Life cycle prediction method and device of cloud resources and computer equipment
Through multi-dimensional data acquisition and feature modeling, resource utilization efficiency characteristics and impact degree characteristics are formed, and the problem of inaccurate prediction of traditional single-dimensional prediction is solved, and more accurate cloud resource life cycle prediction is achieved.
Patent Information
- Application Number
- CN202510184658.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-19
- Publication Date
- 2025-06-06
AI Technical Summary
Traditional technology only relies on a single data dimension to predict the life cycle of cloud resource, resulting in inaccurate prediction results and it is difficult to meet the actual needs of cloud service providers in resource management, maintenance and cost control.
By obtaining the resource usage, resource performance parameters and resource failure records of the target cloud resources, cross-combination and modeling of features, forming resource utilization efficiency characteristics and impact degree characteristics, and input them into the prediction model for life cycle prediction.
It improves the accuracy of cloud resource life cycle prediction, can more comprehensively cover various factors affecting the life cycle of cloud resource, and provides richer information to support accurate prediction.
Smart Images
Figure CN120104326A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of cloud computing technology, and in particular to a method, device and computer equipment for predicting the life cycle of cloud resources. Background Art
[0002] In a cloud computing environment, efficient management and reasonable allocation of cloud resources are crucial. Accurately predicting the life cycle of cloud resources can help cloud service providers plan resources, arrange maintenance, and control costs in advance.
[0003] In traditional technologies, only a single data dimension is often relied upon to predict the life cycle of cloud resources.
[0004] However, this single-dimensional analysis method may result in inaccurate prediction results and make it difficult to meet the actual needs of cloud service providers in terms of resource management, maintenance and cost control. Summary of the invention
[0005] Based on this, it is necessary to provide a cloud resource life cycle prediction method, device and computer equipment that can improve the accuracy of cloud resource life cycle prediction in response to the above technical problems.
[0006] In a first aspect, the present application provides a method for predicting a life cycle of cloud resources, the method comprising:
[0007] Obtain resource usage, resource performance parameters, and resource failure records of target cloud resources;
[0008] Perform cross-combination of resource usage and resource performance parameters to obtain resource utilization efficiency characteristics;
[0009] Model resource performance parameters and resource failure records to obtain impact characteristics that represent the impact of failures on resource usage;
[0010] The resource utilization efficiency characteristics and impact degree characteristics are input into the prediction model to obtain the life cycle of the target cloud resources.
[0011] In one embodiment, resource usage and resource performance parameters are cross-combined to obtain resource utilization efficiency features, including:
[0012] Using resource usage and resource performance parameters as a first data set;
[0013] The minimum support and minimum confidence thresholds configured for the initial association rule mining algorithm are obtained to obtain the configured association rule mining algorithm;
[0014] Based on the configured association rule mining algorithm, scan the first data set to obtain frequent item sets that meet the minimum support;
[0015] Based on frequent item sets, generate association rules that meet the minimum confidence level;
[0016] Based on association rules, the resource usage and resource performance parameters are cross-combined to obtain resource utilization efficiency characteristics.
[0017] In one embodiment, based on the association rules, resource usage and resource performance parameters are cross-combined to obtain resource utilization efficiency features, including:
[0018] Based on the association rules, the resource usage data corresponding to the rule antecedent and the resource performance parameter data corresponding to the rule consequent are multiplied to obtain the resource utilization efficiency characteristics.
[0019] In one embodiment, resource performance parameters and resource fault records are modeled to obtain impact characteristics indicating the impact of the fault on resource usage, including:
[0020] Through the time series analysis algorithm, the resource failure records and resource usage are modeled to obtain the autoregressive integrated moving average model;
[0021] Determine the model parameters of the autoregressive integrated moving average model through the autocorrelation function and the partial autocorrelation function;
[0022] According to the model parameters of the autoregressive integrated moving average model, an impact degree feature representing the impact degree of the fault on resource usage is constructed.
[0023] In one embodiment, determining the model parameters of the autoregressive integrated moving average model through the autocorrelation function and the partial autocorrelation function includes:
[0024] Determine an autocorrelation function graph according to an autocorrelation function value between a fault time series corresponding to a resource fault record and a usage time series data corresponding to a resource usage;
[0025] According to the tailing or truncation in the autocorrelation function graph, the moving average order parameter of the autoregressive integrated moving average model is determined;
[0026] Determine a partial correlation function graph according to a partial correlation function value between a fault time series corresponding to a resource fault record and a usage time series data corresponding to a resource usage;
[0027] According to the tailing or truncation in the partial correlation function graph, the autoregressive order parameter of the autoregressive integrated moving average model is determined;
[0028] Perform differential processing on the fault time series data corresponding to the resource fault record and the usage time series data corresponding to the resource usage to obtain a differential sequence;
[0029] The difference order parameter of the autoregressive integrated moving average model is determined based on the stationarity of the difference series.
