State evaluation method and device for solid state disk, storage medium and program product

Abstract

The invention relates to a state evaluation method and device of a solid state disk, a storage medium and a program product, and belongs to the technical field of storage, and the method comprises the following steps: obtaining related parameter information of the solid state disk, and preprocessing the related parameter information to obtain a first target data feature; according to a joint probability distribution value between the historical merged data features and the health degree, determining a correlation degree between the historical merged data features and the health degree, and according to the correlation degree, screening out a second target data feature from the first target data features; determining a health score corresponding to the second target data feature based on a health score calculation function; and determining a state evaluation result of the solid state disk according to the health score corresponding to the second target data feature, and visually displaying the state evaluation result. According to the method, the accuracy of state evaluation of the solid state disk is improved, and the risk of data loss or server downtime caused by solid state disk faults is effectively reduced.

Description

Technical Field

[0001] The present application relates to the field of storage technology, and in particular to a state assessment method, device, storage medium and program product for a solid state hard disk. Background Art

[0002] With the rapid development of information technology, the demand for data storage continues to grow. SSD (Solid State Disk or Solid State Drive) has been widely used in various fields due to its advantages such as high-speed reading and writing, low power consumption, and strong shock resistance. During the use of SSD, it may gradually enter a sub-healthy state due to various reasons. In this state, although the hard disk may still work normally, the performance may be reduced and there may be potential failure risks. If it is not discovered and handled in time, it may lead to serious consequences such as data loss and system failure. Therefore, how to accurately evaluate the status of the SSD disk to warn of potential problems in advance, and then provide users with enough time to take measures to ensure data security and stable system operation is an issue that needs to be solved urgently. Summary of the invention

[0003] Based on this, it is necessary to provide a state assessment method, device, storage medium and program product for a solid state hard disk that can improve the accuracy of state assessment in response to the above technical problems.

[0004] In a first aspect, a state assessment method for a solid-state hard disk is provided, the method comprising: obtaining relevant parameter information of the solid-state hard disk, preprocessing the relevant parameter information to obtain a first target data feature; determining a correlation between the historical merged data feature and the corresponding health degree based on a joint probability distribution value between the historical merged data feature and the corresponding health degree, and screening out a second target data feature from the first target data feature based on the correlation degree; determining a health score corresponding to the second target data feature based on a health score calculation function; determining a state assessment result of the solid-state hard disk based on the health score corresponding to the second target data feature, and visually displaying the state assessment result.

[0005] Optionally, obtaining relevant parameter information of the solid state drive, and preprocessing the relevant parameter information to obtain the first target data feature includes: Collecting relevant parameter information of the solid-state hard disk, integrating the collected relevant parameter information of the solid-state hard disk, and generating a first target data set; Filling missing values ​​in the target data set to obtain a second target data set; Standardizing the data in the second target data set to obtain a third target data set; Data in the third target data set are merged to generate merged data features in multiple dimensions, and the merged data features in multiple dimensions are defined as the first target data features.

[0006] Optionally, before obtaining the joint probability distribution value between the historical merged data features and the corresponding health degree, the method further includes: Based on the value range, the historical merged data characteristics and health are divided into intervals to obtain the interval division results; Based on the interval division results, the total number of samples is calculated and determined, and the calculation formula includes:

[0007] in, represents the total number of samples, Indicates the number of merged data feature intervals, The number of intervals representing health, Indicates the number of samples in the interval.

[0008] Optionally, a method for obtaining a joint probability distribution value between historical merged data features and corresponding health levels includes: According to the interval combination result of the historical merged data characteristics and the health degree, the joint probability distribution value between the historical merged data characteristics and the corresponding health degree is calculated and determined. The calculation formula includes:

[0009] in, The interval number representing the feature interval of the merged data, Indicates the value range of the merged data feature X in the interval [d, d+1], The interval number indicating the health level, Indicates The merged data characteristics of the dimension, Indicates health, represents the joint probability distribution value, Represents the characteristic value of health in the interval [k, k+1].

[0010] Optionally, the method for determining the correlation between the historical merged data feature and the corresponding health degree includes: The correlation between the historical merged data feature and the corresponding health degree is calculated based on a correlation calculation function, wherein the correlation calculation function includes:

[0011] in, Indicates the degree of association, represents the joint probability distribution value, and Both represent marginal probability distribution values, Indicates the merged data features, Indicates health.

[0012] Optionally, filtering out the second target data feature from the first target data feature according to the correlation degree includes: Sort the correlations in descending order; According to the sorting result, merged data features of a preset number of dimensions are selected from the first target data features as the second target data features.

[0013] Optionally, determining the health score corresponding to the second target data feature based on the health score calculation function includes: A health score calculation function is constructed based on the weighted score item, the data feature historical change trend item and the environmental correction item. The health score calculation function includes:

[0014] in, represents the health score, represents the weighted score, Indicates the historical change trend value of data features. Indicates the environmental correction value; Based on the health score calculation function, a health score corresponding to the second target data feature is calculated and determined.

[0015] Optionally, before determining the health score corresponding to the second target data feature based on the health score calculation function, the method further includes: Based on a smoothing function, mapping conversion processing is performed on the second target data feature, and the smoothing function includes:

[0016]

[0017]

[0018] in, represents the data characteristics after smoothing. Indicates The data features of the dimensions, express The standard deviation of Indicates the preset threshold value, represents the penalty coefficient, represents a constant coefficient; The weight of the data feature in the second target data feature is calculated based on the weight calculation function, wherein the weight calculation function includes:

[0019]

[0020] in, represents the normalized weight, express The sum of the weights corresponding to the dimensions, represents the initial weight, represents the constant coefficient, Represents the weight of the data feature of the i-th dimension.

[0021] Optional, constant coefficient Methods for determining include: Based on the historical database, obtaining a third target data feature corresponding to the second target data feature, and a target health corresponding to the third target data feature; Calculating the health score corresponding to the third target data feature based on the health score calculation function; In response to detecting that the health score corresponding to the third target data feature does not belong to the value range corresponding to the target health level, the constant coefficient Correction is performed until the health score corresponding to the third target data feature falls within the value range corresponding to the target health level.

[0022] Optionally, the method for determining the weighted score includes: The weighted score is calculated and determined based on the weight calculation result and the weighted score calculation function, wherein the weighted score calculation function includes:

[0023] in, represents the weighted score, Indicates the number of dimensions.

[0024] Optionally, the method for calculating the historical change trend value of the data feature includes:

[0025] in, Indicates the historical change trend value of data features. Indicated in The data feature value at the moment, represents the sliding window size, represents the number of dimensions, Indicates the current moment.

[0026] Optionally, the method for calculating the environmental correction value includes:

[0027]

[0028] in, represents the environmental correction value, Represents the characteristic value of environmental data, represents the correction factor, represents the total number of environmental items, represents the basic weight, represents the penalty coefficient, represents the nonlinear index, Indicates the threshold value corresponding to the environmental item.

