshold value, if yes, the pre-stored probe request in the target node is triggered and distributed to the cooperative nodes in parallel, wherein the probe request at least includes the current task state, resource demand and data size, and the probe request is used to query the available resources and processing capacity of other nodes; feedback information sent by the cooperative nodes is received, the cooperative node corresponding to the feedback information received first is defined as the optimal node, the calling request is forwarded to the optimal node through the target node, and a communication link between the optimal node and the user terminal is built; the unique identifier of the optimal node is collected and cached in the target node, when the target node receives the calling request again, the request source is traced back, the unique identifier is issued to the request source, and a data channel between the request source and the optimal node is established; the step of marking the edge nodes comprises: several indexes of the edge nodes are selected, and weights are set; the real-time values of the indexes in the edge nodes are collected, and the load values of each edge node are calculated based on the weights and the real-time values; the step of calculating the load value of the target node comprises: a load fluctuation trend chart is drawn with time as the horizontal coordinate and the load value as the vertical coordinate; attribute information of the edge nodes is collected, labels are generated, and the labels are inserted into the load fluctuation trend chart, wherein the attribute information at least includes deployment position and unique identifier; the step of forwarding the calling request to the optimal node through the target node comprises: the two edge nodes closest to each other are found out to form a mutual aid pair, and a set value is selected from the load values; when the load value of an edge node is greater than the set value, the mutual aid pair is used to re-access the calling request.

2. The method of claim 1, wherein, The method further comprises: a time window is constructed, and the load values are updated after the time window arrives; an adjustment strategy is edited, and the sorting in the distance list is adjusted by using the load values.

3. The method of claim 1, wherein the edge service invocation method is further configured to: The method further comprises: a hierarchical architecture composed of a top layer, a parent layer and a child layer is created, and the access range is divided into several blocks. A correspondence between the hierarchical architecture and the blocks is established, and based on the correspondence, edge nodes in the blocks are mounted into the hierarchical architecture to generate an edge service graph.

4. The method of claim 1, wherein, The step of establishing the data channel between the request source and the optimal node comprises: Via the data channel, a processing task sent by a user terminal is received and split into several task fragments; Based on the edge nodes corresponding to the unique identifiers, the task fragments are distributed into the mutual assistance pairs and the target node.

Images