Multi-timescale CPU load stability evaluation method for test domains

Abstract

The present invention provides a multi-time scale CPU load stability evaluation method for the testing field, which relates to the technical field of CPU load stability evaluation, and aims to more comprehensively capture the dynamic characteristics of the system load, including the following steps: obtaining a multi-scale load ratio based on multiple time scales respectively; obtaining a load stability coefficient according to the multi-scale load ratio, the load stability coefficient is used to describe the stability of the load; obtaining a weighted load index according to the multi-scale load ratio, the weighted load index is used to give different weights to different time scales and obtain the overall load level; and performing system state evaluation according to the load stability coefficient and the weighted load index. The present invention has the advantage of being able to obtain a more accurate and reliable CPU load stability evaluation result.

Description

Technical Field

[0001] The present invention relates to the technical field of CPU load stability evaluation, and in particular to a multi-time scale CPU load stability evaluation method for use in the testing field. Background Art

[0002] In modern computing systems, the volatility of CPU load has a significant impact on system performance and stability.

[0003] Traditional single time window load assessment methods often cannot accurately reflect the actual load status of the system, mainly because the system load often exhibits fluctuation characteristics under multiple time scales. Therefore, the analysis of a single time scale may ignore certain important load characteristics, resulting in deviations in the assessment results.

[0004] Therefore, it is urgent to optimize the CPU load stability evaluation method to more comprehensively capture the dynamic characteristics of the system load and obtain more accurate and reliable CPU load stability evaluation results. Summary of the invention

[0005] The purpose of the present invention is to provide a multi-time scale CPU load stability evaluation method for the testing field, which can capture the dynamic characteristics of the system load more comprehensively.

[0006] The present invention is achieved through the following technical solutions:

[0007] The method for evaluating the CPU load stability at multiple time scales in a test domain includes the following steps:

[0008] Obtain multi-scale load ratios based on multiple time scales respectively;

[0009] Obtaining a load stability coefficient according to the multi-scale load ratio, where the load stability coefficient is used to describe the stability of the load;

[0010] Obtaining a weighted load index according to the multi-scale load ratio, wherein the weighted load index is used to assign different weights to different time scales and obtain an overall load level;

[0011] A system status assessment is performed based on the load stability coefficient and the weighted load index.

[0012] Preferably, the method for obtaining the multi-scale load ratio is:

[0013] ;

[0014] ;

[0015] in, Representative Multiscale load ratio of time scale, Representing a The average CPU load over the time scale, Represents the total number of CPU cores in the cluster. for The total number of time scales selected.

[0016] Preferably, the method for obtaining the load stability coefficient is:

[0017] ;

[0018] in, represents the load stability factor, Representative Multiscale load ratio of time scale, Respectively represent Find the maximum and minimum values ​​of the change.

[0019] Preferably, the method for obtaining the weighted load index is:

[0020] ;

[0021] in, represents the weighted load index, Representative The multiscale load ratio of the time scale, Representative The weight of the time scale, for The total number of time scales selected.

[0022] Preferably, the time scale includes a first time scale, a second time scale and a third time scale;

[0023] The first time scale, the second time scale and the third time scale are 1 minute, 5 minutes and 15 minutes respectively.

[0024] Preferably, the weight is obtained by:

[0025] for Setting the baseline weights;

[0026] Continuous observation The value of: If , then increase the base weight and reduce and / or Value; if , then the increase is reduced on the basis of the benchmark weight and increase the value and / or otherwise, the reference weight is maintained.

[0027] Preferably, the reference weight is: , , ;

[0028] like , , , ;

[0029] like , , , .

[0030] Preferably, the method for evaluating the system status based on the load smoothness coefficient and the weighted load index is: setting multiple thresholds for the load smoothness coefficient and the weighted load index respectively, and performing a combined evaluation based on the size relationship between the load smoothness coefficient and the corresponding threshold and the size relationship between the weighted load index and the corresponding threshold.

[0031] Preferably, the method of setting the multiple thresholds for the combined evaluation is:

[0032] like and , then the system state is stable light load state;

[0033] like and ,or and , then the system state is stable and normal;

[0034] like and ,or and ,or and , then the system status is a state that requires regular monitoring;

[0035] like and ,or and ,or and , then the system status is a state that requires special attention;

[0036] in, represents the load stability factor, Represents the weighted load index.