[0029] Optionally, determining a status evaluation result of the solid state drive according to the health score corresponding to the second target data feature, and visually displaying the status evaluation result includes: Determine, based on a value range of a preset tag, a tag corresponding to a health score corresponding to the second target data feature, wherein the tag is used to describe the health of the solid state drive; The health score corresponding to the second target data feature and the label corresponding to the health score are visualized.

[0030] Optionally, after determining the status assessment result of the solid state drive according to the health score corresponding to the second target data feature, the method further includes: Based on the second target data feature and its corresponding label, a one-to-one mapping relationship is generated and vectorized; The vectorized processing results are stored in the historical database.

[0031] In a second aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: obtaining relevant parameter information of a solid-state drive, preprocessing the relevant parameter information, and obtaining a first target data feature; determining a correlation between the historical merged data feature and the corresponding health degree based on a joint probability distribution value between the historical merged data feature and the corresponding health degree, and screening out a second target data feature from the first target data feature based on the correlation degree; determining a health score corresponding to the second target data feature based on a health score calculation function; determining a status assessment result of the solid-state drive based on the health score corresponding to the second target data feature, and visually displaying the status assessment result.

[0032] According to a third aspect, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: obtaining relevant parameter information of a solid-state drive, preprocessing the relevant parameter information, and obtaining a first target data feature; determining a correlation between the historical merged data feature and the corresponding health degree based on a joint probability distribution value between the historical merged data feature and the corresponding health degree, and filtering out a second target data feature from the first target data feature based on the correlation; determining a health score corresponding to the second target data feature based on a health score calculation function; determining a status evaluation result of the solid-state drive based on the health score corresponding to the second target data feature, and visually displaying the status evaluation result.

[0033] In a fourth aspect, a computer program product is provided, comprising a computer program, which implements the following steps when executed by a processor: obtaining relevant parameter information of a solid-state drive, preprocessing the relevant parameter information, and obtaining a first target data feature; determining a correlation between the historical merged data feature and the corresponding health degree based on a joint probability distribution value between the historical merged data feature and the corresponding health degree, and screening out a second target data feature from the first target data feature based on the correlation degree; determining a health score corresponding to the second target data feature based on a health score calculation function; determining a status assessment result of the solid-state drive based on the health score corresponding to the second target data feature, and visually displaying the status assessment result.

[0034] The above-mentioned state assessment method, device, storage medium and program product of the solid-state hard disk, the method includes: obtaining relevant parameter information of the solid-state hard disk, preprocessing the relevant parameter information, and obtaining a first target data feature; determining the correlation between the historical merged data feature and the corresponding health degree according to the joint probability distribution value between the historical merged data feature and the corresponding health degree, and screening out the second target data feature from the first target data feature according to the correlation degree; determining the health score corresponding to the second target data feature based on the health score calculation function; determining the state assessment result of the solid-state hard disk according to the health score corresponding to the second target data feature, and visually displaying the state assessment result. The application can accurately determine the data feature most associated with the health degree from the multi-dimensional parameter information through the constructed association relationship to determine the health score of the solid-state hard disk, and then effectively determine the change in the health degree of the solid-state hard disk, thereby improving the accuracy of the solid-state hard disk health score calculation, and then improving the accuracy of the solid-state hard disk state assessment, which helps operation and maintenance personnel to understand the health status of the solid-state hard disks in the cluster in advance, and can respond to risk disks in advance, effectively reducing the risk of data loss or server downtime caused by solid-state hard disk failures. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A diagram of an application environment of a method for evaluating a state of a solid state drive in one embodiment; Figure 2 A schematic diagram of a flow chart of a method for evaluating a state of a solid state drive in one embodiment; Figure 3 is another schematic diagram of a flow chart of a method for evaluating a state of a solid state drive in one embodiment; Figure 4 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0037] It should be understood that in the description of the present application, unless the context clearly requires otherwise, words such as "include", "comprises", and the like throughout the specification should be interpreted as including rather than being exclusive or exhaustive; that is, as including but not limited to.

[0038] It should also be understood that the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of this application, unless otherwise specified, "plurality" means two or more.

[0039] It should be noted that the terms "S1", "S2", etc. are only used for the purpose of describing the steps, and do not specifically refer to the order or sequence, nor are they used to limit the present application. They are only for the convenience of describing the method of the present application, and cannot be understood as indicating the order of the steps. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in this field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present application.

[0040] According to the background technology, with the rapid development of information technology, the demand for data storage continues to grow, and solid-state drives have been widely used in various fields. As an important means to ensure the reliability of data storage, SSD disk health scoring technology is also constantly developing and improving. During the use of SSD, it may gradually enter a sub-healthy state due to various reasons. In this state, although the hard disk may still work normally, the performance may decline and there may be potential failure risks. If it is not discovered and handled in time, it may lead to serious consequences such as data loss and system failure. Therefore, health scoring of SSD disks can warn of potential problems in advance and provide users with enough time to take measures to ensure data security and stable system operation.

[0041] Early SSD health scores mainly relied on simple indicators, such as the usage time and read / write times of the hard disk. Although these indicators can reflect the usage of the hard disk to a certain extent, they have limited accuracy in judging the complex health status. With the continuous advancement of technology, more advanced technologies have been applied to SSD health scores. For example, by monitoring the temperature, voltage, current and other parameters of the hard disk, we can have a more comprehensive understanding of the working status of the hard disk. Too high a temperature may mean that the hard disk has poor heat dissipation and there is a risk of failure. Unstable voltage may affect the normal operation of the hard disk. At the same time, using SMART (Self-Monitoring Analysis and Reporting Technology, automatic monitoring analysis and reporting technology, solid-state drives can automatically monitor their own status and provide a series of parameters for users to analyze. These parameters include but are not limited to the number of bad blocks, reallocated sector counts, read error rates, etc., providing a richer data source for health scoring; in recent years, machine learning and artificial intelligence technologies have also begun to play an important role in the field of SSD health scoring. By learning a large amount of hard disk data, machine learning models can automatically identify the health status of hard disks and give corresponding scores. This method can not only improve the accuracy of the score, but also adapt to different hard disk models and usage environments, and has stronger versatility; in actual production environments, disk failures have the characteristics of low probability and complex and diverse. The specific problems are as follows: 1. It is difficult to determine the type of failure: the diversity of failures increases the difficulty of identification; 2. Reliance on manual labeling: a large amount of data needs to be manually labeled, which is costly and inefficient; 3. Scarce data sets: limited relevant failure data further exacerbates the difficulty of model training; 4. Prediction accuracy problems under low computing load: in an environment with limited resources, it is difficult to achieve high-precision prediction of failures.

[0042] Supervised learning methods in machine learning and artificial intelligence technologies rely on accurately labeled data sets for training, and in particular require sufficient positive and negative samples to improve the generalization ability of the model. However, due to the imbalance of positive and negative samples, the existing models can only identify a limited number of fault types and have high requirements for response speed. In addition, the accuracy and sensitivity of disk failure prediction are still relatively low and cannot fully meet actual needs.

