Solid state disk service life prediction method and related equipment

Abstract

The invention provides a solid state disk (SSD) life prediction method and related equipment, and relates to the technical field of SSD life prediction. The method comprises the steps of obtaining performance data of a to-be-predicted solid state disk in a preset time period; identifying the performance data of the to-be-predicted solid state disk in the preset time period, and determining a target service category of the to-be-predicted solid state disk in the preset time period and the performance data of the to-be-predicted solid state disk in the preset time period corresponding to the target service category; processing the performance data in a preset time period corresponding to the target service category through a life prediction model to obtain a life prediction result of the to-be-predicted solid state disk in the target time period; wherein the life prediction model is obtained by training the performance data of the solid state disk in the historical time period and the corresponding historical life value, and the target time period is a time period after the preset time period, so that the accuracy and effectiveness of life prediction of the solid state disk are improved.

Description

Background Art

[0002] Solid-state disk (SSD) is a hard disk made of solid-state electronic storage chip array, which consists of a control unit and a storage unit such as FLASH chip or DRAM chip. Due to the advantages of fast reading and writing, light weight, low energy consumption and small size, SSD is widely used in military, automotive, industrial control, video surveillance, network monitoring, network terminals, electricity, medical, aviation, navigation equipment and many other fields.

[0003] During the use of SSDs, heat dissipation, vibration and other factors may cause SSD failures, resulting in loss of data stored in the SSD or server downtime, which has a great impact on users. It is necessary to predict the life of SSDs and keep track of the life of SSDs in a timely manner.

[0004] In the related art, a simple counting method is usually used to predict the life of SSDs. However, the above life prediction method has the problems of low prediction accuracy and poor reliability.

[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0006] The present disclosure provides a solid state drive life prediction method and related equipment, which at least to a certain extent overcomes the problems of low prediction accuracy and poor reliability in the solid state drive life prediction method in the related art.

[0007] Other features and advantages of the present disclosure will become apparent from the following detailed description, or may be learned in part by the practice of the present disclosure.

[0008] According to one aspect of the present disclosure, a method for predicting the life of a solid-state drive is provided, comprising: obtaining performance data of a solid-state drive to be predicted within a preset time period; identifying the performance data of the solid-state drive to be predicted within the preset time period, determining a target business category of the solid-state drive to be predicted within the preset time period, and performance data within the preset time period corresponding to the target business category; processing the performance data within the preset time period corresponding to the target business category through a life prediction model, and obtaining a life prediction result of the solid-state drive to be predicted within the target time period, wherein the life prediction model is trained based on the performance data of the solid-state drive within a historical time period and the corresponding historical life values, and the target time period is the time period after the preset time period.

[0009] In one embodiment of the present disclosure, the performance data of the solid-state drive to be predicted within a preset time period is identified to determine the target business category of the solid-state drive to be predicted within the preset time period, including: identifying the performance data of the solid-state drive to be predicted within the preset time period by a clustering algorithm or a classification algorithm to obtain the target business category of the solid-state drive to be predicted within the preset time period.

[0010] In one embodiment of the present disclosure, the life prediction model is obtained by training based on the performance data of the training solid-state hard disk in the historical time period and the corresponding historical life value, including: identifying the performance data of the training solid-state hard disk in the historical time period, determining the business category of the training solid-state hard disk in the historical time period, and the performance data in the historical time period corresponding to each of the business categories; using an autoregressive algorithm to construct a life prediction model to be trained based on the performance data in the historical time period corresponding to each of the business categories, with the root mean square error and goodness of fit as the objective function; using the least squares method to estimate the parameters of the life prediction model to be trained, and obtaining the life prediction model to be trained whose value of the objective function meets the preset training conditions as the life prediction model.

[0011] In one embodiment of the present disclosure, the life prediction model includes an ARIMA model and a Holt-winters model, and the business category includes at least one of a strong periodic type, a weak periodic type, a step type, a non-periodic type, and a stable type;

[0012] The autoregressive algorithm is used to construct a life prediction model to be trained based on the performance data in the historical time period corresponding to each of the business categories, with the root mean square error and goodness of fit as the objective function, including:

[0013] Perform differential processing on the performance data of the non-periodic type, the step type, and the weak-periodic type in the historical time period to convert them into stable performance data, and process the stable performance data through the ARIMA model to obtain the corresponding first life prediction result;

[0014] Inputting the performance data of the step type and the strong period type in the historical time period into the Holt-winters model to obtain the corresponding second life prediction result;

[0015] The method of using the least square method to estimate the parameters of the life prediction model to be trained to obtain the life prediction model to be trained when the value of the objective function meets the preset training conditions includes:

[0016] The least squares method is used to perform parameter estimation on the first life prediction result and the second life prediction result to obtain a life prediction model to be trained when the value of the objective function meets the preset training conditions.

[0017] In one embodiment of the present disclosure, the performance data of the solid state drive to be predicted within a preset time period includes: at least one of the number of write times, read and write speeds, error rates, and temperatures of the solid state drive to be predicted.

[0018] In one embodiment of the present disclosure, after the performance data within a preset time period corresponding to the target business category is processed by the life prediction model to obtain the life prediction result of the solid-state hard disk to be predicted within the target time period, the method further includes: outputting the life prediction result, wherein the life prediction result includes a life prediction value and a confidence interval.