[0030] In one embodiment, based on the model parameters of the autoregressive integrated moving average model, an impact degree feature representing the impact degree of the fault on resource usage is constructed, including:
[0031] The resource usage before and after the failure is predicted using the autoregressive integrated moving average model with determined model parameters to obtain the predicted usage;
[0032] According to the deviation between the predicted usage and the resource usage, the deviation dimension feature is obtained;
[0033] Determine the residual series characteristics of the autoregressive integrated moving average model during the fault period;
[0034] The deviation dimension features and residual sequence features are used as impact degree features to represent the impact degree of the fault on resource usage.
[0035] In a second aspect, the present application further provides a cloud resource life cycle prediction device, the device comprising:
[0036] An acquisition module is used to obtain resource usage, resource performance parameters and resource fault records of target cloud resources;
[0037] The first feature generation module is used to perform feature cross-combination on resource usage and resource performance parameters to obtain resource utilization efficiency features;
[0038] The second feature generation module is used to model resource performance parameters and resource fault records to obtain impact degree features representing the impact degree of the fault on resource usage;
[0039] The life cycle prediction module is used to obtain the life cycle of the target cloud resources based on the resource utilization efficiency characteristics and impact degree characteristics input into the prediction model.
[0040] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:
[0041] Obtain resource usage, resource performance parameters, and resource failure records of target cloud resources;
[0042] Perform cross-combination of resource usage and resource performance parameters to obtain resource utilization efficiency characteristics;
[0043] Model resource performance parameters and resource failure records to obtain impact characteristics that represent the impact of failures on resource usage;
[0044] The resource utilization efficiency characteristics and impact degree characteristics are input into the prediction model to obtain the life cycle of the target cloud resources.
[0045] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the following steps are implemented:
[0046] Obtain resource usage, resource performance parameters, and resource failure records of target cloud resources;
[0047] Perform cross-combination of resource usage and resource performance parameters to obtain resource utilization efficiency characteristics;
[0048] Model resource performance parameters and resource failure records to obtain impact characteristics that represent the impact of failures on resource usage;
[0049] The resource utilization efficiency characteristics and impact degree characteristics are input into the prediction model to obtain the life cycle of the target cloud resources.
[0050] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:
[0051] Obtain resource usage, resource performance parameters, and resource failure records of target cloud resources;
[0052] Perform cross-combination of resource usage and resource performance parameters to obtain resource utilization efficiency characteristics;
[0053] Model resource performance parameters and resource failure records to obtain impact characteristics that represent the impact of failures on resource usage;
[0054] The resource utilization efficiency characteristics and impact degree characteristics are input into the prediction model to obtain the life cycle of the target cloud resources.
[0055] The above-mentioned cloud resource life cycle prediction method, device and computer equipment, this application can comprehensively cover various factors that affect the life cycle of cloud resources by obtaining data in multiple dimensions such as resource usage, resource performance parameters and resource fault records of the target cloud resources, and cross-combine the characteristics of resource usage and resource performance parameters to explore the potential relationship between the two, form resource utilization efficiency characteristics, and can understand the efficiency of cloud resources in actual use, providing richer information for accurate prediction; further, modeling resource performance parameters and resource fault records, quantifying the impact of faults on resource usage, and accurately measuring the specific impact of faults on resource usage.
[0056] Then, the resource utilization efficiency characteristics and impact characteristics are input into the prediction model, which can analyze and predict based on multi-dimensional, comprehensive and deeply processed feature data. Compared with the traditional single-dimensional analysis method, this model can more accurately grasp the complex factors involved in the cloud resource life cycle and their interrelationships, thereby greatly improving the accuracy of cloud resource life cycle prediction. BRIEF DESCRIPTION OF THE DRAWINGS
[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.
