Data backup method and device, electronic equipment and nonvolatile storage medium

Abstract

The invention discloses a data backup method and device, electronic equipment and a nonvolatile storage medium. The method comprises the following steps: acquiring target equipment state data of resource pool equipment in a first time period; determining an equipment load parameter of the resource pool equipment in a second time period according to the target equipment state data in the first time period; determining a data variable quantity corresponding to the resource pool equipment, and determining a target data backup strategy according to the data variable quantity and the equipment load parameters; and backing up the data resources in the resource pool equipment according to the target data backup strategy. According to the data backup method and device, the technical problem that data backup is not timely or resources are wasted due to the fact that data backup depends on a fixed period or manual triggering and is lack of flexibility in the prior art is solved.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and more specifically, to a data backup method, apparatus, electronic device, and non-volatile storage medium. Background Technology

[0002] Resource pool data backup is a crucial task in cloud maintenance and is essential for improving the robustness of cloud network systems. Stable and efficient backup and recovery methods ensure that core data is not lost in the event of data corruption due to virus attacks, effectively restore backup data, and guarantee the stable operation of cloud network systems.

[0003] However, cloud resource backup in related technologies often relies on fixed periods or manual triggering, which lacks flexibility and does not take into account dynamic factors such as device load and data popularity. This can easily lead to resource waste or untimely backup. At the same time, the lack of proactive backup mechanisms results in delayed fault detection, making it impossible to predict device failures in advance and affecting business continuity.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This application provides a data backup method, apparatus, electronic device, and non-volatile storage medium to at least solve the technical problem of untimely data backup or resource waste caused by the lack of flexibility in related technologies due to data backup relying on fixed periods or manual triggering.

[0006] According to one aspect of the embodiments of this application, a data backup method is provided, comprising: acquiring target device status data of a resource pool device in a first time period; determining device load parameters of the resource pool device in a second time period based on the target device status data in the first time period, wherein the second time period is the time period immediately following the first time period, and the device load parameters are used to characterize the load status of the resource pool device in the second time period; determining the data change amount corresponding to the resource pool device, and determining a target data backup strategy based on the data change amount and the device load parameters, wherein the data change amount is used to characterize the amount of data resources that have changed in the resource pool device after the last data backup, and the target data backup strategy is used to indicate the backup method under different load states and data change amounts; and backing up the data resources in the resource pool device according to the target data backup strategy.

[0007] Optionally, obtaining the target device status data of the resource pool devices in the first time period includes: using a distributed probe to collect the raw device status data of the resource pool devices, wherein each data collection is spaced one time step apart, and the raw device status data includes at least one of the following: CPU utilization, memory utilization, disk input / output performance parameters, and network traffic; using a sliding window to generate a multi-dimensional time series matrix corresponding to the raw device status data in the first time period to obtain the target device status data.

[0008] Optionally, determining the device load parameters of the resource pool devices in the second time period based on the target device status data in the first time period includes: using a bidirectional long short-term memory network model to analyze the target device status data in the first time period to obtain hidden state features at multiple time steps, wherein the hidden state features are used to characterize the operating state characteristics and potential trends of the resource pool devices at each time step; determining the attention weights corresponding to different time steps, wherein the attention weights are used to characterize the importance of different time steps in predicting device load; and determining the device load parameters of the resource pool devices in the second time period based on the hidden states and attention weights.

[0009] Optionally, determining the target data backup strategy based on the amount of data change and device load parameters includes: monitoring file system events of the resource pool device and determining the amount of data change by analyzing the file system events, wherein file system events include file creation, deletion, and modification events; determining a first data backup strategy as the target data backup strategy when the device load parameter is less than a first load threshold and the amount of data change is greater than a first data volume threshold, wherein the first data backup strategy is used to perform a full backup; determining a second data backup strategy as the target data backup strategy when the device load parameter is less than a second load threshold but not less than the first load threshold and the amount of data change is greater than the second data volume threshold, wherein the first load threshold is less than the second load threshold, the second data volume threshold is less than the first data volume threshold, and the second data backup strategy is used to perform incremental backups; and determining a third data backup strategy as the target data backup strategy when the device load parameter is not less than the second load threshold, wherein the third data backup strategy is used to delay backup tasks.

[0010] Optionally, the method further includes: acquiring device log data of the resource pool devices; determining a first device health characteristic of the resource pool devices based on the device log data; acquiring device operating parameters of the resource pool devices and performing feature extraction on the device operating parameters using a time encoder to obtain a second device health characteristic, wherein the device operating parameters include at least one of the following: device temperature, voltage, and fan speed; determining a device health score corresponding to the resource pool devices based on the first and second device health characteristics, wherein the device health score is used to characterize the health status of the devices and the probability of failure; and determining a preventive backup strategy based on the device health score, wherein the preventive backup strategy is used to indicate the method of proactively backing up data resources before the resource pool devices fail.