[0019] According to another aspect of the present disclosure, a solid-state hard disk life prediction device is also provided, including: a data acquisition module, used to obtain performance data of the solid-state hard disk to be predicted within a preset time period; a category determination module, used to identify the performance data of the solid-state hard disk to be predicted within the preset time period, determine the target business category of the solid-state hard disk to be predicted within the preset time period, and the performance data within the preset time period corresponding to the target business category; a life prediction model, used to process the performance data within the preset time period corresponding to the target business category through the life prediction model, and obtain the life prediction result of the solid-state hard disk to be predicted within the target time period, wherein the life prediction model is trained based on the performance data of the training solid-state hard disk within the historical time period and the corresponding historical life values, and the target time period is the time period after the preset time period.

[0020] According to another aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the above-mentioned solid state drive life prediction method by executing the executable instructions.

[0021] According to another aspect of the present disclosure, 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 above-mentioned solid-state hard disk life prediction method is implemented.

[0022] According to another aspect of the present disclosure, a computer program product is provided, which includes a computer program or computer instructions, and the computer program or the computer instructions are loaded and executed by a processor to enable a computer to implement the above-mentioned solid-state hard disk life prediction method.

[0023] In the disclosed embodiment, the performance data of the solid state drive to be predicted within a preset time period is obtained; the performance data of the solid state drive to be predicted within the preset time period is identified, and the target business category of the solid state drive to be predicted within the preset time period and the performance data within the preset time period corresponding to the target business category are determined; the performance data within the preset time period corresponding to the target business category is processed by the life prediction model to obtain the life prediction result of the solid state drive to be predicted within the target time period; the life prediction model is trained based on the performance data of the training solid state drive within the historical time period and the corresponding historical life value, and the target time period is the time period after the preset time period. The disclosed embodiment constructs a life prediction model through an autoregressive algorithm, which can predict the life decay rate of the solid state drive under different business modes, thereby more accurately predicting the future life, more comprehensively considering various factors affecting the life of the SSD, identifying the degree of influence of different business modes on the disk life, providing more refined input parameters for the prediction model, realizing multi-time scale prediction, and improving the accuracy and reliability of SSD life prediction.

[0024] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the accompanying drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0026] Figure 1 A flow chart of a solid state drive life prediction method provided by an embodiment of the present disclosure is shown;

[0027] Figure 2 Another flow chart of a method for predicting the life of a solid state drive provided by an embodiment of the present disclosure is shown;

[0028] Figure 3 A flow chart of a life prediction model training method provided by an embodiment of the present disclosure is shown;

[0029] Figure 4 A flow chart of another solid state drive life prediction method provided by an embodiment of the present disclosure is shown;

[0030] Figure 5 A flowchart showing an example of a method for predicting the life of a solid state drive provided by an embodiment of the present disclosure;

[0031] Figure 6A schematic diagram showing the structure of a solid state drive life prediction device provided by an embodiment of the present disclosure;

[0032] Figure 7 A structural block diagram of an electronic device provided by an embodiment of the present disclosure is shown;

[0033] Figure 8 A schematic diagram showing the structure of a computer program product provided by an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0034] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that the disclosure will be more comprehensive and complete and to fully convey the concepts of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0035] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0036] For ease of understanding, before introducing the embodiments of the present disclosure, several terms involved in the embodiments of the present disclosure are first explained as follows:

[0037] SSD: Solid State Disk or Solid State Drive, solid state drive;

[0038] Smart technology standard: Self-Monitoring Analysis and Reporting Technology, an automatic hard disk status detection and early warning system and specification;

[0039] ECC correction: a memory error correction principle;

[0040] ULTRA DMA: Ultra Direct Memory Access, advanced direct memory access.

[0041] In order to enable those skilled in the art to better understand the technical solution of the present disclosure, the present disclosure is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present disclosure.

[0042] At present, many servers use SSD hard drives. During the operation of the server, SSD hard drives may malfunction due to problems such as heat dissipation and vibration, resulting in SSD hard drive data loss or server downtime, which brings great inconvenience to users. Therefore, it is necessary to predict the life of SSD hard drives in order to understand the usage of SSD hard drives and replace them in time.

[0043] In related technologies, a simple counting method is used to predict the life of SSD hard disks, such as cumulative write count prediction, write amplification effect prediction, degradation model, health monitoring, etc.

[0044] The cumulative write count prediction predicts the life of the SSD hard disk by tracking the total number of writes to the hard disk. When the total number of writes reaches a threshold, it is predicted that the life of the SSD hard disk will end.

[0045] Write amplification effect prediction is to calculate the difference between the actual amount of written data and the amount of written data requested by the user, and to predict the life of the SSD hard drive by analyzing the write amplification effect.

[0046] The degradation model is to predict the life of the SSD hard disk by establishing a degradation model of the SSD hard disk. The degradation model is usually based on laboratory test data and predicts the life of the SSD hard disk by simulating the degradation of the SSD hard disk under different loads.

[0047] Health monitoring is to predict the life of SSD by detecting the health of the hard disk. Usually, it involves regularly checking the performance parameters of the SSD hard disk, such as read and write speed, error rate, etc., and evaluating the health status of the SSD hard disk through the changes of the above performance parameters.

