An intelligent server management method and system in an edge computing environment

Abstract

An intelligent server management method and system in an edge computing environment, involving the field of electronic digital data processing, in which a dynamic load buffer database is established; the load stability coefficient of each server node is calculated; when the load stability coefficient of the server node is less than a first preset threshold, the server node is marked as a fluctuating node; the node collaborative efficiency value is calculated; the target server candidate set is determined, and the task allocation weight is calculated; the tasks on the fluctuating node are allocated to the server nodes in the target server candidate set; when the real-time load value of the fluctuating node is lower than a third preset threshold for a continuous preset time, the processing efficiency improvement ratio is calculated; for tasks with a processing efficiency improvement ratio greater than a fourth preset threshold, the processing is continued at the current server node; for tasks with a processing efficiency improvement ratio less than or equal to the fourth preset threshold, the processing is moved back to the original fluctuating node. This application is used to reduce communication overhead and business processing delays between nodes.

Description

Technical Field

[0001] The present application belongs to the field of electronic digital data processing, and in particular, relates to an intelligent server management method and system in an edge computing environment. Background Art

[0002] With the rapid development of 5G networks and IoT technologies, edge computing has become an important technical means to solve network latency and bandwidth pressure in cloud computing centers. In an edge computing environment, a large number of server nodes are distributed at the edge of the network to provide local computing and storage services to end users. However, due to the large number of edge servers and their scattered distribution, traditional centralized server management methods are difficult to respond to dynamically changing business needs in a timely manner, resulting in low server resource utilization and easy to cause local server overload when facing sudden business loads.

[0003] In related technologies, the CPU usage, memory usage, network bandwidth and other indicators of each server node can be monitored in real time. When a server is detected to be overloaded, new business requests can be dynamically allocated to server nodes with lower loads. Task migration can be triggered by setting load thresholds, thereby balancing the resource utilization of each server node.

[0004] However, geographically close server nodes often carry related businesses. When a server is overloaded, simply migrating tasks to a less loaded server may destroy this business relevance, causing the business that should have been processed collaboratively on geographically close servers to be dispersed to more distant nodes, increasing the communication overhead between nodes and business processing delays. Summary of the invention

[0005] The present application provides an intelligent server management method and system in an edge computing environment, which are used to reduce communication overhead and business processing delay between nodes.

[0006] In a first aspect, the present application provides an intelligent server management method in an edge computing environment, wherein a dynamic load buffer database is established in a preset central controller, and the dynamic load buffer database is used to store real-time load data of each server node within a preset time period;

[0007] Calculate the load stability coefficient of each server node based on real-time load data. The load stability coefficient is a dimensionless parameter obtained by weighting the standard deviation of processor utilization, memory occupancy, and network bandwidth utilization within a preset time period.

[0008] When the load stability coefficient of the server node is less than a first preset threshold, marking the server node as a fluctuating node;

[0009] Collect network quality parameters between server nodes, and calculate node coordination efficiency values ​​based on network quality parameters and geographical distances between server nodes;

[0010] When the real-time load value of the fluctuating node exceeds the second preset threshold, a target server candidate set is determined based on the node collaborative efficiency value, and a task allocation weight is calculated based on the load stability coefficient of each server node in the target server candidate set;

[0011] Assign tasks on the fluctuation nodes to server nodes in the target server candidate set according to the task allocation weight, and continuously collect real-time load data of the fluctuation nodes;

[0012] When the real-time load value of the fluctuation node is lower than the third preset threshold value for a continuous preset period of time, calculating the processing efficiency improvement ratio of the assigned task on the target server candidate set;

[0013] For tasks where the processing efficiency improvement ratio is greater than a fourth preset threshold, the tasks are kept on the current server node for further processing;

[0014] For tasks whose processing efficiency improvement ratio is less than or equal to the fourth preset threshold, they are migrated back to the original fluctuation node.

[0015] By adopting the above technical solution, a dynamic load buffer database is established in the central controller, and the load stability coefficient is calculated based on the processor utilization rate, memory occupancy rate and network bandwidth utilization rate. The system can accurately identify server nodes with large load fluctuations. The node collaborative efficiency value calculated by combining the network quality parameters and geographical distance between nodes can select the most suitable target server when migrating tasks. The task allocation weight based on the load stability coefficient is adopted so that tasks can be assigned to server nodes with more stable loads and more sufficient resources. By continuously monitoring the load of fluctuating nodes and deciding whether tasks need to be migrated back based on the processing efficiency improvement ratio, it not only ensures the improvement of task processing efficiency, but also avoids unnecessary task migration, so that the system can achieve efficient utilization of server resources while ensuring service quality, and reduce the risk of unbalanced server load in edge computing environments.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, calculating the node coordination efficiency value according to the network quality parameter and the geographical distance between each server node specifically includes:

[0017] Obtain the geographic coordinate information of each server node, and calculate the straight-line distance between nodes based on the geographic coordinate information;

[0018] Collect link delay, packet loss rate, and link available bandwidth between server nodes;

[0019] The node collaboration efficiency value is obtained by normalizing the straight-line distance between nodes, link delay, packet loss rate, and link available bandwidth according to the preset weighting coefficient and then weighting them.

[0020] By adopting the above technical solution, the geographic coordinate information of the server nodes is obtained and the straight-line distance between the nodes is calculated, combined with network quality parameters such as link delay, packet loss rate and link available bandwidth. These parameters are normalized and weighted using preset weighting coefficients. The obtained node collaborative efficiency value not only takes into account the transmission delay caused by physical distance, but also includes the actual impact of network conditions on data transmission. This multi-parameter weighted evaluation method enables the system to more accurately judge the collaborative efficiency between different server nodes. The node collaborative efficiency value calculated in this way can truly reflect the collaborative processing capabilities between nodes and improve the success rate and efficiency of task migration.

[0021] In conjunction with some embodiments of the first aspect, in some embodiments, the task allocation weight is calculated according to the load stability coefficient of each server node in the target server candidate set, specifically including:

[0022] Obtain the current load value and remaining computing resources of each server node in the target server candidate set;

[0023] Calculate the resource reliability of each server node based on the load stability coefficient and the remaining computing resources;

[0024] The resource reliability of each server node is normalized to obtain the corresponding task allocation weight.