[0011] Optionally, determining the first device health feature of the resource pool device based on the device log data includes: using a log analysis model to extract entity fields related to device faults from the device log data, and performing contextual association analysis on the entity fields to obtain the first device health feature. The log analysis model is trained based on a training dataset, which includes: multiple historical device log data and the device fault types corresponding to the historical device log data.

[0012] Optionally, a higher device health score indicates a better health status of the resource pool devices and a lower probability of failure. Based on the device health score, the preventive backup strategy includes: recording abnormal patterns of the resource pool devices and initiating the collection of diagnostic data when the device health score is less than a first score threshold but not less than a second score threshold, wherein the second score threshold is less than the first score threshold; triggering preventive backup and scheduling maintenance work orders when the device health score is less than the second score threshold but not less than a third score threshold, wherein the third score threshold is less than the second score threshold; and automatically isolating the resource pool devices and immediately triggering a full backup when the device health score is less than the third score threshold.

[0013] According to another aspect of the embodiments of this application, a data backup apparatus is also provided, comprising: a data acquisition module, configured to acquire target device status data of a resource pool device in a first time period; a load prediction module, configured to determine device load parameters of the resource pool device in a second time period based on the target device status data in the first time period, wherein the second time period is the time period immediately following the first time period, and the device load parameters are used to characterize the load state of the resource pool device in the second time period; a strategy determination module, configured to determine the data change amount corresponding to the resource pool device, and determine a target data backup strategy based on the data change amount and the device load parameters, wherein the data change amount is used to characterize the amount of data resources that have changed in the resource pool device after the last data backup, and the target data backup strategy is used to indicate the backup method under different load states and data change amounts; and a backup execution module, configured to back up the data resources in the resource pool device according to the target data backup strategy.

[0014] According to another aspect of the embodiments of this application, an electronic device is also provided, including: a memory and a processor, the processor being configured to run a program stored in the memory, wherein the program executes a data backup method during runtime.

[0015] According to another aspect of the embodiments of this application, a non-volatile storage medium is also provided, the non-volatile storage medium including a stored computer program, wherein the device where the non-volatile storage medium is located executes a data backup method by running the computer program.

[0016] According to another aspect of the embodiments of this application, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps of a data backup method.

[0017] In this embodiment, the following methods are employed: acquiring target device status data of the resource pool device in a first time period; determining device load parameters of the resource pool device in a second time period based on the target device status data in the first time period, wherein the second time period is the time period immediately following the first time period, and the device load parameters are used to characterize the load status of the resource pool device in the second time period; determining the data change amount corresponding to the resource pool device, and determining a target data backup strategy based on the data change amount and the device load parameters, wherein the data change amount is used to characterize the amount of data resources that have changed in the resource pool device after the last data backup, and the target data backup strategy is used to indicate the backup method under different load states and data change amounts; and backing up the data resources in the resource pool device according to the target data backup strategy. By dynamically adjusting the backup strategy through deep learning technology, the flexibility and efficiency of resource pool data backup are improved, thereby solving the technical problem of untimely data backup or resource waste caused by the lack of flexibility in related technologies due to data backup relying on fixed periods or manual triggering. Attached Figure Description

[0018] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0019] Figure 1 This is a hardware structure block diagram of a computer terminal (or electronic device) for implementing a data backup method according to an embodiment of this application;

[0020] Figure 2 This is a schematic diagram of a data backup method flow provided according to an embodiment of this application;

[0021] Figure 3 This is a schematic diagram of a method for dynamic backup of cloud resource pool data based on deep learning, according to an embodiment of this application.

[0022] Figure 4 This is a schematic diagram of a data backup device provided according to an embodiment of this application. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0024] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0025] To facilitate a better understanding of the embodiments of this application by those skilled in the art, some technical terms or nouns involved in the embodiments of this application are explained as follows:

[0026] Data backup is a core measure to ensure the security and business continuity of the resource pool. It prevents data loss due to hardware failure, human error, or network attacks, ensures rapid recovery of critical business operations, reduces downtime losses, and is an essential guarantee for the stable operation of the resource pool.

[0027] Long Short-Term Memory (LSTM) network: This is a special type of Recurrent Neural Network (RNN) that addresses the long-term dependency problem of traditional RNNs by introducing a gating mechanism. Its core structure includes a forget gate, an input gate, and an output gate, and it is widely used in fields such as natural language processing and time series prediction.

