Data storage method

Abstract

The invention discloses a data storage method, which relates to the technical field of information storage, and comprises the following steps: acquiring operation environment data to be written into a hardware medium, and evaluating an operation environment influence coefficient of the hardware medium to be written into; acquiring historical storage data of a to-be-written hardware medium, and evaluating an operation performance index and a dynamic load state index of the to-be-written hardware medium; based on the operation performance state index of the to-be-written hardware medium and the operation environment influence coefficient of the to-be-written hardware medium, initializing an internal storage unit health value of the to-be-written hardware medium; based on the dynamic load state index of the to-be-written hardware medium, correcting and initializing the internal storage unit health value of the to-be-written hardware medium, and obtaining a storage unit dynamic health value of the to-be-written hardware medium; and screening out the dynamic health value of the storage unit to be written into the hardware medium, and performing dynamic storage when the dynamic health value meets the storage requirement of the to-be-written data. The method has the advantages that the stability of data storage is improved, and the risk of data loss is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of information storage, and in particular to a data storage and preservation method. Background Art

[0002] Data storage and preservation refers to the process of storing information, data or files in digital form on specific hardware media (such as hard disk, SSD, mechanical hard disk, etc.) so that they can be accessed, retrieved and used at any time in the future.

[0003] At present, since most companies deploy servers early and the benefits of overall server group updates are low, the risk of data loss will increase with the time of server deployment. Therefore, a dynamic storage solution for control data is proposed based on comprehensive analysis of the server's own multi-factor status to reduce the risk of data loss. Summary of the invention

[0004] In order to solve the above technical problems, a data storage and preservation method is provided. This technical solution solves the above-mentioned problem. At present, since most companies deploy servers early and the overall update and iteration benefits of the server group are low, the risk of data loss will increase with the deployment time of the server. Therefore, a dynamic storage solution for control data is proposed based on a comprehensive analysis of the server's own multi-factor status to reduce the risk of data loss.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] A data storage method, comprising:

[0007] Obtaining the operating environment data of the hardware medium to be written, analyzing the impact trend of the operating environment data on the hardware medium to be written, and evaluating the operating environment impact coefficient of the hardware medium to be written;

[0008] Obtain historical storage data to be written into hardware media;

[0009] Analyze the storage quality of the hardware medium to be written based on the historical storage data of the hardware medium to be written, and evaluate the operating performance indicators of the hardware medium to be written;

[0010] Initialize the internal storage unit health value of the hardware medium to be written based on the operating performance status indicator of the hardware medium to be written and the operating environment impact coefficient of the hardware medium to be written;

[0011] Analyze the storage load trend of the hardware medium to be written based on the historical storage data of the hardware medium to be written, and evaluate the dynamic load status index of the hardware medium to be written;

[0012] Based on the dynamic load status indicator of the hardware medium to be written, the internal storage unit health value of the initialized hardware medium to be written is corrected to obtain the storage unit dynamic health value of the hardware medium to be written;

[0013] Based on the dynamic health value of the storage unit of the hardware medium to be written, the dynamic health value of the storage unit of the hardware medium to be written is screened out to meet the storage demand of the data to be written for dynamic storage;

[0014] The initialization of the internal storage unit health value to be written into the hardware medium is specifically:

[0015] R=S·G

[0016] Where R is the initialization health value of each storage unit inside the hardware medium to be written, S is the operating environment impact coefficient of the hardware medium to be written, and the operating performance index of the hardware medium to be written;

[0017] The dynamic health value of the storage unit to be written into the hardware medium is specifically:

[0018] U=F·R

[0019] Wherein, U is the dynamic health value of the storage unit to be written into the hardware medium, and F is the dynamic load status indicator of the storage medium to be stored in the future unit time.