[0058] Figure 1 A schematic diagram of a process flow of a method for predicting a life cycle of cloud resources in one embodiment;
[0059] Figure 2 It is a flowchart of a step of cross-combining characteristics of resource usage and resource performance parameters in one embodiment;
[0060] Figure 3 A flowchart of steps for obtaining a characteristic of the degree of influence representing the degree of influence of a fault on resource usage in one embodiment;
[0061] Figure 4 A schematic flow chart of a step of determining model parameters of an autoregressive integrated moving average model in another embodiment;
[0062] Figure 5 A flowchart of steps for constructing an impact degree feature representing the impact degree of a fault on resource usage in another embodiment;
[0063] Figure 6 is a structural block diagram of a cloud resource life cycle prediction device in one embodiment;
[0064] Figure 7 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION
[0065] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0066] In an exemplary embodiment, Figure 1As shown, a cloud resource life cycle prediction method is provided, and the method is applied to a computer device as an example for illustration, including:
[0067] S101, obtaining resource usage, resource performance parameters and resource failure records of target cloud resources.
[0068] Among them, target cloud resources refer to specific cloud resources that need lifecycle prediction. They can be virtual machines, storage devices, network bandwidth, and any other resources that provide services in the cloud computing environment.
[0069] Among them, resource usage refers to the degree or amount of target cloud resources used at different time points or time periods. For example, for virtual machines, resource usage can be CPU usage, memory usage, disk I / O rate, etc.; for storage devices, resource usage can be the size of the used storage space; for network bandwidth, resource usage is the actual amount of data transmitted.
[0070] Among them, resource performance parameters: reflect the performance of the target cloud resources, these parameters can help evaluate the health and operating efficiency of resources. For example, the CPU main frequency of the virtual machine, memory read and write speed, disk read and write delay; storage device read and write throughput, data redundancy rate; network bandwidth delay, packet loss rate, etc.
[0071] Among them, resource failure records: record the failure information that occurred in the target cloud resources in the past, including the time when the failure occurred, the type of failure (such as hardware failure, software failure), the severity of the failure, and the time to repair the failure.
[0072] Optionally, obtaining this data can usually be achieved through the cloud platform's monitoring system, logging system, or specialized performance monitoring tools.
[0073] S102, cross-combining resource usage and resource performance parameters to obtain resource utilization efficiency characteristics.
[0074] It can be understood that feature cross-combination is a data processing technology that creates new features that are more representative and discriminative by combining different features. In this step, the resource usage and resource performance parameters are cross-combined to mine the potential correlation information between the two.
[0075] For example, new features obtained by cross-combining features are used to measure the utilization efficiency of target cloud resources. For example, the CPU usage rate and the CPU main frequency can be cross-combined to obtain a feature that reflects the utilization degree of the actual CPU processing power; the memory usage rate and the memory read and write speed can be combined to evaluate the utilization efficiency of memory resources. These resource utilization efficiency features can more comprehensively reflect the usage of cloud resources under different performance conditions and provide more valuable information for subsequent life cycle prediction.
[0076] S103, modeling resource performance parameters and resource fault records to obtain impact characteristics representing the impact of the fault on resource usage.
[0077] It can be understood that modeling means: using statistical or machine learning methods to build a mathematical model based on resource performance parameters and resource failure records to describe the relationship between failures and resource usage. Common modeling methods include time series analysis algorithms (such as ARIMA models), regression analysis, and machine learning models (such as decision trees, neural networks, etc.).
[0078] Optionally, features that can quantify the impact of a fault on resource usage are obtained through modeling. For example, by using the ARIMA model to model fault records and resource usage, you can obtain features such as the change trend and magnitude of resource usage before and after the fault occurs. These features can intuitively reflect the impact of the fault on resource usage. By analyzing these impact features, you can predict how resource usage will be affected if a fault occurs in the future.
[0079] S104: Input the resource utilization efficiency characteristics and impact degree characteristics into the prediction model to obtain the life cycle of the target cloud resources.
[0080] The prediction model can be a pre-trained machine learning model, such as a linear regression model, support vector machine, neural network, etc., or a prediction model based on statistical methods. During the training process, these models learn the relationship between resource utilization efficiency characteristics, the impact of faults on resource usage, and the cloud resource life cycle.
[0081] The life cycle of the target cloud resources refers to the entire period of time from the start of use of the target cloud resources to the time when the target cloud resources reach their service life, have irreparable failures, or are eliminated due to no longer meeting business needs.
[0082] Optionally, the resource utilization efficiency characteristics and impact characteristics are input into the prediction model, and the model will predict the life cycle of the target cloud resources based on the rules and patterns it has learned. The prediction results can help cloud service providers to rationally plan resource allocation and management, prepare for resource upgrades and replacements in advance, and improve the efficiency and reliability of cloud resource use.