[0037] The technical solution of the present invention has at least the following advantages and beneficial effects:

[0038] The present invention can capture the changing characteristics of the load at different time scales by obtaining multi-scale load ratios based on multiple time scales, thereby improving the overall understanding of the dynamic characteristics of the load and facilitating more accurate system status assessment;

[0039] The present invention quantifies the stability of the load by introducing a load stability coefficient, and can identify the stability or volatility characteristics of the load based on multiple time scales, providing a basis for maintaining the stability of the system;

[0040] The present invention assigns different weights to weighted load indicators according to the characteristics of different time scales, and can comprehensively consider the contribution of each time scale to the overall load, thereby obtaining more representative load ratio data and improving the accuracy and flexibility of load evaluation;

[0041] The present invention has a reasonable design and comprehensive analysis, and can provide a more accurate and comprehensive analysis method for load monitoring and evaluation, thereby significantly improving the reliability and management efficiency of the system's load analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A schematic flow chart of a multi-time-scale CPU load stability evaluation method for a testing field provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0044] Example 1

[0045] This embodiment provides a multi-time scale CPU load stability evaluation method for the test field, see Figure 1 , including the following steps:

[0046] Obtain multi-scale load ratios based on multiple time scales respectively;

[0047] Obtaining a load stability coefficient according to the multi-scale load ratio, where the load stability coefficient is used to describe the stability of the load;

[0048] Obtaining a weighted load index according to the multi-scale load ratio, wherein the weighted load index is used to assign different weights to different time scales and obtain an overall load level;

[0049] A system status assessment is performed based on the load stability coefficient and the weighted load index.

[0050] Since the system operation status changes dynamically, it is necessary to calculate the CPU load status through multi-time scale CPU load data. Therefore, this embodiment introduces two key indicators: load stability coefficient and weighted load index. The load stability coefficient reflects the fluctuation degree of CPU load, and the weighted load index comprehensively reflects the overall load level of the CPU.

[0051] This embodiment achieves accurate evaluation of the CPU load status by capturing and comprehensively analyzing the CPU load change characteristics of the system at different time scales. This solution does not require a complex data processing process, and can accurately reflect the load characteristics of the system by calculating the load ratio and load stability coefficient of multiple time scales. When the load fluctuates violently, it can be discovered in time and corresponding adjustment measures can be taken; when the load is stable, resource allocation can be optimized to improve system efficiency. The evaluation mechanism of this embodiment effectively avoids the data fluctuations and "false positive" problems caused by a single time scale, takes into account the instantaneous state and historical trends of the system, and integrates multi-time scale information into a single evaluation indicator through a mathematical model, which reduces the cognitive threshold of system monitoring and improves the feasibility of automated operation and maintenance, so that the system can maintain stable operation under complex and changeable load environments, and is suitable for load balancing, resource scheduling, performance optimization and other computing scenarios.

[0052] In this embodiment, the method for obtaining the multi-scale load ratio is:

[0053] ;

[0054] ;

[0055] in, Representative Multiscale load ratio of time scale, Representing a The average CPU load over the time scale, Represents the total number of CPU cores in the cluster. for The total number of time scales selected.

[0056] Next, the method for obtaining the load stability coefficient is preferably:

[0057] ;

[0058] in, represents the load stability factor, Representative Multiscale load ratio of time scale, Respectively represent Find the maximum and minimum values ​​of the change.

[0059] On the other hand, the method for obtaining the weighted load index may be:

[0060] ;

[0061] in, represents the weighted load index, Representative The multiscale load ratio of the time scale, Representative The weight of the time scale, for The total number of time scales selected.

[0062] As a preferred solution of this embodiment, the time scale includes a first time scale, a second time scale and a third time scale;

[0063] The first time scale, the second time scale and the third time scale are 1 minute, 5 minutes and 15 minutes respectively. The three time scales are selected based on the following criteria: 1 minute can reflect the short-term state of the system, corresponding to the rapid response to problem handling; 5 minutes can reflect the medium-term state of the system, corresponding to the troubleshooting time of simple problems; 15 minutes can reflect the long-term state of the system, corresponding to the processing time of complex problems.

[0064] Furthermore, the weight is obtained by:

[0065] for Setting the baseline weights;

[0066] Continuous observation The value of: If , then increase the base weight and reduce and / or Value; if , then the increase is reduced on the basis of the benchmark weight and increase the value and / or otherwise, the reference weight is maintained.

[0067] Specifically, the reference weight may preferably be set to: , , ;

[0068] like , , , ;

[0069] like , , , .