[0020] Preferably, obtaining the operating environment data of the hardware medium to be written, analyzing the influence trend of the operating environment data on the hardware medium to be written, and evaluating the operating environment influence coefficient of the hardware medium to be written specifically includes:

[0021] Standardization processing is performed based on the operating environment data to be written into the hardware medium;

[0022] Based on the standardized manufacturing calibration ideal working environment data of the hardware medium to be written, determine the ideal standardized working environment parameter array of the hardware medium to be written; the manufacturing calibration ideal working environment includes: working temperature, working humidity, and working voltage;

[0023] Performing a difference operation based on an array of ideal standardized working environment parameters of the hardware medium to be written and the operating environment data of the hardware medium to be written, obtaining an operating environment deviation parameter of the hardware medium to be written;

[0024] Calculating an entropy value of the operating environment deviation parameter of the hardware medium to be written based on the operating environment deviation parameter of the hardware medium to be written and an array of ideal standardized working environment parameters of the hardware medium to be written;

[0025] Determine the coefficient of variation of the operating environment deviation parameter according to the entropy value of the operating environment deviation parameter to be written into the hardware medium, and calculate the deviation weight of the operating environment deviation parameter to be written into the hardware medium;

[0026] Calculating the operating environment impact coefficient of the hardware medium to be written according to the deviation weight of the operating environment deviation parameter of the hardware medium to be written and the operating environment deviation parameter of the hardware medium to be written;

[0027] The operating environment impact coefficient of the hardware medium to be written is specifically:

[0028]

[0029] Where S is the operating environment impact coefficient of the hardware medium to be written, E ij is the entropy value of the jth deviation parameter of the ith operating environment to be written into the hardware medium, d ij is the coefficient of variation of the jth deviation parameter of the ith operating environment to be written into the hardware medium, w ij is the deviation weight of the jth deviation parameter of the ith operating environment to be written into the hardware medium, V ij is the jth deviation parameter value of the i-th operating environment to be written into the hardware medium.

[0030] Preferably, analyzing the storage quality of the hardware medium to be written based on the historical storage data of the hardware medium to be written, and evaluating the operating performance index of the hardware medium to be written specifically include:

[0031] Based on the historical storage data to be written into the hardware medium, the storage quality characteristic data in the historical storage data is filtered out according to the unit time, and the storage quality characteristic time series data to be written into the hardware medium is obtained; the storage quality characteristic data includes: read and write speed data, access time data;

[0032] Normalizing the storage quality characteristic time series data to be written into the hardware medium to obtain normalized storage quality characteristic time series data to be written into the hardware medium;

[0033] Based on AR autoregression, a performance change trend prediction model for the hardware media to be written is constructed;

[0034] Substituting the normalized storage quality characteristic time series data of the hardware medium to be written into the performance change trend prediction model of the hardware medium to be written according to the time unit interval, and evaluating the operation performance index of the hardware medium to be written;

[0035] The performance change trend prediction model for the hardware medium to be written is specifically:

[0036]

[0037] In the formula, G is the operating performance index of the hardware medium to be written, x i′t-kis the i′th storage quality characteristic data under the tkth unit time lag learning order to be written into the hardware medium, α0 intercept, α k is the autoregressive coefficient, ε t is the error term under t unit time lag learning order, and q is the total number of lag learning orders.

[0038] Preferably, analyzing the storage load trend of the hardware medium to be written based on the historical storage data of the hardware medium to be written, and evaluating the dynamic load status index of the hardware medium to be written specifically include:

[0039] Based on the historical storage data to be written into the hardware medium, mark the load action event of the storage medium under each unit time, and obtain the characteristic data of the historical load action event of the storage medium; the historical load action event includes: read and write events, IOPS events;

[0040] Constructing a historical load action event curve graph of the medium to be stored according to the historical load action event characteristic data of the storage medium;

[0041] Based on the historical load action event curve graph of the medium to be stored, the peak and valley load trends of the historical load action events are marked to obtain the historical load trend characteristic parameters of the medium to be stored;

[0042] Construct an autoregressive load assessment model for the hardware media to be written;

[0043] Based on the autoregressive load assessment model of the hardware medium to be written, the historical load trend characteristic parameters of the medium to be stored in a unit time are used as the independent variable input, and the dynamic load state index of the medium to be stored in the future unit time is predicted as the dependent variable output;

[0044] The autoregressive load evaluation model for the hardware medium to be written is specifically:

[0045]