[0025] By adopting the above technical solution and taking into account the load stability coefficient of the server node and the remaining computing resources to calculate the resource reliability, the system has established an evaluation index that comprehensively reflects the server's processing capacity and stability. The method of normalizing the resource reliability to obtain the task allocation weight enables the system to reasonably allocate the task load according to the actual processing capacity of the server, making the task allocation more reasonable, fully utilizing the server's computing resources, and ensuring the stability of task processing. The task allocation weight calculated in this way can guide the system to perform more reasonable load balancing and improve the overall resource utilization efficiency.

[0026] In conjunction with some embodiments of the first aspect, in some embodiments, calculating the processing efficiency improvement ratio of the assigned task on the target server candidate set specifically includes:

[0027] Record the average response time and resource consumption when the fluctuation node processes each task;

[0028] Collect the average response time and resource consumption of each task when running on the candidate set of target servers;

[0029] The processing efficiency improvement ratio is calculated based on the average response time ratio and resource consumption ratio of each task before and after migration.

[0030] By adopting the above technical solution, the system can accurately evaluate the actual benefits brought by task migration by calculating the processing efficiency improvement ratio based on the average response time ratio and resource consumption ratio. This efficiency evaluation method based on actual operation data reduces the deviation that may be caused by subjective judgment, allowing the system to make more accurate task migration decisions. The processing efficiency improvement ratio calculated in this way provides a reliable basis for judging whether the task needs to be migrated back to the original server, reducing unnecessary task migration and improving the overall operation efficiency of the system.

[0031] In combination with some embodiments of the first aspect, in some embodiments, after migrating back the tasks whose processing efficiency improvement ratio is less than or equal to a fourth preset threshold to the original fluctuation node, the method further includes:

[0032] Counting the load fluctuation data of the fluctuation node within a preset time period, and generating a load fluctuation time series sequence according to the load fluctuation data;

[0033] Calculate the load change rate and acceleration of the fluctuation node according to the load fluctuation time series;

[0034] When the load change rate exceeds a fifth preset threshold and the load change acceleration is a positive value, extracting task feature data and determining a task priority level;

[0035] Based on the task priority, a target node that meets the preset task resource requirements is selected from the target server candidate set, and a resource reservation channel is established based on the target node.

[0036] By adopting the above technical solution, the load fluctuation data of the fluctuating nodes within the preset time period is counted and a time series is generated. Combined with the calculation of the load change rate and acceleration, the system can accurately grasp the trend and intensity of the load change. When the load change rate exceeds the threshold and the acceleration is positive, it indicates that the load pressure of the fluctuating node is growing rapidly. At this time, by extracting task feature data and determining the task priority level, the system can identify critical tasks with high service quality requirements. Based on the task priority level, a suitable target node is selected from the target server candidate set and a resource reservation channel is established. This preventive resource scheduling mechanism can prepare the required computing resources for important tasks in advance before the load surge causes a decline in service quality, avoiding serious service quality degradation problems during peak load periods. At the same time, it also reduces the system overhead caused by emergency task migration and improves service reliability in edge computing environments.

[0037] In conjunction with some embodiments of the first aspect, in some embodiments, determining the task priority level specifically includes:

[0038] The response time constraints, computing resource consumption, and data interaction frequency of the collection task;

[0039] Perform dependency analysis on tasks and mark tasks that have dependencies on previous tasks as critical path tasks;

[0040] The tasks whose response time constraints are less than the sixth preset threshold in the critical path tasks are marked as the highest priority, and the tasks whose response time constraints are between the sixth preset threshold and the seventh preset threshold are marked as high priority.

[0041] By adopting the above technical solution, tasks with dependencies on previous tasks are marked as critical path tasks, and further subdivided into top priority and high priority according to the urgency of response time constraints. This multi-level priority division method based on task characteristics enables the system to identify tasks that have the greatest impact on the overall service quality. Through the priority mechanism, the system can ensure the service quality of critical tasks when resources are limited, especially marking tasks with the strictest response time constraints as the highest priority, ensuring that these tasks that are most sensitive to latency can obtain sufficient computing resources, thereby improving the service quality assurance capabilities of the edge computing system.

[0042] In combination with some embodiments of the first aspect, in some embodiments, after establishing a resource reservation channel based on the target node, the method further includes:

[0043] Monitor resource occupancy and task processing delay of resource reservation channels;

[0044] Calculate channel utilization efficiency based on resource occupancy rate;

[0045] When the channel utilization efficiency is lower than an eighth preset threshold, releasing a preset number of reserved resources;

[0046] When the task processing delay exceeds the preset delay threshold, the quota of reserved resources is expanded.

[0047] By adopting the above technical solution and continuously monitoring the resource occupancy rate and task processing delay of the resource reservation channel, the system can evaluate the use effect of the reserved resources in real time. The channel utilization efficiency is calculated based on the resource occupancy rate. When the efficiency is lower than the threshold, the preset number of reserved resources is released in time, avoiding excessive reservation and waste of resources. At the same time, when the task processing delay exceeds the preset threshold, the system will actively expand the quota of reserved resources, so that the scale of reserved resources can change adaptively according to the task processing requirements, achieving a balance between the economy and service quality of reserved resources.

[0048] In second aspect, an embodiment of the present application provides an intelligent server management system in an edge computing environment, wherein the intelligent server management system in the edge computing environment comprises: one or more processors and a memory; the memory is coupled to one or more processors, the memory is used to store computer program code, the computer program code comprises computer instructions, and one or more processors call computer instructions to enable the system to execute the method described in the first aspect and any possible implementation method of the first aspect.

[0049] In a third aspect, an embodiment of the present application provides a computer-readable storage medium, comprising instructions, which, when executed on a system, causes the system to execute the method described in the first aspect and any possible implementation of the first aspect.

[0050] In a fourth aspect, an embodiment of the present application provides a computer program product. When the computer program product runs on a system, the system executes the method described in any possible implementation manner in the first aspect.

[0051] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0052] 1. The present application provides an intelligent server management method in an edge computing environment. A dynamic load buffer database is established in the central controller, and the load stability coefficient is calculated based on the processor utilization rate, memory occupancy rate and network bandwidth utilization rate. The system can accurately identify server nodes with large load fluctuations. The node collaborative efficiency value calculated by combining the network quality parameters and geographical distance between nodes can select the most suitable target server when migrating tasks. The task allocation weight based on the load stability coefficient is adopted so that tasks can be assigned to server nodes with more stable loads and more sufficient resources. By continuously monitoring the load of fluctuating nodes and determining whether the task needs to be migrated back based on the processing efficiency improvement ratio, it not only ensures the improvement of task processing efficiency, but also avoids unnecessary task migration, so that the system can achieve efficient utilization of server resources while ensuring service quality, and reduce the risk of unbalanced server load in the edge computing environment.

