A task control method based on real-time communication

Abstract

The application discloses a task control method based on real-time communication, and relates to the field of task control.The method comprises the following steps: determining a node capability score table corresponding to N nodes in a distributed system according to a state information set corresponding to the N nodes; determining a target task publishing scheme according to the node capability score table, wherein the target task publishing scheme comprises task amounts and task distribution sequences respectively undertaken by the N nodes; monitoring resource occupation conditions of each node in the N nodes during publishing of tasks according to the target task publishing scheme, and determining a resource adjustment scheme according to the monitored resource occupation conditions of each node; and monitoring execution times of tasks executed by each node in the N nodes after the execution of the resource adjustment scheme, and adjusting the resources again according to the monitored execution times of the tasks executed by each node. The above scheme can solve the problem that task distribution and resource allocation cannot be dynamically adjusted, thereby leading to poor system stability.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of task control, in particular to a task control method based on real-time communication. BACKGROUND

[0002] In the modern industrial field, real-time task control and efficient scheduling are the core links to ensure system stable operation and resource optimization. With the popularity of distributed systems, industrial monitoring software and data acquisition control play an increasingly important role in multi-node collaboration scenarios. However, in complex environments involving large-scale nodes, traditional methods have exposed many shortcomings.

[0003] Currently, many industrial monitoring software relies on centralized data processing mode, which is difficult to cope with the high-speed update demand of massive information. This mode leads to insufficient timeliness of state monitoring, especially in distributed systems, the delay problem of data acquisition control is particularly prominent. In addition, the low efficiency of command distribution and the uneven allocation of resources further limit the real-time performance and adaptability of the system.

[0004] Specifically, due to the lack of accurate evaluation mechanism for node status, the existing system cannot effectively judge the task execution capability of each node. This makes the task distribution scheme often based on static rules, which is difficult to dynamically adjust to adapt to complex operating environments. At the same time, the real-time monitoring of resource occupation is insufficient, which makes it difficult for the system to maintain stable operation in high-load scenarios. For example, when some nodes have task delays due to insufficient resources, traditional industrial monitoring software and data acquisition control means cannot timely adjust the task distribution strategy or resource allocation scheme.

[0005] In addition, the existing technology also has defects in monitoring the execution time of nodes after task publishing. Once the task allocation is unreasonable or the resource allocation is unbalanced, the system may fall into an inefficient running state, and even trigger a chain reaction, affecting the overall performance. Therefore, how to combine the characteristics of industrial monitoring software and data acquisition control, and build a mechanism that can dynamically adjust task distribution and resource allocation, has become a technical problem to be solved.

[0006] In view of the problem that the related art cannot dynamically adjust task distribution and resource allocation, resulting in poor system stability, no effective solution has been proposed so far.

[0007] Therefore, it is necessary to improve the related art to overcome the defects in the related art. SUMMARY

[0008] The embodiments of the present application provide a task control method based on real-time communication, which at least solves the problem that the existing technology cannot dynamically adjust task distribution and resource allocation, resulting in poor system stability.

[0009] According to an aspect of the embodiments of the present application, a task control method based on real-time communication is provided, which comprises: determining a node capability score table corresponding to N nodes in a distributed system according to a state information set corresponding to the N nodes, wherein the state information set comprises state information of the N nodes acquired in real time, and N is an integer greater than or equal to 2; determining a target task publishing scheme according to the node capability score table, wherein the target task publishing scheme comprises a task amount and a task distribution sequence of the N nodes; monitoring resource occupation of each node in the N nodes in a process of publishing a task according to the target task publishing scheme, and determining a resource adjustment scheme according to the monitored resource occupation of each node, wherein the resource adjustment scheme is used to adjust the resource occupation of the N nodes; and monitoring execution time of each node in the N nodes in executing a task after the resource adjustment scheme is executed, and adjusting the resource again according to the monitored execution time of each node in executing the task.

[0010] In an exemplary embodiment, before the node capability score table corresponding to the N nodes in the distributed system is determined according to the state information set corresponding to the N nodes, the method further comprises: determining a delay node list according to the state information set, wherein the delay node list records an identifier, a node type and a latest state update time of M nodes in the N nodes; each node in the M nodes has a state update time interval exceeding a preset time interval, and M is an integer greater than or equal to 1 and less than or equal to N; the state information of a node comprises a state update record of the node; the acquisition frequency of the state information of the M nodes is updated according to the delay node list; and the state information set is updated according to the state information of the M nodes acquired according to the updated acquisition frequency.

[0011] In an exemplary embodiment, the delay node list is determined according to the state information set, which comprises: determining update frequency data corresponding to the N nodes according to the state information set, wherein the update frequency data comprises an update time interval of each node in the N nodes in a preset time period; determining the M nodes from the N nodes according to the update frequency data; analyzing the state update record of the M nodes by using a data timeliness verification tool, so as to determine a node type of each node in the M nodes according to the continuity and integrity of each node in the M nodes in each state update, and determine a latest state update time of each node, wherein the node type comprises a first delay type node and a second delay type node; and the delay degree of the second delay type node is greater than that of the first delay type node.

[0012] In an example embodiment, the collection frequency of the state information of the M nodes is updated according to the delay node list, including: determining the priority of each node in the M nodes according to the latest state update time of the M nodes in the delay node list; and adjusting the collection frequency of the state information of each node in the M nodes according to the priority of each node in the M nodes.

[0013] In an example embodiment, the node capability score table corresponding to the N nodes in the distributed system is determined according to the state information set corresponding to the N nodes, including: processing the state information set by a big data processing tool to determine the task execution performance data corresponding to each node in the N nodes, wherein the task execution performance data includes task completion time and response speed; the state information of the node includes the running record of the node; the priority of the N nodes in executing tasks is determined according to the task execution performance data corresponding to the N nodes respectively, and the task distribution sequence is determined according to the priority of the N nodes in executing tasks; the state information and the task execution performance data of the N nodes after executing tasks according to the task distribution sequence are associated and matched by a big data information integration tool to determine the capability score of each node in the N nodes, so as to obtain the node capability score table corresponding to the N nodes.

[0014] In an example embodiment, the target task publishing scheme is determined according to the node capability score table, including: determining the task matching degree of each node according to the node state of each node in the N nodes and the capability score of each node in the node capability score table, wherein the node state includes the resource occupation of the node, and the task matching degree is used to represent the matching degree with tasks of different priorities; determining the initial task distribution scheme according to the task matching degree of each node, and adjusting the task amount of each node in the initial task distribution scheme according to the capability score of each node to obtain the adjusted task distribution scheme; updating the task distribution sequence in the adjusted task distribution scheme according to the preset task publishing rule to obtain the target task publishing scheme.

