A server overall performance comparison test method, device, equipment and medium

Abstract

The application provides a kind of server overall performance comparison test method, device, equipment and medium, belong to server performance test technical field, the method steps: the limit stress value when server sample is collected to carry out stress test, and record whether server sample can normally run under limit stress value;Extract the data characteristics corresponding to server sample when testing under safe stress value;Based on the training data set constructed by the data characteristics when running under safe stress value and the corresponding running results of limit stress value;Build BP neural network model, and use training data set to train, obtain data classification model;Extract the data characteristics corresponding to server safe stress value, and input server stress test classification model to data characteristics, predict whether server limit stress value can normally run.The application can predict whether server limit stress value can run by trained server stress test data classification model under server safe stress.

Description

Technical Field

[0001] This invention belongs to the field of server performance testing technology, specifically relating to a comparative testing method, apparatus, equipment, and medium for the overall performance of a server. Background Technology

[0002] With the development of computer and network technologies, information acquisition speeds are increasing rapidly, and service provision is becoming increasingly diversified. Services provided by ordinary computers or devices can no longer meet the demands. Servers, as computer platforms for processing network technologies, have emerged to address this need. Servers can connect to various network systems, such as the Internet, internal organizational networks, inter-organizational networks, and local area networks, providing a wide range of application services to computers connected through these network systems. To ensure the stability of service provision, stress testing of servers has become a crucial testing method. Server stress testing can target key components of the server, such as the CPU, memory, or hard drive, to confirm whether the server can operate normally under a set load.

[0003] Traditional stress testing methods involve testers configuring test scripts based on their experience, and then the test program continuously running the hardware under test according to the script for a set duration, such as 72 hours. Current tests strictly adhere to the configured script until the test ends, regardless of the actual stress or load the hardware under test experiences. However, test stress and hardware load are not constant, often resulting in situations where the stress applied to the server does not match expectations.

[0004] Current stress testing methods, even if testers discover the need to modify stress values, can only wait for the current test to end or interrupt the test altogether before designing the script for the next test. This script adjustment method may require many iterations, extending the testing time and increasing the manpower and time required for script modifications. Furthermore, when applying excessive stress to the server, system crashes often occur due to hardware overload, or even server hardware damage. Moreover, once a system crash or server hardware damage occurs, test data cannot be saved in real time, lacking a basis for subsequent problem tracking.

[0005] Therefore, it is essential to provide a comparative testing method, apparatus, equipment, and medium for server overall performance to address the aforementioned shortcomings. Summary of the Invention

[0006] To address the shortcomings of existing server stress testing methods that strictly adhere to the configured script's set stress, which may result in discrepancies between actual stress and load and expectations, leading to time-consuming and laborious script adjustments, as well as the risk of system crashes and hardware damage from excessive stress, this invention provides a comparative testing method, apparatus, equipment, and medium for overall server performance to solve the aforementioned technical problems.

[0007] In a first aspect, the present invention provides a comparative testing method for the overall performance of a server, comprising the following steps:

[0008] S 1. Collect extreme performance data of a set number of server samples during stress testing, obtain the extreme stress value of the server samples, and record whether the server samples can operate normally under the extreme stress value;

[0009] S 2. Set the safe stress value for the server sample during stress testing according to the preset ratio, and extract the data characteristics of the server sample when testing at the safe stress value;

[0010] S 3. Construct a training dataset based on the data characteristics of the server under safe stress values ​​and the corresponding running results under extreme stress values;

[0011] S4. Build a BP neural network model and train it using the training dataset to obtain a server stress test data classification model;

[0012] S 5. Set the test server to perform stress tests at a safe stress value, extract the corresponding data features, and input the data features into the server stress test classification model to predict whether the test server can operate normally under extreme stress values.

[0013] Furthermore, the specific steps of step S1 are as follows:

[0014] S 11. Set the number of server samples to exceed the threshold;

[0015] S 12. Configure the test script to perform stress tests on the set hardware of each server sample according to the extreme stress value;

[0016] S 13. Record the data characteristics of each server sample at the extreme pressure value and whether the server can operate normally.

[0017] Furthermore, the specific steps of step S12 are as follows:

[0018] S 121. Configure the test script to obtain the ultimate pressure value and the initial pressure value, and divide the interval between the ultimate pressure value and the initial pressure value into segments to obtain the pressure value gradient;

[0019] S 122. Configure the test script to perform stress tests on the set hardware of each server sample for a set duration, starting from the initial stress value and increasing the stress value according to the set stress value gradient, until the limit stress value is reached.