[0028] Temporal attention mechanism: This is a technique in deep learning used to process temporal data. It allows the model to assign different importance or attention to information at different time steps when processing sequential data, enabling the model to more effectively learn the dependencies and importance of different parts of the sequence and adjust its focus as needed.

[0029] LogBERT focuses on log anomaly detection using the advanced pre-trained BERT model. At its core is an end-to-end framework that extracts structured information from raw log data, generates log sequences, and performs modeling and anomaly detection using the BERT model.

[0030] In related technologies, traditional data backup relies on fixed periods or manual triggering, which lacks flexibility and is prone to untimely backup or waste of resources; moreover, traditional data backup methods lack initiative and cannot perform data backup in a timely manner when equipment fails.

[0031] To address the aforementioned issues, this application provides a solution that utilizes deep learning algorithms to formulate a dynamic backup strategy. It analyzes four-dimensional characteristics of device CPU, memory, disk I / O, and network traffic using LSTM and time attention mechanisms, and dynamically adjusts the backup frequency and window based on data change indicators, effectively improving backup flexibility and effectiveness. Simultaneously, a health scoring mechanism is constructed, combining device temperature, voltage, fan speed time-series prediction indicators with log analysis models to comprehensively assess device status and effectively predict device failures. Furthermore, this solution includes a three-layer circuit breaker mechanism, which can perform diagnostic collection, data backup, and device isolation based on the health scoring mechanism, ensuring preventative backups are performed before device failure. Detailed explanations follow.

[0032] According to an embodiment of this application, a method embodiment for data backup is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0033] The methods and embodiments provided in this application can be executed on mobile terminals, computer terminals, or similar computing devices. Figure 1 A hardware block diagram of a computer terminal (or electronic device) for implementing a data backup method is shown. Figure 1 As shown, the computer terminal 10 (or electronic device) may include one or more processors 102 (shown as 102a, 102b, ..., 102n in the figure) 102 (processor 102 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0034] It should be noted that the aforementioned one or more processors 102 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10 (or electronic device). As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0035] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the data backup method in this embodiment. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby implementing the above-mentioned data backup method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0036] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0037] The display may be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 10 (or electronic device).

[0038] Under the above operating environment, this application provides a data backup method. Figure 2 This is a schematic diagram of a data backup method flow provided according to an embodiment of this application, such as... Figure 2 As shown, the method includes the following steps:

[0039] Step S202: Obtain the target device status data of the resource pool device in the first time period;

[0040] Step S204: Based on the target device status data of the first time period, determine the device load parameters of the resource pool device in the second time period. The second time period is the time period immediately following the first time period. The device load parameters are used to characterize the load status of the resource pool device in the second time period.

[0041] Step S206: Determine the amount of data change corresponding to the resource pool device, and determine the target data backup strategy based on the amount of data change and the device load parameters. The amount of data change is used to characterize the amount of data resources that have changed in the resource pool device after the last data backup, and the target data backup strategy is used to indicate the backup method under different load states and data change conditions.

[0042] Step S208: Back up the data resources in the resource pool device according to the target data backup strategy.

[0043] By using the above steps and dynamically adjusting the backup strategy through deep learning technology, the flexibility and efficiency of resource pool data backup are improved. This solves the technical problem of untimely data backup or resource waste caused by the lack of flexibility in related technologies that rely on fixed periods or manual triggering for data backup.

[0044] The data backup method in steps S202 to S208 of the embodiments of this application will be further described below.

[0045] Figure 3 This is a schematic diagram illustrating a method for dynamic backup of cloud resource pool data based on deep learning, according to an embodiment of this application. Figure 3 As shown, this embodiment of the application comprehensively analyzes historical data such as device CPU, memory, disk I / O, and network traffic through an LSTM algorithm combined with an attention mechanism. By combining this with the backup data's popularity, the backup frequency and backup window are dynamically adjusted. This ensures real-time backup while avoiding the backup process consuming device resources and causing disruption to normal business operations. Furthermore, this embodiment of the application constructs a health score model based on device monitoring indicators and log information, and sets up a circuit breaker mechanism to proactively back up data before failure, ensuring the timeliness and integrity of the backup data. A detailed description follows.

[0046] First, the raw sequence is collected, that is, the target device status data of the resource pool devices in the first time period is obtained. The specific steps are as follows.

[0047] In some embodiments of this application, obtaining target device status data of resource pool devices in a first time period includes the following steps: using a distributed probe to collect raw device status data of resource pool devices, wherein each data collection is spaced one time step apart, and the raw device status data includes at least one of the following: CPU utilization, memory utilization, disk input / output performance parameters, and network traffic; using a sliding window to generate a multi-dimensional time series matrix corresponding to the raw device status data in the first time period to obtain the target device status data.