[0043] Furthermore, existing solutions for SSD health scoring can be roughly divided into two categories: one is based on statistical analysis and expert rules, and the other is based on artificial intelligence algorithms. The following are their specific solutions: SSD health scoring method based on statistical analysis and expert rules: For each selected indicator, a reasonable threshold range is set according to empirical values, manufacturer recommendations or a large amount of experimental data. For example, the remaining life percentage is divided into several intervals, such as greater than 80% for excellent, 50%-80% for good, 30%-50% for medium, and less than 30% for poor; the collected SSD indicator data is compared with the set threshold. If the value of an indicator falls within the excellent interval, a higher score is given; if it falls within the poor interval, a lower score is given. The comprehensive health score of the SSD is calculated based on the scores and weights of each indicator. The weighted average method can be used to multiply the score of each indicator by its weight, and then sum them up to get the comprehensive score.

[0044] Based on artificial intelligence algorithms: By collecting a large amount of SSD data, including SMART parameters, performance indicators, usage time, etc., and using machine learning algorithms for training, the trained model can automatically determine the health status of the SSD based on the newly input data and give a score. For example, using algorithms such as random forests and support vector machines to learn from historical data, it can more accurately predict the health status of the SSD.

[0045] The disadvantages of the above existing solutions include: 1. Large limitations: Statistical analysis is usually based on the probability distribution of historical data for inference, but the failure mode of SSD may change with the development of technology and changes in the usage environment, making it difficult for methods based on fixed statistical models to adapt to new situations; 2. Lack of adaptability: Once the statistical model and expert rules are determined, it is often difficult to automatically adjust according to the real-time changes of the SSD in actual applications. For example, when a new SSD model or special usage scenario appears, it may not be possible to timely and accurately evaluate its health status; 3. Difficult to handle complex situations: For complex situations where multiple factors interact, statistical analysis and expert rules may be difficult to comprehensively and accurately consider. For example, the health status of the SSD may be affected by multiple factors such as temperature, read and write frequency, and years of use. The complex relationship between these factors may be difficult to describe with simple statistical models and rules; 4. Traditional artificial intelligence algorithms require a large amount of high-quality data for training. If the data is insufficient or the data quality is not high, the model may be inaccurate. In addition, obtaining sufficient and accurately labeled SSD health data is a challenging task.

[0046] In order to solve the above technical problems, the present application provides a state assessment method, device, equipment and storage medium for solid-state hard disks. By combining multi-source data and artificial intelligence technology, the collection of SMART data and performance data of SSD disks is realized, which can adapt to the health scoring requirements in different work scenarios and is widely applicable to SSD hard disks of different brands, types and usage scenarios. By calculating the correlation between each data feature and the health degree, the data feature with the greatest correlation with the health degree is obtained, and based on the different health scores of the SSD disks, the health of the SSD disks in the server cluster is graded, thereby effectively determining the change in the health degree of the solid-state hard disks, improving the accuracy of the solid-state hard disk health score calculation, and thereby improving the accuracy of the solid-state hard disk state assessment, which helps operation and maintenance personnel to understand the health status of the solid-state hard disks in the cluster in advance, and can take response measures to risky disks in advance, effectively reducing the risk of data loss or server downtime caused by solid-state hard disk failures.

[0047] The state evaluation method of the solid state drive provided in this application can be applied to Figure 1 In the application environment shown, the terminal 102 communicates with the data processing platform set on the server 104 through the network, wherein the terminal 102 can be but not limited to various personal computers, laptops, smart phones, tablet computers and portable wearable devices, and the server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0048] In one embodiment, Figure 2 As shown, a method for evaluating the state of a solid state hard disk is provided. Figure 1 The terminal in is used as an example to illustrate, including the following steps: S1: Obtain relevant parameter information of a solid state drive, and preprocess the relevant parameter information to obtain a first target data feature.

[0049] It should be noted that solid-state drives refer to SSDs, which are hard disks made of solid-state electronic storage chip arrays; relevant parameter information may include performance parameter information such as read and write speeds, response time and IOPS (input / output operations per second), as well as SMART log information such as remapped sector counts, spindle spin-up retries, head calibration retry counts, erase failed block counts, point-to-point error detection counts, uncorrectable error counts, hardware ECC (error checking and correction) correction counts, current to-be-mapped sector counts, offline uncorrectable sector counts and ULTRADMA (Ultra Direct Memory Access) parity error rate.

[0050] In some specific implementations, obtaining relevant parameter information of the solid state drive, and preprocessing the relevant parameter information to obtain the first target data feature includes: Collect relevant parameter information of the solid-state hard disk, integrate the collected relevant parameter information of the solid-state hard disk, and generate a first target data set, wherein, based on the unified standard of the solid-state hard disk, obtain multiple performance status parameters of each training hard disk, and collect the performance indicator data and SMART information of the SSD hard disk by calling the monitoring device deployed on the server cluster. Each time the data is collected, the current latest data of the SSD will be obtained, and these data will be recorded to form a historical usage record. The data collection cycle can be set according to actual needs to ensure that the data is updated in a timely manner by regularly reading the relevant parameter information. After the data collection is completed, the data collected from different sources will be integrated to form a unified data set, such as integrating SMART parameters, performance indicators, device logs and temperature data together and storing them in Figure 3 In the intelligent agent center shown, so that the intelligent agent center can perform comprehensive analysis, such as data cleaning and health scoring; The missing values ​​in the target data set are filled to obtain a second target data set. Specifically, the data collected in the target data set are preprocessed and analyzed to extract useful feature information. For key SSD disk health parameters (such as temperature, remapped sector count, etc.), reasonable values ​​are inferred by combining with the large model to fill in. First, the missing values ​​are processed and inferred based on the existing data context to build a large model agent. For each column, the following prompt word prompt is constructed: [Please fill in the missing {column} value of the data. The value filling is based on the context content. Please execute it according to each column of the table data. If the filling fails, please automatically fill it in with the average value that can be calculated for this column], where, in the context of natural language processing and large language models, "prompt" usually refers to a prompt or guiding input to the model; The data in the second target data set is standardized to obtain the third target data set. Specifically, after the missing value filling is completed, the performance information and SMART information obtained from the SSD disk are standardized, such as normalizing the data features to prepare basic data for determining the health score of the SSD disk. The formula for normalizing the value of each data feature is:

[0051] in, represents the original index, and are the minimum and maximum values ​​of the indicator respectively. To scale values ​​between 0 and 100, both historical reference data and data features used for future predictions need to be standardized; The data in the third target data set are merged to generate merged data features of multiple dimensions, and the merged data features of multiple dimensions are defined as the first target data features. Specifically, the performance indicator information and SMART information after the standardization processing are merged into a unified tabular data, that is, the merged data features of multiple dimensions.