[0048] However, the SSD hard disk life prediction method in the related art has the following disadvantages:

[0049] (1) For the prediction method of cumulative write count: On the one hand, the relationship between write count and life span is not linear. The life span of an SSD hard disk depends not only on the write count, but also on many factors, such as temperature, programming times, erasure times, controller quality, etc. Therefore, predicting the life span of an SSD hard disk only by the cumulative write count will result in inaccurate prediction results. On the other hand, determining an appropriate write count threshold to predict the life span of an SSD hard disk is subjective. Different write count thresholds lead to different life span prediction results. There is no unified standard to determine the write count threshold, making the threshold selection difficult.

[0050] (2) Regarding the write amplification effect: On the one hand, it is difficult to measure accurately. The measurement of the write amplification effect usually relies on the algorithms inside the SSD hard disk. These algorithms may not be transparent enough, making it difficult for external parties to accurately measure and analyze them. On the other hand, changes are difficult to predict. The write amplification effect will change with time and load. Predicting such changes requires complex data analysis and models. The methods in related technologies may not be able to accurately capture these changes.

[0051] (3) For the degradation model, laboratory test data has limitations and may not be representative under actual usage conditions. It is difficult to adapt to different usage scenarios. The degradation model may not be able to adapt to the degradation speed of SSD hard drives in different usage scenarios, resulting in inaccurate prediction results.

[0052] (4) For health monitoring, the performance parameters of health monitoring do not change significantly and rely on regular inspections. In the early life stage of SSD hard drives, the changes in performance parameters may not be obvious, making early life prediction difficult. Health monitoring requires regular inspections, increasing maintenance costs and potential business interruption risks.

[0053] In order to at least partially solve the above technical problems, the present disclosure provides a solid-state hard drive life prediction method, which constructs a life prediction model through an autoregressive algorithm to predict the service life of the solid-state hard drive in a future time period, uses historical data to learn the life decay rate of the SSD hard drive under different business models, and provides more refined input parameters for the life prediction model by analyzing the impact of different business categories on the disk life, thereby more accurately predicting the future life of the SSD hard drive.

[0054] In detail, the present invention obtains the performance data of the solid-state hard disk to be predicted within a preset time period; identifies the performance data of the solid-state hard disk to be predicted within the preset time period, determines the target business category of the solid-state hard disk to be predicted within the preset time period, and the performance data within the preset time period corresponding to the target business category; processes the performance data within the preset time period corresponding to the target business category through the life prediction model, and obtains the life prediction result of the solid-state hard disk to be predicted within the target time period; the life prediction model is trained based on the performance data of the training solid-state hard disk within the historical time period and the corresponding historical life value, and the target time period is the time period after the preset time period. The present invention constructs a life prediction model through an autoregressive algorithm, which can predict the life decay rate of the solid-state hard disk under different business modes, so as to more accurately predict the future life, more comprehensively consider the various factors affecting the life of the SSD, identify the degree of influence of different business modes on the disk life, provide more refined input parameters for the prediction model, realize multi-time scale prediction, and improve the accuracy and reliability of SSD life prediction.

[0055] It should be pointed out that the embodiments of the present invention and the technical features therein may be combined with each other without conflict.

[0056] The present exemplary implementation is described in detail below with reference to the accompanying drawings and embodiments.

[0057] First, a method for predicting the life of a solid state drive is provided in an embodiment of the present disclosure, and the method can be performed by any electronic device with computing and processing capabilities. The method for predicting the life of a solid state drive provided in an embodiment of the present disclosure can be performed by a server, for example, by a solid state drive life prediction device configured on the server.

[0058] Figure 1 A flow chart of a solid state drive life prediction method provided in an embodiment of the present disclosure is shown as follows: Figure 1 As shown, the solid state drive life prediction method provided in the embodiment of the present disclosure includes the following steps:

[0059] S102: Obtain performance data of the solid state drive to be predicted within a preset time period.

[0060] The above-mentioned preset time period may be a time period before the current moment, and the length of the preset time period may be determined according to actual needs. For example, the preset time period may be the performance data of the previous week or the previous two weeks.

[0061] The solid state drive to be predicted may be an enterprise-level SSD, such as a data center, a server, etc.

[0062] In one embodiment, the performance data of the solid state drive to be predicted within a preset time period includes: at least one of the number of write times, read / write speed, error rate, and temperature of the solid state drive to be predicted.

[0063] The number of write times of the solid-state drive to be predicted refers to the erase and write limit of the flash memory chip.

[0064] The number of writes to an SSD is affected by the type of internal flash memory chip. Different types of flash memory chips have different number of writes. Flash memory chips include SLC, MLC, TLC, etc. As the number of writes to an SSD increases, the performance of the flash memory chip gradually decreases, leading to the end of the life of the SSD.

[0065] The number of writes to an SSD is also affected by the working principle of the underlying storage unit and the Write Amplification Factor (WAF).

[0066] The read and write speed of the solid-state drive to be predicted refers to the speed at which a storage device (such as a memory) reads and writes data. It is measured in MB / s and is an important indicator for measuring hard drive performance. The faster the read and write speed, the faster the data transmission and processing speed, which reduces the service life of the SSD.

[0067] The error rate of the solid state drive to be predicted directly affects the service life of the SSD by affecting the stability of the internal flash memory chip and data storage.