[0048] Specifically, device status data such as CPU utilization, memory utilization, disk I / O, and network traffic metrics can be collected in real time using distributed probes. Furthermore, time-series segments can be generated using a sliding window, with the data format being a multi-dimensional time series matrix (i.e., the aforementioned target device status data).

[0049] For example, in this embodiment, the device's CPU utilization, memory utilization, disk I / O, and network traffic can be selected as 4-dimensional features to construct time series data in 5-second units (i.e., sampling frequency of 5 seconds / time). Data from 12 consecutive time units are used as the input feature matrix, with the index set to timestamps, and the data is normalized.

[0050] Subsequently, a bidirectional LSTM network can be used to preprocess the target device status data, while modeling the historical evolution patterns and potential future trends of device indicators, as detailed below.

[0051] In some embodiments of this application, determining the device load parameters of the resource pool device in the second time period based on the target device state data in the first time period includes the following steps: using a bidirectional long short-term memory network model to analyze the target device state data in the first time period to obtain hidden state features at multiple time steps, wherein the hidden state features are used to characterize the operating state characteristics and potential trends of the resource pool device at each time step; determining the attention weights corresponding to different time steps, wherein the attention weights are used to characterize the importance of different time steps in predicting device load; and determining the device load parameters of the resource pool device in the second time period based on the hidden states and attention weights.

[0052] In this embodiment, time series data can be processed by a bidirectional LSTM network. The forward and backward LSTMs capture historical and future dependencies respectively, and gradient pruning is used to prevent explosion. The input shape is a matrix with 4 features and a sequence length of 12, and the output is the hidden state at each time step (i.e., the hidden state features mentioned above).

[0053] For example, a 3-layer LSTM network model can be constructed, with 128 hidden units per layer (hidden layer dimension 128), stacked three times to enhance the model's expressive power. Dropout layers are added between each layer, with a dropout rate of 0.3 to prevent overfitting. During model training, the AdamW optimizer is used, with a learning rate of 0.0001 and weight decay of 0.01 to improve training stability and convergence speed. The learning rate scheduler is CosineAnnealingLR with a period of 50, enabling dynamic adjustment of the learning rate. Finally, the model with the optimal parameters and minimum validation loss during training is saved.

[0054] Meanwhile, since equipment load typically changes periodically, this application also introduces a temporal attention mechanism to enable the model to focus on peak business periods. Attention weights are calculated using a learnable parameter matrix, and a weighted sum is obtained to obtain the context vector. Multi-head attention is supported to capture different temporal patterns, and visualized weights assist in fault tracing.

[0055] Specifically, the relevant formulas for determining the equipment load parameters of the resource pool equipment in the second time period are as follows:

[0056]

[0057] Where, h′ t This represents the LSTM hidden state (including forward and backward passes) used to capture temporal dependencies. In the temporal attention mechanism, the weights for each time step are calculated using the learnable parameter matrix W (64×256 dimensions) and the vector v (64 dimensions), and the attention score is converted into a weight α using the softmax function. t This allows the model to focus on periodic business peaks. c represents the context vector (256 dimensions) output by the attention layer, and W... o This represents the output weight matrix, b o For bias terms, This represents the load at time k in the future (i.e., the equipment load parameters for the second time period).

[0058] For example, the prediction target can be the load status in the next 5 minutes (the second time period). The output is normalized to a value in the range [0,1] by Sigmoid. The two are then weighted and fused, and the final value is used as the basis for decision-making on backup.

[0059] After obtaining the device load parameters for the second time period, to avoid resource waste caused by single-threshold decisions, this embodiment of the application also comprehensively considers dynamic factors such as data change volume. Specifically, it can calculate the change volume in real time by monitoring file system events such as creation, modification, and deletion. It employs a combination of database transaction log analysis to filter out non-critical changes such as temporary files, using the data change volume index as the basis for backup decisions.

[0060] Then, a multi-domain dynamic backup strategy can be formulated based on the predicted load and the amount of changes in equipment data, transforming continuous predicted values ​​into executable backup strategies and significantly saving backup resources. Details are as follows.

[0061] In some embodiments of this application, determining a target data backup strategy based on data change amount and device load parameters includes the following steps: monitoring file system events of the resource pool device and determining the data change amount by analyzing the file system events, wherein the file system events include: file creation, deletion, and modification events; determining a first data backup strategy as the target data backup strategy when the device load parameter is less than a first load threshold and the data change amount is greater than a first data volume threshold, wherein the first data backup strategy is used to perform a full backup; determining a second data backup strategy as the target data backup strategy when the device load parameter is less than a second load threshold but not less than the first load threshold and the data change amount is greater than the second data volume threshold, wherein the first load threshold is less than the second load threshold, the second data volume threshold is less than the first data volume threshold, and the second data backup strategy is used to perform incremental backup; determining a third data backup strategy as the target data backup strategy when the device load parameter is not less than the second load threshold, wherein the third data backup strategy is used to delay backup tasks.

