Solid state disk fault prediction method and system based on artificial intelligence

Abstract

The invention provides a solid state disk fault prediction method and system based on artificial intelligence. The solid state disk fault prediction method comprises the following steps: establishing an association relationship for a target solid state disk; performing running log acquisition on the target solid state disk based on the association relationship to obtain solid state disk log information; analyzing the log information of the solid state disk, and extracting target associated information; respectively performing hardware fault prediction analysis and software fault prediction analysis according to the target associated information by using an artificial intelligence model; and performing fault risk assessment according to the fault prediction analysis data to obtain a fault prediction assessment result. Comprehensive fault prediction of the solid state disk is efficiently realized by means of artificial intelligence, the time delay of fault prediction is reduced, and the situation that when the solid state disk fails, data loss is caused due to the fact that countermeasures cannot be taken in time, normal use of the solid state disk is affected, and customer loss and economic loss are caused due to business failure is avoided. And meanwhile, the fault prediction accuracy is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of solid state disks, in particular to a solid state disk fault prediction method and system based on artificial intelligence. BACKGROUND

[0002] With the rapid development of cloud storage technology and data centers, storage devices have become an important cornerstone of the information society. Solid state disks (SSD) have gradually replaced traditional mechanical hard disks as the mainstream storage medium of computers, servers and data centers due to their fast read and write speed, strong shock resistance, low power consumption and other advantages, and have been widely used in many fields. Although solid state disks (SSD) have high performance and high reliability, once a fault occurs, it will not only cause data loss and bring the risk of data leakage, but also affect business continuity, terminal data services, cause system crashes, and thus directly cause serious economic losses. Therefore, it is very important to predict the fault of the solid state disk.

[0003] At present, when predicting the fault of the solid state disk, a long time is consumed, the prediction of the fault of the solid state disk cannot be fed back in time, and the accuracy of the fault prediction is poor. Therefore, the present application proposes a solid state disk fault prediction method and system based on artificial intelligence, which efficiently realizes comprehensive fault prediction of the solid state disk with the help of artificial intelligence, reduces the time delay of the solid state disk fault prediction, so that measures can be taken in time according to the solid state disk fault prediction result, avoiding the loss of data caused by the failure of the solid state disk when the countermeasures cannot be taken in time, thereby affecting the normal use of the solid state disk, causing the business to be unable to proceed, leading to customer loss and economic loss, and improving the accuracy of the solid state disk fault prediction. SUMMARY

[0004] The present application aims to provide a solid state disk fault prediction method and system based on artificial intelligence to solve the problems raised in the background.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical scheme: a solid state disk fault prediction method based on artificial intelligence, comprising:

[0006] establishing an association relationship for the target solid state disk;

[0007] obtaining the running log of the target solid state disk based on the association relationship to obtain the solid state disk log information;

[0008] parsing the solid state disk log information and extracting the associated information to obtain the target associated information;

[0009] using an artificial intelligence model to respectively perform hardware fault prediction analysis and software fault prediction analysis according to the target associated information to obtain fault prediction analysis data;

[0010] According to the fault prediction analysis data, a fault risk assessment is performed to obtain a fault prediction evaluation result.

[0011] Further, an association relationship is established for the target solid state disk, including:

[0012] Security verification is performed for the target solid state disk, and a trust relationship is established after the security verification is passed, so as to preliminarily determine the association relationship;

[0013] Based on the trust relationship, basic information of the target solid state disk is obtained;

[0014] According to the basic information of the target solid state disk, further association is performed, the basic information of the target solid state disk is accessed and read, and the access path of the running log is determined;

[0015] A running log access reading instruction is generated for the access path, and running log acquisition control information is obtained.

[0016] Further, the solid state disk log information is parsed and associated information is extracted, including:

[0017] The solid state disk log information is parsed and standardized to obtain solid state disk log processing information;

[0018] The model input information is determined according to the artificial intelligence model;

[0019] The model input information is combined with the solid state disk log processing information for first matching analysis, and a first matching result is determined;

[0020] According to the first matching result, target information is extracted from the solid state disk log processing information to obtain first target association information;

[0021] The information amount of the model input information is analyzed in combination with the first target association information, and it is determined whether the number of the first target association information is consistent with the number of the model input data information, to obtain an information amount analysis result;

[0022] According to the information amount analysis result, the model input information is screened in combination with the first target association information, to obtain a model input information screening result;

[0023] The model output information in the model input information screening result is parsed, and a second matching analysis is performed according to the basic information in combination with the solid state disk log processing information, to obtain a second matching result;

[0024] According to the second matching result, target information is extracted from the solid state disk log processing information to obtain second target association information;

[0025] The first target association information and the second target association information are combined as target association information, and the extraction is completed.