[0053] 2. The present application provides an intelligent server management method in an edge computing environment, which counts the load fluctuation data of the fluctuating nodes within a preset time period and generates a time series. Combined with the calculation of the load change rate and acceleration, the system can accurately grasp the trend and intensity of the load change. When the load change rate exceeds the threshold and the acceleration is a positive value, it indicates that the load pressure of the fluctuating node is growing rapidly. At this time, by extracting task feature data and determining the task priority level, the system can identify critical tasks with high service quality requirements. Based on the task priority level, a suitable target node is selected from the target server candidate set and a resource reservation channel is established. This preventive resource scheduling mechanism can prepare the required computing resources for important tasks in advance before the load surge causes a decline in service quality, avoiding the problem of serious service quality degradation during the peak load period. At the same time, it also reduces the system overhead caused by the migration of emergency tasks and improves the service reliability in the edge computing environment.

[0054] 3. This application provides an intelligent server management method in an edge computing environment, which continuously monitors the resource occupancy rate and task processing delay of the resource reservation channel, and the system can evaluate the use effect of the reserved resources in real time. The channel utilization efficiency is calculated according to the resource occupancy rate, and when the efficiency is lower than the threshold, the preset number of reserved resources is released in time to avoid excessive reservation and waste of resources. At the same time, when the task processing delay exceeds the preset threshold, the system will actively expand the quota of reserved resources, so that the scale of reserved resources can change adaptively according to the task processing requirements, achieving a balance between the economy and service quality of reserved resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a flow chart of an intelligent server management method in an edge computing environment in an embodiment of the present application.

[0056] Figure 2 It is a flow chart of a resource reservation method based on load trend analysis in an embodiment of the present application.

[0057] Figure 3 It is a schematic diagram of the physical device structure of an intelligent server management system in an edge computing environment provided in an embodiment of the present application. DETAILED DESCRIPTION

[0058] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification and appended claims of the present application, the singular expressions "one", "a kind of", "said", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations comprising one or more listed items.

[0059] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0060] The following uses an embodiment and combines Figure 1 , describes an intelligent server management method in an edge computing environment in an embodiment of the present application:

[0061] See also Figure 1 , which is a flow chart of an intelligent server management method in an edge computing environment in an embodiment of the present application.

[0062] S101, establishing a dynamic load buffer database in a preset central controller;

[0063] The system establishes a dynamic load buffer database in a preset central controller, and the dynamic load buffer database is used to store the real-time load data of each server node within a preset time period.

[0064] The system establishes a dynamic load buffer database in the preset central controller to store the real-time load data of each server node within a preset time period. The dynamic load buffer database can reflect the load status of each server node in the edge computing environment in real time, providing data support for subsequent intelligent management decisions. In addition to establishing an independent dynamic load buffer database, the system can also store load data in the master database of the central controller, or use distributed database technology to disperse the load data on each server node to improve data reliability and availability.

[0065] When implementing a dynamic load buffer database, the system can use various database management systems, such as relational databases (such as MySQL, Oracle, etc.) or NoSQL databases (such as MongoDB, Cassandra, etc.). When designing a database, it is necessary to consider the characteristics of load data, such as large data volume, frequent updates, and high real-time query requirements, and select appropriate data types and index structures to optimize data storage and access efficiency. At the same time, the system can also set data retention policies to regularly archive or clean up expired load data to control database storage space and maintenance costs.

[0066] S102, calculating the load stability coefficient of each server node according to the real-time load data;

[0067] The system calculates the load stability coefficient of each server node based on real-time load data. The load stability coefficient is a dimensionless parameter obtained by weighting the standard deviation of processor utilization, memory occupancy, and network bandwidth utilization within a preset time period.

[0068] The system calculates the load stability coefficient of each server node based on real-time load data. The load stability coefficient is an indicator that comprehensively measures the volatility of server node load. It takes into account the changes in processor usage, memory occupancy, and network bandwidth utilization within a preset time period, and obtains a dimensionless stability measurement value through weighted calculation. The higher the load stability coefficient, the more stable the load of the server node; conversely, the lower the load stability coefficient, the greater the load fluctuation of the server node, and there may be potential performance bottlenecks or failure risks. In addition to the above three load indicators, the system can also incorporate other indicators that reflect the operating status of the server node, such as disk I / O rate, request response time, etc., to comprehensively evaluate the load stability of the server node.

[0069] When calculating the load stability coefficient, the system first needs to obtain the real-time load data of each server node within a preset time period from the dynamic load buffer database, including the time series of processor usage, memory occupancy, and network bandwidth utilization. Then, for each load indicator, the system calculates its standard deviation within the time period to quantify the degree of fluctuation of the indicator. Next, the system weights and sums the standard deviations of each load indicator according to the preset weighting coefficient to obtain a comprehensive load fluctuation measurement value. Finally, the system normalizes the fluctuation measurement value, and the result is the load stability coefficient of the server node.

[0070] S103, when the load stability coefficient of the server node is less than a first preset threshold, marking the server node as a fluctuating node;

[0071] When the load stability coefficient of a server node is less than the first preset threshold, the system marks the server node as a fluctuating node. A fluctuating node refers to a server node with large load fluctuations and unstable operation. It may face potential problems such as computing resource shortages and network congestion, and requires special attention and regulation. The first preset threshold is the critical value of the load stability coefficient determined based on experience or historical data. When the load stability coefficient of a server node is lower than the threshold, it is considered that the node has entered a fluctuating state. The setting of the threshold needs to balance the detection sensitivity and false alarm rate of load fluctuations. It is necessary to detect real fluctuating nodes in a timely manner, and avoid misjudging normal nodes as fluctuating nodes, resulting in unnecessary management overhead.

[0072] When implementing the marking of fluctuating nodes, the system can add a fluctuation marking field to each server node in the dynamic load buffer database, and use a Boolean value or an enumeration value to represent the fluctuation state of the node. When the load stability coefficient of the server node is less than the first preset threshold, the system sets the fluctuation marking field of the node to true or "fluctuating"; otherwise, it sets it to false or "stable". In this way, the system can quickly obtain the current list of fluctuating nodes through a simple database query, providing a decision-making basis for subsequent task scheduling and resource optimization.

