Resource scheduling optimization method and device based on edge computing

Abstract

The invention provides a resource scheduling optimization method and device based on edge computing, and the method is applied to an edge node, and comprises the steps: determining a cache hit result of a first resource corresponding to request information sent by a user side, and generating a target instruction according to the cache hit result; sending a target instruction to the target device to send the first resource to the user side; wherein the target equipment comprises an edge node, a superior node of the edge node or a central cloud server; the caching mechanism of the edge node is determined based on the access frequency and heat information of different resources. According to the method and the device provided by the invention, the time consumption of cross-region transmission is reduced, so that the transmission delay is reduced; and meanwhile, the node load is dynamically adjusted by utilizing the access frequency and heat of the edge node storage information, so that the utilization efficiency of the limited edge storage space can be improved.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a resource scheduling optimization method and apparatus based on edge computing. Background Technology

[0002] With the large-scale commercialization of 5G (5th Generation Mobile Communication Technology) networks, emerging services such as high-definition video, cloud gaming, AR (Augmented Reality) / VR (Virtual Reality), and connected vehicles are growing rapidly, placing higher demands on network transmission latency and bandwidth.

[0003] In related technologies, the traditional centralized cloud model concentrates user requests to remote data centers for processing, which can easily cause network congestion, high latency, and a decline in user experience. To address this issue, existing technologies also introduce edge computing (MEC) into the network transmission process, which provides computing and storage capabilities at base stations or local data centers close to the user side to achieve localized distribution of some content. However, this edge distribution solution suffers from problems such as low cache hit rate and insufficient resource utilization. Summary of the Invention

[0004] This invention provides a resource scheduling optimization method and apparatus based on edge computing, which solves the defects of existing technologies that rely on base stations or local data centers near the user side to provide computing and storage capabilities to achieve localized distribution of some content, resulting in low cache hit rate and insufficient resource utilization.

[0005] This invention provides a resource scheduling optimization method based on edge computing, applied to edge nodes, comprising: Determine the cache hit result of the first resource corresponding to the request information sent by the user, and generate the target instruction based on the cache hit result; The target instruction is sent to the target device to send the first resource to the user terminal; wherein the target device includes the edge node, the parent node of the edge node, or the central cloud server; the caching mechanism of the edge node is determined based on the access frequency and popularity information of different resources.

[0006] According to the resource scheduling optimization method based on edge computing provided by the present invention, different resources correspond to different storage priorities; the method further includes: Obtain the access frequency of the second resource cached in the edge node; If the access frequency is not lower than the first frequency threshold, update the storage priority of the second resource; If the access frequency is lower than the second frequency threshold, the second resource is removed from the edge node.

[0007] According to a resource scheduling optimization method based on edge computing provided by the present invention, the upper-level node includes at least one, and the upper-level node is adjacent to the edge node; the method further includes: Obtain the resource load information of the edge nodes; If the resource load information is not lower than the first load threshold, the third resource cached by the edge node is sent to the target upper-level node for storage, wherein the target upper-level node is one of at least one upper-level node. If the resource load information is lower than the second load threshold, update the number of the parent node, or update the number of replicas in the parent node.

[0008] According to a resource scheduling optimization method based on edge computing provided by the present invention, the method further includes: Establish a cache index table based on the resources cached by the at least one parent node and the edge node respectively; Perform cross-node queries and resource retrieval tasks based on the cached index table.

[0009] According to the resource scheduling optimization method based on edge computing provided by the present invention, the resource scheduling priority of the edge node, the parent node of the edge node and the central cloud server decreases in that order.

[0010] According to a resource scheduling optimization method based on edge computing provided by the present invention, the user terminals include multiple terminals, and different user terminals correspond to different user priorities; the method further includes: Adjust the transmission bandwidth of the corresponding request information for each user terminal according to the user priority of multiple user terminals.

[0011] According to the resource scheduling optimization method based on edge computing provided by the present invention, the edge nodes and their corresponding upper-level nodes use a point-to-point communication protocol for data interaction.