[0026] Further, the artificial intelligence model comprises: an artificial intelligence first analysis model and an artificial intelligence second analysis model, when the artificial intelligence model is used to perform hardware fault prediction analysis and software fault prediction analysis according to the target correlation information, the target correlation information is sorted into a first target correlation information set and a second target correlation information set according to the artificial intelligence first analysis model and the artificial intelligence second analysis model, the artificial intelligence first analysis model is used to perform hardware fault prediction analysis according to the first target correlation information set, and first fault prediction analysis data is obtained, and the artificial intelligence second analysis model is used to perform software fault prediction analysis according to the second target correlation information set, and second fault prediction analysis data is obtained.

[0027] Further, the artificial intelligence first analysis model is used to perform hardware fault prediction analysis according to the first target correlation information set, comprising:

[0028] Performing hardware composition analysis on the target solid state disk to determine the hardware composition components of the solid state disk;

[0029] Performing effective information identification on the first target correlation information set according to the hardware composition components to determine the effective information of the hardware composition components;

[0030] Performing wear analysis on the hardware composition components according to the effective information of the hardware composition components to obtain first sub-analysis data of the first fault prediction analysis data;

[0031] Performing failure analysis on the hardware composition components according to the effective information of the hardware composition components to obtain second sub-analysis data of the first fault prediction analysis data.

[0032] Further, when the artificial intelligence first analysis model is used to perform hardware fault prediction analysis according to the first target correlation information set, remaining useful life prediction of the target solid state disk is also performed, comprising:

[0033] Using a deep neural network model to pre-train the hard disk drive in the target solid state disk to obtain a hard disk drive remaining useful life prediction initial model;

[0034] Obtaining sample data for the hard disk drive, and selecting a sample with less data amount in the sample data as a target sample;

[0035] Using the target sample to fine-tune the hard disk drive remaining useful life prediction initial model to obtain a hard disk drive remaining useful life prediction optimized model;

[0036] According to the hard disk drive remaining service life prediction optimization model, a target solid state disk remaining service life prediction model is determined, and the target solid state disk remaining service life prediction model is used to predict the remaining service life of the target solid state disk, to obtain the target solid state disk remaining service life prediction data.

[0037] Further, the software failure prediction analysis is performed according to the second target association information set by using the artificial intelligence second analysis model, including:

[0038] The internal behavior of the target solid state disk is analyzed to determine the internal behavior of the target solid state disk.

[0039] According to the internal behavior of the target solid state disk, the behavior data of the second target association information set is obtained, and the internal behavior data of the target solid state disk is obtained.

[0040] According to the internal behavior data of the target solid state disk, the behavior analysis of the target solid state disk is performed to determine the current behavior analysis data.

[0041] Based on the current behavior analysis data, the behavior failure prediction is performed to obtain the second failure prediction analysis data.

[0042] Further, the behavior failure prediction based on the current behavior analysis data includes:

[0043] According to the current behavior analysis data, the behavior trend analysis of the internal behavior is performed in combination with the historical behavior analysis data to obtain the behavior parameter change data of the internal behavior.

[0044] According to the behavior parameter change data of the internal behavior, the change prediction data is obtained.

[0045] The change prediction data is combined with the current behavior analysis data to determine the prediction data, and the second failure prediction analysis data is obtained.

[0046] Further, the fault risk assessment is performed according to the failure prediction analysis data, including:

[0047] According to the failure prediction analysis data, the probability of failure occurrence in a preset time period is calculated according to the failure prediction type, and the failure prediction occurrence probability is obtained.

[0048] According to the failure prediction analysis data, the failure impact analysis is performed, and the failure impact force is evaluated according to the failure impact analysis data, to obtain the failure impact evaluation data.

[0049] The failure prediction occurrence probability is combined with the failure impact evaluation data to determine the risk level, and the data report is generated according to the failure prediction occurrence probability and the failure impact analysis data, to obtain the failure prediction evaluation result.

[0050] The failure prediction result feedback and risk prompt are performed according to the risk level.

[0051] An artificial intelligence-based solid state disk failure prediction system comprises a relationship creation module, an information acquisition module, an information processing module, an intelligent prediction module and a result determination module.

[0052] The relationship creation module is configured to establish an association relationship for a target solid state disk.

[0053] The information acquisition module is configured to acquire a running log of the target solid state disk based on the association relationship to obtain solid state disk log information.

[0054] The information processing module is configured to parse the solid state disk log information and extract associated information to obtain target associated information.

[0055] The intelligent prediction module is configured to match an artificial intelligence model according to the target associated information, and perform hardware failure prediction analysis and software failure prediction analysis on the target associated information corresponding to the matching respectively by using the artificial intelligence model according to the matching combination to obtain failure prediction analysis data.

[0056] The result determination module is configured to perform failure risk assessment according to the failure prediction analysis data to obtain a failure prediction evaluation result.

[0057] The present application uses an artificial intelligence model to efficiently realize comprehensive failure prediction of a solid state disk by means of artificial intelligence, reduces the time delay of solid state disk failure prediction, enables timely determination of failure prediction evaluation results and timely adoption of measures when the failure risk of the solid state disk is high, avoids data loss caused by the inability to take timely countermeasures when the solid state disk fails, thereby affecting the normal use of the solid state disk, causing the business to be unable to proceed, leading to customer loss and economic loss, and also ensuring the accuracy of failure prediction and improving the accuracy of failure prediction evaluation results.