[0015] In an exemplary embodiment, the resource adjustment scheme is determined according to the monitored resource occupancy of each node, including: determining P nodes from the N nodes according to the resource occupancy of each node in the N nodes, wherein the resource occupancy rate of each node in the P nodes is greater than a preset occupancy rate, and P is an integer greater than or equal to 1 and less than or equal to N; determining the resource adjustment scheme as a first resource adjustment scheme, wherein the first resource adjustment scheme is to reduce the task amount of each node in the P nodes in the target task publishing scheme, and increase the reduced task amount to one or more nodes with the lowest resource occupancy rate in the N nodes.

[0016] In an exemplary embodiment, after determining the resource adjustment scheme as the first resource adjustment scheme, the method further includes: in the case that the resource occupancy rate of any target node in the N nodes is greater than the preset occupancy rate after executing the first resource adjustment scheme, obtaining a standby resource from a standby resource pool and allocating the standby resource to the target node; determining the resource adjustment scheme as the first resource adjustment scheme and a second resource adjustment scheme, wherein the second resource adjustment scheme is to allocate the standby resource to the target node.

[0017] In an exemplary embodiment, the resource adjustment is performed again according to the monitored execution time of each node performing a task, including: determining Z nodes from the N nodes, wherein the execution time of each node in the Z nodes is greater than a preset execution time, and Z is an integer greater than or equal to 1 and less than or equal to N; and performing resource adjustment according to the resource configuration state of the Z nodes and the allowed allocation resource information.

[0018] In an exemplary embodiment, after the resource adjustment is performed again according to the monitored execution time of each node performing a task, including: continuously monitoring the stability of task execution of the N nodes by using industrial monitoring software; determining the stability index of each node in the N nodes according to the monitored monitoring data to obtain an execution performance report.

[0019] According to another aspect of the embodiments of the present application, a task control device based on real-time communication is further provided, comprising: a first determining module configured to determine a node capability score table of N nodes in a distributed system according to a state information set corresponding to the N nodes, wherein the state information set comprises state information of the N nodes respectively acquired in real time, and N is an integer greater than or equal to 2; a second determining module configured to determine a target task publishing scheme according to the node capability score table, wherein the target task publishing scheme comprises a task amount and a task distribution sequence respectively undertaken by the N nodes; a third determining module configured to monitor resource occupation of each node of the N nodes in a process of publishing a task according to the target task publishing scheme, and determine a resource adjustment scheme according to the monitored resource occupation of the each node, wherein the resource adjustment scheme is used to adjust the resource occupation of the N nodes; and an adjusting module configured to monitor execution time of the each node of the N nodes in executing a task after the resource adjustment scheme is executed, and perform resource adjustment again according to the monitored execution time of the each node in executing the task.

[0020] According to still another aspect of the embodiments of the present application, a Bluetooth earphone is further provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the task control method based on real-time communication when executing the computer program.

[0021] According to still another aspect of the embodiments of the present application, a computer readable storage medium is further provided, and the computer readable storage medium stores a computer program, wherein the computer program is set to execute the task control method based on real-time communication when running.

[0022] According to still another aspect of the embodiments of the present application, a computer program product is further provided, comprising a computer program, and the computer program is the task control method based on real-time communication when executed by a processor.

[0023] The application generates a node capability score table by acquiring the state information of each node in the distributed system in real time, dynamically formulates a target task release scheme based on the score table, and determines the task amount and distribution sequence of each node, thereby solving the problem that the traditional method cannot dynamically adjust task allocation according to the actual capability of the node, and continuously monitoring the node resource occupation during task execution and formulating a resource adjustment scheme can ensure that the high-load or delay node obtains resource support in time, avoids system instability caused by insufficient resources, and further monitors the node task execution time after executing the resource adjustment to optimize resource allocation again to form a closed-loop feedback mechanism, so as to adapt to the dynamic change demand in a complex environment and significantly improve the stability and adaptability of the system, reduce task delay and resource waste, and accurately evaluate the node capability to realize dynamic optimization of task distribution, improve the resource utilization rate, and ensure the reliable operation of the system in a high-load scenario, and further enhance the adaptive ability of the system through secondary adjustment of the execution time, and finally realize efficient and stable task control and resource management. BRIEF DESCRIPTION OF DRAWINGS

[0024] The accompanying drawings, which are incorporated into and form part of the specification, illustrate embodiments consistent with the present application and, together with the specification, serve to explain the principles of the application.

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed in the embodiments or prior art description will be briefly introduced. Obviously, those skilled in the art can obtain other drawings from these drawings without creative labor.

[0026] Figure 1 is a flow diagram of a task control method based on real-time communication according to an embodiment of the present application;

[0027] Figure 2 is a structural block diagram of a task control device based on real-time communication according to an embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to make the technical personnel in the art better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should belong to the scope of protection of the present application.

[0029] It is to be understood that the terms "first", "second", and the like, used in the description and the claims of the present application as well as the above description of the drawings merely refer to structure that is different, and not necessarily to an order or sequence. It is to be understood that the terms "including", "containing", and / or "having", used in the description and the claims of the present application as well as the above description of the drawings, are used in the inclusive inoperative sense and not the exclusive inoperative sense. It is to be understood that the terms "including", "containing", and / or "having" should be interpreted as "comprising".

[0030] In the present embodiment, a task control method based on real-time communication is provided, which can be applied in a distributed system having a plurality of nodes (for computing or storage), Figure 1 is a flow diagram of a task control method based on real-time communication according to an embodiment of the present application, as shown in Figure 1 The flow includes the following steps S102-S108:

[0031] Step S102: determining a node capability score table corresponding to N nodes in a distributed system according to a state information set corresponding to the N nodes, wherein the state information set includes real-time acquired state information of each of the N nodes, and N is an integer greater than or equal to 2;

[0032] For example, in a multi-node environment, real-time acquisition of the state of each node can be achieved by assuming a distributed system composed of 10 server nodes for processing large-scale data storage tasks. Each node is responsible for different data shards. The state information includes but is not limited to state update records, which contain timestamps and running state data, including but not limited to CPU usage, memory occupancy, and network latency. A preset big data collection tool, such as a script that runs every 5 minutes, is used to collect these data at regular intervals to form a preliminary state information set, i.e., a basic state record. Such a collection frequency can ensure real-time performance while avoiding excessive system burden. Alternatively, after the state information is transmitted to the local cache, it can be classified and organized according to the storage rules. For example, CPU usage is classified as a performance indicator group, and network latency is classified as a communication indicator group, and they are respectively stored in different partitions of the database to form a classified state information group.

[0033] Optionally, in the case of missing CPU data of a certain node in the performance indicator group, the system will reacquire data from the node through the backup acquisition channel, and use the comparison tool to verify the difference between the supplementary data and the existing data to ensure consistency and form a complete state information combination. This dual-channel mechanism significantly improves the reliability of the data.