[0062] Specifically, the output layer integrates LSTM features with data change indicators and uses dynamic threshold decisions. In this embodiment, when the predicted load is <0.3 (first load threshold) and the data change amount is >100MB (first data volume threshold), a full backup is triggered immediately; when 0.3 ≤ load <0.6 (second load threshold) and the data change amount is >50MB (second data volume threshold), an incremental backup is performed; when the load is ≥0.6, the delayed backup decision threshold is dynamically adjusted according to the data importance. The multi-domain dynamic backup strategy formula can be expressed as follows:

[0063]

[0064] While the aforementioned algorithms and strategies address the issue of daily dynamic backup of resource pool devices, they still suffer from delayed device fault detection, making it impossible to predict device failures in advance and resulting in the loss of the latest backup data when a failure occurs. This necessitates a proactive backup mechanism. Therefore, this application also combines a time-series prediction model with a device log analysis model to identify potentially faulty devices in advance and trigger proactive backups before device failure, ensuring the timeliness and integrity of backup data.

[0065] In this embodiment, a multimodal fusion architecture can be adopted, which collects device temperature, voltage, fan speed and other indicators and system logs in real time through embedded probes and log agents, and deploys an LSTM time series model and a BERT log analysis model. The specific steps are as follows.

[0066] In some embodiments of this application, the method further includes the following steps: acquiring device log data of the resource pool device; determining a first device health characteristic of the resource pool device based on the device log data; acquiring device operating parameters of the resource pool device and performing feature extraction on the device operating parameters using a time encoder to obtain a second device health characteristic, wherein the device operating parameters include at least one of the following: device temperature, voltage, and fan speed; determining a device health score corresponding to the resource pool device based on the first and second device health characteristics, wherein the device health score is used to characterize the health status of the device and the probability of failure; and determining a preventive backup strategy based on the device health score, wherein the preventive backup strategy is used to indicate the method of proactively backing up data resources before the resource pool device fails.

[0067] In this embodiment, a log analysis model can be used to analyze and detect abnormal device logs to determine the first device health characteristics of the resource pool devices, as detailed below.

[0068] In some embodiments of this application, determining the first device health feature of a resource pool device based on device log data includes: using a log analysis model to extract entity fields related to device faults from the device log data, and performing contextual association analysis on the entity fields to obtain the first device health feature. The log analysis model is trained based on a training dataset, which includes: multiple historical device log data and the device fault types corresponding to the historical device log data.

[0069] Specifically, the construction process of the log analysis model (LogBERT model) is as follows: collect and preprocess device log data, continue training using the preprocessed device logs on the basis of general BERT, build a log-fault type mapping dataset to fine-tune the classification head, and then deploy the fine-tuned LogBERT model as a gRPC microservice.

[0070] The log analysis model is used to detect abnormal device logs, collect and preprocess device log data, collect device system logs in real time through the Filebeat log agent, use an NLP model to extract structured fields and label fault-related entities, and in the model training phase, on the basis of the general BERT, continue to train using the preprocessed device logs, build a log-fault type mapping dataset to fine-tune the classification head, and deploy the fine-tuned LogBERT model as a gRPC microservice to support concurrent calls.

[0071] This application embodiment collects real-time device operating parameters such as device temperature, voltage, and fan speed, as well as system logs, by deploying embedded probes and log agents. It can calculate device health scores based on multimodal fusion features to assess device status and predict faults. Specifically, based on the first and second device health features, the device health score corresponding to the resource pool device is determined, as shown in the following formula:

[0072]

[0073] HealthScore is the device health score, used to measure device status to trigger backup actions, where n represents the number of modalities, and w i This represents the modal weighting coefficients; the sum of the weighting coefficients should be 1, f. i The modality feature extraction functions are defined as follows: f1 is the transformer time encoder, f2 is the BERT log analysis model, the features of each model are normalized by the Norm function, and finally the superimposed values ​​are standardized by the sigma function to compress the health score to the (0, 1) interval.

[0074] After obtaining the device health score corresponding to the resource pool device, a preventive backup strategy can be further determined based on the device health score. The specific steps are as follows.