[0058] Other features and advantages of the present application will be described in the following description, and some will become apparent from the description, or will be learned from practice of the present application. The purpose and other advantages of the present application can be achieved and obtained by the structure specifically pointed out in the application file.

[0059] The technical solutions of the present application will be further described in detail below with the help of the drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0060] The accompanying drawings are used to provide a further understanding of the present application, and constitute a part of the specification, together with the embodiments of the present application, to explain the present application, and do not constitute a limitation on the present application. In the drawings:

[0061] Figure 1A flowchart of the steps of the solid state disk fault prediction method according to the present application;

[0062] Figure 2 A composition diagram of the artificial intelligence model in the solid state disk fault prediction method according to the present application;

[0063] Figure 3 A first artificial intelligence analysis model step diagram of the artificial intelligence model in the solid state disk fault prediction method according to the present application;

[0064] Figure 4 A second artificial intelligence analysis model step diagram of the artificial intelligence model in the solid state disk fault prediction method according to the present application;

[0065] Figure 5 A diagram of the solid state disk fault prediction system according to the present application. DETAILED DESCRIPTION

[0066] The preferred embodiments of the present application will be described below with reference to the accompanying drawings, and it should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and are not used to limit the present application.

[0067] As shown in Figure 1 , the present application provides a solid state disk fault prediction method based on artificial intelligence, comprising:

[0068] Step one, establishing an association relationship for a target solid state disk;

[0069] Step two, obtaining a running log of the target solid state disk based on the association relationship to obtain solid state disk log information;

[0070] Step three, parsing the solid state disk log information and extracting associated information to obtain target associated information;

[0071] Step four, using an artificial intelligence model to respectively perform hardware fault prediction analysis and software fault prediction analysis according to the target associated information to obtain fault prediction analysis data;

[0072] Step five, performing fault risk assessment according to the fault prediction analysis data to obtain a fault prediction evaluation result.

[0073] In the above technical solution, the target solid state disk is the solid state disk that needs to be fault predicted.

[0074] In the above technical solution, when extracting associated information, fault prediction analysis associated information is determined for hardware fault prediction analysis and software fault prediction analysis respectively, and target information is extracted from the solid state disk log information according to the fault prediction analysis associated information, so as to obtain the target associated information.

[0075] The technical solution above realizes comprehensive fault prediction of the solid state disk by means of the artificial intelligence model of artificial intelligence, reduces the time delay of solid state disk fault prediction, enables timely determination of the fault prediction evaluation result and timely adoption of measures when the risk of solid state disk fault is high, avoids data loss caused by the failure to take timely countermeasures when the solid state disk fails, thereby affecting the normal use of the solid state disk, causing the business to be unable to proceed, leading to customer loss and economic loss, and also guarantees the accuracy of fault prediction and improves the accuracy of the fault prediction evaluation result, and the communication connection between the solid state disk fault prediction and the target solid state disk is realized by establishing the association relationship for the target solid state disk, so that the running log of the target solid state disk is more safely obtained, the current condition of the target solid state disk is understood, thereby the fault prediction can be better based on the current condition of the target solid state disk, the instantaneity of the fault prediction is guaranteed, and the hardware fault prediction analysis or software fault prediction analysis of the artificial intelligence model is only based on the target association information needed, so as to reduce the redundancy of irrelevant information, avoid data information confusion, improve the efficiency of hardware fault prediction analysis or software fault prediction analysis, reduce the error probability, and thereby guarantee the accuracy of the fault prediction evaluation result.

[0076] In one embodiment of the present application, the association relationship is established for the target solid state disk, which includes:

[0077] The target solid state disk is subjected to security verification, and a trust relationship is established after the security verification is passed, so as to preliminarily determine the association relationship;

[0078] The basic information of the target solid state disk is obtained based on the trust relationship;

[0079] The target solid state disk is further associated according to the basic information of the target solid state disk, the basic information of the target solid state disk is accessed, read and analyzed, and the access path of the running log is determined;

[0080] The running log access reading instruction is generated for the access path, and the running log acquisition control information is obtained.

[0081] In the technical solution above, when the target solid state disk is subjected to security verification, bidirectional security verification is performed between the solid state disk fault prediction platform and the target solid state disk, and the security verification is passed when the bidirectional security verification results are both safe.

[0082] In the technical solution above, when the security verification is not passed, a relationship creation exception prompt is given.

[0083] In the technical solution above, the basic information of the target solid state disk includes the size, structure and other composition information of the solid state disk.

[0084] In the technical solution, when the running log of the target solid state disk is acquired based on the association relationship, the running log in the target solid state disk is accessed and read according to the running log access reading instruction based on the running log acquisition control information, so as to obtain the solid state disk log information.