[0083] The above-mentioned cloud resource life cycle prediction method, by obtaining data in multiple dimensions such as resource usage, resource performance parameters and resource fault records of the target cloud resources, can comprehensively cover various factors that affect the life cycle of cloud resources, perform feature cross-combinations on resource usage and resource performance parameters, and mine the potential relationship between the two to form resource utilization efficiency characteristics, which can understand the efficiency of cloud resources in actual use and provide richer information for accurate prediction; further, modeling is performed on resource performance parameters and resource fault records, and the degree of influence of faults on resource usage is quantified to accurately measure the specific degree of influence of faults on resource usage. Then, the resource utilization efficiency characteristics and the degree of influence characteristics are input into the prediction model, and the model can perform analysis and prediction based on the multi-dimensional, comprehensive and deeply processed feature data. Compared with the traditional single-dimensional analysis method, this model can more accurately grasp the complex factors and their relationships involved in the life cycle of cloud resources, thereby greatly improving the accuracy of cloud resource life cycle prediction.
[0084] In an exemplary embodiment, Figure 2 As shown, the resource usage and resource performance parameters are cross-combined to obtain resource utilization efficiency features, including:
[0085] S201, taking resource usage and resource performance parameters as a first data set.
[0086] Among them, resource usage reflects the extent to which cloud resources are used during actual operation, such as CPU utilization, memory occupancy, disk I / O rate, etc.; resource performance parameters reflect the performance indicators of cloud resources themselves, such as CPU main frequency, memory read and write speed, disk capacity, etc.
[0087] Optionally, combine these two types of data into a unified data set.
[0088] For example, this data set contains multiple samples, each of which consists of specific values of resource usage and resource performance parameters. For example, for a virtual machine, a sample may be a collection of data such as its CPU usage is 30%, CPU main frequency is 2.5GHz, memory usage is 50%, memory read and write speed is 1000MB / s, etc. Combining these data into the first data set is the basis for subsequent association rule mining.
[0089] S202, configuring the minimum support and minimum confidence thresholds for the initial association rule mining algorithm to obtain a configured association rule mining algorithm.
[0090] Among them, the association rule mining algorithm is an algorithm used to discover the association relationship between different items in a data set, and common ones include Apriori algorithm, FP-growth algorithm, etc. These algorithms can find association rules that meet certain conditions from the data.
[0091] Among them, the minimum support means that the support measures the frequency of an item set in the data set. Let the number of times item set X appears in data set D be count(X), and the total number of samples in data set D is N, then the support of item set X is support=count(X) / N. The minimum support is a pre-set threshold, and only item sets with support greater than or equal to the threshold will be considered frequent item sets. For example, setting the minimum support to 0.2 means that only item sets with a frequency of more than 20% in the data set will be considered.
[0092] Among them, minimum confidence: confidence is used to measure the reliability of association rules. For association rule X→Y, its confidence confidence(X→Y)= support(X∪Y) / support(X), which represents the proportion of samples that also contain item sets in samples that contain item sets. The minimum confidence is also a pre-set threshold, and only association rules with confidence greater than or equal to the threshold will be retained. For example, setting the minimum confidence to 0.8 means that only when the reliability of the association rule reaches more than 80% will it be recognized.
[0093] The mining algorithm sets the minimum support and minimum confidence thresholds so that the algorithm can filter out meaningful frequent item sets and association rules according to these thresholds during the mining process.
[0094] S203, based on the configured association rule mining algorithm, scan the first data set to obtain frequent item sets that meet the minimum support.
[0095] Optionally, the configured association rule mining algorithm will traverse and analyze all samples in the first data set. The algorithm will count the number of times different item sets appear in the data set and calculate their support.
[0096] Among them, the item sets with support greater than or equal to the minimum support are called frequent item sets. For example, in the cloud resource data set, if it is found that the two items of high CPU usage and high memory usage often appear at the same time, and the support of the item set composed of them exceeds the minimum support threshold, then this item set is a frequent item set. By scanning the first data set, the algorithm can find all frequent item sets that meet the minimum support. These frequent item sets represent the feature combinations that often appear at the same time in the data.
[0097] S204: Generate association rules that meet the minimum confidence level based on the frequent item sets.