[0075] In some embodiments of this application, a higher device health score indicates a better health status of the resource pool device and a lower probability of failure. Determining a preventative backup strategy based on the device health score includes: recording abnormal patterns of the resource pool device and initiating diagnostic data collection when the device health score is less than a first score threshold but not less than a second score threshold, wherein the second score threshold is less than the first score threshold; triggering preventative backup and scheduling maintenance work orders when the device health score is less than the second score threshold but not less than a third score threshold, wherein the third score threshold is less than the second score threshold; and automatically isolating the resource pool device and immediately triggering a full backup when the device health score is less than the third score threshold.

[0076] Specifically, the device status is comprehensively judged through feature fusion network, and a three-level response mechanism is triggered based on real-time health scores to ensure preventive backups are initiated before device failure and to ensure the integrity of backup data. For example, when 0.6 (second threshold score) ≤ score < 0.8 (first threshold score), it is defined as a warning level, abnormal patterns are recorded, and diagnostic data collection is initiated. When 0.4 (third threshold score) ≤ score < 0.6, it is defined as a danger level, triggering preventive backups and scheduling maintenance work orders. When score < 0.4, it is defined as a critical level, and the device is automatically isolated. In this embodiment, data backup tasks can be performed using a Kubernetes Operator.

[0077] This application embodiment sets up a three-layer circuit breaker mechanism, which performs actions such as diagnostic collection, data backup, and device isolation based on device health scores. This ensures preventive backup before device failure and guarantees the integrity of the backup data.

[0078] This application's solution dynamically adjusts backup strategies using deep learning technology, significantly improving the flexibility and efficiency of resource pool data backup. Specifically, by monitoring device status in real time, predicting load conditions, and dynamically deciding on backup methods based on data changes, it avoids the blindness and resource waste of traditional fixed-period backups. Simultaneously, by constructing a health scoring mechanism, combined with device operating parameters and log analysis, it predicts device failures in advance, triggering preventative backups, effectively ensuring data integrity and business continuity, and significantly reducing the risk of data loss due to device failures. Furthermore, the introduction of a three-layer circuit breaker mechanism enables proactive backup before device failure and rapid isolation after failure, further enhancing the system's robustness and security.

[0079] According to an embodiment of this application, an embodiment of a data backup device is also provided. Figure 4 This is a schematic diagram of a data backup device provided according to an embodiment of this application. Figure 4 As shown, the device includes:

[0080] Data acquisition module 40 is used to acquire target device status data of resource pool devices in the first time period;

[0081] The load prediction module 42 is used to determine the device load parameters of the resource pool device in the second time period based on the target device status data in the first time period. The second time period is the time period immediately following the first time period, and the device load parameters are used to characterize the load status of the resource pool device in the second time period.

[0082] The strategy determination module 44 is used to determine the amount of data change corresponding to the resource pool device, and to determine the target data backup strategy based on the amount of data change and the device load parameters. The amount of data change is used to characterize the amount of data resources that have changed in the resource pool device after the last data backup, and the target data backup strategy is used to indicate the backup method under different load states and data change conditions.

[0083] Backup execution module 46 is used to back up data resources in the resource pool device according to the target data backup strategy.

[0084] Optionally, obtaining the target device status data of the resource pool devices in the first time period includes: using a distributed probe to collect the raw device status data of the resource pool devices, wherein each data collection is spaced one time step apart, and the raw device status data includes at least one of the following: CPU utilization, memory utilization, disk input / output performance parameters, and network traffic; using a sliding window to generate a multi-dimensional time series matrix corresponding to the raw device status data in the first time period to obtain the target device status data.

[0085] Optionally, determining the device load parameters of the resource pool devices in the second time period based on the target device status data in the first time period includes: using a bidirectional long short-term memory network model to analyze the target device status data in the first time period to obtain hidden state features at multiple time steps, wherein the hidden state features are used to characterize the operating state characteristics and potential trends of the resource pool devices at each time step; determining the attention weights corresponding to different time steps, wherein the attention weights are used to characterize the importance of different time steps in predicting device load; and determining the device load parameters of the resource pool devices in the second time period based on the hidden states and attention weights.

[0086] Optionally, determining the target data backup strategy based on the amount of data change and device load parameters includes: monitoring file system events of the resource pool device and determining the amount of data change by analyzing the file system events, wherein file system events include file creation, deletion, and modification events; determining a first data backup strategy as the target data backup strategy when the device load parameter is less than a first load threshold and the amount of data change is greater than a first data volume threshold, wherein the first data backup strategy is used to perform a full backup; determining a second data backup strategy as the target data backup strategy when the device load parameter is less than a second load threshold but not less than the first load threshold and the amount of data change is greater than the second data volume threshold, wherein the first load threshold is less than the second load threshold, the second data volume threshold is less than the first data volume threshold, and the second data backup strategy is used to perform incremental backups; and determining a third data backup strategy as the target data backup strategy when the device load parameter is not less than the second load threshold, wherein the third data backup strategy is used to delay backup tasks.