[0020] Furthermore, the specific steps of step S2 are as follows:

[0021] S 21. Pre-set the ratio between the safe stress value and the extreme stress value for server stress testing;

[0022] S 22. Obtain the extreme pressure value of each server sample, and calculate the safe pressure value by combining the ratio of the safe pressure value to the extreme pressure value;

[0023] S 23. Use the security stress value to stress test each server sample and record the data characteristics.

[0024] Furthermore, the specific steps of step S3 are as follows:

[0025] S 31. Obtain the results of whether each server sample can operate normally under extreme pressure values;

[0026] S 32. Obtain the characteristic data of each server sample under stress test at the safe stress value corresponding to the extreme stress value;

[0027] S 33. Establish the correlation between the characteristic data of each server sample under safe pressure value and the results of whether it can operate normally under extreme pressure, and generate a training dataset.

[0028] Furthermore, the specific steps of step S4 are as follows:

[0029] S41. Obtain the number of features N in the training set;

[0030] S42. Construct a BP neural network model, with the number of data features N as the number of input layer nodes, and the result of whether the server can run normally as the number of output layer nodes, and... Let a be the number of hidden layer nodes, and set the activation functions for the hidden and output layers, where a∈[1,10];

[0031] S43. Divide the training dataset into a training data subset and a test data subset;

[0032] S44. Using the feature data under safe pressure values ​​in the training data subset as input and the result of whether it can operate normally under extreme pressure values ​​as output, train the BP neural network model to obtain the initialized BP neural network model;

[0033] S45. Input the feature data under the safe pressure value in the subset of test data into the initial BP neural network model to obtain the prediction result of whether it can operate normally under the extreme pressure value;

[0034] S46. Compare the predicted results of whether the system can operate normally under the ultimate pressure value with the actual results of whether the system can operate normally under the ultimate pressure value in the test subset, and determine whether the error is less than the preset error;

[0035] If so, proceed to step S48;

[0036] If not, proceed to step S47;

[0037] S47. Determine if the maximum number of training iterations has been reached;

[0038] If so, proceed to step S48;

[0039] If not, return to step S44;

[0040] S48. Obtain the server stress test data classification model.

[0041] Furthermore, step S5 is detailed below:

[0042] S 51. Obtain the test server and its maximum stress value;

[0043] S 52. Calculate the safe stress value of the test server based on the ratio between the safe stress value and the extreme stress value of the server stress test;

[0044] S 53. Perform stress tests on the test server using a security stress value and record the data characteristics;

[0045] S 54. Input the data characteristics of the test server under safe stress value into the server stress test data classification model to predict whether the server can operate normally under extreme stress value.

[0046] Secondly, the present invention provides a comparative testing device for the overall performance of a server, comprising:

[0047] The server extreme stress value acquisition module is used to collect extreme performance data of a set number of server samples during stress testing, obtain the extreme stress value of the server samples, and record whether the server samples can operate normally under the extreme stress value.

[0048] The server security stress value acquisition module is used to set the security stress value of the server sample when stress testing according to a preset ratio, and extract the data characteristics of the server sample when testing the security stress value.

[0049] The training dataset construction module is used to construct a training dataset based on the data characteristics of the server under safe stress values ​​and the corresponding running results under extreme stress values.

[0050] The stress test data classification model generation module is used to build a BP neural network model and train it using a training dataset to obtain a server stress test data classification model.

[0051] The server extreme stress test feasibility prediction module is used to set the test server to perform stress tests at a safe stress value, extract the corresponding data features, and input the data features into the server stress test classification model to predict whether the test server can operate normally under extreme stress values.

[0052] Furthermore, the server extreme stress value acquisition module includes:

[0053] The server sample setting unit is used to set server samples that exceed a certain quantity threshold.

[0054] The extreme stress test unit is used to configure test scripts to perform stress tests on the set hardware of each server sample according to the extreme stress value;

[0055] The extreme stress test recording unit is used to record the data characteristics of each server sample at the extreme stress value and whether the server can operate normally.

[0056] Furthermore, the server security stress value acquisition module includes:

[0057] The safe pressure value to limit pressure value ratio setting unit is used to preset the ratio between the safe pressure value and the limit pressure value of the server stress test.