[0098] Optionally, for a frequent item set, association rules and different forms can be generated. For example, for the frequent item set {CPU usage high, memory usage high}, association rules "CPU usage high memory usage high" and "memory usage high CPU usage high" can be generated.
[0099] The confidence of each generated association rule is calculated and compared with the minimum confidence threshold. Only association rules with confidence greater than or equal to the minimum confidence are retained. These retained association rules reflect the feature association relationships with high reliability in the data. For example, if the confidence of the association rule "high CPU usage and high memory usage" meets the minimum confidence requirement, then it can be considered that in the operation of cloud resources, high CPU usage is likely to be accompanied by high memory usage.
[0100] S205 , based on the association rules, cross-combining the resource usage and the resource performance parameters to obtain resource utilization efficiency characteristics.
[0101] Optionally, based on association rules, resource usage and resource performance parameters are cross-combined to obtain resource utilization efficiency characteristics, including: based on association rules, multiplying the resource usage data corresponding to the rule antecedent with the resource performance parameter data corresponding to the rule consequent to obtain resource utilization efficiency characteristics.
[0102] It can be understood that according to the generated association rules, the resource usage data corresponding to the rule antecedent and the resource performance parameter data corresponding to the rule consequent are operated to create new features. This operation can capture the inherent connection between resource usage and resource performance parameters, thereby more comprehensively reflecting the resource utilization efficiency.
[0103] New features obtained by cross-combining features. For example, for the association rule "CPU usage is high and CPU main frequency utilization is high", the CPU usage and CPU main frequency utilization can be multiplied to obtain a new feature, which can more accurately measure the utilization efficiency of CPU resources.
[0104] In this embodiment, this feature cross-combination method based on association rules can mine hidden information in the data and provide more valuable features for subsequent tasks such as cloud resource life cycle prediction.
[0105] In an exemplary embodiment, Figure 3 As shown, resource performance parameters and resource fault records are modeled to obtain impact characteristics that represent the impact of faults on resource usage, including:
[0106] S301, modeling resource failure records and resource usage through a time series analysis algorithm to obtain an autoregressive integrated moving average model.
[0107] Optionally, resource failure records and resource usage data are collected and arranged in chronological order to form a time series.
[0108] The time series analysis algorithm is used to process the above two time series to construct the autoregressive integrated moving average ARIMA model. The time series analysis algorithm is used to process the two time series. The ARIMA model is like a "prediction machine" that can predict the future based on past data patterns. By analyzing the resource failure records and resource usage time series, this "prediction machine" can learn the relationship between failures and resource usage over time, thereby constructing a suitable ARIMA model.
[0109] S302, determining model parameters of the autoregressive integrated moving average model through the autocorrelation function and the partial autocorrelation function.
[0110] Optional, such as Figure 4 As shown, the model parameters of the autoregressive integrated moving average model are determined through the autocorrelation function and the partial autocorrelation function, including:
[0111] S401, determining an autocorrelation function graph according to an autocorrelation function value between a fault time series corresponding to a resource fault record and a usage time series data corresponding to a resource usage.
[0112] Optionally, for the resource failure time series and the resource usage time series, autocorrelation function values of different lag orders between them are calculated, and an autocorrelation function graph is drawn with the lag order as the horizontal axis and the autocorrelation function value as the vertical axis.
[0113] The autocorrelation function understands the correlation between data at different time points in the time series. For the resource failure time series and resource usage time series, we need to calculate their autocorrelation function values at different lag orders. Simply put, the lag order is to compare the data at the current time point with the data at the previous time point. For example, lag 1 is to look at the correlation between the current data and the data at the previous time point.
[0114] S402, determining a moving average order parameter of an autoregressive integrated moving average model according to the tailing or truncation condition in the autocorrelation function graph.
[0115] Optionally, carefully analyze the autocorrelation function graph. There may be two situations: tailing and truncation. Tailing means that the autocorrelation function value gradually decreases with the increase of the lag order, but it will not suddenly become 0; truncation means that the autocorrelation function value suddenly becomes very small and close to 0 after a certain lag order. If truncation occurs, the order at the truncation point is the reference value of the moving average order q. This q value is very important in the ARIMA model, which can help the model better handle the randomness of the data.
[0116] S403, determining a partial correlation function graph according to the partial correlation function value between the fault time series data corresponding to the resource fault record and the usage time series data corresponding to the resource usage.
[0117] Optionally, the partial correlation function values of different lag orders between the resource failure time series and the resource usage time series are calculated. Similarly, the partial correlation function graph is drawn with the lag order as the horizontal axis and the partial correlation function value as the vertical axis.