[0046] Where F is the dynamic load state index of the medium to be stored per unit time in the future, is the characteristic parameter of the historical load trend at the tth unit time, β0 is the intercept, β1, β2, β l are regression coefficients, and ∈ is the error term.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] The present invention proposes a data storage and preservation solution, which analyzes the storage unit initialization health value of the hardware medium to be written in the operating environment and the storage quality state of the hardware medium to be written; performs regression analysis according to the change trend of historical storage data, determines the load state trend of the storage unit of the hardware medium to be written, obtains the dynamic load state index of the hardware medium to be written, and uses the index to correct the initialization health value of each storage unit inside the hardware medium to be written, obtains the dynamic health value of the storage unit of the hardware medium to be written, and provides guidance parameters for real-time data storage; the advantages of the present invention are: improving the stability of data storage and reducing the risk of data loss. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 A flow chart of a data storage method;

[0050] Figure 2 A flow chart of a method for evaluating the operating environment impact coefficient of a hardware medium to be written;

[0051] Figure 3 A flow chart of a method for evaluating the operating performance index of a hardware medium to be written;

[0052] Figure 4 A flow chart of a method for evaluating dynamic load status indicators of hardware media to be written. DETAILED DESCRIPTION

[0053] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0054] Reference Figure 1 As shown, a data storage method includes:

[0055] Obtaining the operating environment data of the hardware medium to be written, analyzing the impact trend of the operating environment data on the hardware medium to be written, and evaluating the operating environment impact coefficient of the hardware medium to be written;

[0056] Obtain historical storage data to be written into hardware media;

[0057] Analyze the storage quality of the hardware medium to be written based on the historical storage data of the hardware medium to be written, and evaluate the operating performance indicators of the hardware medium to be written;

[0058] Initialize the internal storage unit health value of the hardware medium to be written based on the operating performance status indicator of the hardware medium to be written and the operating environment impact coefficient of the hardware medium to be written;

[0059] Analyze the storage load trend of the hardware medium to be written based on the historical storage data of the hardware medium to be written, and evaluate the dynamic load status index of the hardware medium to be written;

[0060] Based on the dynamic load status indicator of the hardware medium to be written, the internal storage unit health value of the initialized hardware medium to be written is corrected to obtain the storage unit dynamic health value of the hardware medium to be written;

[0061] Based on the dynamic health value of the storage unit of the hardware medium to be written, the dynamic health value of the storage unit of the hardware medium to be written is selected to meet the storage demand of the data to be written for dynamic storage;

[0062] The initialization of the internal storage unit health value to be written into the hardware medium is specifically:

[0063] R=S·G

[0064] Where R is the initialization health value of each storage unit inside the hardware medium to be written, S is the operating environment impact coefficient of the hardware medium to be written, and G is the operating performance index of the hardware medium to be written;

[0065] The dynamic health value of the storage unit to be written into the hardware medium is specifically:

[0066] U=F·R

[0067] Wherein, U is the dynamic health value of the storage unit to be written into the hardware medium, and F is the dynamic load status indicator of the storage medium to be stored in the future unit time.

[0068] It should be noted that the storage requirements of the data to be written can be obtained based on the metadata of the data itself, and the purpose of this solution is to solve how to store data safely and optimally. The metadata of the data itself is familiar and undisputed to technicians in this field, so it will not be described in detail here.

[0069] This solution analyzes the initialization health value of the storage unit of the hardware medium to be written in the operating environment and the storage quality status of the hardware medium to be written; performs regression analysis based on the change trend of historical storage data, determines the load state trend of the storage unit of the hardware medium to be written, obtains the dynamic load state index of the hardware medium to be written, and uses the index to correct the initialization health value of each storage unit inside the hardware medium to be written, obtains the dynamic health value of the storage unit of the hardware medium to be written, and provides guidance parameters for real-time data storage; the advantages of the present invention are: improving the stability of data storage and reducing the risk of data loss.