[0052] In the above implementation, the SMART data and performance data of the solid-state drive are collected from multiple sources, and the collected data is preprocessed for subsequent calculation of the health score of the solid-state drive. Based on the collection of multi-dimensional data, this application can adapt to the health score requirements in different work scenarios and can be widely used in SSD hard drives of different brands, types, and usage scenarios.

[0053] S2: Determine the correlation between the historical merged data features and the corresponding health levels according to the joint probability distribution value between the historical merged data features and the corresponding health levels, and select the second target data features from the first target data features according to the correlation.

[0054] It should be noted that the joint probability distribution value refers to the probability that two or more random variables take specific values ​​at the same time; the correlation is used to describe the correlation between two data. The larger the correlation, the greater the contribution of one data to the other.

[0055] In some specific implementations, before obtaining the joint probability distribution value between the historical merged data features and the corresponding health levels, the method further includes: Based on the value range, the historical merged data features and health are divided into intervals to obtain the interval division results. Specifically, a batch of historical data of SSD health disk scores is obtained, including the collected performance indicator information, SMART features and corresponding health. H is defined to represent health, and the performance indicator information and SMART information are merged into a unified tabular data. The merged data feature of the jth dimension is defined as For historical data features, such as SMART feature information, it can be divided into several intervals according to its value range, and the health can also be divided into specific intervals. The number of interval divisions can be set according to actual needs. For example, the health can include four levels: excellent, good, medium and poor. The corresponding division intervals are [80, 100] for excellent, [50, 80] for good, [30, 50] for medium, and less than 30 for poor. The SMART features and performance index values ​​can be obtained by using the equal spacing method. The number of intervals is set to 10, that is, d takes a value of 1~10; Based on the interval division results, the total number of samples is calculated and determined, and the calculation formula includes:

[0056] in, represents the total number of samples, Indicates the number of merged data feature intervals, that is, the number of intervals of data after any dimension is merged. The number of intervals representing health, Indicates the number of samples in the interval.

[0057] In some specific implementations, a method for obtaining a joint probability distribution value between historical merged data features and corresponding health levels includes: According to the interval combination result of the historical merged data features and the health degree, that is, according to each interval combination of the historical merged data features and the health degree, the joint probability distribution value between the historical merged data features and the corresponding health degree is calculated and determined. The calculation formula includes:

[0058] in, The interval number representing the feature interval of the merged data, Indicates the value range of the merged data feature X in the interval [d, d+1], where d is 1 to 10. The interval number indicating the health level, Indicates The merged data characteristics of the dimension, Indicates health, represents the joint probability distribution value, It represents the characteristic value of health in the interval [k, k+1], including four levels: excellent, good, medium and poor. The value range of health score is defined as 0~100, then the characteristic value range of health is [80, 100], [50, 80), [30, 50) and less than 30. For example, assuming there are 1000 samples, There are 200 samples in the interval [20, 40) and H in the interval [40, 60). Then .

[0059] In some specific implementations, the dependency relationship between the merged data features of each dimensional merge and the health degree is calculated based on the obtained joint probability distribution value. Specifically, the method for determining the correlation between the historical merged data features and the corresponding health degree includes: The correlation between the historical merged data feature and the corresponding health degree is calculated based on a correlation calculation function, wherein the correlation calculation function includes:

[0060] in, Indicates the degree of association, represents the joint probability distribution value, and Both represent marginal probability distribution values, Indicates the merged data features, Indicates health.

[0061] In some specific implementations, selecting the second target data feature from the first target data feature according to the association degree includes: The associations are sorted in descending order. Specifically, The bigger, the For the target variable The greater the contribution, based on this, the correlation degrees are sorted so that the data features of the first N dimensions can be selected from large to small as the candidate feature set; According to the sorting result, merged data features of a preset number of dimensions are screened out from the first target data feature as the second target data feature, wherein the preset number can be set according to actual needs. The preferred value of this application is 11, that is, based on the data features of the first 11 dimensions in the sorting result, the corresponding merged data feature is selected from the first target data feature as the second target data feature.

[0062] In the above implementation, by constructing the association relationship, the data features most associated with the health status in the preprocessed data features are accurately captured, and the health score is calculated based on the associated data, which can effectively obtain the changes in the SSD health status captured by each collected data feature, thereby improving the accuracy of the health score result.

[0063] In some embodiments, the method further comprises: Obtaining the identifier of the solid-state drive to which the first target data feature belongs, and detecting whether there is a solid-state drive with the same identifier and its corresponding fourth target data feature in a historical database, wherein the historical database is used to store a mapping relationship generated by the historical second target data feature and its corresponding solid-state drive identifier; If it exists, then the solid-state hard disk with the same identification and its corresponding fourth target data feature are extracted from the historical database. Based on the similarity algorithm, the fifth target data feature that is most similar to the fourth target data feature is extracted from the first target data feature, and the likelihood value between the fourth target data feature and the fifth target data feature is calculated. The specific calculation formula is:

[0064] in, represents the likelihood value, Indicates the fourth target data feature data feature values, Indicates the fifth target data feature data feature values, the attributes of the two data features used for calculation have the same meaning; In response to detecting that the likelihood value is greater than a preset threshold, using the fourth target data feature as the second target data feature for subsequent calculation of a health score of the solid state drive; If not, the correlation is calculated according to the above steps, and the second target data feature is determined according to the correlation for subsequent calculation of the health score of the solid state drive.

[0065] In the above implementation, by calculating the likelihood value between the current data feature and the historical data feature in the historical database, the relevant data features that do not need to calculate the correlation are determined, based on which the evaluation efficiency of the solid state drive status can be improved.

[0066] S3: Determine the health score corresponding to the second target data feature based on the health score calculation function.

[0067] It should be noted that the health score calculation function includes a weighted score item, a data feature historical change trend item, and an environmental correction item.

[0068] In some specific implementations, before determining the health score corresponding to the second target data feature based on the health score calculation function, the method further includes: Define the combined data features of the preset number of dimensions obtained above as [X 1 , ..., X f ], based on a smoothing function, mapping and transforming the second target data feature so that all data features are smooth, wherein the smoothing function includes:

[0069]

[0070] in, represents the data characteristics after smoothing. Indicates The data features of the dimension, i ranges from 1 to f. If the preset number is 11, then i ranges from 1 to 11, which is the same as N. express The standard deviation of Furthermore, in order to more clearly reflect the changes in health, such as a certain parameter will significantly reduce the health score after exceeding a certain value, a threshold is set for each data feature. , if the parameter exceeds the threshold, the health score will be greatly affected. Therefore, the data features are adjusted based on the threshold. The specific adjustment formula is:

[0071] in, Indicates the preset threshold value, represents the penalty coefficient, Represents constant coefficient, penalty coefficient It is a key parameter in the health scoring system that controls the degree of change in the health score after exceeding the threshold. It is usually constructed based on the risk level of the parameter, historical failure data, equipment life, and the needs of the scoring system. It can be optimized in a data-driven way. , ensure that the impact of each SMART parameter on disk health can be reasonably reflected in the scoring formula, based on expert experience, smooth transition: smaller Values, such as =1.5~2, rapid change: larger Values, such as =4~5; Furthermore, based on a weight calculation function, the weight of the data feature in the second target data feature is calculated, and the weight calculation function includes:

[0072]

[0073] in, represents the normalized weight, express The sum of the weights corresponding to the dimensions. If N=11, , Represents the initial weight, which is generally 1. It represents the constant coefficient, and its initial value is set to 0.2. It can be modified according to the subsequent health comparison results. Represents the weight of the data feature of the i-th dimension.