[0068] The temperature of the solid state drive to be predicted affects the performance and service life of the SSD. The higher the temperature, the lower the read and write speed of the SSD, the lower the data reliability, the faster the aging process of the internal components, and the higher the failure rate.

[0069] In one embodiment, the method further includes preprocessing the collected performance data of the solid-state drive to be predicted within a preset time period, such as removing outliers, filling missing values, normalizing, etc., to facilitate subsequent model reasoning.

[0070] S104, identifying the performance data of the solid state drive to be predicted within a preset time period, determining the target service category of the solid state drive to be predicted within the preset time period, and the performance data within the preset time period corresponding to the target service category.

[0071] In one embodiment, the attenuation fluctuation of the SSD hard disk can reflect the business mode, and the curve shape of the service life of the SSD hard disk corresponding to different business modes is different.

[0072] In one embodiment, the target service type may be one or more of a strong periodic type, a weak periodic type, a step type, a non-periodic type, and a stable type.

[0073] Among them, the strong cycle refers to the time period or period when the SSD hard disk experiences high-intensity use and high-frequency write / erase (P / E) cycles. When the performance data of the SSD hard disk belongs to the strong cycle type, it will accelerate the wear of the SSD hard disk and reduce the service life of the SSD hard disk.

[0074] Weak cycle refers to the period of time or period when the SSD is used less intensively or written / erased less frequently. When the performance data of the SSD belongs to the weak cycle type, the wear of the SSD is small, and accordingly, the service life of the SSD is longer during the weak cycle.

[0075] Acyclic refers to the fact that the performance data of the SSD hard disk has no obvious cyclic characteristics within a certain period of time.

[0076] The step type refers to the performance data of the SSD hard disk showing a step-like upward or step-like downward trend within a preset time period.

[0077] Stable type means that the performance data of the SSD hard disk fluctuates within a certain range of average usage intensity within a preset time period, and the regularity is not obvious.

[0078] When the solid-state drive to be predicted includes multiple target business categories within a preset time period, the preset time period can be divided into multiple sub-time periods according to the target business categories, and the performance data within each sub-time period can be determined to obtain the performance data within the preset time period corresponding to the multiple target business categories, thereby providing a data basis for the life prediction of the SSD hard drive.

[0079] S106. Process the performance data within a preset time period corresponding to the target business category through a life prediction model to obtain a life prediction result of the solid-state hard disk to be predicted within the target time period; wherein the life prediction model is trained based on the performance data of the solid-state hard disk within a historical time period and the corresponding historical life values, and the target time period is the time period after the preset time period.

[0080] In one embodiment, the life prediction model is a trained model. The performance data within a preset time period corresponding to the target service category is input into the life prediction model, and the life prediction result within the target time period is output, that is, the life prediction model can characterize the corresponding relationship between the performance data within the preset time period and the life prediction result within the target time period.

[0081] It should be noted that the target time period can be a time period after the preset time period. The length of the target time period is determined according to the forecast requirements, for example, the life of the SSD hard disk is predicted after 30 days, 60 days and half a year.

[0082] In one embodiment, the life prediction model may include multiple sub-models, each sub-model processes performance data of a target business category, and the output results of each sub-model are weighted and summed to obtain the life prediction result of the life prediction model.

[0083] In the disclosed embodiment, the performance data of the solid state drive to be predicted within a preset time period is obtained; the performance data of the solid state drive to be predicted within the preset time period is identified, and the target business category of the solid state drive to be predicted within the preset time period and the performance data within the preset time period corresponding to the target business category are determined; the performance data within the preset time period corresponding to the target business category is processed by the life prediction model to obtain the life prediction result of the solid state drive to be predicted within the target time period; the life prediction model is trained based on the performance data of the training solid state drive within the historical time period and the corresponding historical life value, and the target time period is the time period after the preset time period. The disclosed embodiment constructs a life prediction model through an autoregressive algorithm, which can predict the life decay rate of the solid state drive under different business modes, thereby more accurately predicting the future life, more comprehensively considering various factors affecting the life of the SSD, identifying the degree of influence of different business modes on the disk life, providing more refined input parameters for the prediction model, realizing multi-time scale prediction, and improving the accuracy and reliability of SSD life prediction.

[0084] Figure 2 Another flow chart of a method for predicting the life of a solid state drive provided by an embodiment of the present disclosure is shown. Figure 1 Based on the embodiment, S104 is further defined as S1042, which is defined in a manner of identifying the service category of the performance data. Figure 2 As shown, in one embodiment, the above S1042 identifies the performance data of the solid state drive to be predicted within a preset time period, and determines the target service category of the solid state drive to be predicted within the preset time period, including:

[0085] S1042: Identify the performance data of the solid state drive to be predicted within a preset time period by using a clustering algorithm or a classification algorithm to obtain a target service category of the solid state drive to be predicted within the preset time period.

[0086] Clustering algorithm is an unsupervised learning method that can divide a data set into multiple groups or clusters with similar characteristics. The performance data of the SSD to be predicted within a preset time period are grouped according to their characteristics, so that the performance data within the same cluster is as similar as possible, and the performance data of different clusters are as different as possible. The clustering algorithm can be one or more of K-Means, K-Medoids, hierarchical clustering, DBSCAN, etc.