[0070] Reference Figure 2As shown, obtaining the operating environment data of the hardware medium to be written, analyzing the influence trend of the operating environment data on the hardware medium to be written, and evaluating the operating environment influence coefficient of the hardware medium to be written specifically include:

[0071] Standardization processing is performed based on the operating environment data to be written into the hardware medium;

[0072] Based on the standardized manufacturing calibration ideal working environment data of the hardware medium to be written, determine the ideal standardized working environment parameter array of the hardware medium to be written; the manufacturing calibration ideal working environment includes: working temperature, working humidity, and working voltage;

[0073] Performing a difference operation based on an array of ideal standardized working environment parameters of the hardware medium to be written and the operating environment data of the hardware medium to be written, obtaining an operating environment deviation parameter of the hardware medium to be written;

[0074] Calculating an entropy value of the operating environment deviation parameter of the hardware medium to be written based on the operating environment deviation parameter of the hardware medium to be written and an array of ideal standardized working environment parameters of the hardware medium to be written;

[0075] Determine the coefficient of variation of the operating environment deviation parameter according to the entropy value of the operating environment deviation parameter to be written into the hardware medium, and calculate the deviation weight of the operating environment deviation parameter to be written into the hardware medium;

[0076] Calculating the operating environment impact coefficient of the hardware medium to be written according to the deviation weight of the operating environment deviation parameter of the hardware medium to be written and the operating environment deviation parameter of the hardware medium to be written;

[0077] The operating environment impact coefficient of the hardware medium to be written is specifically:

[0078]

[0079] Where S is the operating environment impact coefficient of the hardware medium to be written, E ij is the entropy value of the jth deviation parameter of the ith operating environment to be written into the hardware medium, d ij is the coefficient of variation of the jth deviation parameter of the ith operating environment to be written into the hardware medium, w ij is the deviation weight of the jth deviation parameter of the ith operating environment to be written into the hardware medium, V ij is the jth deviation parameter value of the i-th operating environment to be written into the hardware medium.

[0080] It should be noted that the manufacturing calibration ideal working environment data are all obtained through heterogeneous sensors deployed in the server;

[0081] This solution processes actual and ideal operating environment data in a standardized manner, calculates deviation parameters, and uses the entropy method to analyze the degree of deviation variation to determine the weight. Finally, a comprehensive evaluation is performed to obtain the operating environment impact coefficient of the hardware medium, providing key data for the subsequent optimized writing and stable operation of the hardware medium.

[0082] Reference Figure 3 As shown, based on the historical storage data of the hardware medium to be written, the storage quality of the hardware medium to be written is analyzed, and the operating performance indicators of the hardware medium to be written are evaluated, which specifically include:

[0083] Based on the historical storage data to be written into the hardware medium, the storage quality characteristic data in the historical storage data is filtered out according to the unit time, and the storage quality characteristic time series data to be written into the hardware medium is obtained; the storage quality characteristic data includes: read and write speed data, access time data;

[0084] Normalizing the storage quality characteristic time series data to be written into the hardware medium to obtain normalized storage quality characteristic time series data to be written into the hardware medium;

[0085] Based on AR autoregression, a performance change trend prediction model for the hardware media to be written is constructed;

[0086] Substituting the normalized storage quality characteristic time series data of the hardware medium to be written into the performance change trend prediction model of the hardware medium to be written according to the time unit interval, and evaluating the operation performance index of the hardware medium to be written;

[0087] The performance change trend prediction model for the hardware medium to be written is specifically:

[0088]

[0089] In the formula, G is the operating performance index of the hardware medium to be written, x i′t-k is the i′th storage quality characteristic data under the tkth unit time lag learning order to be written into the hardware medium, α0 intercept, α k is the autoregressive coefficient, ε t is the error term under t unit time lag learning order, and q is the total number of lag learning orders.

[0090] This solution analyzes the storage quality characteristic data such as read and write speed, access time, etc. in the historical storage data to be written to the hardware medium, and normalizes it after time series processing. It then uses the AR autoregressive model to build a performance change trend prediction model to evaluate the operating performance indicators of the hardware medium to be written, and improve the quantitative reference indicators for subsequent data storage.