[0085] The association relationship between the target solid state disk is established in the manner of establishing trust before establishing connection, the information communication security between the target solid state disk is guaranteed, the target solid state disk is prevented from being damaged or causing data information theft of the target solid state disk, the security of the target solid state disk is guaranteed, and the trust relationship between the target solid state disk is established through the security verification of the target solid state disk, so that the forgery, tampering or man-in-the-middle attack is effectively eliminated, and the basic information of the target solid state disk is acquired, so that the general situation of the target solid state disk is known, the running log access reading instruction can access and read the running log more accurately, and the over-access or invalid access reading phenomenon is avoided, and the running log acquisition based on the association relationship for the target solid state disk is guaranteed.

[0086] In one embodiment of the present application, the solid state disk log information is parsed and the association information is extracted, comprising:

[0087] The solid state disk log information is parsed and standardized to obtain solid state disk log processing information;

[0088] The model input information is determined according to the artificial intelligence model;

[0089] The model input information is combined with the solid state disk log processing information for first matching analysis to determine the first matching result;

[0090] The target information is extracted in the solid state disk log processing information according to the first matching result to obtain the first target association information;

[0091] The information quantity analysis is performed on the model input information combined with the first target association information to determine whether the number of the first target association information is consistent with the number of the model input data information, and the information quantity analysis result is obtained;

[0092] The model input information is screened according to the information quantity analysis result combined with the first target association information, and the model input information screening result is obtained;

[0093] The basic information of the model output information in the model input information screening result is parsed, and the second matching analysis is performed on the model input information combined with the solid state disk log processing information according to the basic information, and the second matching result is obtained;

[0094] The target information is extracted in the solid state disk log processing information according to the second matching result to obtain the second target association information;

[0095] The first target associated information and the second target associated information are combined as target associated information, and the extraction is completed.

[0096] In the technical solution, when the solid state disk log information is parsed and normalized, the solid state disk log contained information and the information form are determined, and then the solid state disk log contained information is normalized in combination with the information form to convert the solid state disk log contained information into a recognizable and readable form, so that the solid state disk log processing information is obtained.

[0097] In the technical solution, when the model input information is screened in combination with the first target associated information according to the information amount analysis result, if the information amount analysis result is that the number of the first target associated information is consistent with the number of the model input data information, the model input information does not need to be screened in combination with the first target associated information, and if the information amount analysis result is that the number of the first target associated information is inconsistent with the number of the model input data information, the model input information is screened in combination with the first target associated information.

[0098] In the technical solution, when the model input information is screened in combination with the first target associated information, the information in the model input information that is not corresponding to the first target associated information is screened out to obtain the model input information screening result.

[0099] In the technical solution, when the model output information in the model input information screening result is parsed, it is determined that the model output information in the model input information screening result is obtained by calculation and analysis of which basic information, and then the basic information is combined with the solid state disk log processing information to perform matching analysis again.

[0100] The technical solution realizes the standardization of the solid state disk log information by parsing and normalizing the solid state disk log information, so that the solid state disk log information is recognizable and readable, avoids the situation that some information cannot be normally recognized and read due to the difference in the information form of the solid state disk log contained information, affects the fault prediction of the solid state disk, and guarantees the performance of the solid state disk fault prediction.

[0101] In one embodiment of the present application, as Figure 2As shown, the artificial intelligence model includes: an artificial intelligence first analysis model and an artificial intelligence second analysis model, when the artificial intelligence model is used to perform hardware fault prediction analysis and software fault prediction analysis according to target association information, the target association information is sorted into a first target association information set and a second target association information set according to the artificial intelligence first analysis model and the artificial intelligence second analysis model, the artificial intelligence first analysis model is used to perform hardware fault prediction analysis according to the first target association information set, first fault prediction analysis data is obtained, and the artificial intelligence second analysis model is used to perform software fault prediction analysis according to the second target association information set, second fault prediction analysis data is obtained.

[0102] In the above technical solution, the fault prediction analysis data includes: the first fault prediction analysis data and the second fault prediction analysis data.

[0103] In the above technical solution, the artificial intelligence first analysis model is used to perform hardware fault prediction analysis according to the first target association information set, and at the same time, the artificial intelligence second analysis model is used to perform software fault prediction analysis according to the second target association information set.

[0104] In the above technical solution, when the target association information is sorted into the first target association information set and the second target association information set according to the artificial intelligence first analysis model and the artificial intelligence second analysis model, the target association information can exist only in the first target association information set, only in the second target association information set, or in both the first target association information set and the second target association information set, and the existence of the target association information in the first target association information set and the second target association information set depends on whether it belongs to the model input information of the artificial intelligence first analysis model and the artificial intelligence second analysis model.

[0105] The above technical solution realizes synchronous analysis of hardware fault prediction and software fault prediction through the artificial intelligence first analysis model and the artificial intelligence second analysis model, so that comprehensive fault prediction of the solid state disk can be efficiently performed, so that measures can be taken in a timely manner according to the fault prediction result of the solid state disk, data loss caused by the failure of the solid state disk without timely measures is avoided, and the normal use of the solid state disk is affected, so that the business cannot be performed, leading to customer loss and economic loss. Moreover, the fault prediction analysis is performed with the aid of artificial intelligence, which is accurate and less likely to have fault prediction analysis errors.