[0073] S104, collecting network quality parameters between server nodes, and calculating node coordination efficiency values ​​according to the network quality parameters and the geographical distances between the server nodes;

[0074] The system collects network quality parameters between server nodes, and calculates the node coordination efficiency value based on the network quality parameters and the geographical distance between each server node. Specifically: the geographical coordinate information of each server node is obtained, and the straight-line distance between nodes is calculated based on the geographical coordinate information;

[0075] Collect link delay, packet loss rate, and link available bandwidth between server nodes;

[0076] The node collaboration efficiency value is obtained by normalizing the straight-line distance between nodes, link delay, packet loss rate, and link available bandwidth according to the preset weighting coefficient and then weighting them.

[0077] The system collects network quality parameters between server nodes and calculates the node coordination efficiency value based on the network quality parameters and the geographical distance between each server node. The node coordination efficiency value is a comprehensive indicator to measure the collaborative working ability between different server nodes. It comprehensively considers the network transmission performance and physical distance factors between nodes. Network quality parameters include network delay, packet loss rate, available bandwidth, etc., which reflect the timeliness and reliability of data interaction between nodes; while geographical distance affects the physical limit and propagation delay of network transmission. The higher the node coordination efficiency value, the stronger the coordination ability between nodes, which can share computing and storage resources more efficiently and complete task processing collaboratively.

[0078] In the process of calculating the node collaborative efficiency value, the system first needs to obtain the geographic location information of the server node. The latitude and longitude coordinates of the node can be obtained through GPS positioning, IP address resolution, etc. Then, the system calculates the straight-line distance between each node based on the latitude and longitude coordinates using the spherical distance formula. Next, the system collects network quality parameters between server nodes, such as network delay, packet loss rate, available bandwidth, etc., which can be obtained in real time through active detection or passive monitoring. Finally, the system weights the distance between nodes and each network quality parameter according to the preset weighting coefficient to obtain the node collaborative efficiency value. Among them, the distance weight reflects the degree of influence of the geographical location on the collaborative efficiency, while the network quality weight reflects the relative importance of different network indicators to the collaborative efficiency.

[0079] S105. When the real-time load value of the fluctuating node exceeds a second preset threshold, a target server candidate set is determined based on the node collaborative efficiency value, and a task allocation weight is calculated based on the load stability coefficient of each server node in the target server candidate set;

[0080] When the real-time load value of the fluctuating node exceeds the second preset threshold, the system determines the target server candidate set based on the node coordination efficiency value, and calculates the task allocation weight according to the load stability coefficient of each server node in the target server candidate set, specifically including:

[0081] Obtain the current load value and remaining computing resources of each server node in the target server candidate set;

[0082] Calculate the resource reliability of each server node based on the load stability coefficient and the remaining computing resources;

[0083] The resource reliability of each server node is normalized to obtain the corresponding task allocation weight.

[0084] When the real-time load value of the fluctuating node exceeds the second preset threshold, the system determines the target server candidate set based on the node collaborative efficiency value, and calculates the task allocation weight according to the load stability coefficient of each server node in the candidate set. The second preset threshold is the warning line for the fluctuating node load. When the load value exceeds this threshold, it means that the fluctuating node faces the risk of shortage of computing resources and needs support from other stable nodes. The system first selects nodes with high collaborative efficiency with the fluctuating nodes from all stable nodes as the target server candidate set, and then further evaluates the load stability and computing power of the candidate nodes, calculates the task allocation weight, and provides a decision-making basis for subsequent task migration and load balancing.

[0085] When determining the candidate set of target servers, the system comprehensively considers two factors: the node collaborative efficiency value and the load stability coefficient. Specifically, the system first sorts all stable nodes from high to low according to the collaborative efficiency value with the fluctuating nodes, and selects the top N nodes with the highest collaborative efficiency as the preliminary candidate set. The value of N can be set according to actual conditions, ensuring that there are enough candidate nodes to support task allocation, while controlling the size of the candidate set to avoid an overly complicated allocation process. Next, the system further screens nodes with a load stability coefficient higher than a certain threshold from the preliminary candidate set as the final target server candidate set. Load stability is an important factor in ensuring the efficiency of task execution. Selecting nodes with stable loads can minimize the risk of task failure or delays.

[0086] When calculating the task allocation weight, the system comprehensively evaluates the load stability and computing power of the candidate nodes. First, the system obtains the current load status information of the candidate nodes (such as the usage of CPU, memory, network and other resources) and the idle computing resource capacity to evaluate the computing power of the nodes. Then, based on the load stability coefficient and the idle computing resource capacity, the system calculates the comprehensive capability index of each candidate node as a reference for task allocation. Finally, the system normalizes the comprehensive capability indicators of all candidate nodes to obtain the task allocation weight. The higher the weight of the node, the more stable its load, the more sufficient its computing resources, and the more suitable it is for taking on new task allocation.

[0087] S106, allocating tasks on the fluctuation nodes to server nodes in the target server candidate set according to task allocation weights, and continuously collecting real-time load data of the fluctuation nodes;

[0088] The system allocates tasks on the fluctuating nodes to server nodes in the target server candidate set according to the task allocation weight, and continuously collects real-time load data of the fluctuating nodes. The task allocation weight reflects the load stability and computing power of the candidate nodes. The higher the weight of the node, the greater the proportion of tasks allocated to it. By controlling the distribution ratio of tasks by weight, the load can be balanced among multiple server nodes to avoid situations where individual nodes are overloaded while other nodes have idle resources. At the same time, after the task is allocated, the system also needs to continuously monitor the load status of the fluctuating nodes. On the one hand, it evaluates the effect of task allocation, and on the other hand, it promptly discovers the recovery of the load of the fluctuating nodes to prepare for subsequent task migration.

[0089] When implementing task allocation, the system can use scheduling algorithms such as weighted polling or weighted random. Taking the weighted polling algorithm as an example, the system calculates the number of tasks that can be assigned to each candidate node in a scheduling cycle based on the task allocation weights of each candidate node. Then, the system sorts the tasks to be assigned on the fluctuating nodes according to the expected execution time, and gives priority to tasks with shorter execution times to relieve the load pressure of the fluctuating nodes as soon as possible. Next, the system assigns the sorted tasks to each candidate node in turn, and the number of tasks assigned to each node is proportional to its weight. When a node is assigned a sufficient number of tasks in the current scheduling cycle, the system skips the node until all tasks are assigned. In this way, through weighted polling, the system can achieve balanced distribution of tasks among multiple nodes while taking into account the execution efficiency of tasks.

