Search

CN-121996435-A - Auxiliary decision-based intelligent scheduling method and system for server resources

CN121996435ACN 121996435 ACN121996435 ACN 121996435ACN-121996435-A

Abstract

The invention relates to the technical field of server resource scheduling, and discloses an intelligent server resource scheduling method and system based on auxiliary decision. The method comprises the steps of setting an execution ending time point of a task in an abnormal execution category in task execution information as a reference time point, analyzing a change trend of a basic parameter set to extract an inflection point, dividing a plurality of servers according to a load evaluation value generated by combining the inflection point and a time difference to form a schedulable server set, selecting a server from the schedulable server set to be matched with the task to be executed and completing task distribution, calculating an adaptation score of a current task and the current server, and judging whether the current task is migrated to the server in the schedulable server set for execution. The method divides the servers by analyzing the variation trend of the execution parameters of the abnormal tasks, improves the accuracy of server state evaluation, judges whether to execute migration on the current task by adapting the scores, and ensures stable execution of the tasks.

Inventors

  • YANG ZHEN

Assignees

  • 北京安联通科技有限公司
  • 杨桢

Dates

Publication Date
20260508
Application Date
20260410

Claims (10)

  1. 1. The server resource intelligent scheduling method based on the auxiliary decision is characterized by comprising the following steps of: Dynamically generating a load evaluation value for each of a plurality of servers to divide the plurality of servers and form a schedulable server set, wherein the generation of the load evaluation value is based on analysis of execution parameters of abnormal tasks in task execution information; When the task to be executed exists, selecting a server from the schedulable server set to be matched with the task to be executed so as to complete task distribution; For the current task being executed, calculating the adaptation scores of the current task and the server where the current task is currently located, and determining that the current task is continuously executed at the server where the current task is currently located or is migrated to the server in the schedulable server set for execution according to the adaptation scores; The step of determining that the current task continues to be executed or migrated includes executing server migration if the adaptation score is lower than a preset adaptation score threshold, migrating the current task to a server in a schedulable server set for execution, repeatedly calculating the adaptation score at preset time intervals if the adaptation score is not lower than the adaptation score threshold, executing server migration if the adaptation score is lower than the adaptation score threshold in any one of continuous preset times of calculation, and otherwise, determining that the current task continues to be executed at the current server.
  2. 2. The intelligent scheduling method for server resources based on auxiliary decision-making according to claim 1, wherein the step of generating a load evaluation value for each of the plurality of servers comprises obtaining task execution information and classifying the tasks into a normal execution class or an abnormal execution class according to a task end state; The method comprises the steps of establishing a basic parameter set based on total running time, accumulated energy consumption and abnormal occurrence frequency of an abnormal task in a preset period before a reference time point, extracting inflection points in variation trend of each parameter in the basic parameter set in the preset period based on the basic parameter set, calculating time difference between the reference time point corresponding to the inflection points and the reference time point, and generating a load evaluation value according to the inflection points and the time difference.
  3. 3. The intelligent scheduling method for server resources based on auxiliary decision-making according to claim 2, wherein the task execution information comprises task attribution server information, execution duration, power consumption in running and task ending state, total running time is calculated according to the execution duration, accumulated power consumption is calculated according to the power consumption in running, and abnormal occurrence frequency is calculated according to occurrence frequency of abnormal execution categories.
  4. 4. The intelligent scheduling method for server resources based on auxiliary decision-making according to claim 2, wherein the step of generating the load evaluation value further comprises constructing graphic data of each parameter changing with time based on a basic parameter set, calculating a change rate of the graphic data to determine a change trend, determining an inflection point in the change trend when the change rate changes from a positive value to a non-positive value or from a negative value to a non-negative value, and calculating the load evaluation value through a preset function by using a parameter change amplitude and a time difference corresponding to the inflection point as input parameters.
  5. 5. The intelligent scheduling method for server resources based on auxiliary decision-making according to claim 1, wherein the step of dividing the plurality of servers and forming a schedulable server set comprises the steps of setting a load evaluation upper limit value, dividing the corresponding servers into schedulable states when the load evaluation value corresponding to the servers is lower than the load evaluation upper limit value, dividing the corresponding servers into in-load states when the load evaluation value corresponding to the servers is not lower than the load evaluation upper limit value, and collecting all the servers in the schedulable states to form the schedulable server set.
  6. 6. The intelligent scheduling method for server resources based on auxiliary decision-making according to claim 5, wherein before all servers in a schedulable state are assembled to form a schedulable server set, the method further comprises the steps of obtaining a server temperature change value and a power consumption change deviation of the server in the schedulable state in a preset analysis period, classifying the corresponding server as a resource to be checked if the server temperature change value exceeds a preset temperature threshold or the power consumption change deviation exceeds a preset power consumption threshold, and otherwise, assembling the corresponding server to the schedulable server set.
  7. 7. The intelligent scheduling method for server resources based on auxiliary decision-making according to claim 6, wherein for the servers classified as the resources to be verified, the method further comprises obtaining a current temperature and a current power consumption variation deviation of the servers at a preset verification time point, and adding the servers into the schedulable server set if the current temperature of the servers is not higher than a temperature threshold and the current power consumption variation deviation does not exceed a power consumption threshold.
  8. 8. An auxiliary decision-making-based server resource intelligent scheduling system is characterized by comprising the following modules: The load evaluation module is used for dynamically generating a load evaluation value for each of the plurality of servers, and dividing the plurality of servers according to the load evaluation value to form a schedulable server set, wherein the generation of the load evaluation value is obtained based on the analysis of the execution parameters of abnormal tasks in the task execution information; The task allocation module is used for selecting a server from the schedulable server set to be matched with the task to be executed when the task to be executed exists; And the task migration decision module is used for calculating the adaptation score of the current task being executed and the server where the current task is located, and determining that the current task is continuously executed or migrated to the server in the schedulable server set for execution according to the adaptation score.
  9. 9. The intelligent scheduling system for server resources based on decision-assist as set forth in claim 8 wherein the task migration decision module determines that the current task continues to execute or migrate comprises executing a server migration if the adaptation score is below a preset adaptation score threshold, migrating the current task to a server in the set of schedulable servers for execution, repeatedly calculating the adaptation score at preset time intervals if the adaptation score is not below the adaptation score threshold, and executing the server migration if the adaptation score is below the adaptation score threshold any of a number of consecutive preset calculations, otherwise determining that the current task continues to execute at the current server.
  10. 10. The intelligent scheduling system for server resources based on decision-assist as set forth in claim 8, wherein the load assessment module generates a load assessment value for each of the plurality of servers, the load assessment module comprising obtaining task execution information and classifying the tasks into a normal execution class or an abnormal execution class according to a task end state; The method comprises the steps of establishing a basic parameter set based on total running time, accumulated energy consumption and abnormal occurrence frequency of an abnormal task in a preset period before a reference time point, extracting inflection points in variation trend of each parameter in the basic parameter set in the preset period based on the basic parameter set, calculating time difference between the reference time point corresponding to the inflection points and the reference time point, and generating a load evaluation value according to the inflection points and the time difference.