[0012] This invention also provides a resource scheduling optimization device based on edge computing, applied to edge nodes, comprising: The request-response module is used to determine the cache hit result of the first resource corresponding to the request information sent by the user, and generate a target instruction based on the cache hit result. The scheduling module is used to send the target instruction to the target device to send the first resource to the user terminal; wherein the target device includes the edge node, the parent node of the edge node, or the central cloud server; the caching mechanism of the edge node is determined based on the access frequency and popularity information of different resources.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the resource scheduling optimization method based on edge computing as described above.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the resource scheduling optimization method based on edge computing as described above.

[0015] The resource scheduling optimization method and apparatus based on edge computing provided by this invention determines edge nodes through access frequency and popularity information of different resources, calculates the cache hit result of the first resource corresponding to the request information sent by the user terminal, generates a target instruction based on the cache hit result, and sends the target instruction to the target device to send the first resource to the user terminal. This reduces the time consumption of cross-regional transmission, thereby reducing transmission latency. At the same time, by dynamically adjusting the node load based on the access frequency and popularity of the information stored on the edge nodes, the utilization efficiency of limited edge storage space can be improved. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is one of the flowcharts of the resource scheduling optimization method based on edge computing provided by the present invention.

[0018] Figure 2 This is a schematic diagram of the dynamic cache adjustment and coordination mechanism provided by the present invention.

[0019] Figure 3 This is a schematic diagram of the system provided by the present invention.

[0020] Figure 4 This is the second flowchart of the resource scheduling optimization method based on edge computing provided by the present invention.

[0021] Figure 5This is a schematic diagram of the resource scheduling optimization device based on edge computing provided by the present invention.

[0022] Figure 6 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0024] The following is combined with Figures 1-5 This invention describes a resource scheduling optimization method and apparatus based on edge computing.

[0025] Figure 1 This is one of the flowcharts illustrating the resource scheduling optimization method based on edge computing provided by the present invention, such as... Figure 1 As shown, when applied to edge nodes, this method includes the following steps: Step 110: Determine the cache hit result of the first resource corresponding to the request information sent by the user, and generate the target instruction based on the cache hit result.

[0026] In this step, the first resource refers to the specific content request initiated by the user terminal through the 5G network, such as high-definition video clips, cloud gaming rendering frames, or AR / VR data packets.

[0027] In this step, the user terminal includes mobile phones, tablets, in-vehicle terminals, wearable devices, or IoT sensors.

[0028] In this step, edge nodes are deployed on MEC servers, carrier local data centers, or access routers near 5G base stations.

[0029] In this step, the target instruction is a decision logic signal generated based on the search results. For example, if a match is found, the instruction points to local reading; if a match is not found, the instruction points to initiating a back-to-origin request to the superior node or the central cloud.

[0030] In this embodiment, after receiving a request, the access edge node (such as an MEC server) determines whether the resource exists in the local physical storage by searching the local storage directory or a preset index table.

[0031] For example, in a 5G smart stadium scenario, a large number of user terminals simultaneously request to watch the live replay video of a certain game (i.e., the first resource). When the edge node receives a user's request information (such as a URL or content hash ID), it will immediately match it in the local "cache content list". If the match is successful (cache hit): the system determines that the first resource already exists locally, and then generates a "local retrieval instruction" to directly retrieve the video stream from the local disk; if the match fails (cache miss): the system determines that the first resource is missing. At this time, the system will further query the preset index table or generate a "cooperative retrieval instruction" or a "request instruction from the superior" according to the network topology. This instruction contains the target address of the request (such as the IP address of the nearest superior edge node or central cloud).

[0032] Step 120: Send a target instruction to the target device to send the first resource to the user terminal; wherein, the target device includes an edge node, the parent node of the edge node, or a central cloud server; the caching mechanism of the edge node is determined based on the access frequency and popularity information of different resources.

[0033] In this step, the target device is the node that actually provides the content. This device can be an edge node, a neighboring node of the edge node, or a central cloud server. If the content is matched locally on the edge node, the target device is the current edge node. If the content is not matched, the target device is the requested parent node or the central cloud.

[0034] In this embodiment, the identifier of the requested content (i.e., the first resource) can be matched with the edge node cache directory. If the match is successful, it is determined that the cache has been hit; otherwise, it is determined that the cache has not been hit.

[0035] In this embodiment, during or after resource transmission, edge nodes do not blindly cache all data. Instead, they dynamically decide whether to retain a copy of the resource and for how long based on a "time decay + heat prediction" strategy.