[0034] In one embodiment, when performing the aggregation process for the complete state information combination, assuming that the integration tool finds that two nodes report conflicting network delay data in the same time period, one is 50 milliseconds and the other is 80 milliseconds. At this time, the data with the newer timestamp can be selected as the final result according to the timestamp priority rule. This rule effectively avoids the confusion caused by data conflicts and ensures the accuracy of the data set.

[0035] It should be noted that the distributed mechanism of multi-node state real-time acquisition can reduce the risk of single-point failure through distributed acquisition. For example, if a certain node is temporarily unavailable, the backup channel can obtain indirect data from adjacent nodes to ensure the continuity of the overall monitoring. Classified storage facilitates subsequent analysis, such as quickly locating performance bottlenecks. The data integration and conflict resolution mechanism further improves the reliability of the data, providing a solid foundation for system optimization. For example, in actual application, assuming that the CPU usage of a certain node is continuously higher than 90%, the classified state information group can quickly locate the problem, and whether it is caused by network delay or not can be analyzed by combining the complete state information combination. This whole-process design from acquisition to integration not only improves the monitoring efficiency of industrial monitoring software, but also provides accurate basis for fault troubleshooting and resource scheduling, significantly enhancing the stability and response capability of the system.

[0036] In one exemplary embodiment, before performing the above step S102, the method further includes steps S11-S13:

[0037] Step S11: determining a delay node list according to the state information set, wherein the delay node list records the identification, node type and latest state update time of M nodes in the N nodes; each node in the M nodes has a state update time interval exceeding a preset time interval, and M is an integer greater than or equal to 1 and less than or equal to N; the state information of a node includes the state update record of the node;

[0038] In one exemplary embodiment, the above step S11 can be performed by steps S111-S113:

[0039] Step S111: determining update frequency data corresponding to the N nodes according to the state information set, wherein the update frequency data includes the update time interval of each node in the N nodes in a preset time period;

[0040] Step S112: determining the M nodes from the N nodes according to the update frequency data;

[0041] In an exemplary embodiment, in a multi-node cooperation scenario, the processing of the state information set and the identification of the delayed nodes can be analyzed in detail from multiple perspectives. For the timestamp comparison of the state update records, assuming that in a distributed storage system consisting of 8 nodes, the state update records of each node contain timestamps and running state data. The timestamp comparison tool checks the update time interval of each node within the past 1 hour one by one to obtain the update frequency data. If a node has an update interval of 15 minutes, which is obviously more than the preset threshold of 10 minutes, under the normal condition that each node updates its state every 5 minutes, the node is marked as a potential delayed node (i.e., the M nodes mentioned above). This comparison method can quickly screen out nodes with abnormal updates and lay the foundation for subsequent analysis.

[0042] Optionally, for the step of identifying abnormal delayed nodes, the industrial monitoring software can be used to further confirm the problem. Assuming that a marked node has multiple update interval exceeding conditions within the last 24 hours, the industrial monitoring software analyzes its state log and finds that the node frequently goes offline in a specific time period. In this way, it is verified again that it is a delayed node. The detailed monitoring method of the industrial monitoring software helps to accurately locate the problem node and avoid misjudgment.

[0043] Step S113: using a data timeliness verification tool to analyze the state update records of the M nodes to determine the node type of each node in the M nodes according to the continuity and integrity of each state update of each node in the M nodes, and to determine the latest state update time of each node, wherein the node type includes a first delayed type node and a second delayed type node; the delay degree of the second delayed type node is greater than the delay degree of the first delayed type node.

[0044] In an exemplary embodiment, the data timeliness verification tool analyzes the latest state update records of the delayed nodes. Assuming that a node has a long update interval, but its state update records show that the update data is complete and continuous each time, which may be due to a small amount of tasks resulting in a low update frequency, and it is classified as a non-urgent delayed node. Another node has intermittent update records, and the latest update data is missing, and it is classified as an urgent delayed node. This classification method can distinguish different types of delay problems and provide a basis for subsequent processing.

[0045] For example, node N3 is marked as an emergency delay node with a last update time of 2 hours ago, while node N5 is a non-emergency delay node with an update time of 30 minutes ago. Such a delay node list facilitates managers to quickly understand the status of each node and take targeted measures.

[0046] It should be noted that the implementation methods of each of the above subjects are closely related in a distributed system, from timestamp comparison to the generation of the final delay node list, forming a complete delay node identification process. Each link pays attention to detailed analysis and classification processing of data to ensure accurate marking of problem nodes. This whole-process design can provide clear guidance for system maintenance, especially in a multi-node collaboration environment, which helps to quickly respond to potential problems and improve overall operational efficiency.

[0047] Step S12: updating the collection frequency of the state information of the M nodes according to the delay node list;

[0048] Step S13: updating the state information set according to the state information of the M nodes obtained by the updated collection frequency.

[0049] In an exemplary embodiment, the above step S12 can be implemented by: determining the priority of each node in the M nodes in the delay node list according to the last state update time of each node; and adjusting the collection frequency of the state information of each node in the M nodes according to the priority of each node in the M nodes.

[0050] For example, in the delay node management scenario of a distributed storage system, assume a system composed of 10 nodes, of which 3 nodes are marked as delay nodes. Assume that the last state update time of node N1 is 3 hours ago, that of node N2 is 1 hour ago, and that of node N3 is 5 hours ago, then the task priority of N3 is the highest, that of N1 is the second, and that of N2 is the lowest. Further, the frequency adjustment tool will dynamically manage the data collection frequency. Assume that node N3 in the high-priority queue originally collects data every 30 minutes, and the tool will adjust it to collect data every 10 minutes, while the low-priority node N2 remains to collect data every 30 minutes. Through this adjustment, the system can pay more attention to the state changes of serious delay nodes and optimize resource allocation. Alternatively, for the adjusted collection frequency, the industrial data collection tool will perform periodic grabbing of the state information of the delay nodes. Assume that node N3 grabs state data every 10 minutes after adjusting the frequency, and obtains the latest state information when performing the update task. Further analyze the integrity of the state information set, if the data coverage rate reaches more than 90% of the preset, it is considered to meet the requirements. This periodic grabbing method can ensure the timeliness of the data and provide a reliable basis for subsequent updates.

[0051] It should be noted that in a distributed system, data collection control is a key means to achieve efficient monitoring of node status and optimal allocation of resources. Through data collection control, the system can dynamically adjust the data collection frequency of high-priority nodes (such as N3), for example, from every 30 minutes to every 10 minutes, so as to pay more attention to the status changes of severely delayed nodes in a timely manner. At the same time, low-priority nodes (such as N2) maintain a lower collection frequency to reduce resource consumption. Data collection control allocates collection resources reasonably according to the results of task priority sorting and frequency adjustment tools, avoiding resource waste or insufficient monitoring caused by uniform frequency for all nodes. This mechanism ensures that the system can quickly respond to problems of high-priority nodes, and optimizes the overall resource configuration through data collection control, improving the stability and efficiency of the system.