[0087] Optionally, the data backup device is further configured to: acquire device log data of the resource pool device, and determine a first device health characteristic of the resource pool device based on the device log data; acquire device operating parameters of the resource pool device, and perform feature extraction on the device operating parameters using a time encoder to obtain a second device health characteristic, wherein the device operating parameters include at least one of the following: device temperature, voltage, and fan speed; determine a device health score corresponding to the resource pool device based on the first device health characteristic and the second device health characteristic, wherein the device health score is used to characterize the health status of the device and the probability of failure; and determine a preventive backup strategy based on the device health score, wherein the preventive backup strategy is used to indicate the method of proactively backing up data resources before the resource pool device fails.

[0088] Optionally, determining the first device health feature of the resource pool device based on the device log data includes: using a log analysis model to extract entity fields related to device faults from the device log data, and performing contextual association analysis on the entity fields to obtain the first device health feature. The log analysis model is trained based on a training dataset, which includes: multiple historical device log data and the device fault types corresponding to the historical device log data.

[0089] Optionally, a higher device health score indicates a better health status of the resource pool devices and a lower probability of failure. Based on the device health score, the preventive backup strategy includes: recording abnormal patterns of the resource pool devices and initiating the collection of diagnostic data when the device health score is less than a first score threshold but not less than a second score threshold, wherein the second score threshold is less than the first score threshold; triggering preventive backup and scheduling maintenance work orders when the device health score is less than the second score threshold but not less than a third score threshold, wherein the third score threshold is less than the second score threshold; and automatically isolating the resource pool devices and immediately triggering a full backup when the device health score is less than the third score threshold.

[0090] It should be noted that each module in the above data backup device can be a program module (e.g., a set of program instructions to implement a specific function) or a hardware module. For the latter, it can take the following forms, but is not limited to them: each of the above modules is represented by a processor, or the functions of each of the above modules are implemented by a processor.

[0091] It should be noted that the data backup device provided in this embodiment can be used to perform... Figure 2 The data backup method shown above is also applicable to the embodiments of this application, and will not be repeated here.

[0092] This application embodiment also provides a non-volatile storage medium, which includes a stored computer program. The device containing the non-volatile storage medium executes the following data backup method by running the computer program: acquiring target device status data of a resource pool device in a first time period; determining device load parameters of the resource pool device in a second time period based on the target device status data in the first time period, wherein the second time period is the time period immediately following the first time period, and the device load parameters characterize the load state of the resource pool device in the second time period; determining the data change amount corresponding to the resource pool device, and determining a target data backup strategy based on the data change amount and the device load parameters, wherein the data change amount characterizes the amount of data resources that have changed in the resource pool device since the last data backup, and the target data backup strategy indicates the backup method under different load states and data change amounts; and backing up the data resources in the resource pool device according to the target data backup strategy.

[0093] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the data backup method described in various embodiments of this application: acquiring target device status data of a resource pool device in a first time period; determining device load parameters of the resource pool device in a second time period based on the target device status data in the first time period, wherein the second time period is the time period immediately following the first time period, and the device load parameters are used to characterize the load status of the resource pool device in the second time period; determining the data change amount corresponding to the resource pool device, and determining a target data backup strategy based on the data change amount and the device load parameters, wherein the data change amount is used to characterize the amount of data resources that have changed in the resource pool device after the last data backup, and the target data backup strategy is used to indicate the backup method under different load states and data change amounts; and backing up the data resources in the resource pool device according to the target data backup strategy.

[0094] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0095] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0096] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0097] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0098] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0099] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0100] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A data backup method, characterized in that, include: Obtain the target device status data of the resource pool devices in the first time period; Based on the target device status data in the first time period, the device load parameters of the resource pool device in the second time period are determined, wherein the second time period is the time period immediately following the first time period, and the device load parameters are used to characterize the load status of the resource pool device in the second time period. The amount of data change corresponding to the resource pool device is determined, and a target data backup strategy is determined based on the amount of data change and the device load parameters. The amount of data change is used to characterize the amount of data resources that have changed in the resource pool device after the last data backup, and the target data backup strategy is used to indicate the backup method under different load states and data change conditions. The data resources in the resource pool device are backed up according to the target data backup strategy.