[0074] In some specific implementations, in addition to the collected performance indicators and SMART indicator parameters, historical trends and the environment in which the server cluster is located are also important factors affecting the health of the SSD disk. Therefore, when calculating the health score, in addition to weighted summation of the performance indicator parameters and SMART indicator parameters, it is also necessary to consider the impact of historical trend changes and the environment in which the server cluster is located on the SSD disk. Based on the weight calculation result and the weighted score calculation function, the weighted calculation score of the performance indicator parameters and SMART parameters is calculated and determined. The weighted score calculation function includes:

[0075] in, represents the weighted score, represents the number of dimensions, Represents the data features after threshold adjustment.

[0076] For example, if is 11, then , ,... is the weight of each data feature, , ..., It is the data feature after threshold adjustment, and its value ranges from 0 to 100.

[0077] In some specific implementations, when calculating the health score, not only the current data feature value should be considered, but also its historical change trend should be referred to. If a data feature shows a trend of rapid deterioration in the near future, even if its current data feature value does not exceed the threshold, a lower health score should be given. Based on this, the historical trend of each data feature can be calculated by moving average or weighted moving average, and incorporated into the scoring formula. Specifically, the calculation method of the historical change trend value of the data feature includes:

[0078] in, Indicates the historical change trend value of data features. Indicated in The data feature value at the moment, represents the sliding window size, represents the number of dimensions, Indicates the current moment.

[0079] In some specific implementations, environmental factors have a significant impact on the life of the solid state drive. For example, in a high temperature environment, the solid state drive may be more prone to failure. In order to reflect this, an environmental correction item is introduced to correct the health score through sensor data (such as temperature, humidity, etc.). The environmental correction item can be calculated based on the degree of deviation of the environmental parameters. The specific calculation method of the environmental correction value includes:

[0080]

[0081] in, represents the environmental correction value, Represents the characteristic value of environmental data, which may include the solid state drive temperature E collected by the rack sensor 1 , room temperature E 2 、Humidity in the equipment room E 3 , vibration intensity E 4 and air velocity E 5 , represents the correction factor, represents the total number of environmental items, Indicates the basic weight. According to expert experience, the basic weights set include temperature basic weight of 1.2, vibration basic weight of 1.0, humidity basic weight of 1.1, and other basic weights are 1.0 by default. It represents the penalty coefficient. According to expert experience, the temperature penalty coefficient can be set to 2.5, the vibration penalty coefficient to 1.8, the humidity penalty coefficient to 1.5, and the other penalty coefficients are set to 1.0 by default. Represents the nonlinear index, the preferred value is 1.8, Indicates the threshold value corresponding to the environmental item, preferably, the solid state drive temperature E 1 The temperature threshold is 50, the room temperature E 2 The temperature threshold is 27, the humidity threshold is 45, the vibration intensity threshold is 0.5, and the air flow speed threshold is 2.5.

[0082] In some specific implementations, determining the health score corresponding to the second target data feature based on the health score calculation function includes: A health score calculation function is constructed based on the weighted score item, the data feature historical change trend item and the environmental correction item. The health score calculation function includes:

[0083] in, represents the health score, represents the weighted score, Indicates the historical change trend value of data features. Indicates the environmental correction value; Based on the health score calculation function, a health score corresponding to the second target data feature is calculated and determined.

[0084] In some specific embodiments, based on the health score calculation function, before determining the health score corresponding to the second target data feature, due to the initial constant coefficient The value is set to 0.2, which does not necessarily meet the accuracy requirements of the real score. Therefore, the constant coefficient in the weighting formula needs to be Correct as needed, where the constant coefficient Methods for determining include: Based on the historical database, a third target data feature corresponding to the second target data feature and a target health corresponding to the third target data feature are obtained, wherein the standardized data is structured by using a large model, and the data in the historical database is efficiently retrieved by using a similarity algorithm (such as cosine similarity), and the data feature that is most similar to the second target data feature is determined as the third target data feature; Based on the health score calculation function, the health score corresponding to the third target data feature is calculated, that is, the reference health score corresponding to the third target data feature is determined by calculating the health score calculation function constructed above. _参照 ; Check the reference health score _参照Whether it is within the score range of the target health degree corresponding to the third target data feature, as described above, the score range of the health degree is excellent - [80, 100], good - [50, 80], medium - [30, 50) and poor - less than 30, in response to detecting that the health score corresponding to the third target data feature does not belong to the value range corresponding to the target health degree, the constant coefficient Correct until the health score corresponding to the third target data feature falls within the value range corresponding to the target health level. Specifically, if _参照 If the score is not within the target health label range, the constant coefficient needs to be Make adjustments if _参照 Greater than the reference label upper limit S _max value, reduce the constant coefficient , and recalculate _参照 , until _参照 Falls within the value range of the reference health label [S _min , S _max ], if _参照 Less than the lower limit S of the value range of the reference health label _min value, increase the constant coefficient , and recalculate _参照 , until _参照 Falls within the value range of the reference health label [S _min , S _max ], where the constant coefficient is reduced or increased by the same amount according to the preset constant value , if the preset constant value is 0.1, the initial constant coefficient is 0.2, then the first constant coefficient Increase or decrease by 0.1, which is 0.3 or 0.1, and the second constant coefficient after the increase or decrease Increase or decrease by 0.1 to 0.4 or 0, and so on.

[0085] When the constant coefficient When it is determined, the preprocessed second target data feature is input into the health score calculation function to obtain the health score of the solid state drive.

[0086] In the above implementation, historical trends and environmental factors are comprehensively considered, and historical factors and external environmental factors that affect the health of the SSD disk are included in the health score, so as to further improve the accuracy of the SSD disk health score. The relevant parameters in the calculation process of each weighted indicator can be adaptively adjusted and updated, so as to continuously optimize the health score model, and have higher sensitivity to new models and new brands of SSD disks, so as to effectively adapt to the health scoring requirements of different brands, different working conditions, and different computer rooms and clusters, thereby improving the intelligence of the health scoring process and the accuracy of the health scoring results.

[0087] S4: Determine a status assessment result of the solid state drive according to the health score corresponding to the second target data feature, and visualize the status assessment result.

[0088] It should be noted that the status assessment result of the solid state drive is the health level determined based on the health score.