[0106] In one embodiment of the present application, as shown in Figure 3 The artificial intelligence first analysis model is used to perform hardware fault prediction analysis according to the first target association information set, including:

[0107] A1, hardware composition analysis is performed on the target solid state disk to determine the hardware composition components of the solid state disk;

[0108] A2, the effective information of the hardware component is determined by identifying the effective information of the first target associated information set according to the hardware component;

[0109] A3, the first sub-analysis data of the first fault prediction analysis data is obtained by performing wear analysis on the hardware component according to the effective information of the hardware component;

[0110] A4, the second sub-analysis data of the first fault prediction analysis data is obtained by performing failure analysis on the hardware component according to the effective information of the hardware component.

[0111] In the above technical solution, the first fault prediction analysis data includes: the first sub-analysis data of the first fault prediction analysis data and the second sub-analysis data of the first fault prediction analysis data.

[0112] In the above technical solution, the hardware components of the solid state disk include: controller, flash chip, etc.

[0113] The above technical solution realizes fault prediction analysis of the hardware components of the solid state disk by using artificial intelligence first analysis model to perform hardware fault prediction analysis according to the first target associated information set, clearly determines the current situation and fault prediction condition of the hardware components of the solid state disk, so that the hardware components of the solid state disk can be predicted in advance before failure occurs, and then the corresponding measures can be taken in time for the failure of the hardware components of the solid state disk, thereby reducing the loss caused by the failure of the hardware components of the solid state disk.

[0114] In one embodiment of the present application, when the artificial intelligence first analysis model is used to perform hardware fault prediction analysis according to the first target associated information set, the remaining service life of the target solid state disk is also predicted, which includes:

[0115] A deep neural network model is used to pre-train the hard disk drive in the target solid state disk to obtain a hard disk drive remaining service life prediction initial model;

[0116] Sample data is obtained for the hard disk drive, and a small amount of data in the sample data is selected as a target sample;

[0117] The target sample is used to fine-tune the hard disk drive remaining service life prediction initial model to obtain a hard disk drive remaining service life prediction optimization model;

[0118] The target solid state disk remaining service life prediction model is determined according to the hard disk drive remaining service life prediction optimization model, and the target solid state disk remaining service life prediction model is used to predict the remaining service life of the target solid state disk to obtain the remaining service life prediction data of the target solid state disk.

[0119] In the technical solution, when the deep neural network model is pre-trained for the hard disk drive in the target solid state disk, the neural network model Transformer based on the attention mechanism is used to pre-train the remaining useful life prediction of the hard disk drive in the solid state disk, so that the deep neural network model can pre-train the remaining useful life prediction of the hard disk drive in the solid state disk, and obtain the initial model for predicting the remaining useful life of the hard disk drive.

[0120] In the technical solution, when the deep neural network model is pre-trained for the hard disk drive in the target solid state disk, the neural network model Transformer based on the attention mechanism is used to pre-train the remaining useful life prediction of the hard disk drive in the solid state disk, so that the deep neural network model can pre-train the remaining useful life prediction of the hard disk drive in the solid state disk, and obtain the initial model for predicting the remaining useful life of the hard disk drive.

[0121] In the technical solution, when the deep neural network model is pre-trained for the hard disk drive in the target solid state disk, the neural network model Transformer based on the attention mechanism is used to pre-train the remaining useful life prediction of the hard disk drive in the solid state disk, so that the deep neural network model can pre-train the remaining useful life prediction of the hard disk drive in the solid state disk, and obtain the initial model for predicting the remaining useful life of the hard disk drive.

[0122] The technical scheme avoids the influence of insufficient remaining service life of the target solid state disk on the normal use of the solid state disk during use, and pre-trains the hard disk drive in the target solid state disk by using a deep neural network model, so as to fully utilize the characteristics of large models, large data and large calculations, improve the performance of the hard disk drive remaining service life prediction model, and make the hard disk drive remaining service life prediction optimization model better predict the remaining service life of the target solid state disk, thereby improving the accuracy of the remaining service life prediction data of the target solid state disk.

[0123] In one embodiment of the present application, as shown in Figure 4 The software failure prediction analysis based on the second target correlation information set by using the artificial intelligence second analysis model includes:

[0124] B1, internal behavior analysis is performed on the target solid state disk to determine the internal behavior of the target solid state disk.

[0125] B2, behavior data acquisition is performed on the second target correlation information set according to the internal behavior of the target solid state disk to obtain internal behavior data of the target solid state disk.

[0126] B3, behavior analysis is performed on the target solid state disk according to the internal behavior data of the target solid state disk to determine current behavior analysis data.

[0127] B4, behavior failure prediction is performed based on the current behavior analysis data to obtain second failure prediction analysis data.

[0128] In the above technical scheme, the internal behavior of the target solid state disk includes access behavior, drive behavior, etc.