[0052] In an exemplary embodiment, the above step S102 can be implemented by the following steps S21-S23:

[0053] Step S21: processing the set of state information by a big data processing tool to determine the task execution performance data corresponding to each of the N nodes, wherein the task execution performance data includes task completion time, response speed; the state information of the node includes the running record of the node;

[0054] For example, in the delay node management scenario of a distributed storage system, the processing and analysis of the set of state information can be performed by the big data processing tool to sort the running records of each node in layers. Assuming that a system contains 8 nodes, the big data processing tool will classify and sort the running records of the nodes according to time period, task type, etc. dimensions, and extract historical task execution performance data related to task execution, such as task completion time, response speed, etc.

[0055] Step S22: determining the priority of the N nodes in executing tasks according to the task execution performance data corresponding to each of the N nodes, and determining a task distribution sequence according to the priority of the N nodes in executing tasks;

[0056] For example, assuming that the historical record of node N4 shows that its average task completion time is 2 hours, while that of node N5 is 4 hours, the preliminary ability assessment result will reflect that the execution ability of N4 is stronger than that of N5. This hierarchical arrangement helps to clearly understand the performance differences of each node. In the matching process of the preliminary ability assessment result and the preset threshold, the data comparison tool will compare the assessment result with the preset standard. Assuming that the preset threshold is that the task completion time should not exceed 3 hours, then the assessment result of node N5 is below the threshold, which needs further attention. The big data information screening tool will extract the state information of N5, including the reason for the delay, the resource occupation, etc., to determine the priority ranking of N5 in the instruction distribution. Assuming that the delay of N5 is caused by insufficient resources, the tool will give priority to N5 to allocate resources to support it.

[0057] For example, in the task allocation link, the task allocation tool will dynamically adjust the order of instruction distribution according to the priority ranking. Assuming that there are 3 nodes in the system that need to process tasks, N5 has the highest priority, N4 has the second highest priority, and N6 has the lowest priority, the tool will preferentially distribute tasks to N5 to ensure that its resource problem is solved. The adjusted task distribution sequence is N5, N4, N6, which ensures that the key node is given priority to support.

[0058] Step S23: Through the big data information integration tool, the state information and task execution performance data of the N nodes after the N nodes execute the tasks according to the task distribution sequence are associated and matched, and the ability score of each of the N nodes is determined to obtain a node ability score table corresponding to the N nodes.

[0059] For example, assuming that the execution efficiency of N5 is improved after the task distribution, and the completion time is shortened to 2.5 hours, the big data information integration tool will integrate the state information and efficiency data of N5 to generate a score table, which shows that the suitability of N5 has improved. In the score table, the score of N4 may be 85, the score of N5 may be 75, and the score of N6 may be 90, and the system can determine the suitability of each node in subsequent instruction distribution according to the score table. This integration method helps to provide a reference for subsequent task allocation.

[0060] It should be noted that from the overall process, each of the above links is closely related to the management of the delayed node, from data arrangement to ability assessment, to task distribution and score table generation, forming a complete analysis and optimization chain. Assuming that the delay problem of node N5 is significantly alleviated after optimization, the overall task processing efficiency of the system is improved. This whole-process management method not only can solve the node problem in a targeted manner, but also can provide data support for system resource allocation to ensure the rationality of task distribution.

[0061] Step S104: determining a target task publishing scheme according to the node capability score table, wherein the target task publishing scheme has a task amount and a task distribution order for each of the N nodes;

[0062] In an exemplary embodiment, the above step S104 can be implemented by the following steps S31-S33:

[0063] Step S31: determining a task matching degree of each of the N nodes according to a node state of each of the N nodes and a capability score of each of the N nodes in the node capability score table, wherein the node state comprises a resource occupation of the node, and the task matching degree is used to represent a matching degree with tasks of different priorities;

[0064] It should be noted that in a distributed system, the capability of each node needs to be quantitatively evaluated through a node capability score table first. The score table is generated based on multi-dimensional data such as historical performance, current resource occupation, and task execution efficiency of the node. At the same time, the industrial monitoring software collects node state information of each node, which mainly includes resource occupation (such as CPU utilization, memory usage, network bandwidth, etc.). These state information combined with the node capability score can further analyze the matching degree between each node and tasks of different priorities.

[0065] The task matching degree is a key indicator to measure whether a node is suitable for processing a specific priority task. For example, assuming that the capability score of node N1 is 75, which is lower than the preset threshold of 80, but its current resource occupation is low, it is considered more suitable for processing high-priority tasks to improve running efficiency. While node N2 has a score of 85, which is higher than the threshold, but if its resource occupation is high, it is more suitable for assigning low-priority tasks to avoid overload. This matching method not only considers the capability score of the node, but also combines its current state, thereby achieving dynamic balance.

[0066] Step S32: determining an initial task distribution scheme according to the task matching degree of each of the nodes, and adjusting the task amount of each of the nodes in the initial task distribution scheme according to the capability score of each of the nodes to obtain an adjusted task distribution scheme;

[0067] For example, based on the calculation result of the above task matching degree, the system can initially formulate a task distribution scheme. The task distribution scheme in this stage may be generated based on static rules or empirical values, but needs to be further optimized to adapt to actual needs. For example, in a system containing 6 nodes, the initial scheme may allocate 10% of the task amount to node N1, 20% to node N3, and so on.

[0068] Next, the system will adjust the task allocation ratio according to the node capability score table. For example, node N1 scores 75 points, although it is initially allocated 10% of the task, but due to its low score and sufficient resources, the system may increase its task allocation to 15% to provide more support. For node N3, which scores 90 points, is initially allocated 20% of the task, but considering its high resource occupation, it may reduce its task allocation to 15% to avoid overload. This adjustment can make the resource allocation more in line with the actual ability of the node, ensuring the rationality of task allocation.

[0069] Step S33: updating the task distribution order in the adjusted task distribution scheme according to the preset task publishing rule to obtain the target task publishing scheme.

[0070] It should be noted that after completing the task adjustment, the system will further optimize the task distribution order according to the preset task publishing rule. The preset task publishing rule clearly defines the logic of task distribution order, for example: low-level marked nodes prefer to handle critical tasks. High-score nodes prefer to handle ordinary tasks. Dynamic sorting is combined with task priority and node state.