[0089] In some specific implementations, determining a status evaluation result of the solid state drive according to a health score corresponding to the second target data feature, and visually displaying the status evaluation result includes: Based on the value range of the preset label, determine the label corresponding to the health score corresponding to the second target data feature. The label is used to describe the health of the solid state drive. For example, if the value range of the health score is [80, 100], the health is excellent, and the label is excellent. This step gives the health score of the solid state drive through Prompt and maps the health score to the corresponding label. Prompt is specifically: "Based on the calculated total score, map the total score to the label of "excellent", "good", "medium" or "poor". The rule is: map the total health score to one of the four labels: [80, 100] points are labeled as excellent; [50, 80) points are labeled as good; [30, 50) points are labeled as medium; and less than 30 points are labeled as poor. Please output the SSD health score and label according to these rules"; The health score corresponding to the second target data feature and the label corresponding to the health score are visualized, specifically, Figure 3 As shown, the SSD health score and label are interacted with the front-end display interface and output in a visual form.

[0090] In some specific implementations, after determining the status assessment result of the solid state drive according to the health score corresponding to the second target data feature, the method further includes: Based on the second target data feature and its corresponding label, a one-to-one mapping relationship is generated and vectorized; The vectorized processing results are stored in the historical database.

[0091] For example, if the dimension is 11, it includes 11 types of feature data, S 1 ~S 11 These are 11 data features selected based on the correlation. The data features of SSDs of different brands are slightly different. 1 ~S 11 The data features are shown in Table 1. The data with health score labels of SSD disks of the same brand are output in the following form: {X=[attribute name 1: N 1 , attribute name 2: N 2 , ..., attribute name 11: N 11 ], Y = "health"}, health includes four levels: excellent, good, medium and poor. According to the health score mapping described in the above rules, greater than or equal to 80 points are excellent, greater than or equal to 50 and less than 80 points are good, greater than or equal to 30 and less than 50 points are medium, and less than 30 points are poor. The data features and labels of the SSD disk are presented in a vectorized form, such as using a large model (such as an embedding model) to convert certain key structural features into vector representations. For example: "Reallocated_Sector_Ct = 120 -> embedding vector [0.12, 0.34, 0.56, ...], PE_cycles = 600 -> embedding vector [0.23, 0.45, 0.67, ...]" After the data features are vectorized, the corresponding label vectorization results are combined to generate a one-to-one mapping relationship and store it in the vector database for later use in constant coefficients. The value is corrected.

[0092] Table 1: Data characteristics table.

[0093]

[0094] In the above implementation, by visually displaying the health score results, the results can be presented to the operation and maintenance personnel more intuitively, thereby improving the work efficiency of the operation and maintenance personnel. At the same time, the health score results are stored in a database for continuous updating and optimization of the health score model, which can further improve the accuracy of the health score results.

[0095] In the above-mentioned state assessment method of the solid-state hard disk, the method includes: obtaining relevant parameter information of the solid-state hard disk, preprocessing the relevant parameter information, and obtaining a first target data feature; determining the correlation between the historical merged data feature and the corresponding health degree according to the joint probability distribution value between the historical merged data feature and the corresponding health degree, and screening out the second target data feature from the first target data feature according to the correlation degree; determining the health score corresponding to the second target data feature based on the health score calculation function; determining the state assessment result of the solid-state hard disk according to the health score corresponding to the second target data feature, and visually displaying the state assessment result. The present application can accurately determine the data feature most associated with the health degree from the multi-dimensional parameter information through the constructed association relationship to determine the health score of the solid-state hard disk, and then effectively determine the change in the health degree of the solid-state hard disk, thereby improving the accuracy of the solid-state hard disk health score calculation, and then improving the accuracy of the solid-state hard disk state assessment, which helps operation and maintenance personnel to understand the health status of the solid-state hard disks in the cluster in advance, and can respond to risk disks in advance, effectively reducing the risk of data loss or server downtime caused by solid-state hard disk failures.

[0096] It should be understood that although Figure 2-Figure 3 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2-Figure 3 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0097] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 4As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a state evaluation method of a solid-state hard disk is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0098] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0099] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: S1: Acquire relevant parameter information of the solid state drive, and preprocess the relevant parameter information to obtain a first target data feature; S2: determining the correlation between the historical merged data feature and the corresponding health degree according to the joint probability distribution value between the historical merged data feature and the corresponding health degree, and selecting the second target data feature from the first target data feature according to the correlation degree; S3: Determine the health score corresponding to the second target data feature based on the health score calculation function; S4: Determine a status assessment result of the solid state drive according to the health score corresponding to the second target data feature, and visualize the status assessment result.

[0100] In one embodiment, when the processor executes the computer program, the processor further implements the following steps: Collecting relevant parameter information of the solid-state hard disk, integrating the collected relevant parameter information of the solid-state hard disk, and generating a first target data set; Filling missing values ​​in the target data set to obtain a second target data set; Standardizing the data in the second target data set to obtain a third target data set; Data in the third target data set are merged to generate merged data features in multiple dimensions, and the merged data features in multiple dimensions are defined as the first target data features.

[0101] In one embodiment, when the processor executes the computer program, the processor further implements the following steps: Based on the value range, the historical merged data characteristics and health are divided into intervals to obtain the interval division results; Based on the interval division results, the total number of samples is calculated and determined, and the calculation formula includes:

[0102] in, represents the total number of samples, Indicates the number of merged data feature intervals, The number of intervals representing health, Indicates the number of samples in the interval.

[0103] In one embodiment, when the processor executes the computer program, the processor further implements the following steps: According to the interval combination result of the historical merged data characteristics and the health degree, the joint probability distribution value between the historical merged data characteristics and the corresponding health degree is calculated and determined. The calculation formula includes:

[0104] in, The interval number representing the feature interval of the merged data, Indicates the value range of the merged data feature X in the interval [d, d+1], The interval number indicating the health level, Indicates The merged data characteristics of the dimension, Indicates health, represents the joint probability distribution value, Represents the characteristic value of health in the interval [k, k+1].

[0105] In one embodiment, when the processor executes the computer program, the processor further implements the following steps: The correlation between the historical merged data feature and the corresponding health degree is calculated based on a correlation calculation function, wherein the correlation calculation function includes:

[0106] in, Indicates the degree of association, represents the joint probability distribution value, and Both represent marginal probability distribution values, Indicates the merged data features, Indicates health.

[0107] In one embodiment, when the processor executes the computer program, the processor further implements the following steps: Sort the correlations in descending order; According to the sorting result, merged data features of a preset number of dimensions are selected from the first target data features as the second target data features.

[0108] In one embodiment, when the processor executes the computer program, the processor further implements the following steps: A health score calculation function is constructed based on the weighted score item, the data feature historical change trend item and the environmental correction item. The health score calculation function includes:

[0109] in, represents the health score, represents the weighted score, Indicates the historical change trend value of data features. Indicates the environmental correction value; Based on the health score calculation function, a health score corresponding to the second target data feature is calculated and determined.