Description

Auxiliary decision-based intelligent scheduling method and system for server resources Technical Field The invention belongs to the technical field of server resource scheduling, and particularly relates to an intelligent server resource scheduling method and system based on auxiliary decision. Background With the development of cloud computing, big data and artificial intelligence technology, servers are used as core infrastructures for carrying data processing, storage and computing, the status of the servers in a modern information technology system is increasingly prominent, and in order to ensure the stability and high efficiency of various online services, enterprise applications and computing tasks, how to reasonably schedule and manage resources of server clusters has become a key technical link for ensuring service quality and improving operation benefits. The existing server resource scheduling technology depends on single and represented indexes in the running state evaluation, such as task execution frequency, to judge the load condition of the server, the evaluation mode cannot distinguish the inherent complexity and resource consumption of the task, so that the scheduling decision is not aligned, the existing technology lacks guarantee on the continuity of the task in the scheduling execution layer, when the running task is migrated from one server to another server, seamless connection of the execution state cannot be maintained, the task is interrupted, the performance is suddenly reduced or the restarting is required, the continuity and the user experience of the service are affected, the existing technology lacks self-adaptive capability on the dynamically changed service requirement, the service request driven by the terminal user naturally has burstiness and uncertainty, the dynamic resource requirement cannot be matched in real time, the resource waste and the overall utilization rate of the system are low. In view of the above, the present invention provides an auxiliary decision-based intelligent scheduling method and system for server resources. Disclosure of Invention The invention aims to provide an intelligent scheduling method and system for server resources based on auxiliary decision making, which are used for solving the technical problems of frequent switching of server resources, poor execution continuity of running tasks and reduced running efficiency in the prior art. To achieve the above objective, an embodiment of the present invention provides an intelligent scheduling method for server resources based on auxiliary decision, including the following steps: an intelligent scheduling method of server resources based on auxiliary decision-making comprises the following steps: Dynamically generating a load evaluation value for each of a plurality of servers to divide the plurality of servers and form a schedulable server set, wherein the generation of the load evaluation value is based on analysis of execution parameters of abnormal tasks in task execution information; When the task to be executed exists, selecting a server from the schedulable server set to be matched with the task to be executed so as to complete task distribution; For the current task being executed, calculating the adaptation scores of the current task and the server where the current task is currently located, and determining that the current task is continuously executed at the server where the current task is currently located or is migrated to the server in the schedulable server set for execution according to the adaptation scores; The step of determining that the current task continues to be executed or migrated includes executing server migration if the adaptation score is lower than a preset adaptation score threshold, migrating the current task to a server in a schedulable server set for execution, repeatedly calculating the adaptation score at preset time intervals if the adaptation score is not lower than the adaptation score threshold, executing server migration if the adaptation score is lower than the adaptation score threshold in any one of continuous preset times of calculation, and otherwise, determining that the current task continues to be executed at the current server. Preferably, the step of generating the load evaluation value for each of the plurality of servers includes acquiring task execution information and classifying the tasks into a normal execution class or an abnormal execution class according to a task end state; The method comprises the steps of establishing a basic parameter set based on total running time, accumulated energy consumption and abnormal occurrence frequency of an abnormal task in a preset period before a reference time point, extracting inflection points in variation trend of each parameter in the basic parameter set in the preset period based on the basic parameter set, calculating time difference between the reference time point corresponding to the inf