[0087] The classification algorithm can classify and label the performance data of the SSD to be predicted, establish a corresponding classification model based on the existing data samples, and automatically classify the new unlabeled performance data into predefined categories. The classification algorithm can use one or more of the following methods: decision tree, random forest, support vector machine, naive Bayes, K nearest neighbor, logistic regression, neural network, gradient boosting decision tree, etc.

[0088] In the embodiments of the present disclosure, the present disclosure analyzes the performance data of the solid-state drive to be predicted within a preset time period and identifies the business category through a clustering algorithm and an inter-classification algorithm, thereby improving data processing efficiency and improving the accuracy and efficiency of SSD hard drive life prediction.

[0089] Figure 3 A flow chart of a life prediction model training method provided by an embodiment of the present disclosure is shown. Figure 3 As shown, the above life prediction model is obtained based on the performance data of the training solid state drive in the historical time period and the corresponding historical life value, including:

[0090] S302, identifying the performance data of the training solid state drive in the historical time period, determining the service category of the training solid state drive in the historical time period, and the performance data in the historical time period corresponding to each service category;

[0091] S304, using an autoregressive algorithm, based on the performance data in the historical time period corresponding to each business category, constructing a life prediction model to be trained, with root mean square error and goodness of fit as the objective function, the life prediction model includes an ARIMA model and a Holt-winters model;

[0092] S306, using the least square method to estimate the parameters of the life prediction model to be trained, and obtaining the life prediction model to be trained when the value of the objective function meets the preset training conditions as the life prediction model.

[0093] In one embodiment, when training the life prediction model, a data set can be constructed using performance data of the training solid-state drive in a historical time period. The data set includes a training set and a validation set. The ratio of performance data in the training set and the validation set can be 8:2, 7:3, etc.

[0094] The performance data of the training solid-state drive in a historical time period may include historical write times, historical read and write speeds, historical error rates, historical temperatures, etc. The performance data collected in the historical time period is preprocessed, for example, outliers are removed, missing values ​​are filled, and normalization is performed to ensure data quality.

[0095] A clustering algorithm or a classification algorithm may be used to identify the performance data of the training solid-state hard disk within a historical time period, and to divide the performance data within the historical time period into several business types to facilitate the subsequent training and reasoning of the life prediction model.

[0096] It should be noted that the business category of the training solid-state drive in the historical time period also includes at least one of a strong periodic type, a weak periodic type, a step type, a non-periodic type, and a stable type.

[0097] In one embodiment, an autoregressive algorithm is used to construct a life prediction model to be trained, with inputs such as the number of historical write times and business categories, and outputs a life prediction value for a period of time in the future.

[0098] The life prediction model includes the ARIMA model, which consists of three parts: autoregression, difference and moving average. It is suitable for time series data that is stationary or can be converted to a stationary type through difference.

[0099] In one embodiment, the performance data in the historical time period of the non-periodic, step-type, and weak-periodic types can be differentially processed and converted into stable time series data, and then the stable performance data can be processed through the ARIMA model to obtain the corresponding first life prediction result.

[0100] The life prediction model also includes the Holt-winters model, which also includes three parts: level, trend and seasonality. It is suitable for time series data with obvious trend and seasonal patterns. The performance data in the historical time period of the ladder type and strong cycle type is input into the Holt-winters model to obtain the corresponding second life prediction results.

[0101] In one embodiment, the target loss may be a root mean square error and a goodness of fit, wherein the root mean square error, also known as root mean square deviation, is a measure of the difference between a life prediction result and a historical life value.

[0102] Life expectancy prediction results The root mean square deviation RMSD corresponding to the estimated historical life value y is expressed as follows:

[0103]

[0104] Where MSE is the mean square error and m is the number of life prediction results.

[0105] The goodness of fit indicates how well the regression line obtained by the life prediction model fits the historical life value. The coefficient of determination R 2 Characterize and determine the coefficient R 2 The maximum value is 1. The coefficient of determination R 2is also close to 1, indicating that the better the regression line fits the historical life value, the lower the determination coefficient R 2 The smaller it is, the worse the regression line fits the historical life value.

[0106] The least squares method is used to estimate the parameters of the first life prediction results and the second life prediction results, and the life prediction model to be trained is obtained when the value of the objective function meets the preset training conditions. The least squares method estimates the parameters of the life prediction model to be trained by minimizing the sum of squares of errors to find the best function match for the data.

[0107] In another embodiment, the life prediction result of the life prediction model to be trained is obtained by weighted averaging the first life prediction result and the second life prediction result, and then the value of the objective function of the model is obtained according to the life prediction result of the life prediction model to be trained and the historical life value. When the value of the objective function does not exceed the preset loss threshold, it indicates that the value of the objective function meets the preset training conditions, and the trained life prediction model is obtained. When the value of the objective function exceeds the preset loss threshold, it indicates that the value of the objective function does not meet the preset training conditions, and the model parameters of the life prediction model are continuously adjusted and retrained.

[0108] In a feasible implementation, the performance data within one year of history can be used to train the life prediction model to be trained, and the optimal model can be selected through cross-validation, residual analysis, autocorrelation test, etc., and the model parameters, such as lag order, seasonal order, etc., can be adjusted to improve the accuracy of life prediction.