[0071] For example, assume that node N1 is marked as low-level priority sorting, but it is currently resource sufficient and needs to handle critical tasks, it can be arranged in the front row of the distribution order. High-score nodes such as N3 may be arranged at the back to ensure the overall stability of the system. The final generated target task publishing scheme not only reflects the dynamic balance of node state and task demand, but also guarantees the priority logic of task execution. In addition, the system also considers the historical performance and current state of the node, reasonably arranges the task quantity, avoids the occurrence of repeated problems, and further improves the accuracy of task distribution and the stability of the overall operation of the system.

[0072] In an exemplary embodiment, assume that node N1 has improved its task completion efficiency after adjustment, and its score gradually approaches the threshold from 75 points, the system will adjust its allocation ratio again according to real-time data. This dynamic adjustment mechanism can continuously optimize resource allocation efficiency.

[0073] Step S106: monitoring the resource occupation of each of the N nodes during the process of publishing tasks according to the target task publishing scheme, and determining a resource adjustment scheme according to the monitored resource occupation of each node, wherein the resource adjustment scheme is used to adjust the resource occupation of the N nodes;

[0074] In an exemplary embodiment, the above determination of the resource adjustment scheme according to the monitored resource occupation of each node can be realized by the following steps S41-S42:

[0075] Step S41: determining P nodes from the N nodes according to the resource occupation of each node in the N nodes, wherein the resource occupation rate of each node in the P nodes is greater than a preset occupation rate, and P is an integer greater than or equal to 1 and less than or equal to N;

[0076] It should be noted that in a distributed system, the resource occupation of each node needs to be obtained through a big data collection tool first. These data include but are not limited to CPU utilization, memory usage, network bandwidth, and other indicators. Through real-time collection of the resource occupation of the nodes by industrial monitoring software, a detailed resource distribution data table can be generated.

[0077] Next, the system will compare the resource occupation rate of each node with the preset balancing threshold (such as 70%). If the resource occupation rate of a node exceeds the threshold, it will be marked as a high-load node. For example, in a system containing 5 nodes, assuming that the resource occupation rate of node N1 is 90%, and the resource occupation rates of the other nodes are 40%, 50%, 60%, and 30% respectively, node N1 is included in the list of high-load nodes (i.e. the above-mentioned P nodes) because it exceeds the preset threshold of 70%. The core of this step is to quickly locate the nodes with unbalanced resource allocation through core indicator extraction and comparison, ensuring the accuracy of the subsequent adjustment direction.

[0078] Step S42: determining that the resource adjustment scheme is a first resource adjustment scheme, wherein the first resource adjustment scheme is to reduce the task quantity of each node in the P nodes in the target task publishing scheme, and to increase the reduced task quantity to one or more nodes with the lowest resource occupation rate in the N nodes.

[0079] It should be noted that once the P high-load nodes are determined, the system will develop a first resource adjustment scheme. The core idea of this scheme is to alleviate the pressure on high-load nodes by redistributing task quantities and improve the balance of overall resource utilization.

[0080] Specifically, the first resource adjustment scheme requires reducing the task quantity of the P high-load nodes in the target task publishing scheme. For example, assuming that node N1 is marked as a high-load node, and its current task quantity accounts for 30%, the system will reduce its task quantity to 20% according to the rules. This reduction operation aims to reduce the workload of high-load nodes and avoid task delays or failures due to insufficient resources.

[0081] Meanwhile, the system will increase the amount of tasks transferred from the high-load node to the node or nodes with the lowest resource occupancy rate. For example, in the above example, the resource occupancy rate of node N2 is only 40%, which is much lower than that of other nodes. Therefore, the system may transfer part of the tasks from node N1 to node N2. This task transfer not only effectively relieves the pressure on the high-load node, but also fully utilizes the idle resources of the low-load node, thereby optimizing the allocation of resources.

[0082] It should be noted that the adjustment ratio of the amount of tasks needs to be calculated in combination with the resource adjustment rule. For example, assuming that the amount of tasks of node N1 is reduced from 30% to 20%, the reduced 10% amount of tasks may be evenly distributed to nodes N2 and N5 with lower resource occupancy rates (each increased by 5%). This adjustment method ensures the rationality and fairness of task allocation, avoiding the situation where a single node bears too much new tasks.

[0083] Finally, the first resource adjustment scheme achieves the relief of the pressure on the high-load node and the improvement of the resource utilization of the low-load node by dynamically adjusting the amount of tasks. This step not only improves the overall performance of the system, but also lays a foundation for subsequent more detailed resource adjustment.

[0084] In an exemplary embodiment, when the first resource adjustment scheme is verified again, the big data processing tool can perform in-depth analysis on the task distribution and resource occupancy data. Assuming that the occupancy rate of node N1 after adjustment is reduced to 75%, which is still slightly higher than the threshold, the system will further adjust the task matching priority of node N1 and transfer part of the low-priority tasks to other nodes. This secondary verification mechanism can ensure the rationality of resource allocation, avoid potential risks caused by single adjustment, and improve the stability of task execution.

[0085] For example, in the process of determining the final resource adjustment result, the system will comprehensively consider the historical load data and the current state of the nodes. Assuming that the occupancy rate of node N1 stabilizes below the threshold after multiple adjustments, the system will record the adjustment process as a reference for subsequent optimization. This comprehensive consideration not only improves the accuracy of resource adjustment, but also provides a guarantee for the stability of long-term operation of the system. Through dynamic adjustment and verification, the resource allocation scheme can better adapt to the changes in the state of the nodes.

[0086] In an exemplary embodiment, after determining that the resource adjustment scheme is the first resource adjustment scheme, the method further includes the following steps S43-S44:

[0087] Step S43: After executing the first resource adjustment scheme, if the resource occupancy rate of any target node in the N nodes is greater than the preset occupancy rate, a standby resource is obtained from a standby resource pool, and the standby resource is allocated to the target node.

[0088] It should be noted that after the first resource adjustment scheme is executed, if it is found that the resource occupancy rate of some target nodes among the N nodes is still greater than the preset occupancy rate (for example, the resource occupancy rate of node A is still 70%, which is higher than the threshold of 65%), further measures need to be taken to relieve the pressure of these target nodes. At this time, the system will enter step S43, that is, obtaining standby resources from the standby resource pool and allocating them to the target nodes.

[0089] Specifically, the standby resource pool is a set of idle resources prepared by the system in advance to deal with sudden situations or resource shortages. Through the resource pool management tool, the available resource data in the standby resource pool can be queried. For example, there are 3 available resource units in the resource pool, each with different performance parameters, and the system will select a set of standby resources that meet the conditions according to the needs of the target nodes. This selection process ensures that the allocated resources can meet the actual needs of the target nodes.

