Data processing method and device and electronic equipment

Abstract

The invention discloses a data processing method and device and electronic equipment. The method comprises the steps of obtaining actual performance data of a plurality of performance reference indexes at a current moment in a process that a target service is executed; for the plurality of performance reference indexes, determining data trend characteristics, data periodicity characteristics and data change rate characteristics according to the actual performance data and the actual historical performance data; analyzing and processing the data trend feature, the data periodicity feature and the data change rate feature corresponding to each performance reference index to obtain a target feature; inputting the target features into a target performance parameter prediction model, and outputting predicted performance data corresponding to the plurality of performance reference indexes at the prediction moment; according to the prediction performance data and the actual performance data of each performance reference index, the resource data allocated to the target service at the prediction moment is determined, so that the flexibility of resource allocation is ensured, reasonable allocation of system resources is realized, and the system is enabled to operate stably, efficiently and normally.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a data processing method, device and electronic equipment. Background Art

[0002] In the management and control process of smart grids, the Dianhong operating system, also known as the power industry implementation operating system, is used to coordinate the management and scheduling of various industrial equipment such as substation switchgear, generator sets, energy storage equipment, various sensors and actuators.

[0003] By rationally allocating system resources corresponding to the Dianhong operating system, the efficient operation of various industrial equipment can be ensured. However, in the prior art, the system resources are mainly allocated through static scheduling, that is, the allocation of preset system resources is achieved through a fixed execution cycle, which not only fails to meet the flexibility requirements for resource allocation and scheduling, but also fails to ensure that system resources are effectively allocated and utilized, affecting the normal operation of the system. Summary of the invention

[0004] The present invention provides a data processing method, device and electronic equipment, which ensure the flexibility of resource allocation, realize the rational allocation of system resources, and enable the system to operate normally and stably and efficiently.

[0005] According to one aspect of the present invention, there is provided a data processing method, the method comprising:

[0006] During the execution of the target service, actual performance data of multiple performance reference indicators at the current moment are obtained, wherein the multiple performance reference indicators include at least two of a CPU instantaneous occupancy indicator, a memory instantaneous usage indicator, a disk read / write indicator, a network average flow indicator, a system average response time indicator, a process indicator, and a temperature and energy consumption indicator;

[0007] For multiple performance reference indicators, determine data trend characteristics, data periodicity characteristics, and data change rate characteristics corresponding to the performance reference indicators based on actual performance data under the performance reference indicators and actual historical performance data corresponding to multiple historical moments before the current moment;

[0008] By analyzing and processing the data trend characteristics, data periodicity characteristics and data change rate characteristics corresponding to each performance reference indicator, the target characteristics to be input into the target performance parameter prediction model are obtained;

[0009] Input the target features into the target performance parameter prediction model, and output the predicted performance data corresponding to multiple performance reference indicators at the prediction time;

[0010] The resource data allocated to the target service at the prediction time is determined according to the predicted performance data and the actual performance data corresponding to each performance reference indicator.

[0011] According to another aspect of the present invention, there is provided a data processing device, the device comprising:

[0012] A data acquisition module is used to acquire actual performance data of multiple performance reference indicators at the current moment during the execution of the target business, wherein the multiple performance reference indicators include at least two of a CPU instantaneous occupancy indicator, a memory instantaneous usage indicator, a disk read / write indicator, a network average flow indicator, a system average response time indicator, a process indicator, and a temperature and energy consumption indicator;

[0013] A feature determination module is used to determine, for multiple performance reference indicators, data trend features, data periodicity features, and data change rate features corresponding to the performance reference indicators based on actual performance data under the performance reference indicators and actual historical performance data corresponding to multiple historical moments before the current moment;

[0014] A feature analysis module is used to obtain target features to be input into the target performance parameter prediction model by analyzing and processing the data trend features, data periodicity features and data change rate features corresponding to each performance reference indicator;

[0015] The performance data prediction module is used to input the target features into the target performance parameter prediction model and output the predicted performance data corresponding to the multiple performance reference indicators at the prediction time;

[0016] The resource data determination module is used to determine the resource data allocated to the target business at the prediction time according to the predicted performance data and the actual performance data corresponding to each performance reference indicator.

[0017] According to another aspect of the present invention, there is provided an electronic device, the electronic device comprising:

[0018] at least one processor; and

[0019] a memory communicatively connected to at least one processor; wherein,

[0020] The memory stores a computer program that can be executed by at least one processor. The computer program is executed by at least one processor so that the at least one processor can execute the data processing method of any embodiment of the present invention.

[0021] According to another aspect of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the data processing method of any embodiment of the present invention when executed.

[0022] According to another aspect of the present invention, there is provided a computer program product, comprising a computer program, wherein the computer program implements the data processing method according to any embodiment of the present invention when executed by a processor.