[0090] S107, when the real-time load value of the fluctuation node is lower than the third preset threshold value for a continuous preset period of time, calculating the processing efficiency improvement ratio of the assigned task on the target server candidate set;

[0091] When the real-time load value of the fluctuation node is lower than the third preset threshold value for a continuous preset period of time, the system calculates the processing efficiency improvement ratio of the assigned tasks on the target server candidate set, specifically: recording the average response time and resource consumption when the fluctuation node processes each task;

[0092] Collect the average response time and resource consumption of each task when running on the candidate set of target servers;

[0093] The processing efficiency improvement ratio is calculated based on the average response time ratio and resource consumption ratio of each task before and after migration.

[0094] When the real-time load value of the fluctuating node is lower than the third preset threshold for a continuous preset period of time, the system calculates the processing efficiency improvement ratio of the tasks assigned to the candidate set of target servers. The third preset threshold is usually lower than the second preset threshold and is used to determine whether the load of the fluctuating node has returned to normal. The continuous preset time is to prevent misjudgment caused by short-term load fluctuations. Only when the load remains stable for a period of time is the fluctuating node considered to have returned to normal and task migration can be considered. The processing efficiency improvement ratio reflects the change in execution efficiency after the task is migrated from the fluctuating node to the candidate node. It comprehensively considers the two factors of task response time and resource utilization, and is an important indicator for evaluating the effectiveness of task allocation.

[0095] When calculating the processing efficiency improvement ratio, the system first needs to record the performance indicators of the task when it is executed on the fluctuating node, including the average response time and resource consumption. The average response time reflects the time overhead of the entire life cycle of the task from submission to completion, and the resource consumption reflects the occupation of computing, storage, network and other resources during the task execution. These indicators can be obtained through log analysis, performance monitoring and other means. Then, the system collects the performance indicators of the task after execution on the candidate node, including the average response time and resource consumption. Next, the system calculates the ratio of response time and resource consumption before and after task migration, and then performs weighted summation to obtain the comprehensive processing efficiency improvement ratio. Among them, a response time ratio less than 1 indicates accelerated task execution, a resource consumption ratio less than 1 indicates improved resource utilization, and a processing efficiency improvement ratio greater than 1 indicates that task allocation has achieved positive results.

[0096] S108, for tasks where the processing efficiency improvement ratio is greater than a fourth preset threshold, keep processing at the current server node;

[0097] For tasks whose processing efficiency improvement ratio is greater than the fourth preset threshold, the system chooses to continue to execute on the current server node. The fourth preset threshold is a key indicator for evaluating the effect of task allocation. When the processing efficiency improvement ratio of a task exceeds this threshold, it indicates that migrating the task from the fluctuating node to the candidate node can significantly improve the execution performance of the task. Therefore, the system tends to maintain the existing task allocation status to obtain continuous performance optimization benefits. The setting of the threshold requires balancing the task execution efficiency and the migration overhead, and selecting a suitable compromise point. The higher the threshold, the more significant the required performance improvement, and the longer the task will remain executed on the candidate node.

[0098] When implementing the task retention strategy, the system needs to update the task status information and monitoring indicators. Specifically, the system changes the assignment nodes of these tasks from the original fluctuating nodes to the actual candidate nodes, and records the time point when the tasks start to run stably on the candidate nodes. At the same time, the system continues to monitor the execution of these tasks on the candidate nodes, and regularly collects key indicators such as task response time and resource utilization to continuously evaluate the effectiveness of task assignment. In addition, considering that the load on the candidate node may rise sharply during the continuous operation of the task, the system also needs to set the maximum time limit for task retention. When the task continues to run on the candidate node for more than the maximum time limit, the system needs to re-evaluate the task assignment status and perform secondary migration if necessary to prevent the candidate node from being overloaded and affecting the overall service quality.

[0099] S109: For tasks whose processing efficiency improvement ratio is less than or equal to a fourth preset threshold, migrate them back to the original fluctuation node.

[0100] For tasks whose processing efficiency improvement ratio is less than or equal to the fourth preset threshold, the system migrates them from the candidate node back to the original fluctuating node for execution. The execution efficiency of these tasks on the candidate node is not significantly improved, and the benefits of continuing to run on the candidate node cannot offset the overhead of task migration, so migrating them back to the original node is a better choice. Task migration can free up the computing resources of the candidate node and provide more scheduling space for the allocation and execution of other tasks. At the same time, task migration also helps to restore the workload of the fluctuating node, maintain its load at a reasonable level, and avoid wasting computing resources in a low-load state for a long time.

[0101] When implementing task relocation, the system needs to develop a detailed migration plan and execution process. First, the system screens out the list of tasks that need to be relocated based on the processing efficiency improvement ratio of the tasks, and sorts them according to the priority and resource requirements of the tasks. Then, the system communicates with the wave node to confirm its current load status and available resources, and evaluates whether it has the conditions to receive the relocated tasks. If the load of the wave node is normal and the resources are sufficient, the system will start the task relocation process; if the wave node does not yet have the conditions to receive, the system can choose to temporarily store the task on the candidate node or other temporary node, waiting for the right time to relocate. Before officially executing the relocation, the system also needs to perform necessary data synchronization and state transfer with the candidate node and the wave node to ensure that the task can be smoothly migrated from the candidate node to the wave node and continue to execute without interruption or data loss.

[0102] In the above embodiment, a dynamic load buffer database is established in the central controller, and the load stability coefficient is calculated based on the processor utilization rate, memory occupancy rate and network bandwidth utilization rate. The system can accurately identify server nodes with large load fluctuations. The node collaborative efficiency value calculated by combining the network quality parameters and geographical distance between nodes can select the most suitable target server when migrating tasks. The task allocation weight based on the load stability coefficient is adopted so that tasks can be assigned to server nodes with more stable loads and more sufficient resources. By continuously monitoring the load of fluctuating nodes and determining whether the task needs to be migrated back based on the processing efficiency improvement ratio, it not only ensures the improvement of task processing efficiency, but also avoids unnecessary task migration, so that the system can achieve efficient utilization of server resources while ensuring service quality, and reduce the risk of unbalanced server load in edge computing environments.

[0103] In order to further improve the predictability and initiative of the above-mentioned server management method, the embodiment of the present application also provides a resource reservation method based on load trend analysis. This method performs real-time monitoring and trend analysis on the load changes of the fluctuating nodes, identifies potential risks in advance before the load increases abnormally, and establishes a corresponding resource reservation mechanism based on the priority characteristics of the task. Figure 2 , a resource reservation method based on load trend analysis in an embodiment of the present application is described:

[0104] See also Figure 2 , which is a flow chart of a resource reservation method based on load trend analysis in an embodiment of the present application.