[0118] The partial correlation function is similar to the autocorrelation function, but it can exclude the influence of data at other time points in the middle and more directly reflect the correlation between the current data and the data at a previous time point. Similarly, the partial correlation function values of the resource failure time series and the resource usage time series at different lag orders are calculated.
[0119] S404, determining the autoregressive order parameter of the autoregressive integrated moving average model according to the tailing or truncation in the partial correlation function graph.
[0120] Optionally, analyze the tailing or truncation of the partial correlation function graph. If truncation occurs, the order at the truncation point is the reference value of the autoregressive order p. The p-value allows the ARIMA model to take into account the impact of past data on current data.
[0121] S405 , performing differential processing on the fault time series data corresponding to the resource fault record and the usage time series data corresponding to the resource usage to obtain a differential sequence.
[0122] Optionally, many time series data may not be very stable at the beginning, that is, the statistical characteristics of the data such as the mean and variance will change over time. The difference processing is to make the time series stable so that the ARIMA model can work better. Perform difference operations on the resource failure time series and the resource usage time series respectively. For example, the first-order difference is to subtract the data at the previous time point from the data at the current time point, and the new series obtained is the difference series.
[0123] S406, based on the stationarity of the difference sequence, the difference order parameter of the autoregressive integrated moving average model.
[0124] Optionally, analyze the statistical characteristics of the difference sequence (such as whether the mean and variance are stable) or use statistical test methods (such as ADF test) to determine the stationarity of the sequence. If the sequence reaches stationarity after d-order differences, the difference order is d.
[0125] S303: constructing an impact degree feature representing the impact degree of the fault on resource usage according to the model parameters of the autoregressive integrated moving average model.
[0126] Optional, such as Figure 5 As shown, based on the model parameters of the autoregressive integrated moving average model, the impact degree characteristics representing the impact degree of the fault on resource usage are constructed, including:
[0127] S501, predicting resource usage before and after a fault occurs using an autoregressive integrated moving average model with determined model parameters to obtain a predicted usage.
[0128] Optional, model prediction: Use the ARIMA model with determined parameters (p, d, q) to predict resource usage before and after the failure to obtain predicted usage.
[0129] S502, obtaining a deviation dimension feature according to the deviation between the predicted usage and the resource usage.
[0130] Optionally, the predicted usage is compared with the actual resource usage, and the deviation between the two is calculated, and the deviation is used as a feature of a dimension.
[0131] S503, determining the residual sequence characteristics of the autoregressive integrated moving average model during the fault period.
[0132] Optionally, during the failure period, the model's predicted value minus the actual value is the residual. The residual can reflect the part that the model did not predict. Then analyze the characteristics of the residual series, such as the mean and variance of the residual, which can further reflect the performance of the model and the complex impact of the failure on resource usage.
[0133] S504: Use the deviation dimension feature and the residual sequence feature as impact degree features indicating the impact degree of the fault on resource usage.
[0134] Optionally, the deviation dimension feature and the residual sequence feature are combined to form a comprehensive feature that represents the impact of the fault on resource usage.
[0135] In this embodiment, by collecting resource failure records and resource usage data and building an ARIMA model, we can deeply explore the inherent laws of failures and resource usage over time. For example, there may be a complex correlation between the number of server failures and CPU usage. The ARIMA model can learn this dynamic relationship, thereby providing a solid foundation for subsequent analysis and prediction.
[0136] Time series analysis algorithms are particularly suitable for processing data arranged in chronological order. Resource failure records and resource usage data have obvious time characteristics. The ARIMA model can make full use of the time information in these data and accurately capture the trend, seasonality, and periodicity of the data. Compared with traditional static data analysis methods, it can better reflect the actual situation.
[0137] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.
[0138] Based on the same inventive concept, the embodiment of the present application also provides a cloud resource lifecycle prediction device for implementing the cloud resource lifecycle prediction method involved above. The implementation solution provided by the device to solve the problem is similar to the implementation solution recorded in the above method, so the specific limitations in the embodiments of one or more cloud resource lifecycle prediction devices provided below can refer to the limitations of the cloud resource lifecycle prediction method above, and will not be repeated here.