[0058] The safety pressure value calculation unit is used to obtain the limit pressure value of each server sample and calculate the safety pressure value by combining the ratio between the safety pressure value and the limit pressure value.

[0059] The security stress test recording unit is used to perform stress tests on each server sample using security stress values ​​and record data characteristics.

[0060] Furthermore, the training dataset construction module includes:

[0061] The extreme pressure value operation result acquisition unit is used to obtain the result of whether each server sample can operate normally under extreme pressure value;

[0062] The security stress test feature data acquisition unit is used to acquire feature data of each server sample under stress test at the corresponding security stress value of the extreme stress value.

[0063] The training dataset generation unit is used to establish the correlation between the feature data of each server sample under safe stress value and the results of whether it can operate normally under extreme stress value, and to generate the training dataset.

[0064] Furthermore, the stress test data classification model generation module includes:

[0065] The data feature quantity acquisition unit is used to acquire the number N of data features in the training set;

[0066] The BP neural network model building unit is used to build a BP neural network model, with the number of data features N as the number of input layer nodes, and the result of whether the server can run normally as the number of output layer nodes. Let a be the number of hidden layer nodes, and set the activation functions for the hidden and output layers, where a∈[1,10];

[0067] Training dataset partitioning unit, used to divide the training dataset into training data subsets and test data subsets;

[0068] The model training unit is used to train the BP neural network model by taking the feature data under the safe pressure value in the training data subset as input and the result of whether it can operate normally under the extreme pressure value as output, so as to obtain the initialized BP neural network model.

[0069] The model's preliminary prediction unit is used to input the feature data under the safe pressure value in the subset of test data into the initial BP neural network model to obtain the prediction result of whether it can operate normally under the extreme pressure value;

[0070] The model error verification unit is used to compare the predicted results of whether the system can operate normally under the extreme pressure value with the actual results of whether the system can operate normally under the extreme pressure value in the test subset, and to determine whether the error is less than the preset error.

[0071] The training count determination unit is used to determine whether the maximum number of training counts has been reached when the model prediction error is greater than or equal to the preset error.

[0072] The server stress test data classification model generation unit is used to obtain a server stress test data classification model when the model prediction error is greater than the preset error or the maximum number of training iterations is reached.

[0073] Furthermore, the server extreme stress test feasibility prediction module includes:

[0074] The test server acquisition unit is used to acquire the test server and its maximum stress value.

[0075] The safe stress value calculation unit is used to calculate the safe stress value of the test server based on the ratio between the safe stress value and the extreme stress value during server stress testing.

[0076] The server under test security stress testing unit is used to stress test the server using security stress values ​​and record data characteristics.

[0077] The server under test extreme stress test result prediction unit is used to input the data characteristics of the server under stress test at a safe stress value into the server stress test data classification model to predict whether the server can operate normally under extreme stress value.

[0078] Thirdly, the present invention provides an apparatus including a processor and a memory;

[0079] The memory is used to store a computer program, and the processor is used to retrieve and run the computer program from the memory, causing the device to perform the method described in the first aspect above.

[0080] Fourthly, the present invention provides a storage medium,

[0081] The storage medium stores instructions that, when run on a computer, cause the computer to perform the method described in the first aspect above.

[0082] The beneficial effects of this invention are as follows:

[0083] The comparative testing method, apparatus, equipment, and medium for overall server performance provided by this invention can predict whether a server can operate under extreme pressure by testing under safe server pressure using a pre-trained server stress test data classification model. This avoids the problem of system crashes caused by excessive hardware load at extreme pressure values, or even damage to the server hardware. It is characterized by high efficiency, accuracy, and low cost.

[0084] Furthermore, the design principle of this invention is reliable, the structure is simple, and it has a very wide range of application prospects.

[0085] Therefore, it is evident that the present invention has outstanding substantive features and significant progress compared with the prior art, and the beneficial effects of its implementation are also obvious. Attached Figure Description

[0086] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0087] Figure 1 This is a flowchart of Embodiment 1 of the comparative test method for the overall performance of the server of the present invention.

[0088] Figure 2This is a flowchart of Embodiment 2 of the comparative test method for the overall performance of the server of the present invention.

[0089] Figure 3 This is a schematic diagram of the comparative testing device for the overall performance of the server of this invention. Detailed Implementation

[0090] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.