[0036] In this embodiment, an artificial intelligence prediction model can be deployed in the edge node to predict future access hotspots based on historical request data, that is, to obtain the popularity value of different resources, and thus cache the content that may be accessed in advance according to the popularity value of the resources.

[0037] For example, if the target instruction points to the local device, the target device is identified as an edge node. The video stream is sent to the user terminal via the 5G base station, and the "access popularity statistics" data of the video is updated in the background to increase its count and maintain its high priority in the cache replacement algorithm. If the target instruction points to the upper-level node or the central cloud server, the edge node continues to send requests to the upper-level node or the central cloud server. The target device responds to the request and sends the video stream back. When the video stream is sent to the user through the edge node, the edge node will analyze whether the content belongs to "high-frequency hot content" using an "artificial intelligence prediction model". The edge node will "mirrorize" and store the video stream in the local cache while forwarding it to the user and establish a relevant index. Conversely, if the resource is determined to be unpopular content, it will be forwarded without storage or only temporarily cached and then quickly removed.

[0038] The resource scheduling optimization method based on edge computing proposed in this embodiment can be applied to scenarios such as high-definition video playback, cloud gaming, augmented reality / virtual reality, vehicle networking, and smart city monitoring.

[0039] The resource scheduling optimization method based on edge computing provided in this invention determines edge nodes based on access frequency and popularity information of different resources, calculates the cache hit result of the first resource corresponding to the request information sent by the user terminal, generates a target instruction based on the cache hit result, and sends the target instruction to the target device to send the first resource to the user terminal. This reduces the time consumption of cross-regional transmission, thereby reducing transmission latency. At the same time, by dynamically adjusting the node load based on the access frequency and popularity of the information stored on the edge nodes, the utilization efficiency of limited edge storage space can be improved.

[0040] In some embodiments, different resources correspond to different storage priorities; the method further includes: obtaining the access frequency of a second resource cached in the edge node; updating the storage priority of the second resource if the access frequency is not lower than a first frequency threshold; and removing the second resource from the edge node if the access frequency is lower than a second frequency threshold.

[0041] In this embodiment, the first frequency threshold and the second frequency threshold can be set according to user needs. The first frequency threshold and the second frequency threshold can be the same or different.

[0042] In this embodiment, the second resource can be content data (such as video clips, game assets, etc.) that has already been stored in the local cache of the edge node.

[0043] In this embodiment, the system quantifies the level of willingness to retain cached content; the higher the priority, the less likely the content is to be overwritten or deleted.

[0044] In this embodiment, the access frequency edge node periodically counts the number of times the resource is requested per unit time.

[0045] In this embodiment, content accessed more frequently than a threshold is cached first, while content accessed less frequently is gradually evicted.

[0046] Specifically, the first frequency threshold is used as the criterion for judging popular content, and the second frequency threshold is used as the criterion for eliminating unpopular content. This is used in the specific execution process of the "time decay + popularity prediction" strategy. It can solve the problem of limited storage resources at edge nodes and ensure that limited storage space always serves the most valuable (hottest) content through dynamic selection and elimination.

[0047] In a feasible implementation, for example in a 5G vehicle-to-everything (V2X) scenario, the edge node (MEC server) caches a large number of high-precision map data packets and real-time traffic video (secondary resource). The system sets a first frequency threshold of "1000 times / hour" and a second frequency threshold of "10 times / hour". For high-frequency content: when the "real-time accident warning video" of a certain road segment is frequently requested by passing vehicles and the access frequency reaches 1200 times / hour (not lower than the first frequency threshold), the system determines that it is high-popularity content and immediately updates its storage priority to "high" or "locked state".

[0048] For low-frequency content: For another piece of "traffic history records from last month", due to the expiration of timeliness, the access frequency drops to 5 times / hour (below the second frequency threshold). After the system identifies this, it lowers its storage priority to "lowest" or directly marks it as "to be released", and immediately removes the resource from the edge node, thereby freeing up storage space to store new hot data.

[0049] The resource scheduling optimization method based on edge computing provided in this invention obtains the access frequency of a second resource cached in an edge node; updates the storage priority of the second resource when the access frequency is not lower than a first frequency threshold; and removes the second resource from the edge node when the access frequency is lower than a second frequency threshold. This prevents edge nodes from being filled with invalid or outdated data. By utilizing a dynamic adjustment mechanism of "retaining high-frequency content for a long time and gradually replacing low-frequency content," the cache hit rate of edge nodes is further improved, the utilization efficiency of storage resources is increased, and users can obtain millisecond-level low-latency responses when requesting hot content.