[0139] In an exemplary embodiment, Figure 6 As shown, a cloud resource life cycle prediction device is provided, comprising:
[0140] The acquisition module 11 is used to obtain the resource usage, resource performance parameters and resource fault records of the target cloud resources;
[0141] The first feature generation module 12 is used to perform feature cross-combination on resource usage and resource performance parameters to obtain resource utilization efficiency features;
[0142] The second feature generation module 13 is used to model resource performance parameters and resource fault records to obtain an impact degree feature representing the impact degree of the fault on resource usage;
[0143] The life cycle prediction module 14 is used to obtain the life cycle of the target cloud resources according to the resource utilization efficiency characteristics and the impact degree characteristics input into the prediction model.
[0144] In one of the embodiments, the first feature generation module 12 is further used to: use the resource usage and the resource performance parameter as the first data set;
[0145] The minimum support and minimum confidence thresholds configured for the initial association rule mining algorithm are obtained to obtain the configured association rule mining algorithm;
[0146] Based on the configured association rule mining algorithm, scan the first data set to obtain frequent item sets that meet the minimum support;
[0147] Based on frequent item sets, generate association rules that meet the minimum confidence level;
[0148] Based on association rules, the resource usage and resource performance parameters are cross-combined to obtain resource utilization efficiency characteristics.
[0149] In one embodiment, the first feature generation module 12 is further used to: based on the association rule, multiply the resource usage data corresponding to the rule antecedent and the resource performance parameter data corresponding to the rule consequent to obtain the resource utilization efficiency feature.
[0150] In one embodiment, the second feature generation module 13 is further used to:
[0151] Through the time series analysis algorithm, the resource failure records and resource usage are modeled to obtain the autoregressive integrated moving average model;
[0152] Determine the model parameters of the autoregressive integrated moving average model through the autocorrelation function and the partial autocorrelation function;
[0153] According to the model parameters of the autoregressive integrated moving average model, an impact degree feature representing the impact degree of the fault on resource usage is constructed.
[0154] In one embodiment, the second feature generation module 13 is further used to:
[0155] Determine an autocorrelation function graph according to an autocorrelation function value between a fault time series corresponding to a resource fault record and a usage time series data corresponding to a resource usage;
[0156] According to the tailing or truncation in the autocorrelation function graph, the moving average order parameter of the autoregressive integrated moving average model is determined;
[0157] Determine a partial correlation function graph according to a partial correlation function value between a fault time series corresponding to a resource fault record and a usage time series data corresponding to a resource usage;
[0158] According to the tailing or truncation in the partial correlation function graph, the autoregressive order parameter of the autoregressive integrated moving average model is determined;
[0159] Perform differential processing on the fault time series data corresponding to the resource fault record and the usage time series data corresponding to the resource usage to obtain a differential sequence;
[0160] The difference order parameter of the autoregressive integrated moving average model is determined based on the stationarity of the difference series.
[0161] In one embodiment, the second feature generation module 13 is further used to:
[0162] The resource usage before and after the failure is predicted using the autoregressive integrated moving average model with determined model parameters to obtain the predicted usage;
[0163] According to the deviation between the predicted usage and the resource usage, the deviation dimension feature is obtained;
[0164] Determine the residual series characteristics of the autoregressive integrated moving average model during the fault period;
[0165] The deviation dimension features and residual sequence features are used as impact degree features to represent the impact degree of the fault on resource usage.
[0166] Each module in the above-mentioned cloud resource life cycle prediction device can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.
[0167] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 7As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data of the life cycle prediction method of cloud resources. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for predicting the life cycle of cloud resources is implemented.
[0168] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0169] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:
[0170] Obtain resource usage, resource performance parameters, and resource failure records of target cloud resources;
[0171] Perform cross-combination of resource usage and resource performance parameters to obtain resource utilization efficiency characteristics;
[0172] Model resource performance parameters and resource failure records to obtain impact characteristics that represent the impact of failures on resource usage;
[0173] The resource utilization efficiency characteristics and impact degree characteristics are input into the prediction model to obtain the life cycle of the target cloud resources.
[0174] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:
[0175] Obtain resource usage, resource performance parameters, and resource failure records of target cloud resources;
[0176] Perform cross-combination of resource usage and resource performance parameters to obtain resource utilization efficiency characteristics;
[0177] Model resource performance parameters and resource failure records to obtain impact characteristics that represent the impact of failures on resource usage;
[0178] The resource utilization efficiency characteristics and impact degree characteristics are input into the prediction model to obtain the life cycle of the target cloud resources.