[0129] In the above technical scheme, when the behavior analysis is performed on the target solid state disk according to the internal behavior data of the target solid state disk, the behavior parameter analysis is performed on the internal behavior of the solid state disk according to the logical page / block, and the behavior parameter includes FTL collapse, firmware bug, drive timeout, etc.

[0130] The above technical scheme realizes the failure prediction of the internal behavior of the target solid state disk by using the artificial intelligence second analysis model to perform software failure prediction analysis based on the second target correlation information set, so as to determine the current internal behavior of the target solid state disk, predict the failure occurrence of the internal behavior, and further reduce the influence of internal behavior abnormalities on the performance of the solid state disk.

[0131] In one embodiment of the present application, the behavior failure prediction based on the current behavior analysis data includes:

[0132] The behavior trend analysis of the internal behavior is performed on the current behavior analysis data combined with the historical behavior analysis data, and the behavior parameter change data of the internal behavior is obtained.

[0133] The change prediction is performed according to the behavior parameter change data of the internal behavior, and change prediction data is obtained.

[0134] The prediction data determination is performed by combining the change prediction data with the current behavior analysis data, and second fault prediction analysis data is obtained.

[0135] In the above technical solution, when the behavior trend analysis of the internal behavior is performed on the current behavior analysis data combined with the historical behavior analysis data, the current behavior analysis data is combined with the historical behavior analysis data according to the behavior parameters to perform the behavior trend change analysis, for each behavior parameter, the number of behavior analysis data is determined, when the number of behavior analysis data is not greater than a preset threshold, the change analysis calculation is performed on the behavior analysis data of the behavior parameters in the adjacent two behavior processes, the behavior parameter change data of the internal behavior is obtained, when the number of behavior analysis data is greater than the preset threshold, the change analysis combination is determined according to the behavior analysis data, the behavior analysis data is combined with the previous n behavior analysis data according to the combination determination rule, the data analysis combination of the multiple behavior parameters is obtained, the combination data determination is performed on the data analysis combination respectively, and the behavior parameter change data analysis calculation is performed between the adjacent combinations according to the combination data, and the behavior parameter change data of the internal behavior is obtained. Here, the value of n is less than the preset threshold.

[0136] The above technical solution combines the historical behavior analysis data to clearly show the overall behavior parameter change of the internal behavior when performing the behavior fault prediction, provides more sufficient data support for the behavior fault prediction, and improves the accuracy of the behavior fault prediction.

[0137] In one embodiment of the present application, the fault risk assessment is performed according to the fault prediction analysis data, including:

[0138] The probability of fault occurrence in a preset time period is calculated according to the fault prediction analysis data according to the fault prediction type, and the fault prediction occurrence probability is obtained.

[0139] The fault impact analysis is performed according to the fault prediction analysis data, and the fault impact assessment is performed on the fault impact analysis data, and the fault impact assessment data is obtained.

[0140] The risk level is determined by combining the fault prediction occurrence probability with the fault impact assessment data, and the data report is generated according to the fault prediction occurrence probability and the fault impact analysis data, and the fault prediction assessment result is obtained.

[0141] The fault prediction result feedback and risk prompt are performed according to the risk level.

[0142] In the technical solution, the preset time period is determined according to the frequency of fault prediction of the solid state disk, for example, if the frequency of fault prediction of the solid state disk is once a week, the preset time period is one week, and the preset time period can be appropriately shortened based on one week, for example, five days or six days.

[0143] In the technical solution, when the fault influence analysis data is evaluated according to the fault influence evaluation rule, for example, according to the fault influence analysis data, if the solid state disk cannot be used when the fault occurs, the fault influence evaluation data is high influence, if the local function of the solid state disk cannot be used, the fault influence evaluation data is medium influence, and if the use of the solid state disk is not affected, the fault influence evaluation data is low influence.

[0144] In the technical solution, when the risk level is determined by combining the fault prediction occurrence probability with the fault influence evaluation data, the fault prediction occurrence probability is evaluated to determine the fault prediction occurrence probability level, the fault prediction occurrence probability level is combined with the fault influence evaluation data for comprehensive analysis, when the fault prediction occurrence probability level is high probability and the fault influence evaluation data is high influence, the risk level is extremely high, when the fault prediction occurrence probability level is high probability or the fault influence evaluation data is high influence, the risk level is high, when the fault prediction occurrence probability level is medium probability or the fault influence evaluation data is medium influence, the risk level is medium, and when the fault prediction occurrence probability level is low probability and the fault influence evaluation data is low influence, the risk level is low.

[0145] In the technical solution, the fault prediction result feedback and risk prompt are performed according to the risk level, and different risk levels correspond to different risk prompt modes.