[0050] In some embodiments, the parent node includes at least one, and the parent node is adjacent to the edge node; the method further includes: obtaining resource load information of the edge node; if the resource load information is not lower than a first load threshold, sending the third resource cached by the edge node to a target parent node for storage, wherein the target parent node is one of at least one parent node; if the resource load information is lower than a second load threshold, updating the number of parent nodes, or updating the number of replicas in the parent node.

[0051] In this embodiment, the first load threshold and the second load threshold can be set according to user needs. The first load threshold and the second load threshold can be the same or different.

[0052] In this embodiment, the upper-level node can be the aggregation layer edge node or regional center located one level above the current access edge node in the network topology and which is physically or logically close.

[0053] In this embodiment, the third resource refers to content data that is currently stored locally on the edge node, but may need to be migrated or backed up in order to balance the load.

[0054] In this embodiment, resource load information may include the storage space utilization rate of the edge node, CPU computing load, or I / O throughput.

[0055] This embodiment uses a first load threshold as a high load warning line and a second load threshold as a low load idle line to establish a bidirectional elastic scaling mechanism based on load pressure. This mechanism dynamically adjusts the cache location and number of copies of content according to the node load, thereby preventing edge nodes from crashing due to overload or wasting resources due to idleness.

[0056] In a feasible implementation, such as in an industrial park scenario where 5G augmented reality (AR) assisted maintenance is used, the edge node on the base station side is responsible for caching high-definition 3D device models (third resources). For high-load scenarios, when a large number of devices in the park are updated simultaneously, causing the storage occupancy rate of the edge node to reach 85% (not lower than the first load threshold), the temporarily less accessed third resources (such as 3D models of older models of devices) in the local cache are migrated and sent to the target upper-level node (such as the park's aggregation server) for storage, and the local space is released to accommodate new popular data. This mechanism ensures that the local node will not stop service due to "warehouse overload", and the data only moves up one layer and does not flow back to the distant central cloud.

[0057] In addition, for low-load scenarios (resource contraction), when the edge node load drops to 15% (below the second load threshold) during nighttime shutdown, it indicates that the system is idle. At this time, in order to save upper-layer resources, the system will instruct the upper-level node to update the number of replicas in the upper-level node (for example, deleting temporary redundant replicas created by the upper-level node to offload the load), or update the number of upper-level nodes in the software-defined networking (SDN) architecture (disconnecting unnecessary upper-level backup node connections), thereby reducing the operation and maintenance costs and energy consumption of the entire edge computing network.

[0058] The resource scheduling optimization method based on edge computing provided in this invention obtains the resource load information of edge nodes; when the resource load information is not lower than a first load threshold, the third resource cached by the edge node is sent to the target upper-level node for storage; when the resource load information is lower than a second load threshold, the number of upper-level nodes or the number of replicas in the upper-level nodes is updated. This avoids network congestion caused by single-point overload and also ensures that even when local space is insufficient, user requests can still be quickly responded to by the nearest upper-level node through inter-level collaborative storage, maintaining a low-latency user experience.

[0059] In some embodiments, the method further includes: establishing a cache index table based on the resources cached by at least one parent node and the edge node; and performing cross-node query and resource acquisition tasks based on the cache index table.

[0060] In this embodiment, the cache index table is a logical mapping table used to record the distribution of storage resources of different nodes in the network (including the current edge node and its corresponding parent node); wherein, the cache index table includes fields such as resource ID, storage node location (IP address), resource popularity and replica status.

[0061] Specifically, the system collects and aggregates cache directory information of edge nodes and their parent nodes in real time or periodically through a cache index table. When the local cache is not hit, the system no longer blindly broadcasts the request upwards level by level. Instead, it first queries the index table to accurately locate the specific node where the target resource is located (which may be an adjacent edge node or a specific parent node), and initiates a data retrieval task in a targeted manner, thus realizing the collaboration and information sharing of different edge nodes.

[0062] In some embodiments, edge nodes and their corresponding parent nodes use a point-to-point communication protocol for data interaction.