[0105] S201, collecting statistics on load fluctuation data of fluctuation nodes within a preset time period, and generating a load fluctuation time series according to the load fluctuation data;

[0106] The system collects statistics on the load fluctuation data of the fluctuation nodes within a preset time period, and generates a load fluctuation time series based on the load fluctuation data. The load fluctuation data reflects how the node load changes over time, and can include indicators in multiple dimensions such as CPU usage, memory occupancy, and network bandwidth utilization. The length of the preset time period can be set according to actual needs and system characteristics. It must be long enough to capture the overall change trend of the load, but short enough to ensure the real-time nature and processing efficiency of the data. The system can obtain load fluctuation data through regular sampling or real-time monitoring, and organize it into a load fluctuation time series in chronological order to provide a data basis for subsequent trend analysis.

[0107] When generating the load fluctuation time series, the system can use a variety of data processing and feature extraction techniques. For example, the system can denoise and smooth the original load fluctuation data, remove the influence of accidental factors and outliers, and extract the main trend information of load changes. Common denoising and smoothing methods include moving average, exponential smoothing, wavelet transform, etc. In addition, the system can also sample and compress the load fluctuation time series to reduce the amount of data and computing overhead. For example, the system can use strategies such as equal interval sampling or adaptive sampling to select representative time points and load values ​​to construct the time series; the system can also use time series compression algorithms such as PAA (Piecewise Aggregate Approximation) and SAX (Symbolic Aggregate Approximation) to convert the original load fluctuation series into a more concise and efficient representation.

[0108] S202, calculating the load change rate and acceleration of the fluctuation node according to the load fluctuation time series;

[0109] The system calculates the load change rate and acceleration of the fluctuating node based on the load fluctuation time series. The load change rate reflects the magnitude of the change in the node load per unit time, which indicates how fast the load increases or decreases; the load change acceleration reflects the trend of the load change rate, which indicates whether the load increase or decrease trend is accelerating or slowing down. By calculating the load change rate and acceleration, the system can monitor and warn of abnormal changes in node load in real time, take timely countermeasures before the load suddenly increases or decreases, and improve the stability and reliability of the system.

[0110] When calculating the load change rate and acceleration, the system can use methods such as numerical differentiation and curve fitting. Specifically, for the load change rate, the system can use the finite difference method to calculate the load difference between two adjacent time points in the load fluctuation time series, and then divide it by the time interval to obtain the average rate of load change; for the load change acceleration, the system can apply the finite difference method again on the basis of the load change rate, calculate the rate difference between two adjacent time points, and then divide it by the time interval to obtain the rate of change of the load change rate, that is, the acceleration. In addition, the system can also use curve fitting methods such as polynomial fitting or spline interpolation to map the load fluctuation time series to a continuous and smooth function curve, and then obtain the analytical expressions of the load change rate and acceleration by taking derivatives to improve the accuracy and efficiency of the calculation.

[0111] S203, when the load change rate exceeds the fifth preset threshold value and the load change acceleration is a positive value, extracting task feature data and determining the task priority level;

[0112] When the load change rate exceeds the fifth preset threshold and the load change acceleration is a positive value, the system extracts the task feature data and determines the task priority level, wherein determining the task priority level specifically includes:

[0113] The response time constraints, computing resource consumption, and data interaction frequency of the collection task;

[0114] Perform dependency analysis on tasks and mark tasks that have dependencies on previous tasks as critical path tasks;

[0115] The tasks whose response time constraints are less than the sixth preset threshold in the critical path tasks are marked as the highest priority, and the tasks whose response time constraints are between the sixth preset threshold and the seventh preset threshold are marked as high priority.

[0116] When the load change rate exceeds the fifth preset threshold and the load change acceleration is positive, the system extracts the task feature data and determines the task priority. The fifth preset threshold is a key indicator for determining whether there is an abnormal growth trend in the node load. When the load change rate exceeds this threshold and the load change acceleration is positive, it indicates that the node is facing the risk of a sharp increase in load and that countermeasures must be taken urgently. In order to migrate critical tasks to nodes with sufficient resources in a timely manner before the service quality deteriorates due to excessive load, the system needs to determine the priority of the task based on the characteristics and importance of the task, and give more guarantees to high-priority tasks in subsequent resource reservation and task scheduling.

[0117] In the process of determining the priority of tasks, the system comprehensively considers multiple characteristic dimensions of the tasks, including response time constraints, computing resource consumption, and data interaction frequency. Response time constraints reflect the real-time requirements of tasks. The stricter the constraints, the higher the priority of the task; computing resource consumption reflects the demand of tasks for CPU, memory, storage and other resources. The greater the consumption, the higher the priority of the task; data interaction frequency reflects the dependency between tasks and other tasks or external systems. The higher the frequency, the higher the priority of the task. In addition, the system will also perform dependency analysis on tasks, identify tasks on the critical path, that is, those tasks that have a significant impact on the entire task chain or business process, and give them higher priority.

[0118] When determining the priority of a task, the system can use methods such as multi-level thresholds and weighted scoring. For example, the system can preset two response time constraint thresholds, the sixth preset threshold and the seventh preset threshold, mark the task with a response time constraint less than the sixth threshold as the highest priority, mark the task with a response time constraint between the sixth threshold and the seventh threshold as a high priority, and mark the remaining tasks as ordinary priority. Similarly, the system can also set corresponding thresholds based on computing resource consumption and data interaction frequency to grade tasks. In addition to threshold comparison, the system can also use a weighted scoring method to assign different weights to each feature dimension of the task, and then calculate the weighted sum or weighted average to obtain a comprehensive priority score for the task, and determine the final priority of the task based on the score ranking.

[0119] In the process of determining the priority of tasks, the feature data of some tasks may be incomplete or inaccurate, affecting the reliability of priority evaluation. To solve this problem, the system can introduce default value filling and data verification mechanisms. Specifically, for missing feature data, such as the lack of response time constraints or computing resource consumption, the system can set reasonable default values, such as average response time or median resource consumption, based on the type of task, historical data and other information. For inaccurate feature data, such as abnormally high or low values, the system can perform data verification and correction to limit it to a reasonable range. For example, the system can count the distribution characteristics of indicators such as response time and resource consumption based on the historical execution records of the task, and truncate or replace the data that obviously deviates from the distribution. Through default value filling and data verification, the system can effectively make up for the lack of task feature data and improve the accuracy and robustness of task priority evaluation.