[0109] Figure 4 A flow chart of another method for predicting the life of a solid state drive provided by an embodiment of the present disclosure is shown. Figure 1 Based on the embodiment, S108 is added after S106 to limit the usage scenarios of the life prediction results. Figure 4 As shown, after obtaining the life prediction result of the solid state drive to be predicted within the target time period in S106, the solid state drive life prediction method of the embodiment of the present disclosure further includes:

[0110] S108. Outputting the life prediction result, which includes the life prediction value and the confidence interval.

[0111] The life prediction value refers to the remaining life value of the solid-state drive to be predicted within the target time period, or the remaining life percentage.

[0112] The confidence interval is an estimated interval of a population parameter constructed with sample statistics. It can show the extent to which the true value of the parameter has a certain probability of falling around the measurement result, and give the degree of credibility of the measured value of the life prediction value.

[0113] The confidence interval is an interval composed of the upper and lower confidence limits of the life of the SSD hard disk. For example, for a given set of sample data, the mean is μ and the standard deviation is σ, then the 100×(1-α)% confidence interval of the mean of the overall data is (μ-Z α / 2 σ,μ+Z α / 2 σ), where α is the coverage area of ​​the non-confidence level in the normal distribution, and Zα / 2 is the corresponding standard score. The confidence level is 1-α, and if α=0.05, then the confidence level is 95%.

[0114] The display method of the life prediction results can be determined according to actual needs, and can be displayed in the form of life curves, charts, etc.

[0115] In one embodiment, based on the life prediction results, the business department can make preparations such as data migration and system backup in advance. The IT department can repair and replace disk resources in advance based on the life prediction results to reduce business interruption time and improve system reliability.

[0116] In order to deepen the understanding of the solid state drive life prediction method provided by the present disclosure, the following Figure 5 Provide explanation.

[0117] The method for predicting the life of a solid state drive provided by the embodiment of the present disclosure includes the following steps:

[0118] S501, data input, input the performance data and historical life value of the training SSD hard disk in the historical time period; collect the performance data of the SSD hard disk, including the number of writes, read and write speeds, error rates, temperatures, etc., and pre-process the collected data, such as removing outliers, filling missing values, normalizing, etc., to facilitate subsequent model training.

[0119] S502, business category identification, identifying large and small business models through clustering algorithms and classification algorithms, and determining the business category of the training solid state drive;

[0120] S502. Use an autoregressive algorithm to build a life prediction model. Use an autoregressive algorithm (such as a seasonal ARIMA model) to build a life prediction model. The model input is historical write times, business model and other features, and the output is the life prediction value in the future.

[0121] S503, conduct model training and optimization; use one year of historical data to train the prediction model, select the optimal model through cross-validation and other methods, and adjust model parameters, such as lag order, seasonal order, etc., to improve prediction accuracy.

[0122] S504, multi-time scale prediction: Based on the trained model, the SSD lifespan is predicted after 30 days, 60 days and half a year.

[0123] S505. Output the life prediction results, which include the life prediction value and the corresponding confidence interval. This is to facilitate the business department and IT department to make corresponding migration and maintenance decisions. Decision support and resource management: Based on the prediction results, the business department can make preparations such as data migration and system backup in advance. The IT department can repair and replace disk resources in advance based on the prediction results to reduce business interruption time and improve system reliability.

[0124] In this disclosure, a multi-factor comprehensive SSD life prediction model is constructed, using an autoregressive algorithm, combined with multiple factors such as the number of writes, read and write speeds, error rates, and temperature to build a more comprehensive life prediction model. Differentiated business model analysis is performed to provide refined input parameters for the prediction model by identifying different business models and usage scenarios, and distinguishing the impact of large and small businesses on the disk life.

[0125] The present disclosure can provide multi-time scale prediction capabilities, provide life predictions for 30 days, 60 days and half a year, provide business departments and IT departments with a time window for advance preparation and decision-making, use historical data for model training and optimization, select the optimal model parameters through cross-validation and other methods, and improve prediction accuracy. According to the prediction results, the business department can migrate data and back up the system in advance, and the IT department can repair and replace disk resources in advance, reduce business interruption time, and improve system reliability. In addition, through the confidence interval of the life prediction results: output the prediction results and their confidence intervals to help relevant departments make more reliable migration and maintenance decisions.

[0126] Based on the same inventive concept, the embodiments of the present disclosure also provide a solid state drive life prediction device, such as the following embodiments. Since the principle of solving the problem in the device embodiment is similar to that in the above method embodiment, the implementation of the device embodiment can refer to the implementation of the above method embodiment, and the repeated parts will not be repeated.

[0127] Figure 6 FIG. 1 is a schematic diagram of a solid state drive life prediction device provided by an embodiment of the present disclosure. Figure 6 As shown, the solid state drive life prediction device of this embodiment includes a data acquisition module 610 , a category determination module 620 and a life prediction module 630 .

[0128] The data acquisition module 610 is used to acquire the performance data of the solid state drive to be predicted within a preset time period;

[0129] The category determination module 620 is used to identify the performance data of the solid state drive to be predicted within a preset time period, determine the target service category of the solid state drive to be predicted within the preset time period, and the performance data within the preset time period corresponding to the target service category;

[0130] The life prediction module 630 is used to process the performance data within a preset time period corresponding to the target business category through a life prediction model to obtain a life prediction result of the solid-state hard disk to be predicted within the target time period, wherein the life prediction model is trained based on the performance data of the training solid-state hard disk within a historical time period and the corresponding historical life value, and the target time period is the time period after the preset time period.