[0179] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:
[0180] Obtain resource usage, resource performance parameters, and resource failure records of target cloud resources;
[0181] Perform cross-combination of resource usage and resource performance parameters to obtain resource utilization efficiency characteristics;
[0182] Model resource performance parameters and resource failure records to obtain impact characteristics that represent the impact of failures on resource usage;
[0183] The resource utilization efficiency characteristics and impact degree characteristics are input into the prediction model to obtain the life cycle of the target cloud resources.
[0184] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.
[0185] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.
[0186] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.
[0187] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A method for predicting the life cycle of cloud resources, characterized in that: The method comprises: Obtain resource usage, resource performance parameters, and resource failure records of target cloud resources; Performing a cross-combination of the resource usage and the resource performance parameter to obtain a resource utilization efficiency feature; Modeling the resource performance parameters and resource fault records to obtain impact characteristics that represent the impact of the fault on resource usage; The resource utilization efficiency characteristics and the impact degree characteristics are input into a prediction model to obtain the life cycle of the target cloud resources.
2. The method according to claim 1, characterized in that: The cross-combination of the resource usage and the resource performance parameter to obtain a resource utilization efficiency feature includes: taking the resource usage and the resource performance parameter as a first data set; The minimum support and minimum confidence thresholds configured for the initial association rule mining algorithm are obtained to obtain the configured association rule mining algorithm; Based on the configured association rule mining algorithm, scanning the first data set to obtain frequent item sets that meet the minimum support; Based on the frequent item sets, generating association rules that satisfy the minimum confidence; Based on the association rule, the resource usage and the resource performance parameter are cross-combined to obtain a resource utilization efficiency feature.
3. The method according to claim 2, characterized in that The resource usage and the resource performance parameter are cross-combined based on the association rule to obtain resource utilization efficiency features, including: Based on the association rule, the resource usage data corresponding to the rule antecedent and the resource performance parameter data corresponding to the rule consequent are multiplied to obtain the resource utilization efficiency feature.
4. The method according to claim 1, characterized in that The modeling of the resource performance parameters and resource fault records to obtain the impact degree characteristics representing the impact degree of the fault on resource usage includes: Through the time series analysis algorithm, the resource failure records and resource usage are modeled to obtain the autoregressive integrated moving average model; Determining the model parameters of the autoregressive integrated moving average model through the autocorrelation function and the partial autocorrelation function; According to the model parameters of the autoregressive integrated moving average model, an impact degree feature representing the impact degree of the fault on resource usage is constructed.
5. The method according to claim 4, characterized in that Determining the model parameters of the autoregressive integrated moving average model through the autocorrelation function and the partial autocorrelation function includes: Determining an autocorrelation function graph according to an autocorrelation function value between a fault time series corresponding to the resource fault record and a usage time series data corresponding to the resource usage; Determine the moving average order parameter of the autoregressive integrated moving average model according to the tailing or truncation in the autocorrelation function graph; Determine a partial correlation function graph according to a partial correlation function value between a fault time series corresponding to a resource fault record and a usage time series data corresponding to the resource usage; Determining the autoregressive order parameter of the autoregressive integrated moving average model according to the tailing or truncation in the partial correlation function graph; Performing differential processing on the fault time series data corresponding to the resource fault record and the usage time series data corresponding to the resource usage to obtain a differential sequence; According to the stationarity of the difference sequence, the difference order parameter of the autoregressive integrated moving average model.
6. The method according to claim 4, characterized in that The constructing, based on the model parameters of the autoregressive integrated moving average model, an impact degree feature representing the impact degree of the fault on resource usage includes: The resource usage before and after the failure is predicted using the autoregressive integrated moving average model with determined model parameters to obtain the predicted usage; Obtaining a deviation dimension feature according to a deviation between the predicted usage and the resource usage; Determining residual sequence characteristics of the autoregressive integrated moving average model during the fault period; The deviation dimension feature and the residual sequence feature are used as impact degree features indicating the impact degree of the fault on resource usage.
7. A cloud resource life cycle prediction device, characterized in that: The device comprises: An acquisition module is used to obtain resource usage, resource performance parameters and resource fault records of target cloud resources; A first feature generation module is used to perform feature cross-combination on the resource usage and the resource performance parameter to obtain a resource utilization efficiency feature; A second feature generation module is used to model the resource performance parameters and resource fault records to obtain an impact degree feature representing the impact degree of the fault on resource usage; The life cycle prediction module is used to obtain the life cycle of the target cloud resource based on the resource utilization efficiency characteristics and the impact degree characteristics input into the prediction model.
8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.