[0120] S204: Based on the task priority, select a target node that meets the preset task resource requirements from the target server candidate set, and establish a resource reservation channel based on the target node.

[0121] Based on the task priority, the system selects the target node that meets the preset task resource requirements from the target server candidate set, and establishes a resource reservation channel based on the target node. The target server candidate set is a potential migration target node selected based on factors such as load stability and computing power. The system needs to further select the optimal target node from the candidate set to meet the resource requirements and performance requirements of high-priority tasks. At the same time, in order to ensure that the task can obtain stable and continuous resource guarantees after migration, the system also needs to establish a dedicated resource reservation channel on the target node to provide isolation and priority access mechanisms for the operation of the task.

[0122] When selecting the target node, the system comprehensively considers the resource requirements of the task and the resource status of the node. Specifically, the system first estimates the scale and ratio of CPU, memory, storage and other resources required for the task based on the priority level of the task, computing resource consumption and other characteristics, and forms a description of the preset task resource requirements. Then, the system collects the resource usage of each node in the candidate set of target servers in real time, including CPU utilization, memory occupancy, storage space utilization, etc., and matches and compares them with the preset task resource requirements. The system selects nodes with relatively low resource utilization and resource scale and ratio that can meet the task requirements as candidate target nodes, and sorts and selects the candidate target nodes based on factors such as node load stability and network quality, and finally determines the optimal target node.

[0123] When establishing a resource reservation channel, the system can use virtualization, containerization and other technologies to create independent resource partitions and access control mechanisms on the target node. For example, the system can use a virtual machine or container engine to create a dedicated virtual computing environment for the migration task, and configure the corresponding CPU, memory, storage and other resource quotas to ensure that the task can monopolize the necessary computing resources without interference from other tasks or processes. At the same time, the system can also establish a virtual private network channel between nodes through network slicing, QoS and other technologies to provide bandwidth guarantee and priority scheduling mechanism for task data transmission. In addition, the system can use resource orchestration and scheduling frameworks, such as Kubernetes, Mesos, etc., to centrally manage and dynamically adjust the resource reservation channel, and timely optimize resource configuration and allocation strategies according to the actual operation of the task and load changes, so as to improve resource utilization efficiency.

[0124] In the above embodiment, the load fluctuation data of the fluctuating node within a preset time period is counted and a time series is generated. Combined with the calculation of the load change rate and acceleration, the system can accurately grasp the trend and intensity of the load change. When the load change rate exceeds the threshold and the acceleration is a positive value, it indicates that the load pressure of the fluctuating node is growing rapidly. At this time, by extracting task feature data and determining the task priority level, the system can identify critical tasks with high service quality requirements. Based on the task priority level, a suitable target node is selected from the target server candidate set and a resource reservation channel is established. This preventive resource scheduling mechanism can prepare the required computing resources for important tasks in advance before the load surge causes a decline in service quality, avoiding serious service quality degradation problems during peak load periods. At the same time, it also reduces the system overhead caused by emergency task migration and improves service reliability in edge computing environments.

[0125] Further, in another embodiment, after establishing the resource reservation channel based on the target node, the method further includes:

[0126] Monitor resource occupancy and task processing delay of resource reservation channels;

[0127] Calculate channel utilization efficiency based on resource occupancy rate;

[0128] When the channel utilization efficiency is lower than an eighth preset threshold, releasing a preset number of reserved resources;

[0129] When the task processing delay exceeds the preset delay threshold, the quota of reserved resources is expanded.

[0130] In this embodiment, after establishing the resource reservation channel, the system also introduces a dynamic adjustment mechanism to elastically scale and optimize the configuration of reserved resources according to the actual operation status of the channel to improve resource utilization efficiency and task execution performance.

[0131] Specifically, the system continuously monitors the operating indicators of the resource reservation channel, including resource occupancy rate and task processing delay. Resource occupancy rate reflects the actual use of reserved resources, such as CPU occupancy rate, memory occupancy rate, etc., which measures the degree of resource consumption of tasks in the channel; task processing delay reflects the execution efficiency of tasks in the channel, such as task response time, task completion time, etc., which measures the service quality and user experience of tasks in the channel. By collecting and analyzing these two types of indicators in real time, the system can comprehensively evaluate the operating status and effect of the resource reservation channel.

[0132] After obtaining the resource occupancy data, the system further calculates the channel utilization efficiency. Channel utilization efficiency is a comprehensive indicator that associates resource occupancy with the scale of reserved resources, and is used to measure the utilization degree and input-output ratio of reserved resources. Specifically, the system can divide the actual amount of occupied resources by the total amount of reserved resources to obtain a utilization efficiency value in the form of a percentage. The higher the utilization efficiency, the more accurate the resource allocation is; conversely, the lower the utilization efficiency, which means that the reserved resources are wasted or idle, and need to be properly adjusted and recycled.

[0133] When the channel utilization efficiency is lower than the eighth preset threshold, the system automatically releases a preset number of reserved resources. The eighth preset threshold is the critical point that triggers resource release. When the utilization efficiency continues to be lower than this threshold for a period of time, the system believes that the current resource reservation scale is too large and there is room for optimization. Therefore, the system allocates a portion of the current reserved resources, such as 10% or 20%, removes it from the resource reservation channel, and releases it for use by other tasks or systems. By dynamically releasing reserved resources, the system can reduce resource waste and improve overall resource utilization.

[0134] In contrast, when the task processing delay exceeds the preset delay threshold, the system will automatically expand the quota of reserved resources. The preset delay threshold is a key indicator for measuring task execution performance. When the task processing delay continues to exceed the threshold for a period of time, it indicates that the current resource reservation scale is insufficient and cannot meet the performance requirements of the task. Therefore, the system allocates additional resources from the available resource pool, such as increasing the CPU and memory quota by 20% or 50%, and adds them to the resource reservation channel to expand the scale of reserved resources. By dynamically expanding reserved resources, the system can provide more computing power for critical tasks, alleviate performance bottlenecks, and improve task execution efficiency and service quality.

[0135] In the process of dynamically adjusting reserved resources, the system can adopt a gradual adjustment strategy to avoid system oscillation caused by overly aggressive resource allocation or recycling. For example, when the resource release condition is triggered, the system can first release a small part of the reserved resources, such as 5%, and then continue to observe for a period of time. If the channel utilization efficiency is still low, the next 5% will be released until the preset total release amount is reached or the utilization efficiency returns to a reasonable level. Similarly, when expanding reserved resources, the system can also adopt a gradual expansion method, increasing a small part of the resource quota each time, and then evaluating the changes in task processing delays before deciding whether to continue expanding. Through gradual adjustment, the system can optimize resource allocation more smoothly and reduce the impact on the business.