[0131] It should be noted that the above data acquisition module 610, category determination module 620 and life prediction module 630 correspond to S102 to S106 in the method embodiment, and the examples and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the contents disclosed in the above method embodiment. It should be noted that the above modules, as part of the device, can be executed in a computer system such as a set of computer executable instructions.

[0132] In one embodiment, the category determination module 620 is used to identify the performance data of the solid state drive to be predicted within a preset time period by a clustering algorithm or a classification algorithm to obtain the target service category of the solid state drive to be predicted within the preset time period.

[0133] It should be noted that the business category includes at least one of a strong cyclical type, a weak cyclical type, a step type, a non-cyclical type, and a stable type.

[0134] In one embodiment, the device also includes a model training module not shown in the accompanying drawings, and the model training module is used to identify the performance data of the training solid-state hard disk in the historical time period, determine the business category of the training solid-state hard disk in the historical time period, and the performance data in the historical time period corresponding to each business category; using an autoregressive algorithm, according to the performance data in the historical time period corresponding to each business category, construct a life prediction model to be trained, with the root mean square error and goodness of fit as the objective function, the life prediction model includes an ARIMA model and a Holt-winters model; using the least squares method to estimate the parameters of the life prediction model to be trained, and obtaining the life prediction model to be trained when the value of the objective function meets the preset training conditions as the life prediction model.

[0135] It should be noted that the performance data of the solid state drive to be predicted within a preset time period includes at least one of the number of write times, read and write speed, error rate, and temperature of the solid state drive to be predicted.

[0136] In one embodiment, the device also includes a result output module not shown in the accompanying drawings, and the result output module is used to output a life prediction result after obtaining the life prediction result of the solid-state hard disk to be predicted within the target time period, and the life prediction result includes a life prediction value and a confidence interval.

[0137] In the disclosed embodiment, the performance data of the solid state drive to be predicted within a preset time period is obtained; the performance data of the solid state drive to be predicted within the preset time period is identified, and the target business category of the solid state drive to be predicted within the preset time period and the performance data within the preset time period corresponding to the target business category are determined; the performance data within the preset time period corresponding to the target business category is processed by the life prediction model to obtain the life prediction result of the solid state drive to be predicted within the target time period; the life prediction model is trained based on the performance data of the training solid state drive within the historical time period and the corresponding historical life value, and the target time period is the time period after the preset time period. The disclosed embodiment constructs a life prediction model through an autoregressive algorithm, which can predict the life decay rate of the solid state drive under different business modes, thereby more accurately predicting the future life, more comprehensively considering various factors affecting the life of the SSD, identifying the degree of influence of different business modes on the disk life, providing more refined input parameters for the prediction model, realizing multi-time scale prediction, and improving the accuracy and reliability of SSD life prediction.

[0138] It will be appreciated by those skilled in the art that various aspects of the present invention may be implemented as a system, method or program product. Therefore, various aspects of the present invention may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to herein as a "circuit", "module" or "system".

[0139] Refer to the following Figure 7 An electronic device 700 according to this embodiment of the present invention will be described. Figure 7 The electronic device 700 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0140] like Figure 7 As shown, the electronic device 700 is in the form of a general computing device. The components of the electronic device 700 may include but are not limited to: at least one processing unit 710, at least one storage unit 720, and a bus 730 connecting different system components (including the storage unit 720 and the processing unit 710).

[0141] The storage unit stores program codes, which can be executed by the processing unit 710, so that the processing unit 710 performs the steps of various exemplary embodiments of the present invention described in the above “Exemplary Method” section of this specification. For example, the processing unit 710 can perform the following steps: Figure 1 As shown in the embodiment of the present disclosure, the performance data of the solid-state hard disk to be predicted within a preset time period is obtained; the performance data of the solid-state hard disk to be predicted within the preset time period is identified, and the target business category of the solid-state hard disk to be predicted within the preset time period, and the performance data within the preset time period corresponding to the target business category are determined; the performance data within the preset time period corresponding to the target business category is processed by a life prediction model to obtain a life prediction result of the solid-state hard disk to be predicted within the target time period; wherein the life prediction model is trained based on the performance data of the training solid-state hard disk within the historical time period and the corresponding historical life value, and the target time period is the time period after the preset time period.

[0142] The storage unit 720 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 7201 and / or a cache storage unit 7202 , and may further include a read-only storage unit (ROM) 7203 .

[0143] The storage unit 720 may also include a program / utility 7204 having a set (at least one) of program modules 7205, such program modules 7205 including but not limited to: an operating system, one or more application programs, other program modules and program data, each of which or some combination may include the implementation of a network environment.

[0144] Bus 730 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0145] The electronic device 700 may also communicate with one or more external devices 740 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), one or more devices that enable a user to interact with the system, and / or any device that enables the electronic device 700 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 760. Furthermore, the system may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 760. Figure 7 As shown, the network adapter 760 communicates with other modules of the electronic device 700 via the bus 730. It should be understood that although Figure 7 Not shown, other hardware and / or software modules may be used in conjunction with electronic device 700, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0146] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0147] In an exemplary embodiment of the present disclosure, a computer-readable storage medium is further provided, which may be a readable signal medium or a readable storage medium. Figure 8 A schematic diagram of a computer-readable storage medium in an embodiment of the present disclosure is shown. Figure 8 As shown, a program product capable of implementing the above method of the present disclosure is stored on the computer-readable storage medium 800. In some possible implementations, various aspects of the present disclosure may also be implemented in the form of a program product, which includes a program code. When the program product is run on a terminal device, the program code is used to execute the steps according to various exemplary implementations of the present disclosure described in the above “Exemplary Method” section of this specification.