[0063] In this embodiment, a point-to-point communication protocol (P2P Protocol) is established between the edge node and the upper-level node, realizing a lightweight data transmission standard.

[0064] Specifically, the cache index table is a distributed cache index table. In this embodiment, by introducing a distributed cache index table, each edge node shares cache directory information through a lightweight communication protocol. When a node receives a missed request, it can quickly determine whether the target content is stored on adjacent nodes, thereby reducing the proportion of cross-level requests.

[0065] This implementation method can maintain a high cache hit rate and low access latency even under high user traffic and high concurrency. Test results show that the method of this invention can improve the cache hit rate by more than 20% and reduce the average transmission latency by more than 30%, making it particularly suitable for scenarios with stringent network performance requirements, such as high-definition video playback, cloud gaming, AR / VR, and vehicle networking. In a feasible implementation, such as in a real-time cloud gaming rendering scenario, the user terminal connects to "edge node A". To improve response speed, the system constructs a joint cache index table covering "edge node A" and its directly connected "upper-level aggregation node B". Simultaneously, edge node A and upper-level node B synchronize their respective cache list summaries every few seconds and update the index table. The table records that game asset package X is stored on node A (local), while rare high-texture background package Y is stored on node B (upper-level). When the user enters a new scene in the game and needs to load background package Y, edge node A first searches locally. If it finds that it is missing, it queries the cache index table and finds that background package Y is located on upper-level node B. Then, edge node A directly generates a targeted retrieval command and sends it to node B to quickly retrieve background package Y and forward it to the user, without having to query other unrelated nodes or wait for scheduling from the central cloud.

[0066] Figure 2 This is a schematic diagram of the dynamic cache adjustment and coordination mechanism provided by the present invention. Figure 2 In the illustrated embodiment, the edge nodes include multiple nodes (corresponding to three edge nodes A), which record the cached content list, access popularity, and node complexity status of each edge node, and construct a shared cache index table through a lightweight communication protocol. The cached content list is a distributed cache index table that includes elements such as content ID, storage node location, access popularity, and number of replicas. The system adjusts the load and resource priority of each edge node according to a dynamic adjustment strategy. This dynamic adjustment strategy is formulated based on mechanisms such as popularity prediction, load balancing, and replica management.

[0067] The resource scheduling optimization method based on edge computing provided in this invention establishes a cache index table by using the resources cached by at least one upper-level node and the edge node respectively; and performs cross-node query and resource acquisition tasks according to the cache index table, which can obtain content with the shortest logical path and avoid invalid signaling interaction and redundant data transmission.

[0068] In some embodiments, the resource scheduling priority of the edge node, the parent node of the edge node, and the central cloud server decreases in that order.

[0069] In this embodiment, when a user initiates a content request, the system selects the decision order rules for the response nodes. For example, the system executes a tiered query strategy: first, it tries to call the resources in the edge node (highest priority) that is physically closest; second, it tries the resources in the next closest parent node (medium priority); and finally, it considers the resources in the remote central cloud server (lowest priority).

[0070] For example, in 5G virtual reality (VR) panoramic live streaming scenarios, bandwidth and latency are extremely sensitive; when a user wearing a VR headset requests a panoramic video stream, the scheduling system strictly executes requests according to priority: The system first initiates scheduling to the MEC server (i.e., edge node) deployed on the 5G base station side. Due to the closest proximity, if there is caching at this location, the response latency is only a few milliseconds, perfectly ensuring a VR experience without dizziness. Only when the edge node on the base station side reports a "miss" will the system activate a secondary strategy to request resources from the upper-level node of the metropolitan area network aggregation layer. Although the latency is slightly increased, it is still better than going back to the origin server. Finally, only when the first two levels cannot provide resources (such as unpopular old videos) will the system send the request to the central cloud server thousands of kilometers away, and the central cloud server will return the requested resources to the user.

[0071] Figure 3 This is a schematic diagram of the system provided by the present invention. Figure 3 In the illustrated embodiment, the system includes a user terminal (corresponding to the user end), edge nodes, upper-level edge nodes (corresponding to the upper-level nodes of the edge nodes), and a central cloud (central cloud server). The user terminal sends a content request stream to the target device via 5G network access. If the request is found in the local cache of an edge node, the edge node queries and responds to the user terminal with the requested content. If the request is not found in the local cache but is found in the upper-level edge node, the upper-level edge node performs a cross-node query and resource retrieval using a collaborative cache index and then sends the requested content back to the user terminal. If the request is not found in the upper-level edge node, the resource is retrieved from the global content library via the central cloud and sent to the user terminal.