[0070] Because through continuous observation The value can identify the system load change trend. For example, when When the system load is increasing, it indicates that the system load is increasing. At this time, we should pay more attention to short-term load changes. Therefore, we should increase the base weight based on the baseline weight. If the value of When the system load is decreasing, it indicates that the system load is decreasing. At this time, we should consider the long-term load condition more. Therefore, we should increase the weight based on the baseline weight. The value of .

[0071] On this basis, the method for evaluating the system status according to the load smoothness coefficient and the weighted load index is: setting multiple thresholds for the load smoothness coefficient and the weighted load index respectively, and performing a combined evaluation based on the size relationship between the load smoothness coefficient and the corresponding threshold and the size relationship between the weighted load index and the corresponding threshold.

[0072] Based on load stabilization factor and weighted load index , a comprehensive assessment of the system status can be performed. When , it indicates that the CPU load is highly stable; when When , it means that the CPU load is relatively stable; when When , it means that the CPU load is generally stable; when When , it indicates that the CPU load fluctuates slightly; when When , it means that the CPU load fluctuates violently. At the same time, through the weighted load index The system load level can be determined when When the load is light, When It is in heavy load state.

[0073] Therefore, the method of setting the multiple thresholds to perform the combined evaluation may be:

[0074] like and , then the system state is stable light load state;

[0075] like and ,or and , then the system state is stable and normal;

[0076] like and ,or and ,or and , then the system status is a state that requires regular monitoring;

[0077] like and ,or and ,or and , then the system status is a state that requires special attention;

[0078] in, represents the load stability factor, Represents the weighted load index.

[0079] Based on the combined evaluation results of the above two indicators, the system can adopt corresponding adjustment strategies, such as increasing the monitoring frequency, starting load balancing, optimizing resource allocation, etc.

[0080] In summary, by collecting data at multiple time scales and evaluating the system status by combining the load stability coefficient and weighted load index, the system load can be comprehensively considered in multiple dimensions, which helps to have a more comprehensive understanding of the system's operating status, discover potential problems in a timely manner, and achieve a more comprehensive and reliable CPU load stability evaluation.

[0081] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A multi-time scale CPU load stability evaluation method for a test domain, characterized in that: The following steps are involved: Obtain multi-scale load ratios based on multiple time scales respectively; Obtaining a load stability coefficient according to the multi-scale load ratio, where the load stability coefficient is used to describe the stability of the load; Obtaining a weighted load index according to the multi-scale load ratio, wherein the weighted load index is used to assign different weights to different time scales and obtain an overall load level; Performing system status evaluation according to the load stability coefficient and the weighted load index; The method for obtaining the multi-scale load ratio is: ； ； in, Representative Multiscale load ratio of time scale, Representing a The average CPU load over the time scale, Represents the total number of CPU cores in the cluster. for The total number of time scales selected; The method for obtaining the load stability coefficient is: ； in, represents the load stability factor, Representative Multiscale load ratio of time scale, Respectively represent Find the maximum and minimum values ​​of the change; The method for evaluating the system status based on the load stability coefficient and the weighted load index is: setting multiple thresholds for the load stability coefficient and the weighted load index respectively, and performing a combined evaluation based on the size relationship between the load stability coefficient and the corresponding threshold and the size relationship between the weighted load index and the corresponding threshold.

2. The multi-time scale CPU load stability evaluation method for testing field according to claim 1, characterized in that: The method for obtaining the weighted load index is: ； in, represents the weighted load index, Representative The multiscale load ratio of the time scale, Representative The weight of the time scale, for The total number of time scales selected.

3. The multi-time scale CPU load stability evaluation method for testing field according to claim 2, characterized in that: The time scale includes a first time scale, a second time scale and a third time scale; The first time scale, the second time scale and the third time scale are 1 minute, 5 minutes and 15 minutes respectively.

4. The multi-time scale CPU load stability evaluation method for testing field according to claim 3, characterized in that: The method for obtaining the weight is: for Set the base weight; Continuous observation The value of: If , then increase the base weight and reduce and / or Value; if , then reduce the and increase the value and / or otherwise, the reference weight is maintained.

5. The method for evaluating CPU load stability at multiple time scales for use in a test field according to claim 4, characterized in that: The reference weights are: , , ; like , , , ; like , , , .

6. The multi-time scale CPU load stability evaluation method for testing field according to claim 1, characterized in that: The method of setting the multiple thresholds to perform the combined evaluation is: like and , then the system state is stable light load state; like and ,or and , then the system state is stable and normal; like and ,or and ,or and , then the system status is a state that requires regular monitoring; like and ,or and ,or and , then the system status is a state that requires special attention; in, represents the load stability factor, represents the weighted load index.

Images