[0148] More specific examples of computer-readable storage media in the present disclosure may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0149] In the present disclosure, a computer readable storage medium may include a data signal propagated in baseband or as part of a carrier wave, wherein a readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A readable signal medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0150] Alternatively, the program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0151] In a specific implementation, the program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0152] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.

[0153] In addition, although the steps of the method in the present disclosure are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0154] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a mobile terminal, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0155] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. The present disclosure is intended to cover any variations, uses or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the appended claims.

Claims

1. A method for predicting the life of a solid state drive, characterized in that: include: Obtaining performance data of the solid state drive to be predicted within a preset time period; Identify the performance data of the solid-state drive to be predicted within a preset time period, determine the target service category of the solid-state drive to be predicted within the preset time period, and the performance data within the preset time period corresponding to the target service category; Processing the performance data within a preset time period corresponding to the target service category through a life prediction model to obtain a life prediction result of the solid state drive to be predicted within the target time period; The life prediction model is obtained based on the performance data of the solid state drive in a historical time period and the corresponding historical life value, and the target time period is the time period after the preset time period.

2. The solid state drive life prediction method according to claim 1, characterized in that: The identifying the performance data of the solid state drive to be predicted within a preset time period and determining the target service category of the solid state drive to be predicted within the preset time period includes: The performance data of the solid-state hard disk to be predicted within a preset time period is identified by a clustering algorithm or a classification algorithm to obtain a target service category of the solid-state hard disk to be predicted within the preset time period.

3. The solid state drive life prediction method according to claim 1, characterized in that: The life prediction model is obtained by training the performance data of the solid state drive in a historical time period and the corresponding historical life value, and includes: Identify the performance data of the training solid-state drive in a historical time period, determine the service category of the training solid-state drive in the historical time period, and the performance data in the historical time period corresponding to each service category; Using an autoregressive algorithm, a life prediction model to be trained is constructed based on the performance data in the historical time period corresponding to each of the business categories, with the root mean square error and goodness of fit as the objective function; The least square method is used to estimate the parameters of the life prediction model to be trained, and the life prediction model to be trained is obtained when the value of the objective function meets the preset training conditions, which is used as the life prediction model.

4. The solid state drive life prediction method according to claim 3, characterized in that: The life prediction model includes an ARIMA model and a Holt-winters model, and the business category includes at least one of a strong cyclical type, a weak cyclical type, a step type, a non-cyclical type, and a stable type; The autoregressive algorithm is used to construct a life prediction model to be trained based on the performance data in the historical time period corresponding to each of the business categories, with the root mean square error and goodness of fit as the objective function, including: The performance data of the non-periodic type, the step type, and the weak-periodic type in the historical time period are differentially processed to convert them into stable performance data, and the stable performance data are processed by the ARIMA model to obtain the corresponding first life prediction result; Inputting the performance data of the step type and the strong period type in the historical time period into the Holt-winters model to obtain the corresponding second life prediction result; The method of using the least square method to estimate the parameters of the life prediction model to be trained to obtain the life prediction model to be trained when the value of the objective function meets the preset training conditions includes: The least squares method is used to perform parameter estimation on the first life prediction result and the second life prediction result to obtain a life prediction model to be trained when the value of the objective function meets the preset training conditions.

5. The solid state drive life prediction method according to claim 1, characterized in that: The performance data of the solid state drive to be predicted within a preset time period includes at least one of the number of write times, read / write speed, error rate, and temperature of the solid state drive to be predicted.

6. The method for predicting the life of a solid state drive according to any one of claims 1 to 5, characterized in that: After obtaining the life prediction result of the solid state hard disk to be predicted within the target time period, the method further includes: The life prediction result is output, wherein the life prediction result includes a life prediction value and a confidence interval.

7. A solid state hard disk life prediction device, characterized in that: include: A data acquisition module, used to acquire performance data of the solid state drive to be predicted within a preset time period; A category determination module, used to identify the performance data of the solid-state drive to be predicted within a preset time period, determine the target service category of the solid-state drive to be predicted within the preset time period, and the performance data within the preset time period corresponding to the target service category; A life prediction model is used to process the performance data within a preset time period corresponding to the target business category through the life prediction model to obtain a life prediction result of the solid-state hard disk to be predicted within the target time period, wherein the life prediction model is trained based on the performance data of the training solid-state hard disk within a historical time period and the corresponding historical life value, and the target time period is the time period after the preset time period.

8. An electronic device, characterized in that: include: processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the solid state drive life prediction method as described in any one of claims 1-6 by executing the executable instructions.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the solid state drive life prediction method as described in any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that The computer program product includes a computer program or computer instructions, and the computer program or computer instructions are loaded and executed by a processor to enable a computer to implement the solid state drive life prediction method as described in any one of claims 1-6.

Images