[0091] Example 1:

[0092] To address the issue that server stress testing strictly adheres to the configured script's settings, which may result in discrepancies between actual stress and load and expectations, leading to time-consuming and laborious script adjustments, and the risk of system crashes and hardware damage due to excessive stress, the following measures are proposed. Figure 1 As shown, this invention provides a comparative testing method for the overall performance of a server, comprising the following steps:

[0093] S 1. Collect extreme performance data of a set number of server samples during stress testing, obtain the extreme stress value of the server samples, and record whether the server samples can operate normally under the extreme stress value;

[0094] S 2. Set the safe stress value for the server sample during stress testing according to the preset ratio, and extract the data characteristics of the server sample when testing at the safe stress value;

[0095] S 3. Construct a training dataset based on the data characteristics of the server under safe stress values ​​and the corresponding running results under extreme stress values;

[0096] S4. Build a BP neural network model and train it using the training dataset to obtain a server stress test data classification model;

[0097] S 5. Set the test server to perform stress tests at a safe stress value, extract the corresponding data features, and input the data features into the server stress test classification model to predict whether the test server can operate normally under extreme stress values.

[0098] Example 2:

[0099] To address the issue that server stress testing strictly adheres to the configured script's settings, which may result in discrepancies between actual stress and load and expectations, leading to time-consuming and laborious script adjustments, and the risk of system crashes and hardware damage due to excessive stress, the following measures are proposed. Figure 2 As shown, this invention provides a comparative testing method for the overall performance of a server, comprising the following steps:

[0100] S1. Collect extreme performance data of a set number of server samples during stress testing, obtain the extreme stress value of the server samples, and record whether the server samples can operate normally under the extreme stress value; the specific steps of step S1 are as follows:

[0101] S 11. Set the number of server samples to exceed the threshold;

[0102] S 12. Configure the test script to perform stress tests on the set hardware of each server sample according to the extreme stress value;

[0103] S 13. Record the data characteristics of each server sample at the extreme pressure value and whether the server can operate normally;

[0104] S2. Set the safe stress value for the server sample during stress testing according to the preset ratio, and extract the data features corresponding to the server sample when testing at the safe stress value; the specific steps of step S2 are as follows:

[0105] S 21. Pre-set the ratio between the safe stress value and the extreme stress value for server stress testing;

[0106] S 22. Obtain the extreme pressure value of each server sample, and calculate the safe pressure value by combining the ratio of the safe pressure value to the extreme pressure value;

[0107] S 23. Use the security stress value to stress test each server sample and record the data characteristics;

[0108] S3. Construct a training dataset based on the data characteristics of the server under safe stress and the corresponding running results under extreme stress; the specific steps of step S3 are as follows:

[0109] S 31. Obtain the results of whether each server sample can operate normally under extreme pressure values;

[0110] S 32. Obtain the characteristic data of each server sample under stress test at the safe stress value corresponding to the extreme stress value;

[0111] S 33. Establish the correlation between the feature data of each server sample under safe pressure value and the results of whether it can operate normally under extreme pressure, and generate a training dataset;

[0112] S4. Build a BP neural network model and train it using the training dataset to obtain a server stress test data classification model; the specific steps of step S4 are as follows:

[0113] S41. Obtain the number of features N in the training set;

[0114] S42. Construct a BP neural network model, with the number of data features N as the number of input layer nodes, and the result of whether the server can run normally as the number of output layer nodes, and... Let a be the number of hidden layer nodes, and set the activation functions for the hidden and output layers, where a∈[1,10];

[0115] S43. Divide the training dataset into a training data subset and a test data subset;

[0116] S44. Using the feature data under safe pressure values ​​in the training data subset as input and the result of whether it can operate normally under extreme pressure values ​​as output, train the BP neural network model to obtain the initialized BP neural network model;

[0117] S45. Input the feature data under the safe pressure value in the subset of test data into the initial BP neural network model to obtain the prediction result of whether it can operate normally under the extreme pressure value;

[0118] S46. Compare the predicted results of whether the system can operate normally under the ultimate pressure value with the actual results of whether the system can operate normally under the ultimate pressure value in the test subset, and determine whether the error is less than the preset error;

[0119] If so, proceed to step S48;

[0120] If not, proceed to step S47;

[0121] S47. Determine if the maximum number of training iterations has been reached;

[0122] If so, proceed to step S48;

[0123] If not, return to step S44;