[0090] Once the set of standby resources that meet the conditions is determined, the system will use the allocation process tool to match these standby resources with the target nodes. For example, assume that after target node A is allocated a standby resource unit, its resource occupancy rate is expected to drop to 60%, which is lower than the preset threshold of 65%, thereby significantly relieving the pressure. This matching method not only improves the running stability of the target nodes, but also provides basic data for subsequent verification.

[0091] It should be noted that the allocation process of standby resources needs to be dynamically adjusted in combination with the specific state of the target nodes and the availability of the resource pool. For example, if the task volume of target node A surges during peak hours, the system can avoid the impact of task delays on the overall system through timely feedback from the monitoring tool and rapid response of the resource pool. In addition, if the standby resource pool is insufficient, an expansion scheme can also be considered, such as transferring part of the low-priority tasks of the target nodes to other nodes to further balance the load distribution.

[0092] Step S44: determining that the resource adjustment scheme is a first resource adjustment scheme and a second resource adjustment scheme, wherein the second resource adjustment scheme is the allocation of standby resources to the target nodes.

[0093] It should be noted that after step S43 is completed, the system needs to determine the final resource adjustment scheme. At this time, the resource adjustment scheme includes two parts: the first resource adjustment scheme and the second resource adjustment scheme. The first resource adjustment scheme refers to relieving the pressure of high-load nodes by reallocating task volume, while the second resource adjustment scheme refers to allocating standby resources to target nodes.

[0094] Specifically, the first resource adjustment scheme has been implemented in step S42 by downgrading the task amount of high-load nodes and increasing the task amount to nodes with lower resource occupancy rates, achieving preliminary resource balancing. However, if the resource occupancy rate of the target node is still higher than the preset threshold, the second resource adjustment scheme needs to be further implemented, i.e., allocating a standby resource to the target node.

[0095] For example, in the node management scenario of a distributed storage system, assume that after the first resource adjustment scheme, the resource occupancy rate of node A decreases from 85% to 70%, but is still higher than the preset threshold of 65%. At this time, the system will allocate a standby resource unit from the standby resource pool to node A, so that its resource occupancy rate is further reduced to below 60%. This double adjustment mechanism ensures the rationality of resource allocation and avoids potential problems caused by single adjustment.

[0096] Finally, the system will comprehensively analyze the load distribution information of the nodes and the utilization of the standby resources, and record the entire adjustment process as a subsequent reference. For example, assume that node A runs stably after multiple adjustments, and the standby resource pool still has redundancy, the system will optimize the management efficiency of the resource pool to support long-term stable operation. In addition, from a business perspective, this mechanism is particularly important in the distributed storage scenario, which can guarantee the continuity of data access and the overall performance of the system.

[0097] In summary, step S44 clearly defines the complete content of the resource adjustment scheme, including both the first resource adjustment scheme achieved by task redistribution and the second resource adjustment scheme achieved by standby resource allocation. This double strategy ensures that the system can flexibly cope with various complex scenarios, improving resource utilization efficiency and task execution stability.

[0098] It should be noted that in the process of formulating the resource adjustment scheme, data acquisition control plays a crucial role. First, through data acquisition control, the resource occupation of each node is monitored in real time, and P high-load nodes with resource occupation rate greater than the preset occupation rate are screened out. For example, in a system containing N nodes, the data acquisition control tool will continuously collect the resource usage data such as CPU and memory of each node, and if it is found that some nodes (such as P) exceed the preset threshold, they will be marked as high-load nodes. Then, when determining the first resource adjustment scheme, data acquisition control will combine the target task publishing scheme, reduce the task quantity of P high-load nodes, and allocate the reduced task quantity to one or more nodes with the lowest resource occupation rate. This adjustment relies on the dynamic data support provided by data acquisition control to ensure the rationality of task redistribution. Further, after the execution of the first resource adjustment scheme, if the resource occupation rate of the target node is still higher than the preset value, data acquisition control will continue to play a role. By obtaining standby resources from the standby resource pool and allocating them to these target nodes, a second resource adjustment scheme is formed. Throughout the process, data acquisition control is not only used for real-time monitoring of node status, but also provides a basis for the rational allocation of standby resources. The repeated data acquisition control mechanism ensures the accuracy of the resource adjustment scheme and the stability of the system.

[0099] Step S108: After executing the resource adjustment scheme, the execution time of each node in the N nodes for executing a task is monitored, and resource adjustment is performed again according to the monitored execution time of each node for executing a task.

[0100] In an example embodiment, the above step S108 includes steps S51-S52:

[0101] Step S51: Determine Z nodes from the N nodes, wherein the execution time of each node in the Z nodes is greater than a preset execution time, and Z is an integer greater than or equal to 1 and less than or equal to N;

[0102] It should be noted that in a distributed system, the execution time and feedback state of each node need to be collected in real time through industrial monitoring software first. These data include instruction response time, task completion status, and other key indicators.

[0103] Next, the system will compare the execution time of each node with the preset threshold (such as 1.2 seconds). If the execution time of a node exceeds the threshold, it will be marked as a node to be optimized. For example, in a system containing 5 nodes, assuming that the execution time of node B is 2.5 seconds, which is much higher than the preset threshold of 1.2 seconds, node B is included in the list of nodes to be optimized.

[0104] Finally, Z nodes are selected from all N nodes, where the execution time of each node is greater than the preset execution time. These nodes constitute the list of nodes to be optimized. The core of this step is to quickly locate the performance bottleneck node through core index extraction and comparison, ensuring the accuracy of the subsequent adjustment direction.

[0105] Step S52: Resource adjustment according to the resource configuration state of the Z nodes and the allowed resource information.

[0106] It should be noted that once the Z nodes to be optimized are determined, the system needs to query the current resource configuration status of these nodes. Through the resource management tool, the available resource information can be obtained from the configuration data. For example, assume that node B currently has a resource configuration showing that its load is concentrated on storage tasks, while the resource utilization of other nodes in the system is low, and the resource management tool can preliminarily determine the range of available resources.

[0107] Next, the system will use the load balancing tool to check the execution efficiency of each node. Based on the temporarily adjusted load distribution data, it is evaluated whether the execution efficiency of each node meets the preset standard. For example, assume that the execution time of node B is reduced to 1.8 seconds after adjustment, but it is still lower than the preset efficiency threshold of 1.5 seconds, and the system will allocate additional resources from the standby resource pool to support the task processing of node B.

[0108] If the preliminary adjustment fails to meet the preset standard, a dynamic adjustment mechanism is further adopted. For example, the system will add a resource unit from the standby resource pool to support the task processing of node B, and update the parameter scheme. This dynamic adjustment mechanism can flexibly respond to node performance fluctuations, ensuring the continuity and stability of task execution.

[0109] Finally, the stability of the instruction execution is continuously tracked through industrial monitoring software. For example, assume that the stability index of node B is reduced from the previous fluctuation rate of 0.3 to 0.1 under the support of the newly added resources, which meets the preset requirement range. The industrial monitoring software will continuously record the relevant data to determine whether further optimization is needed, and finally determine the resource configuration scheme.