[0146] The technical solution presents the fault prediction evaluation result of the solid state disk in a more intuitive way by evaluating the fault risk according to the fault prediction analysis data, and combines the risk prompt to improve the vigilance of the relevant personnel, so that the relevant personnel can not only intuitively know the current fault prediction occurrence probability and fault influence analysis data of the solid state disk, but also can improve vigilance through the risk prompt under different risk levels, so that the relevant personnel can take corresponding measures according to the fault prediction evaluation result of the solid state disk in time, thereby avoiding data loss caused by failure to take measures in time when the solid state disk fails, and affecting the normal use of the solid state disk, so that the business cannot be carried out, leading to customer loss and economic loss.

[0147] As shown in Figure 5 The embodiment of the present application provides a solid state disk fault prediction system based on artificial intelligence, which comprises a relationship creation module, an information acquisition module, an information processing module, an intelligent prediction module and a result determination module.

[0148] The relationship creation module is configured to establish an association relationship for the target solid state disk.

[0149] The information acquisition module is configured to acquire running logs of the target solid state disk based on the association relationship to obtain solid state disk log information.

[0150] The information processing module is configured to analyze the solid state disk log information and extract associated information to obtain target associated information.

[0151] The intelligent prediction module is configured to match an artificial intelligence model according to the target associated information, and perform hardware fault prediction analysis and software fault prediction analysis on the target associated information corresponding to the matching according to the matched artificial intelligence model to obtain fault prediction analysis data.

[0152] The result determination module is configured to perform fault risk assessment according to the fault prediction analysis data to obtain a fault prediction evaluation result.

[0153] In the above technical solution, the relationship creation module, the information acquisition module, the information processing module, the intelligent prediction module, and the result determination module are connected in sequence.

[0154] The above technical solution uses an artificial intelligence model to efficiently realize comprehensive fault prediction of a solid state disk through the intelligent prediction module, reduces the time delay of solid state disk fault prediction, enables timely determination of a fault prediction evaluation result and timely adoption of measures when the risk of a solid state disk fault is high, avoids data loss caused by the inability to take timely countermeasures when a solid state disk fails, thereby affecting the normal use of the solid state disk, causing the business to be unable to proceed, leading to customer loss and economic loss, and also guarantees the accuracy of fault prediction and improves the precision of the fault prediction evaluation result. Moreover, the relationship creation module establishes an association relationship for the target solid state disk to realize communication connection between the solid state disk fault prediction and the target solid state disk, so that the information acquisition module can more safely acquire the running logs of the target solid state disk and understand the current status of the target solid state disk, thereby enabling the intelligent prediction module to better perform fault prediction based on the current status of the target solid state disk, guaranteeing the immediacy of fault prediction, and the information processing module analyzes the solid state disk log information and extracts associated information, so that the artificial intelligence model in the intelligent prediction module can only perform hardware fault prediction analysis or software fault prediction analysis according to the required target associated information, reducing redundant irrelevant information, avoiding data information disorder, improving the efficiency of hardware fault prediction analysis or software fault prediction analysis, reducing the error probability, and thereby guaranteeing the precision of the fault prediction evaluation result determined by the result determination module.

[0155] The first and second in the present disclosure are only referred to different application stages.

[0156] Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

[0157] It is to be understood that the disclosure is not limited to the precise structures herein described and illustrated in the drawings, and that various modifications and changes can be made without departing from its scope. The scope of the disclosure is limited only by the claims that follow.

Claims

1. A method for predicting a failure of a solid state drive based on artificial intelligence, the method comprising: The method comprises the following steps: establishing an association relationship for a target solid state disk; based on the association relationship, obtaining the running log of the target solid state disk to obtain the solid state disk log information; analyzing the solid state disk log information and extracting the associated information to obtain the target associated information; using an artificial intelligence model to perform hardware fault prediction analysis and software fault prediction analysis based on the target associated information to obtain fault prediction analysis data; performing fault risk assessment according to the fault prediction analysis data to obtain a fault prediction evaluation result. 2.The solid state drive failure prediction method of claim 1, wherein, The method for establishing an association relationship for a target solid state disk comprises the following steps: performing security verification on the target solid state disk, and establishing a trust relationship after the security verification is passed to preliminarily determine the association relationship; obtaining the basic information of the target solid state disk based on the trust relationship; further associating according to the basic information of the target solid state disk, accessing and reading the basic information of the target solid state disk to determine the access path of the running log; generating a running log access reading instruction for the access path to obtain running log acquisition control information. 3.The solid state drive failure prediction method of claim 1, wherein, The method for analyzing the solid state disk log information and extracting the associated information comprises the following steps: analyzing and normalizing the solid state disk log information to obtain solid state disk log processing information; determining model input information according to the artificial intelligence model; performing first matching analysis on the model input information combined with the solid state disk log processing information to determine a first matching result; extracting target information in the solid state disk log processing information according to the first matching result to obtain first target associated information; performing information quantity analysis on the model input information combined with the first target associated information to determine whether the number of the first target associated information is consistent with the number of model input data information, and obtaining an information quantity analysis result; performing filtering on the model input information according to the information quantity analysis result combined with the first target associated information to obtain a model input information filtering result; performing basic information analysis on the model output information in the model input information filtering result, and performing second matching analysis on the solid state disk log processing information combined with the basic information to obtain a second matching result; extracting target information in the solid state disk log processing information according to the second matching result to obtain second target associated information; combining the first target associated information and the second target associated information as the target associated information to complete the extraction. 4.The solid state drive failure prediction method of claim 1, wherein, The artificial intelligence model comprises an artificial intelligence first analysis model and an artificial intelligence second analysis model. When the artificial intelligence model is used to perform hardware fault prediction analysis and software fault prediction analysis based on the target associated information, the target associated information is sorted into a first target associated information set and a second target associated information set according to the artificial intelligence first analysis model and the artificial intelligence second analysis model. The artificial intelligence first analysis model is used to perform hardware fault prediction analysis based on the first target associated information set to obtain first fault prediction analysis data. The artificial intelligence second analysis model is used to perform software fault prediction analysis based on the second target associated information set to obtain second fault prediction analysis data.