[0072] The resource scheduling optimization method based on edge computing provided in this embodiment of the invention compresses the data transmission path to the maximum extent by setting the resource scheduling priority of edge nodes, the parent nodes of edge nodes and the central cloud server in descending order, ensuring that most high-frequency traffic is absorbed by edge nodes, greatly reducing the average transmission latency, and thus improving the response speed of the entire distribution system.

[0073] In some embodiments, the user terminal includes multiple user terminals, and different user terminals correspond to different user priorities; the method further includes: adjusting the transmission bandwidth of the request information corresponding to each user terminal according to the user priorities corresponding to the multiple user terminals.

[0074] In this embodiment, user priority can be a weighted identifier determined based on the user's subscription level and service type.

[0075] In this embodiment, when processing concurrent requests, the edge node uses network slicing or QoS (Quality of Service) technology to dynamically allocate downlink data throughput capacity to ensure that high-priority user requests are responded to first, which can solve the problem of fairness and efficiency in resource allocation in network congestion or bandwidth-limited scenarios.

[0076] In a feasible implementation, such as in a public Wi-Fi hotspot coverage area of ​​a 5G smart city, an edge node simultaneously receives content requests from multiple user terminals, at which point bandwidth resources tend to be saturated. The system detects that user A is a doctor's terminal performing a remote medical consultation (high priority, extremely sensitive business), while user B is an ordinary entertainment user downloading a high-definition movie (low priority). The system performs bandwidth scheduling according to the priority strategy, that is, it limits user B's download requests, compresses the bandwidth allocated to it to the basic level, and ensures that it does not occupy too many channels. At the same time, the freed-up bandwidth resources are allocated to user A, giving it a dedicated high-speed channel to ensure that the real-time transmission of medical images is smooth and without delay.

[0077] The edge computing-based resource scheduling optimization method provided in this invention adjusts the transmission bandwidth of request information corresponding to each user terminal by adjusting the user priority of multiple user terminals, ensuring that the experience quality of critical services and core users is not disturbed, and enhancing the system's intelligent scheduling capability and service quality assurance capability under resource-constrained conditions.

[0078] Figure 4 This is the second flowchart of the resource scheduling optimization method based on edge computing provided by this invention. Figure 4 In the illustrated embodiment, under a 5G network environment, a user terminal (such as a mobile phone, tablet, or in-vehicle device) initiates a video content request (i.e., the request information sent by the user terminal) to the nearest edge node via a wireless access network. After receiving the request, the edge node first searches for the target content (i.e., the request information) in its local cache directory. If the local cache is hit, the result is directly returned to the user terminal, with the entire process having a delay of only a few milliseconds. If the cache is not hit, a query is initiated to the upper-level edge node. If the target content exists on the upper-level node, it is returned to the requester and cached locally. If the upper-level node still does not hit the cache, the data is requested from the central cloud and transmitted back to the user after transmission is completed, while also being stored in the edge node.

[0079] The resource scheduling optimization device based on edge computing provided by the present invention will be described below. The resource scheduling optimization device based on edge computing described below and the resource scheduling optimization method based on edge computing described above can be referred to in correspondence.

[0080] Figure 5 This is a schematic diagram of the resource scheduling optimization device based on edge computing provided by the present invention, as shown below. Figure 5 As shown, the resource scheduling optimization device based on edge computing is applied to edge nodes and includes a request response module 510 and a scheduling module 520.

[0081] The request-response module 510 is used to determine the cache hit result of the first resource corresponding to the request information sent by the user, and generate the target instruction based on the cache hit result. The scheduling module 520 is used to send target instructions to the target device to send the first resource to the user terminal; wherein, the target device includes an edge node, the parent node of the edge node, or a central cloud server; the caching mechanism of the edge node is determined based on the access frequency and popularity information of different resources.