[0124] S48. Obtain the server stress test data classification model;

[0125] S5. Set the test server to perform stress testing at a safe stress value, extract the corresponding data features, and input the data features into the server stress test classification model to predict whether the test server can operate normally under extreme stress value; the specific steps of step S5 are as follows:

[0126] S 51. Obtain the test server and its maximum stress value;

[0127] S 52. Calculate the safe stress value of the test server based on the ratio between the safe stress value and the extreme stress value of the server stress test;

[0128] S 53. Perform stress tests on the test server using a security stress value and record the data characteristics;

[0129] S 54. Input the data characteristics of the test server under safe stress value into the server stress test data classification model to predict whether the server can operate normally under extreme stress value.

[0130] Example 3:

[0131] To address the issue that server stress testing strictly adheres to the configured script's settings, which may result in discrepancies between actual stress and load and expectations, leading to time-consuming and laborious script adjustments, and the risk of system crashes and hardware damage due to excessive stress, the following measures are proposed. Figure 2 As shown, this invention provides a comparative testing method for the overall performance of a server, comprising the following steps:

[0132] S1. Collect extreme performance data of a set number of server samples during stress testing, obtain the extreme stress value of the server samples, and record whether the server samples can operate normally under the extreme stress value; the specific steps of step S1 are as follows:

[0133] S 11. Set the number of server samples to exceed the threshold;

[0134] S12. Configure the test script to perform stress tests on the hardware settings of each server sample according to the extreme stress value; the specific steps of step S12 are as follows:

[0135] S 121. Configure the test script to obtain the ultimate pressure value and the initial pressure value, and divide the interval between the ultimate pressure value and the initial pressure value into segments to obtain the pressure value gradient;

[0136] S 122. Configure the test script to perform stress tests on the set hardware of each server sample for a set duration, starting from the initial stress value and increasing the stress value according to the set stress value gradient, until the limit stress value is reached.

[0137] S 13. Record the data characteristics of each server sample at the extreme stress value and whether the server can operate normally; for example, the data characteristics of server stress test use five data characteristics: CPU utilization, memory utilization, hard disk utilization, server throughput, and response time.

[0138] S2. Set the safe stress value for the server sample during stress testing according to the preset ratio, and extract the data features corresponding to the server sample when testing at the safe stress value; the specific steps of step S2 are as follows:

[0139] S 21. Pre-set the ratio between the safe stress value and the maximum stress value for the server stress test; for example, pre-set the safe stress value to be 80% of the maximum stress value.

[0140] S 22. Obtain the extreme pressure value of each server sample, and calculate the safe pressure value by combining the ratio of the safe pressure value to the extreme pressure value;

[0141] S 23. Use security stress values ​​to perform stress tests on each server sample and record data characteristics; for example, record the values ​​of five data characteristics: CPU utilization, memory utilization, hard disk utilization, server throughput, and response time.

[0142] S3. Construct a training dataset based on the data characteristics of the server under safe stress and the corresponding running results under extreme stress; the specific steps of step S3 are as follows:

[0143] S 31. Obtain the results of whether each server sample can operate normally under extreme pressure values;

[0144] S 32. Obtain the characteristic data of each server sample under stress test at the safe stress value corresponding to the extreme stress value;

[0145] S 33. Establish the correlation between the feature data of each server sample under safe pressure value and the results of whether it can operate normally under extreme pressure, and generate a training dataset;

[0146] S4. Build a BP neural network model and train it using the training dataset to obtain a server stress test data classification model; the specific steps of step S4 are as follows:

[0147] S41. Obtain the number of features N in the training set;

[0148] S42. Construct a BP neural network model, with the number of data features N as the number of input layer nodes, and the result of whether the server can run normally as the number of output layer nodes, and... Let N be the number of hidden layer nodes and set the activation functions for the hidden layer and the output layer, where a∈[1,10]; for example, based on five data features such as CPU utilization, memory utilization, hard disk utilization, server throughput, and response time, N=5, and the preferred number of hidden layer nodes is L=4; the activation functions for the hidden layer and the output layer are Tansig functions;

[0149] S43. Divide the training dataset into a training data subset and a test data subset;

[0150] S44. Using the feature data under safe pressure values ​​in the training data subset as input and the result of whether it can operate normally under extreme pressure values ​​as output, train the BP neural network model to obtain the initialized BP neural network model;