2. The data backup method according to claim 1, characterized in that, The target device status data obtained from the resource pool devices in the first time period includes: Distributed probes are used to collect raw device status data of the resource pool devices. Each data collection is spaced one time step apart. The raw device status data includes at least one of the following: CPU utilization, memory utilization, disk input / output performance parameters, and network traffic. Using a sliding window, a multi-dimensional time series matrix corresponding to the original device status data within the first time period is generated to obtain the target device status data.

3. The data backup method according to claim 2, characterized in that, Based on the target device status data in the first time period, the device load parameters of the resource pool device in the second time period are determined as follows: A bidirectional long short-term memory network model is used to analyze the state data of the target device in the first time period to obtain hidden state features at multiple time steps. The hidden state features are used to characterize the operating state features and potential trends of the resource pool device at the time step. Determine the attention weights corresponding to different time steps, wherein the attention weights are used to characterize the importance of different time steps in predicting device load; Based on the hidden state and the attention weight, the device load parameters of the resource pool device in the second time period are determined.

4. The data backup method according to claim 1, characterized in that, Based on the data change and the device load parameters, the target data backup strategy is determined as follows: Monitor file system events of the resource pool device, and determine the amount of data change by analyzing the file system events, wherein the file system events include: file creation, deletion and modification events; When the device load parameter is less than the first load threshold and the data change amount is greater than the first data amount threshold, the first data backup strategy is determined as the target data backup strategy, wherein the first data backup strategy is used to perform a full backup. When the device load parameter is less than the second load threshold but not less than the first load threshold, and the data change amount is greater than the second data volume threshold, the second data backup strategy is determined as the target data backup strategy, wherein the first load threshold is less than the second load threshold, the second data volume threshold is less than the first data volume threshold, and the second data backup strategy is used to perform incremental backup. If the device load parameter is not less than the second load threshold, a third data backup strategy is determined as the target data backup strategy, wherein the third data backup strategy is used to delay backup tasks.

5. The data backup method according to claim 1, characterized in that, The method further includes: Obtain the device log data of the resource pool device, and determine the first device health characteristic of the resource pool device based on the device log data; The device operating parameters of the resource pool device are obtained, and the device operating parameters are extracted using a time encoder to obtain a second device health feature. The device operating parameters include at least one of the following: device temperature, voltage, and fan speed. Based on the first device health characteristics and the second device health characteristics, a device health score corresponding to the resource pool device is determined, wherein the device health score is used to characterize the health status of the device and the probability of failure. Based on the device health score, a preventive backup strategy is determined, wherein the preventive backup strategy is used to indicate the method of proactively backing up data resources before the resource pool device fails.

6. The data backup method according to claim 5, characterized in that, Based on the device log data, the first device health characteristics of the resource pool devices are determined to include: A log analysis model is used to extract entity fields related to device faults from the device log data, and context association analysis is performed on the entity fields to obtain the first device health features. The log analysis model is trained based on a training dataset, which includes: multiple historical device log data and the device fault types corresponding to the historical device log data.

7. The data backup method according to claim 5, characterized in that, The higher the device health score, the better the health status of the resource pool device and the lower the probability of failure. Based on the device health score, the preventative backup strategy includes: If the device health score is less than a first score threshold but not less than a second score threshold, the abnormal mode of the resource pool device is recorded, and the collection of diagnostic data is initiated, wherein the second score threshold is less than the first score threshold. If the device health score is less than the second score threshold but not less than the third score threshold, a preventative backup is triggered and a maintenance work order is scheduled, wherein the third score threshold is less than the second score threshold. If the device health score is less than the third score threshold, the resource pool device is automatically isolated and a full backup is immediately triggered.

8. A data backup device, characterized in that, include: The data acquisition module is used to acquire the target device status data of the resource pool devices in the first time period; The load prediction module is used to determine the device load parameters of the resource pool device in a second time period based on the target device status data in the first time period, wherein the second time period is the time period immediately following the first time period, and the device load parameters are used to characterize the load status of the resource pool device in the second time period. The strategy determination module is used to determine the amount of data change corresponding to the resource pool device, and determine the target data backup strategy based on the amount of data change and the device load parameters. The amount of data change is used to characterize the amount of data resources that have changed in the resource pool device after the last data backup, and the target data backup strategy is used to indicate the backup method under different load states and data change conditions. The backup execution module is used to back up the data resources in the resource pool device according to the target data backup strategy.

9. An electronic device, characterized in that, include: A memory and a processor, the processor being configured to run a program stored in the memory, wherein the program, when running, performs the data backup method according to any one of claims 1 to 7.

10. A non-volatile storage medium, characterized in that, The non-volatile storage medium includes a stored computer program, wherein the device containing the non-volatile storage medium executes the data backup method according to any one of claims 1 to 7 by running the computer program.

11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the data backup method according to any one of claims 1 to 7.

Images