5. The solid state drive failure prediction method of claim 4, wherein, The method for using the artificial intelligence first analysis model to perform hardware fault prediction analysis based on the first target associated information set comprises the following steps: The hardware composition of the target solid state disk is analyzed, and the hardware composition components of the solid state disk are determined; According to the hardware composition components, the effective information of the first target association information set is identified, and the effective information of the hardware composition components is determined; According to the effective information of the hardware composition components, the wear analysis of the hardware composition components is carried out, and the first sub-analysis data of the first fault prediction analysis data is obtained; According to the effective information of the hardware composition components, the failure analysis of the hardware composition components is carried out, and the second sub-analysis data of the first fault prediction analysis data is obtained.

6. The solid state drive failure prediction method of claim 5, wherein, When the first analysis model of artificial intelligence is used to analyze the hardware fault prediction according to the first target association information set, the remaining service life of the target solid state disk is also predicted, including: A deep neural network model is used to pre-train the hard disk drive in the target solid state disk to obtain a hard disk drive remaining service life prediction initial model; Sample data is obtained for the hard disk drive, and samples with less data are selected as target samples in the sample data; The target sample is used to fine-tune the hard disk drive remaining service life prediction initial model to obtain a hard disk drive remaining service life prediction optimization model; According to the hard disk drive remaining service life prediction optimization model, a target solid state disk remaining service life prediction model is determined, and the target solid state disk remaining service life prediction model is used to predict the remaining service life of the target solid state disk to obtain the remaining service life prediction data of the target solid state disk.

7. The solid state drive failure prediction method of claim 4, wherein, The second analysis model of artificial intelligence is used to analyze the software fault prediction according to the second target association information set, including: Internal behavior analysis is performed on the target solid state disk to determine the internal behavior of the target solid state disk; According to the internal behavior of the target solid state disk, behavior data is obtained for the second target association information set to obtain internal behavior data of the target solid state disk; According to the internal behavior data of the target solid state disk, behavior analysis is performed on the target solid state disk to determine the current behavior analysis data; Based on the current behavior analysis data, behavior fault prediction is performed to obtain the second fault prediction analysis data.

8. The solid state drive failure prediction method of claim 7, wherein, Based on the current behavior analysis data, behavior fault prediction is performed, including: Internal behavior trend analysis is performed on the current behavior analysis data combined with historical behavior analysis data to obtain behavior parameter change data of the internal behavior; According to the behavior parameter change data of the internal behavior, change prediction is performed to obtain change prediction data; The change prediction data is combined with the current behavior analysis data to determine the prediction data, and the second fault prediction analysis data is obtained. 9.The solid state drive failure prediction method of claim 1, wherein, According to the fault prediction analysis data, the fault risk assessment is carried out, including: According to the fault prediction analysis data, the probability of fault occurrence in a preset time period is calculated according to the fault prediction type, and the fault prediction occurrence probability is obtained; According to the fault prediction analysis data, the fault influence analysis is carried out, and the fault influence analysis data is used to evaluate the fault influence, and the fault influence evaluation data is obtained; The fault prediction occurrence probability is combined with the fault influence evaluation data to determine the risk level, and the fault prediction evaluation result is obtained by generating a data report according to the fault prediction occurrence probability and the fault influence analysis data; The fault prediction result feedback and risk prompt are performed according to the risk level. 10.A system for predicting failure of a solid state drive based on artificial intelligence, the system comprising: The solid state disk fault prediction system comprises a relationship creation module, an information acquisition module, an information processing module, an intelligent prediction module and a result determination module. The relationship creation module is configured to establish an association relationship for a target solid state disk. The information acquisition module is configured to acquire running logs of the target solid state disk based on the association relationship to obtain solid state disk log information. The information processing module is configured to analyze the solid state disk log information and extract associated information to obtain target associated information. The intelligent prediction module is configured to match an artificial intelligence model according to the target associated information, and perform hardware fault prediction analysis and software fault prediction analysis on the target associated information corresponding to the matching according to the matched artificial intelligence model to obtain fault prediction analysis data. The result determination module is configured to perform fault risk assessment according to the fault prediction analysis data to obtain a fault prediction evaluation result.

Images