[0151] S45. Input the feature data under the safe pressure value in the subset of test data into the initial BP neural network model to obtain the prediction result of whether it can operate normally under the extreme pressure value;

[0152] S46. Compare the predicted results of whether the system can operate normally under the ultimate pressure value with the actual results of whether the system can operate normally under the ultimate pressure value in the test subset, and determine whether the error is less than the preset error;

[0153] If so, proceed to step S48;

[0154] If not, proceed to step S47;

[0155] S47. Determine if the maximum number of training iterations has been reached;

[0156] If so, proceed to step S48;

[0157] If not, return to step S44;

[0158] S48. Obtain the server stress test data classification model;

[0159] S5. Set the test server to perform stress testing at a safe stress value, extract the corresponding data features, and input the data features into the server stress test classification model to predict whether the test server can operate normally under extreme stress value; the specific steps of step S5 are as follows:

[0160] S 51. Obtain the test server and its maximum stress value;

[0161] S 52. Calculate the safe stress value of the test server based on the ratio between the safe stress value and the extreme stress value of the server stress test;

[0162] S 53. Perform stress tests on the test server using a security stress value and record the data characteristics;

[0163] S 54. Input the data characteristics of the test server under safe stress value into the server stress test data classification model to predict whether the server can operate normally under extreme stress value.

[0164] Example 4:

[0165] To address the issue that server stress testing strictly adheres to the configured script's settings, which may result in discrepancies between actual stress and load and expectations, leading to time-consuming and laborious script adjustments, and the risk of system crashes and hardware damage due to excessive stress, the following measures are proposed. Figure 3 As shown, the present invention provides a comparative testing device for the overall performance of a server, comprising:

[0166] The server extreme stress value acquisition module is used to collect extreme performance data of a set number of server samples during stress testing, obtain the extreme stress value of the server samples, and record whether the server samples can operate normally under the extreme stress value.

[0167] The server security stress value acquisition module is used to set the security stress value of the server sample when stress testing according to a preset ratio, and extract the data characteristics of the server sample when testing the security stress value.

[0168] The training dataset construction module is used to construct a training dataset based on the data characteristics of the server under safe stress values ​​and the corresponding running results under extreme stress values.

[0169] The stress test data classification model generation module is used to build a BP neural network model and train it using a training dataset to obtain a server stress test data classification model.

[0170] The server extreme stress test feasibility prediction module is used to set the test server to perform stress tests at a safe stress value, extract the corresponding data features, and input the data features into the server stress test classification model to predict whether the test server can operate normally under extreme stress values.

[0171] Example 5:

[0172] To address the issue that server stress testing strictly adheres to the configured script's settings, which may result in discrepancies between actual stress and load and expectations, leading to time-consuming and laborious script adjustments, and the risk of system crashes and hardware damage due to excessive stress, the following measures are proposed. Figure 3 As shown, the present invention provides a comparative testing device for the overall performance of a server, comprising:

[0173] The server extreme stress value acquisition module is used to collect extreme performance data of a set number of server samples during stress testing, obtain the extreme stress value of the server samples, and record whether the server samples can operate normally under the extreme stress value; the server extreme stress value acquisition module includes:

[0174] The server sample setting unit is used to set server samples that exceed a certain quantity threshold.

[0175] The extreme stress test unit is used to configure test scripts to perform stress tests on the set hardware of each server sample according to the extreme stress value;

[0176] The extreme stress test recording unit is used to record the data characteristics of each server sample at the extreme stress value and whether the server can operate normally;

[0177] The server security stress value acquisition module is used to set the security stress value for server samples during stress testing according to a preset ratio, and to extract the data characteristics of the server samples during the security stress test. The server security stress value acquisition module includes:

[0178] The safe pressure value to limit pressure value ratio setting unit is used to preset the ratio between the safe pressure value and the limit pressure value of the server stress test.

[0179] The safety pressure value calculation unit is used to obtain the limit pressure value of each server sample and calculate the safety pressure value by combining the ratio between the safety pressure value and the limit pressure value.

[0180] The security stress test recording unit is used to perform stress tests on each server sample using security stress values ​​and record data characteristics.

[0181] The training dataset construction module is used to construct a training dataset based on the data characteristics of the server under safe stress and the corresponding running results under extreme stress. The training dataset construction module includes:

[0182] The extreme pressure value operation result acquisition unit is used to obtain the result of whether each server sample can operate normally under extreme pressure value;

[0183] The security stress test feature data acquisition unit is used to acquire feature data of each server sample under stress test at the corresponding security stress value of the extreme stress value.