[0110] In one embodiment, when the processor executes the computer program, the processor further implements the following steps: Based on a smoothing function, mapping conversion processing is performed on the second target data feature, and the smoothing function includes:

[0111]

[0112]

[0113] in, represents the data characteristics after smoothing. Indicates The data features of the dimensions, express The standard deviation of Indicates the preset threshold value, represents the penalty coefficient, represents a constant coefficient; The weight of the data feature in the second target data feature is calculated based on the weight calculation function, wherein the weight calculation function includes:

[0114]

[0115] in, represents the normalized weight, express The sum of the weights corresponding to the dimensions, represents the initial weight, represents the constant coefficient, Represents the weight of the data feature of the i-th dimension.

[0116] In one embodiment, when the processor executes the computer program, the processor further implements the following steps: Based on the historical database, obtaining a third target data feature corresponding to the second target data feature, and a target health corresponding to the third target data feature; Calculating the health score corresponding to the third target data feature based on the health score calculation function; In response to detecting that the health score corresponding to the third target data feature does not belong to the value range corresponding to the target health level, the constant coefficient Correction is performed until the health score corresponding to the third target data feature falls within the value range corresponding to the target health level.

[0117] In one embodiment, when the processor executes the computer program, the processor further implements the following steps: The weighted score is calculated and determined based on the weight calculation result and the weighted score calculation function, wherein the weighted score calculation function includes:

[0118] in, represents the weighted score, Indicates the number of dimensions.

[0119] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0120] in, Indicates the historical change trend value of data features. Indicated in The data feature value at the moment, represents the sliding window size, represents the number of dimensions, Indicates the current moment.

[0121] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0122]

[0123] in, represents the environmental correction value, Represents the characteristic value of environmental data, represents the correction factor, represents the total number of environmental items, represents the basic weight, represents the penalty coefficient, represents the nonlinear index, Indicates the threshold value corresponding to the environmental item.

[0124] In one embodiment, when the processor executes the computer program, the processor further implements the following steps: Determine, based on a value range of a preset tag, a tag corresponding to a health score corresponding to the second target data feature, wherein the tag is used to describe the health of the solid state drive; The health score corresponding to the second target data feature and the label corresponding to the health score are visualized.

[0125] In one embodiment, when the processor executes the computer program, the processor further implements the following steps: Based on the second target data feature and its corresponding label, a one-to-one mapping relationship is generated and vectorized; The vectorized processing results are stored in the historical database.

[0126] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: S1: Acquire relevant parameter information of the solid state drive, and preprocess the relevant parameter information to obtain a first target data feature; S2: determining the correlation between the historical merged data feature and the corresponding health degree according to the joint probability distribution value between the historical merged data feature and the corresponding health degree, and selecting the second target data feature from the first target data feature according to the correlation degree; S3: Determine the health score corresponding to the second target data feature based on the health score calculation function; S4: Determine a status assessment result of the solid state drive according to the health score corresponding to the second target data feature, and visualize the status assessment result.

[0127] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: Collecting relevant parameter information of the solid-state hard disk, integrating the collected relevant parameter information of the solid-state hard disk, and generating a first target data set; Filling missing values ​​in the target data set to obtain a second target data set; Standardizing the data in the second target data set to obtain a third target data set; Data in the third target data set are merged to generate merged data features in multiple dimensions, and the merged data features in multiple dimensions are defined as the first target data features.

[0128] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: Based on the value range, the historical merged data characteristics and health are divided into intervals to obtain the interval division results; Based on the interval division results, the total number of samples is calculated and determined, and the calculation formula includes:

[0129] in, represents the total number of samples, Indicates the number of merged data feature intervals, The number of intervals representing health, Indicates the number of samples in the interval.

[0130] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: According to the interval combination result of the historical merged data characteristics and the health degree, the joint probability distribution value between the historical merged data characteristics and the corresponding health degree is calculated and determined. The calculation formula includes:

[0131] in, The interval number representing the feature interval of the merged data, Indicates the value range of the merged data feature X in the interval [d, d+1], The interval number indicating the health level, Indicates The merged data characteristics of the dimension, Indicates health, represents the joint probability distribution value, Represents the characteristic value of health in the interval [k, k+1].

[0132] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: The correlation between the historical merged data feature and the corresponding health degree is calculated based on a correlation calculation function, wherein the correlation calculation function includes:

[0133] in, Indicates the degree of association, represents the joint probability distribution value, and Both represent marginal probability distribution values, Indicates the merged data features, Indicates health.

[0134] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: Sort the correlations in descending order; According to the sorting result, merged data features of a preset number of dimensions are selected from the first target data features as the second target data features.

[0135] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: A health score calculation function is constructed based on the weighted score item, the data feature historical change trend item and the environmental correction item. The health score calculation function includes:

[0136] in, represents the health score, represents the weighted score, Indicates the historical change trend value of data features. Indicates the environmental correction value; Based on the health score calculation function, a health score corresponding to the second target data feature is calculated and determined.

[0137] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: Based on a smoothing function, mapping conversion processing is performed on the second target data feature, and the smoothing function includes:

[0138]

[0139]

[0140] in, represents the data characteristics after smoothing. Indicates The data features of the dimensions, express The standard deviation of Indicates the preset threshold value, represents the penalty coefficient, represents a constant coefficient; The weight of the data feature in the second target data feature is calculated based on the weight calculation function, wherein the weight calculation function includes:

[0141]

[0142] in, represents the normalized weight, express The sum of the weights corresponding to the dimensions, represents the initial weight, represents the constant coefficient, Represents the weight of the data feature of the i-th dimension.

[0143] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: Based on the historical database, obtaining a third target data feature corresponding to the second target data feature, and a target health corresponding to the third target data feature; Calculating the health score corresponding to the third target data feature based on the health score calculation function; In response to detecting that the health score corresponding to the third target data feature does not belong to the value range corresponding to the target health level, the constant coefficient Correction is performed until the health score corresponding to the third target data feature falls within the value range corresponding to the target health level.

[0144] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: The weighted score is calculated and determined based on the weight calculation result and the weighted score calculation function, wherein the weighted score calculation function includes:

[0145] in, represents the weighted score, Indicates the number of dimensions.

[0146] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0147] in, Indicates the historical change trend value of data features. Indicated in The data feature value at the moment, represents the sliding window size, represents the number of dimensions, Indicates the current moment.

[0148] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0149]

[0150] in, represents the environmental correction value, Represents the characteristic value of environmental data, represents the correction factor, represents the total number of environmental items, represents the basic weight, represents the penalty coefficient, represents the nonlinear index, Indicates the threshold value corresponding to the environmental item.

[0151] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: Determine, based on a value range of a preset tag, a tag corresponding to a health score corresponding to the second target data feature, wherein the tag is used to describe the health of the solid state drive; The health score corresponding to the second target data feature and the label corresponding to the health score are visualized.

[0152] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented: Based on the second target data features and their corresponding labels, a one-to-one mapping relationship is generated and vectorized; The vectorized processing results are stored in the historical database.