[0136] In the above embodiment, the resource occupancy rate and task processing delay of the resource reservation channel are continuously monitored, and the system can evaluate the use effect of the reserved resources in real time. The channel utilization efficiency is calculated according to the resource occupancy rate, and when the efficiency is lower than the threshold, the preset number of reserved resources is released in time, thereby avoiding excessive reservation and waste of resources. At the same time, when the task processing delay exceeds the preset threshold, the system will actively expand the quota of reserved resources, so that the scale of reserved resources can be adaptively changed according to the task processing requirements, thereby achieving a balance between the economy of reserved resources and service quality.

[0137] The following describes the system in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , which is a schematic diagram of the physical device structure of an intelligent server management system in an edge computing environment provided in an embodiment of the present application.

[0138] It should be noted that Figure 3 The structure of the system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0139] like Figure 3As shown, the system includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 to the random access memory (RAM) 303, such as executing the method in the above embodiment. In RAM 303, various programs and data required for system operation are also stored. CPU 301, ROM 302 and RAM 303 are connected to each other through a bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0140] The following components are connected to the I / O interface 305: an input section 306 including a camera, an infrared sensor, etc.; an output section 307 including a liquid crystal display (LCD) and a speaker, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0141] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present invention are performed.

[0142] It should be noted that the computer-readable medium shown in the embodiment of the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable computer program. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.

[0143] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Among them, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0144] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the system described in the above embodiment; or may exist independently without being assembled into the system. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of a system, the system implements the method provided in the above embodiment.

[0145] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, a person of ordinary skill in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0146] As used in the above embodiments, the term "when..." may be interpreted as "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted as "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

[0147] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk), etc.

[0148] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. An intelligent server management method in an edge computing environment, characterized in that: include: Establishing a dynamic load buffer database in a preset central controller, wherein the dynamic load buffer database is used to store real-time load data of each server node within a preset time period; Calculate the load stability coefficient of each server node according to the real-time load data, where the load stability coefficient is a dimensionless parameter obtained by weighting the standard deviation of the processor usage rate, the memory occupancy rate, and the network bandwidth utilization rate within the preset time period; When the load stability coefficient of the server node is less than a first preset threshold, marking the server node as a fluctuating node; Collecting network quality parameters between the server nodes, and calculating node coordination efficiency values ​​according to the network quality parameters and the geographical distances between the server nodes; When the real-time load value of the fluctuation node exceeds a second preset threshold, a target server candidate set is determined based on the node collaborative efficiency value, and a task allocation weight is calculated according to a load stability coefficient of each server node in the target server candidate set; Allocating the tasks on the fluctuation nodes to the server nodes in the target server candidate set according to the task allocation weights, and continuously collecting real-time load data of the fluctuation nodes; When the real-time load value of the fluctuation node is lower than a third preset threshold value for a continuous preset period of time, calculating the processing efficiency improvement ratio of the assigned task on the target server candidate set; For the task where the processing efficiency improvement ratio is greater than a fourth preset threshold, keep processing at the current server node; For the tasks whose processing efficiency improvement ratio is less than or equal to the fourth preset threshold, they are migrated back to the original fluctuation node.

2. The method according to claim 1, characterized in that The calculating the node coordination efficiency value according to the network quality parameter and the geographical distance between each of the server nodes specifically includes: Obtaining geographic coordinate information of each of the server nodes, and calculating the straight-line distance between the nodes according to the geographic coordinate information; Collecting link delay, data packet loss rate and link available bandwidth between the server nodes; The straight-line distance between the nodes, the link delay, the data packet loss rate and the link available bandwidth are normalized and weighted according to a preset weighting coefficient to obtain a node coordination efficiency value.

3. The method according to claim 1, characterized in that The step of calculating the task allocation weight according to the load stability coefficient of each server node in the target server candidate set specifically includes: Obtaining the current load value and remaining computing resources of each server node in the target server candidate set; Calculating the resource reliability of each of the server nodes based on the load stability coefficient and the remaining computing resources; The resource reliability of each server node is normalized to obtain the corresponding task allocation weight.

4. The method according to claim 1, characterized in that: The calculation of the processing efficiency improvement ratio of the assigned tasks on the target server candidate set specifically includes: Record the average response time and resource consumption of the fluctuation node when processing each task; Collecting the average response time and resource consumption of each task when it is run on the candidate set of target servers; The processing efficiency improvement ratio is calculated based on the average response time ratio and resource consumption ratio of each task before and after the migration.

5. The method according to claim 1, characterized in that: After the task whose processing efficiency improvement ratio is less than or equal to the fourth preset threshold is migrated back to the original fluctuation node, the method further includes: Counting the load fluctuation data of the fluctuation node within a preset time period, and generating a load fluctuation time series according to the load fluctuation data; Calculating the load change rate and acceleration of the fluctuation node according to the load fluctuation time series; When the load change rate exceeds a fifth preset threshold and the load change acceleration is a positive value, extracting task feature data and determining a task priority level; Based on the task priority level, a target node that meets the preset task resource requirements is selected from the target server candidate set, and a resource reservation channel is established based on the target node.

6. The method according to claim 5, characterized in that Determining the task priority level specifically includes: The response time constraints, computing resource consumption, and data interaction frequency of the collection task; Perform dependency analysis on the tasks, and mark the tasks that have dependencies on previous tasks as critical path tasks; The tasks whose response time constraints are less than the sixth preset threshold in the critical path tasks are marked as the highest priority, and the tasks whose response time constraints are between the sixth preset threshold and the seventh preset threshold are marked as the high priority.

7. The method according to claim 5, characterized in that After establishing the resource reservation channel based on the target node, the method further includes: Monitoring resource occupancy and task processing delay of the resource reservation channel; Calculate channel utilization efficiency according to the resource occupancy rate; When the channel utilization efficiency is lower than an eighth preset threshold, releasing a preset number of reserved resources; When the task processing delay exceeds a preset delay threshold, the quota of the reserved resources is expanded.

8. An intelligent server management system in an edge computing environment, characterized in that: The system comprises: One or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the system to execute the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a system, the system is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is run on a system, the system is caused to execute the method according to any one of claims 1 to 7.

Images