[0184] The training dataset generation unit is used to establish the correlation between the feature data of each server sample under safe stress value and the results of whether it can operate normally under extreme stress, and to generate the training dataset.

[0185] The stress test data classification model generation module is used to build a BP neural network model and train it using a training dataset to obtain a server stress test data classification model. The stress test data classification model generation module includes:

[0186] The data feature quantity acquisition unit is used to acquire the number N of data features in the training set;

[0187] The BP neural network model building unit is used to build a BP neural network model, with the number of data features N as the number of input layer nodes, and the result of whether the server can run normally as the number of output layer nodes. Let a be the number of hidden layer nodes, and set the activation functions for the hidden and output layers, where a∈[1,10];

[0188] Training dataset partitioning unit, used to divide the training dataset into training data subsets and test data subsets;

[0189] The model training unit is used to train the BP neural network model by taking the feature data under the safe pressure value in the training data subset as input and the result of whether it can operate normally under the extreme pressure value as output, so as to obtain the initialized BP neural network model.

[0190] The model's preliminary prediction unit is used to input the feature data under the safe pressure value in the subset of test data into the initial BP neural network model to obtain the prediction result of whether it can operate normally under the extreme pressure value;

[0191] The model error verification unit is used to compare the predicted results of whether the system can operate normally under the extreme pressure value with the actual results of whether the system can operate normally under the extreme pressure value in the test subset, and to determine whether the error is less than the preset error.

[0192] The training count determination unit is used to determine whether the maximum number of training counts has been reached when the model prediction error is greater than or equal to the preset error.

[0193] The server stress test data classification model generation unit is used to obtain the server stress test data classification model when the model prediction error is greater than the preset error or the maximum number of training times is reached.

[0194] The server extreme stress test feasibility prediction module is used to set the test server to perform stress tests at a safe stress value, extract corresponding data features, and input the data features into the server stress test classification model to predict whether the test server can operate normally under extreme stress values. The server extreme stress test feasibility prediction module includes:

[0195] The test server acquisition unit is used to acquire the test server and its maximum stress value.

[0196] The safe stress value calculation unit is used to calculate the safe stress value of the test server based on the ratio between the safe stress value and the extreme stress value during server stress testing.

[0197] The server under test security stress testing unit is used to stress test the server using security stress values ​​and record data characteristics.

[0198] The server under test extreme stress test result prediction unit is used to input the data characteristics of the server under stress test at a safe stress value into the server stress test data classification model to predict whether the server can operate normally under extreme stress value.

[0199] Example 6:

[0200] This invention provides a device, including a processor and a memory;

[0201] The memory is used to store computer programs, and the processor is used to call and run the computer programs from the memory, so that the device performs the methods described in Embodiment 1, Embodiment 2, or Embodiment 3.

[0202] Example 7:

[0203] This invention provides a storage medium,

[0204] The storage medium stores instructions that, when run on a computer, cause the computer to perform the method described in any one of Embodiment 1, Embodiment 2, or Embodiment 3.

[0205] Although the present invention has been described in detail with reference to the accompanying drawings and preferred embodiments, the invention is not limited thereto. Various equivalent modifications or substitutions can be made to the embodiments of the invention by those skilled in the art without departing from the spirit and essence of the invention, and such modifications or substitutions should all be within the scope of the invention. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the invention should also be covered within the protection scope of the invention. Therefore, the protection scope of the invention should be determined by the scope of the claims.