[0082] The edge computing-based resource scheduling optimization device provided in this embodiment of the invention determines edge nodes based on access frequency and popularity information of different resources, calculates the cache hit result of the first resource corresponding to the request information sent by the user terminal, generates a target instruction based on the cache hit result, and sends the target instruction to the target device to send the first resource to the user terminal. This reduces the time consumption of cross-regional transmission, thereby reducing transmission latency. At the same time, by dynamically adjusting the node load based on the access frequency and popularity of the information stored on the edge nodes, the utilization efficiency of limited edge storage space can be improved.

[0083] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6As shown, the electronic device may include a processor 610, a communications interface 620, a memory 630, and a communication bus 640. The processor 610, communications interface 620, and memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute a resource scheduling optimization method based on edge computing. This method is applied to edge nodes and includes: determining the cache hit result of a first resource corresponding to a request sent by a user terminal, and generating a target instruction based on the cache hit result; sending the target instruction to a target device to send the first resource to the user terminal; wherein the target device includes an edge node, a parent node of the edge node, or a central cloud server; the caching mechanism of the edge node is determined based on the access frequency and popularity information of different resources.

[0084] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0085] On the other hand, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program is implemented to perform the resource scheduling optimization method based on edge computing provided by the above methods. The method is applied to an edge node and includes: determining the cache hit result of a first resource corresponding to a request information sent by a user terminal, and generating a target instruction based on the cache hit result; sending the target instruction to a target device to send the first resource to the user terminal; wherein the target device includes an edge node, a parent node of the edge node, or a central cloud server; the caching mechanism of the edge node is determined based on the access frequency and popularity information of different resources.

[0086] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0087] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A resource scheduling optimization method based on edge computing, applied to edge nodes, characterized in that, include: Determine the cache hit result of the first resource corresponding to the request information sent by the user, and generate the target instruction based on the cache hit result; The target instruction is sent to the target device to send the first resource to the user terminal; wherein the target device includes the edge node, the parent node of the edge node, or the central cloud server; the caching mechanism of the edge node is determined based on the access frequency and popularity information of different resources.

2. The resource scheduling optimization method based on edge computing according to claim 1, characterized in that, Different resources correspond to different storage priorities; the method further includes: Obtain the access frequency of the second resource cached in the edge node; If the access frequency is not lower than the first frequency threshold, update the storage priority of the second resource; If the access frequency is lower than the second frequency threshold, the second resource is removed from the edge node.

3. The resource scheduling optimization method based on edge computing according to claim 1, characterized in that, The parent node includes at least one, and the parent node is adjacent to the edge node; the method further includes: Obtain the resource load information of the edge nodes; If the resource load information is not lower than the first load threshold, the third resource cached by the edge node is sent to the target upper-level node for storage, wherein the target upper-level node is one of at least one upper-level node. If the resource load information is lower than the second load threshold, update the number of the parent node, or update the number of replicas in the parent node.

4. The resource scheduling optimization method based on edge computing according to claim 3, characterized in that, The method further includes: Establish a cache index table based on the resources cached by the at least one parent node and the edge node respectively; Perform cross-node queries and resource retrieval tasks based on the cached index table.

5. The resource scheduling optimization method based on edge computing according to claim 1, characterized in that, The resource scheduling priority of the edge node, the parent node of the edge node, and the central cloud server decreases in that order.

6. The resource scheduling optimization method based on edge computing according to claim 1, characterized in that, The user terminals include multiple terminals, each corresponding to a different user priority; the method further includes: Adjust the transmission bandwidth of the corresponding request information for each user terminal according to the user priority of multiple user terminals.

7. The resource scheduling optimization method based on edge computing according to any one of claims 1-6, characterized in that, The edge nodes and their corresponding parent nodes use a point-to-point communication protocol for data exchange.

8. A resource scheduling optimization device based on edge computing, applied to edge nodes, characterized in that, include: The request-response module is used to determine the cache hit result of the first resource corresponding to the request information sent by the user, and generate a target instruction based on the cache hit result. The scheduling module is used to send the target instruction to the target device to send the first resource to the user terminal; wherein the target device includes the edge node, the parent node of the edge node, or the central cloud server; the caching mechanism of the edge node is determined based on the access frequency and popularity information of different resources.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the resource scheduling optimization method based on edge computing as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the resource scheduling optimization method based on edge computing as described in any one of claims 1 to 7.

Images