[0091] Reference Figure 4As shown, based on the historical storage data of the hardware medium to be written, the storage load trend of the hardware medium to be written is analyzed, and the dynamic load status indicators of the hardware medium to be written are evaluated, which specifically include:

[0092] Based on the historical storage data to be written into the hardware medium, mark the load action event of the storage medium under each unit time, and obtain the characteristic data of the historical load action event of the storage medium; the historical load action event includes: read and write events, IOPS events;

[0093] Constructing a historical load action event curve graph of the medium to be stored according to the historical load action event characteristic data of the storage medium;

[0094] Based on the historical load action event curve graph of the medium to be stored, the peak and valley load trends of the historical load action events are marked to obtain the historical load trend characteristic parameters of the medium to be stored;

[0095] Construct an autoregressive load assessment model for the hardware media to be written;

[0096] Based on the autoregressive load assessment model of the hardware medium to be written, the historical load trend characteristic parameters of the medium to be stored in a unit time are used as the independent variable input, and the dynamic load state index of the medium to be stored in the future unit time is predicted as the dependent variable output;

[0097] The autoregressive load evaluation model for the hardware medium to be written is specifically:

[0098]

[0099] Where F is the dynamic load state index of the medium to be stored per unit time in the future, is the characteristic parameter of the historical load trend at the tth unit time, β0 is the intercept, β1, β2, β l are regression coefficients, and ∈ is the error term.

[0100] It is understandable that since the storage medium in the server will have different degrees of stability waves in the internal storage unit under different load conditions, this solution analyzes the historical storage data to be written to the hardware medium, marks and obtains the load action event characteristic data of the storage medium (such as read and write events, IOPS events), and then constructs a historical load action event curve graph, and marks the peak and valley trend characteristic parameters of the historical load. Based on these characteristic parameters, an autoregressive load assessment model is constructed to predict the dynamic load state indicators of the storage medium in the future unit time, so as to achieve early assessment and prediction of the load state of the storage medium, thereby improving storage efficiency.

[0101] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A data storage method, characterized in that: include: Obtaining the operating environment data of the hardware medium to be written, analyzing the impact trend of the operating environment data on the hardware medium to be written, and evaluating the operating environment impact coefficient of the hardware medium to be written; Obtain historical storage data to be written into hardware media; Analyze the storage quality of the hardware medium to be written based on the historical storage data of the hardware medium to be written, and evaluate the operating performance indicators of the hardware medium to be written; Initialize the internal storage unit health value of the hardware medium to be written based on the operating performance status indicator of the hardware medium to be written and the operating environment impact coefficient of the hardware medium to be written; Analyze the storage load trend of the hardware medium to be written based on the historical storage data of the hardware medium to be written, and evaluate the dynamic load status index of the hardware medium to be written; Based on the dynamic load status indicator of the hardware medium to be written, the internal storage unit health value of the initialized hardware medium to be written is corrected to obtain the storage unit dynamic health value of the hardware medium to be written; Based on the dynamic health value of the storage unit of the hardware medium to be written, the dynamic health value of the storage unit of the hardware medium to be written is screened out to meet the storage demand of the data to be written for dynamic storage; The initialization of the internal storage unit health value to be written into the hardware medium is specifically: R=S·R Where R is the initialization health value of each storage unit inside the hardware medium to be written, S is the operating environment impact coefficient of the hardware medium to be written, and the operating performance index of the hardware medium to be written; The dynamic health value of the storage unit to be written into the hardware medium is specifically: U=F·R Wherein, U is the dynamic health value of the storage unit to be written into the hardware medium, and F is the dynamic load status indicator of the storage medium to be stored in the future unit time.