[0110] In addition, as an extended solution, if the standby resource pool is insufficient, the system can also transfer part of the non-urgent tasks to other low-load nodes through task migration. For example, part of the tasks of node B are transferred to nodes C and D, so as to further balance the system load and improve the overall running efficiency. This way not only provides more flexibility for resource management, but also effectively avoids the problem of overall performance decline caused by the overload of a single node.

[0111] In an exemplary embodiment, after resource adjustment according to the monitored execution time of each node performing tasks, the method further comprises: continuously monitoring the stability of task execution of the N nodes by industrial monitoring software; determining the stability index of each node in the N nodes according to the monitored monitoring data, to obtain an execution performance report.

[0112] It should be noted that after completing the resource adjustment of the execution time of each node performing tasks, the system enters the continuous monitoring phase of performance stability. In this phase, the industrial monitoring software tracks the stability of task execution of the N nodes in real time, ensuring that the adjusted resource configuration can meet the system operation requirements.

[0113] The industrial monitoring software continuously collects task execution data of each node, including instruction response time, task completion status and other key indicators. Through analysis of these data, a detailed monitoring data set is generated for evaluating the stability performance of each node.

[0114] Based on the monitoring data, the system calculates and determines the stability index of each node. For example, assume that the execution time of node A is stabilized at 1.4 seconds after adjustment, and the volatility rate is reduced from 0.25 to 0.1, which indicates that its stability has reached the preset requirement range. The calculation of the stability index considers factors such as the mean, variance and fluctuation amplitude of the execution time, ensuring a comprehensive reflection of the actual running state of the node.

[0115] According to the stability index of all nodes, the system generates a final execution performance report. This report not only records the stability performance of each node, but also provides performance evaluation results of the overall system. For example, if the execution time of all nodes is lower than the preset threshold (such as 1.5 seconds) and the volatility rate meets the requirements, it means that the current resource configuration scheme is effective.

[0116] If the stability index of some nodes still does not meet the preset standard, the system will trigger further optimization process. For example, it may be necessary to reassign tasks or call additional resources in the standby resource pool to further improve node performance. This closed-loop mechanism ensures that the system can adapt flexibly in complex environments and avoid affecting the overall performance due to single node problems.

[0117] In summary, through continuous tracking by industrial monitoring software and quantitative evaluation of stability index, the system can timely discover potential problems and take measures to ensure the continuity of task execution and the smoothness of data access, providing reliable operation guarantee for distributed storage scenarios.

[0118] The above steps generate a node capability score table by acquiring state information of each node in the distributed system in real time, dynamically formulate a target task release scheme based on the score table, and determine the task amount and distribution sequence of each node, which solves the problem that the traditional method cannot dynamically adjust task allocation according to the actual capability of the node. Meanwhile, continuously monitoring the node resource occupation during task execution and formulating a resource adjustment scheme can ensure that the high-load or delay node obtains resource support in time, avoid system instability caused by insufficient resources, and further monitor the node task execution time after executing the resource adjustment to optimize resource allocation again to form a closed-loop feedback mechanism, so as to adapt to dynamic changes in complex environments and significantly improve the stability and adaptability of the system, reduce task delay and resource waste. The application can accurately evaluate the node capability to realize dynamic optimization of task distribution, improve the resource utilization rate, ensure the reliable operation of the system in a high-load scenario, and further enhance the adaptive ability of the system through secondary adjustment of the execution time, and finally realize efficient and stable task control and resource management.

[0119] Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be realized by means of software and the necessary general hardware platform, of course, it can also be realized by hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical application of the application can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes a plurality of instructions for making a terminal device (which can be a mobile phone, computer, server, or network device) execute the method of each embodiment of the application.

[0120] In the present embodiment, a task control device based on real-time communication is also provided, which is used to implement the above-mentioned embodiments and preferred embodiments, and will not be described again. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, hardware or a combination of software and hardware is also possible and is contemplated.

[0121] Figure 2 is a structural block diagram of a task control device based on real-time communication according to an embodiment of the application, and the device comprises:

[0122] The first determination module 202 is configured to determine a node capability score table corresponding to N nodes in a distributed system according to a state information set corresponding to the N nodes, wherein the state information set comprises real-time acquired state information of each of the N nodes, and N is an integer greater than or equal to 2;

[0123] The second determining module 204 is configured to determine a target task publishing scheme according to the node capability score table, wherein the target task publishing scheme includes the amount of tasks undertaken by each of the N nodes and the task distribution sequence.

[0124] The third determining module 206 is configured to monitor the resource occupation of each of the N nodes during the process of publishing tasks according to the target task publishing scheme, and determine a resource adjustment scheme according to the monitored resource occupation of each of the N nodes, wherein the resource adjustment scheme is used to adjust the resource occupation of the N nodes.

[0125] The adjusting module is configured to monitor the execution time of each of the N nodes for executing tasks after the execution of the resource adjustment scheme, and perform resource adjustment again according to the monitored execution time of each of the N nodes for executing tasks.

[0126] The above device generates a node capability score table by real-time acquisition of state information of each node in a distributed system, dynamically formulates a target task publishing scheme based on the score table, and determines the amount of tasks undertaken by each node and the task distribution sequence, thereby solving the problem that the traditional method cannot dynamically adjust task allocation according to the actual capability of a node. Meanwhile, continuous monitoring of the resource occupation of a node during task execution and formulating a resource adjustment scheme can ensure that a high-load or delayed node obtains resource support in time, avoid system instability caused by insufficient resources, and further monitor the task execution time of a node after the execution of the resource adjustment to optimize resource allocation again, thereby forming a closed-loop feedback mechanism to adapt to dynamic changes in complex environments and significantly improve the stability and adaptability of the system, reduce task delay and resource waste, accurately evaluate the capability of a node to achieve dynamic optimization of task distribution, improve the resource utilization rate, and ensure reliable operation of the system in a high-load scenario. Furthermore, the second adjustment of the execution time further enhances the adaptive ability of the system, and finally realizes efficient and stable task control and resource management.

[0127] Embodiments of the present application also provide a computer readable storage medium having a computer program stored therein, wherein the computer program is configured to execute the steps in any of the method embodiments described above when running.

[0128] Optionally, in the present embodiment, the above storage medium can be configured to store a computer program for executing the following steps.

[0129] In an example embodiment, the computer readable storage medium described above can include, but is not limited to, a U disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store computer programs.

[0130] The specific examples in the embodiments can refer to the examples described in the above embodiments and example implementations, which will not be repeated here.