Claims

1. A comparative testing method for the overall performance of a server, characterized in that, Includes the following steps: S1. Collect extreme performance data of a set number of server samples during stress testing, obtain the extreme stress value of the server samples, and record whether the server samples can operate normally under the extreme stress value. S2. Set the safe stress value for the server sample during stress testing according to the preset ratio, and extract the data features corresponding to the server sample when testing at the safe stress value; The specific steps of step S2 are as follows: S21. Pre-set the ratio between the safe stress value and the extreme stress value for server stress testing; S22. Obtain the extreme pressure value of each server sample, and calculate the safe pressure value by combining the ratio of the safe pressure value to the extreme pressure value; S23. Perform stress tests on each server sample using the security stress value and record the data characteristics; S3. Construct a training dataset based on the data characteristics of the server under safe stress values ​​and the corresponding running results under extreme stress values; The specific steps of step S3 are as follows: S31. Obtain the results of whether each server sample can operate normally under extreme pressure values; S32. Obtain the characteristic data of each server sample under stress test at the safe stress value corresponding to the extreme stress value; S33. Establish the correlation between the feature data of each server sample under safe pressure value and the results of whether it can operate normally under extreme pressure, and generate a training dataset; S4. Build a BP neural network model and train it using the training dataset to obtain a server stress test data classification model; S5. Set the test server to perform stress tests at a safe stress value, extract the corresponding data features, and input the data features into the server stress test classification model to predict whether the test server can operate normally under extreme stress values. The specific steps of step S5 are as follows: S51. Obtain the test server and its maximum stress value; S52. Calculate the safe stress value of the test server based on the ratio between the safe stress value and the extreme stress value of the server stress test; S53. Perform stress tests on the test server using a security stress value and record the data characteristics; S54. Input the data characteristics of the test server under safe stress value into the server stress test data classification model to predict whether the server can operate normally under extreme stress value.

2. The comparative testing method for overall server performance as described in claim 1, characterized in that, The specific steps of step S1 are as follows: S11. Set the number of server samples to exceed the threshold; S12. Configure the test script to perform stress tests on the hardware settings of each server sample according to the extreme stress value; S13. Record the data characteristics of each server sample at the extreme pressure value and whether the server can operate normally.

3. The comparative testing method for overall server performance as described in claim 2, characterized in that, The specific steps of step S12 are as follows: S121. Configure the test script to obtain the ultimate pressure value and the initial pressure value, and divide the interval between the ultimate pressure value and the initial pressure value into segments to obtain the pressure value gradient; S122. Configure the test script to perform stress tests on the set hardware of each server sample for a set duration, starting from the initial stress value and increasing the stress value according to the set stress value gradient, until the limit stress value is reached.

4. The comparative testing method for overall server performance as described in claim 1, characterized in that, The specific steps of step S4 are as follows: S41. Obtain the number of features N in the training set; S42. Construct a BP neural network model, with N as the number of data features as the number of input layer nodes, and the result of whether the server can run normally as the number of output layer nodes, and L= +a represents the number of hidden layer nodes, and sets the activation functions for the hidden and output layers, where a∈[1,10]; S43. Divide the training dataset into a training data subset and a test data subset; S44. Using the feature data under safe pressure values ​​in the training data subset as input and the result of whether it can operate normally under extreme pressure values ​​as output, train the BP neural network model to obtain the initialized BP neural network model; S45. Input the feature data under the safe pressure value in the subset of test data into the initial BP neural network model to obtain the prediction result of whether it can operate normally under the extreme pressure value; S46. Compare the predicted results of whether the system can operate normally under the ultimate pressure value with the actual results of whether the system can operate normally under the ultimate pressure value in the test subset, and determine whether the error is less than the preset error. If so, proceed to step S48; If not, proceed to step S47; S47. Determine if the maximum number of training iterations has been reached; If so, proceed to step S48; If not, return to step S44; S48. Obtain the server stress test data classification model.

5. A comparative testing apparatus for the overall performance of a server, comprising performing the method described in any one of claims 1-4, characterized in that, include: The server extreme stress value acquisition module is used to collect extreme performance data of a set number of server samples during stress testing, obtain the extreme stress value of the server samples, and record whether the server samples can operate normally under the extreme stress value. The server security stress value acquisition module is used to set the security stress value of the server sample when stress testing according to a preset ratio, and extract the data characteristics of the server sample when testing the security stress value. The training dataset construction module is used to construct a training dataset based on the data characteristics of the server under safe stress values ​​and the corresponding running results under extreme stress values. The stress test data classification model generation module is used to build a BP neural network model and train it using a training dataset to obtain a server stress test data classification model. The server extreme stress test feasibility prediction module is used to set the test server to perform stress tests at a safe stress value, extract the corresponding data features, and input the data features into the server stress test classification model to predict whether the test server can operate normally under extreme stress values.

6. A device, characterized in that, Including processor and memory; The memory is used to store a computer program, and the processor is used to retrieve and run the computer program from the memory, causing the device to perform the method described in any one of claims 1-4.

7. A storage medium, characterized in that, The storage medium stores instructions that, when run on a computer, cause the computer to perform the method described in any one of claims 1-4.

Images