[0153] In one embodiment, a computer program product is provided, the computer program product comprising a computer program, the computer program when executed by a processor implements the following steps: S1: Acquire relevant parameter information of the solid state drive, and preprocess the relevant parameter information to obtain a first target data feature; S2: determining the correlation between the historical merged data feature and the corresponding health degree according to the joint probability distribution value between the historical merged data feature and the corresponding health degree, and selecting the second target data feature from the first target data feature according to the correlation degree; S3: Determine the health score corresponding to the second target data feature based on the health score calculation function; S4: Determine a status assessment result of the solid state drive according to the health score corresponding to the second target data feature, and visualize the status assessment result.

[0154] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0155] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0156] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application.

Claims

1. A method for evaluating the state of a solid state drive, characterized in that: The method comprises: Acquire relevant parameter information of the solid state disk, and preprocess the relevant parameter information to obtain a first target data feature; Determine the correlation between the historical merged data feature and the corresponding health degree according to the joint probability distribution value between the historical merged data feature and the corresponding health degree, and select the second target data feature from the first target data feature according to the correlation degree; Determining a health score corresponding to the second target data feature based on a health score calculation function; According to the health score corresponding to the second target data feature, a status assessment result of the solid state drive is determined, and the status assessment result is visualized.

2. The method for evaluating the state of a solid state drive according to claim 1, wherein: Acquiring relevant parameter information of the solid state drive, and preprocessing the relevant parameter information to obtain the first target data feature includes: Collecting relevant parameter information of the solid-state hard disk, integrating the collected relevant parameter information of the solid-state hard disk, and generating a first target data set; Filling missing values ​​in the target data set to obtain a second target data set; Standardizing the data in the second target data set to obtain a third target data set; Data in the third target data set are merged to generate merged data features in multiple dimensions, and the merged data features in multiple dimensions are defined as the first target data features.

3. The method for evaluating the state of a solid state drive according to claim 1, wherein: Before obtaining the joint probability distribution value between the historical merged data features and the corresponding health degree, the method further includes: Based on the value range, the historical merged data characteristics and health are divided into intervals to obtain the interval division results; Based on the interval division results, the total number of samples is calculated and determined, and the calculation formula includes: in, represents the total number of samples, Indicates the number of merged data feature intervals, The number of intervals representing health, Indicates the number of interval samples.

4. The method for evaluating the state of a solid state drive according to claim 3, wherein: The method for obtaining the joint probability distribution value between the historical merged data features and the corresponding health degree includes: According to the interval combination result of the historical merged data characteristics and the health degree, the joint probability distribution value between the historical merged data characteristics and the corresponding health degree is calculated and determined. The calculation formula includes: in, The interval number representing the feature interval of the merged data, Indicates the value range of the merged data feature X in the interval [d, d+1], The interval number indicating the health level, Indicates The merged data characteristics of the dimension, Indicates health, represents the joint probability distribution value, Represents the characteristic value of health in the interval [k, k+1].

5. The method for evaluating the state of a solid state drive according to claim 1, wherein: The method for determining the correlation between the historical combined data feature and the corresponding health degree includes: The correlation between the historical merged data feature and the corresponding health degree is calculated based on a correlation calculation function, wherein the correlation calculation function includes: in, Indicates the degree of association, represents the joint probability distribution value, and Both represent marginal probability distribution values, Indicates the merged data features, Indicates health.

6. The method for evaluating the state of a solid state drive according to claim 1, characterized in that: According to the correlation degree, filtering out the second target data feature from the first target data feature includes: Sort the correlations in descending order; According to the sorting result, merged data features of a preset number of dimensions are selected from the first target data features as the second target data features.

7. The method for evaluating the state of a solid state drive according to claim 1, wherein: Determining the health score corresponding to the second target data feature based on the health score calculation function includes: A health score calculation function is constructed based on the weighted score item, the data feature historical change trend item and the environmental correction item. The health score calculation function includes: in, represents the health score, represents the weighted score, Indicates the historical change trend value of data features. Indicates the environmental correction value; Based on the health score calculation function, a health score corresponding to the second target data feature is calculated and determined.

8. The method for evaluating the state of a solid state drive according to claim 7, wherein: Before determining the health score corresponding to the second target data feature based on the health score calculation function, the method further includes: Based on a smoothing function, mapping conversion processing is performed on the second target data feature, and the smoothing function includes: in, represents the data characteristics after smoothing. Indicates The data features of the dimensions, express The standard deviation of Indicates the preset threshold value, represents the penalty coefficient, represents a constant coefficient; The weight of the data feature in the second target data feature is calculated based on the weight calculation function, wherein the weight calculation function includes: in, represents the normalized weight, express The sum of the weights corresponding to the dimensions, represents the initial weight, represents the constant coefficient, Represents the weight of the data feature of the i-th dimension.

9. The method for evaluating the state of a solid state drive according to claim 8, characterized in that: Constant coefficient Methods for determining include: Based on the historical database, obtaining a third target data feature corresponding to the second target data feature, and a target health corresponding to the third target data feature; Calculating the health score corresponding to the third target data feature based on the health score calculation function; In response to detecting that the health score corresponding to the third target data feature does not belong to the value range corresponding to the target health level, the constant coefficient Correction is performed until the health score corresponding to the third target data feature falls within the value range corresponding to the target health level.

10. The method for evaluating the state of a solid state drive according to claim 9, characterized in that: The method for determining the weighted score includes: The weighted score is calculated and determined based on the weight calculation result and the weighted score calculation function, wherein the weighted score calculation function includes: in, represents the weighted score, Indicates the number of dimensions.

11. The method for evaluating the state of a solid state drive according to claim 9, wherein: The method for calculating the historical change trend value of the data feature includes: in, Indicates the historical change trend value of data features. Indicated in The data feature value at the moment, represents the sliding window size, represents the number of dimensions, Indicates the current moment.

12. The method for evaluating the state of a solid state drive according to claim 9, wherein: The calculation method of the environmental correction value includes: in, represents the environmental correction value, Represents the characteristic value of environmental data, represents the correction factor, represents the total number of environmental items, represents the basic weight, represents the penalty coefficient, represents the nonlinear index, Indicates the threshold value corresponding to the environmental item.

13. The method for evaluating the state of a solid state drive according to claim 1, wherein: Determining a status evaluation result of the solid state drive according to the health score corresponding to the second target data feature, and visually displaying the status evaluation result includes: Determine, based on a value range of a preset tag, a tag corresponding to a health score corresponding to the second target data feature, wherein the tag is used to describe the health of the solid state drive; The health score corresponding to the second target data feature and the label corresponding to the health score are visualized.

14. The method for evaluating the state of a solid state drive according to claim 13, wherein: After determining the status assessment result of the solid state drive according to the health score corresponding to the second target data feature, the method further includes: Based on the second target data feature and its corresponding label, a one-to-one mapping relationship is generated and vectorized; The vectorized processing results are stored in the historical database.

15. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 14 is implemented.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 14 is implemented.

17. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 14 is implemented.

Images