[0131] The embodiments of the present application also provide a computer program product, which includes a computer program, and the computer program is used to execute the steps in any of the method embodiments described above when executed by a processor.

[0132] The embodiments of the present application also provide another computer program product, which includes a non-volatile computer readable storage medium, and the non-volatile computer readable storage medium stores a computer program, and the computer program is used to execute the steps in any of the method embodiments described above when executed by a processor.

[0133] The embodiments of the present application also provide an electronic device, which includes a memory and a processor, the memory stores a computer program, and the processor is configured to execute the steps in any of the method embodiments described above by using the computer program.

[0134] The specific examples in the embodiments can refer to the examples described in the above embodiments and example implementations, which will not be repeated here.

[0135] Obviously, those skilled in the art should understand that the modules or steps of the present application described above can be realized by using general computing devices, which can be concentrated on a single computing device or distributed on a network composed of multiple computing devices, and they can be realized by using program codes executable by computing devices, so that they can be stored in storage devices and executed by computing devices, and in some cases, the steps shown or described can be executed in different orders, or they can be manufactured into individual integrated circuit modules, or multiple modules or steps can be manufactured into a single integrated circuit module. Therefore, the present application is not limited to any specific combination of hardware and software.

[0136] The above only describes the preferred embodiments of the present application, and it should be noted that, for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A task control method based on real-time communication, characterized by, The method comprises the following steps: determining a node capability score table corresponding to N nodes in a distributed system according to a set of state information corresponding to the N nodes, wherein the set of state information comprises real-time acquired state information of each of the N nodes, and N is an integer greater than or equal to 2; determining a target task publishing scheme according to the node capability score table, wherein the target task publishing scheme comprises a task amount and a task distribution sequence of each of the N nodes; monitoring resource occupation of each of the N nodes in a process of publishing tasks according to the target task publishing scheme, and determining a resource adjustment scheme according to the monitored resource occupation of each of the N nodes, wherein the resource adjustment scheme is used to adjust the resource occupation of the N nodes; monitoring execution time of each of the N nodes in executing tasks after the resource adjustment scheme is executed, and adjusting resources again according to the monitored execution time of each of the N nodes in executing tasks; wherein determining a node capability score table corresponding to N nodes in a distributed system according to a set of state information corresponding to the N nodes comprises: processing the set of state information by a big data processing tool to determine task execution performance data corresponding to each of the N nodes, wherein the task execution performance data comprises task completion time and response speed, and the state information of a node comprises a running record of the node; determining a priority of the N nodes in executing tasks according to the task execution performance data corresponding to each of the N nodes, and determining a task distribution sequence according to the priority of the N nodes in executing tasks; associating and matching state information and task execution performance data of the N nodes after the N nodes execute tasks according to the task distribution sequence by a big data information integration tool to determine a capability score of each of the N nodes, so as to obtain the node capability score table corresponding to the N nodes.

2. The method according to claim 1, characterized in that, Before determining a node capability score table corresponding to N nodes in a distributed system according to a set of state information corresponding to the N nodes, the method further comprises: determining a delay node list according to the set of state information, wherein the delay node list records an identifier, a node type and a latest state update time of M nodes of the N nodes, and each of the M nodes has a state update time interval exceeding a preset time interval, M is an integer greater than or equal to 1 and less than or equal to N, and the state information of a node comprises a state update record of the node; updating a collection frequency of collecting state information of the M nodes according to the delay node list; updating the set of state information according to the state information of the M nodes collected according to the updated collection frequency.

3. The method of claim 2, wherein, Determining a delay node list according to the set of state information comprises: determining update frequency data corresponding to the N nodes according to the set of state information, wherein the update frequency data comprises an update time interval of each of the N nodes in a preset time period; determining the M nodes from the N nodes according to the update frequency data. The data time verification tool is used to analyze the state update records of the M nodes to determine the node type of each of the M nodes according to the continuity and integrity of each of the M nodes each time the state is updated, and to determine the latest state update time of each of the M nodes, wherein the node type includes a first delay type node and a second delay type node; the delay degree of the second delay type node is greater than the delay degree of the first delay type node.

4. The method of claim 2, wherein, According to the delay node list, the acquisition frequency of the state information of the M nodes is updated, including: According to the latest state update time of the M nodes in the delay node list, the priority of each of the M nodes is determined; According to the priority of each of the M nodes, the acquisition frequency of the state information of each of the M nodes is adjusted.

5. The method of claim 1, wherein, According to the node capability score table, a target task publishing scheme is determined, including: According to the node state of each of the N nodes and the capability score of each of the N nodes in the node capability score table, the task matching degree of each of the N nodes is determined, wherein the node state includes the resource occupation of the node, and the task matching degree is used to represent the matching degree with tasks of different priorities; According to the task matching degree of each of the N nodes, an initial task distribution scheme is determined, and the task amount of each of the N nodes in the initial task distribution scheme is adjusted according to the capability score of each of the N nodes, to obtain an adjusted task distribution scheme; According to a preset task publishing rule, the task distribution order in the adjusted task distribution scheme is updated to obtain the target task publishing scheme.

6. The method of claim 1, wherein, According to the monitored resource occupation of each of the N nodes, a resource adjustment scheme is determined, including: According to the resource occupation of each of the N nodes, P nodes are determined from the N nodes, wherein the resource occupation rate of each of the P nodes is greater than a preset occupation rate, and P is an integer greater than or equal to 1 and less than or equal to N; The resource adjustment scheme is determined as a first resource adjustment scheme, wherein the first resource adjustment scheme is to reduce the task amount of each of the P nodes in the target task publishing scheme, and to increase the reduced task amount to one or more nodes with the lowest resource occupation rate in the N nodes.

7. The method of claim 6, wherein, After determining the resource adjustment scheme as the first resource adjustment scheme, the method further includes: After executing the first resource adjustment scheme, if the resource occupation rate of a target node in the N nodes is greater than the preset occupation rate, a standby resource is obtained from a standby resource pool and allocated to the target node; The resource adjustment scheme is determined as a first resource adjustment scheme and a second resource adjustment scheme, wherein the second resource adjustment scheme is that the standby resource is allocated to the target node.

8. The method of claim 1, wherein, According to the monitored execution time of each of the N nodes in executing tasks, resource adjustment is performed again, including: determining Z nodes from the N nodes, wherein an execution time of each node in the Z nodes is greater than a preset execution time, and Z is an integer greater than or equal to 1 and less than or equal to N; adjusting resources according to a resource configuration state of the Z nodes and allowed resource information.

9. The method of claim 1, wherein, After adjusting the resources again according to the monitored execution time of each node in executing a task, the method comprises: continuously monitoring stability of task execution of the N nodes by using industrial monitoring software; determining a stability index of each node in the N nodes according to monitored monitoring data, and obtaining an execution performance report.

Images