2. A data storage method according to claim 1, characterized in that: Obtaining the operating environment data of the hardware medium to be written, analyzing the impact trend of the operating environment data on the hardware medium to be written, and evaluating the operating environment impact coefficient of the hardware medium to be written specifically include: Standardization processing is performed based on the operating environment data to be written into the hardware medium; Based on the standardized manufacturing calibration ideal working environment data of the hardware medium to be written, determining an array of ideal standardized working environment parameters of the hardware medium to be written; the manufacturing calibration ideal working environment includes: working temperature, working humidity, and working voltage; Performing a difference operation based on an array of ideal standardized working environment parameters of the hardware medium to be written and the operating environment data of the hardware medium to be written, obtaining an operating environment deviation parameter of the hardware medium to be written; Calculating an entropy value of the operating environment deviation parameter of the hardware medium to be written based on the operating environment deviation parameter of the hardware medium to be written and an array of ideal standardized working environment parameters of the hardware medium to be written; Determine the coefficient of variation of the operating environment deviation parameter according to the entropy value of the operating environment deviation parameter to be written into the hardware medium, and calculate the deviation weight of the operating environment deviation parameter to be written into the hardware medium; The operating environment impact coefficient of the hardware medium to be written is calculated according to the deviation weight of the operating environment deviation parameter of the hardware medium to be written and the operating environment deviation parameter of the hardware medium to be written.

3. A data storage method according to claim 2, characterized in that: The operating environment impact coefficient of the hardware medium to be written is specifically: Where S is the operating environment impact coefficient of the hardware medium to be written, E ij is the entropy value of the jth deviation parameter of the ith operating environment to be written into the hardware medium, d ij is the coefficient of variation of the jth deviation parameter of the ith operating environment to be written into the hardware medium, w ij is the deviation weight of the jth deviation parameter of the ith operating environment to be written into the hardware medium, V ij is the jth deviation parameter value of the i-th operating environment to be written into the hardware medium.

4. A data storage method according to claim 3, characterized in that: According to the historical storage data of the hardware medium to be written, the storage quality of the hardware medium to be written is analyzed, and the operating performance indicators of the hardware medium to be written are evaluated, including: Based on the historical storage data to be written into the hardware medium, the storage quality characteristic data in the historical storage data is filtered out according to the unit time, and the storage quality characteristic time series data to be written into the hardware medium is obtained; the storage quality characteristic data includes: read and write speed data, access time data; Normalizing the storage quality characteristic time series data to be written into the hardware medium to obtain normalized storage quality characteristic time series data to be written into the hardware medium; Based on AR autoregression, a performance change trend prediction model for the hardware media to be written is constructed; The normalized storage quality characteristic time series data of the hardware medium to be written is substituted into the performance change trend prediction model of the hardware medium to be written according to the time unit interval to evaluate the operation performance index of the hardware medium to be written.

5. A data storage method according to claim 4, characterized in that: The performance change trend prediction model to be written into the hardware medium is specifically: In the formula, G is the operating performance index of the hardware medium to be written, x i′t-k is the i′th storage quality characteristic data under the tkth unit time lag learning order to be written into the hardware medium, α0 intercept, α k is the autoregressive coefficient, ε t is the error term under t unit time lag learning order, and q is the total number of lag learning orders.

6. A data storage method according to claim 5, characterized in that: According to the historical storage data of the hardware medium to be written, the storage load trend of the hardware medium to be written is analyzed, and the dynamic load status indicators of the hardware medium to be written are evaluated, including: Based on the historical storage data to be written into the hardware medium, mark the load action event of the storage medium under each unit time, and obtain the characteristic data of the historical load action event of the storage medium; the historical load action event includes: read and write events, IOPS events; Constructing a historical load action event curve graph of the medium to be stored according to the historical load action event characteristic data of the storage medium; Based on the historical load action event curve graph of the medium to be stored, the peak and valley load trends of the historical load action events are marked to obtain the historical load trend characteristic parameters of the medium to be stored; Construct an autoregressive load assessment model for the hardware media to be written; Based on the autoregressive load assessment model of the hardware medium to be written, the historical load trend characteristic parameters of the medium to be stored in unit time are used as independent variable input, and the dynamic load state index of the medium to be stored in unit time in the future is predicted as the dependent variable output.

7. A data storage method according to claim 6, characterized in that: The autoregressive load evaluation model for the hardware medium to be written is specifically: Where F is the dynamic load state index of the medium to be stored per unit time in the future, is the characteristic parameter of the historical load trend at the tth unit time, β0 is the intercept, β1, β2, β l are regression coefficients, and ∈ is the error term.

Images