[0023] The technical solution of the embodiment of the present invention obtains the actual performance data of multiple performance reference indicators at the current moment during the execution of the target business, and determines the data trend characteristics, data periodicity characteristics and data change rate characteristics corresponding to the performance reference indicators according to the actual performance data under each performance reference indicator and the actual historical performance data corresponding to multiple historical moments before the current moment. By analyzing and processing the data trend characteristics, data periodicity characteristics and data change rate characteristics corresponding to each performance reference indicator, the target characteristics to be input into the target performance parameter prediction model are obtained, providing accurate input characteristics for subsequent prediction processing. The target characteristics are input into the target performance parameter prediction model, and the predicted performance data corresponding to the multiple performance reference indicators at the prediction moment are output. According to the predicted performance data and the actual performance data corresponding to each performance reference indicator, the resource data allocated to the target business at the prediction moment is determined, so as to realize the resource allocation processing of the target business through the resource data, and realize the intelligent scheduling of system resources. It solves the problems of low flexibility and ineffective utilization of system resources caused by static allocation of system resources, effectively improves system response speed, system resource utilization and system stability, especially in complex and high-load industrial environments, ensures that the system can operate stably and efficiently, reduces the risk of system failures caused by overload or insufficient resources, and enhances the reliability and scalability of the overall system.

[0024] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0026] Figure 1 is a flow chart of a data processing method provided by an embodiment of the present invention;

[0027] Figure 2 is a flow chart of a data processing method provided by an embodiment of the present invention;

[0028] Figure 3 is a structural schematic diagram of a data processing device provided by an embodiment of the present invention;

[0029] Figure 4 It is a structural schematic diagram of an electronic device for implementing the data processing method of an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] Embodiment 1

[0033] Figure 1 This is a flow chart of a data processing method provided by the first embodiment of the present invention. This embodiment can be applied to the situation where the performance data of the target business is analyzed to achieve reasonable allocation of system resources to the target business. The method can be executed by a data processing device, which can be implemented in the form of hardware and / or software. The data processing device can be configured in an electronic device such as a mobile phone, a computer or a server. Figure 1 As shown, the method includes:

[0034] S110 . During the execution of the target service, actual performance data of multiple performance reference indicators at the current moment are obtained.

[0035] Among them, the multiple performance reference indicators include at least two of the CPU instantaneous occupancy indicator, memory instantaneous usage indicator, disk read and write indicator, network average traffic indicator, system average response time indicator, process indicator, temperature and energy consumption indicator.

[0036] The embodiments of the present invention can be applied to a scenario in which business data is interacted between a Dianhong operating system and at least one industrial device associated therewith, and system resources are allocated for a target business associated with the current industrial device based on the Dianhong operating system.

[0037] Among them, the Dianhong operating system, also known as the power industry implementation operating system, is used to coordinate and manage various industrial equipment such as switchgear, generator sets, energy storage equipment, and various sensors and actuators in substations. When the current industrial equipment sends a business processing request to the Dianhong operating system, the Dianhong operating system executes the target business corresponding to the business processing request when receiving the business processing request. That is, the target business can be understood as the business that the Dianhong operating system currently needs to process.

[0038] Performance reference indicators can be used to characterize the system resource usage under the corresponding indicators when executing the target business. The actual performance data of the CPU instantaneous occupancy indicator, that is, the CPU instantaneous occupancy, can be understood as the ratio of the CPU occupied by non-idle tasks at the current moment. Optionally, the CPU instantaneous occupancy can be viewed through the performance monitoring tool corresponding to the Dianhong operating system. The actual performance data of the memory instantaneous utilization indicator, that is, the memory instantaneous utilization, can be understood as the ratio of the memory used at the current moment to the total memory. The actual performance data of the disk read and write indicator, that is, the disk read and write speed, can be understood as the number of inputs and outputs per unit time. The actual performance data of the network average traffic indicator, that is, the network average traffic, can be understood as the average number of data frames transmitted per unit time. Optionally, the network average traffic can be determined by the duration between the current moment and the preset historical moment and the total number of data frames transmitted by the network corresponding to the duration. The actual performance data of the system average response time indicator, that is, the system average response time, can be understood as the average time for the Dianhong operating system to process requests. Optionally, the average system response time corresponding to processing each request may be determined according to the time between the current moment and a preset historical moment and the number of business requests corresponding to the time.

[0039] The actual performance data of the process indicator can be used to characterize the system process status at the current moment. Optionally, the running process at the current moment can be set to 1, and the completed process or dormant process can be set to 0, so as to determine the actual performance data of the process indicator according to the status of all processes corresponding to the Dianhong operating system at the current moment. The actual performance data of the temperature indicator can be understood as the device temperature corresponding to the device deployed with the Dianhong operating system at the current moment. The actual performance data of the energy consumption indicator can be understood as the remaining power of the device deployed with the Dianhong operating system at the current moment.

[0040] Specifically, when the Dianhong operating system executes the target business, the actual performance data of at least two performance reference indicators including the instantaneous CPU occupancy index, the instantaneous memory usage index, the disk read and write index, the average network traffic index, the average system response time index, the process index, and the temperature and energy consumption index are obtained at the current moment, and the obtained actual performance data is cleaned to obtain the cleaned actual performance data, and the cleaned actual performance data is standardized to determine the characteristics that can reflect the system status of the Dianhong operating system through the standardized actual performance data.

[0041] S120. For multiple performance reference indicators, determine data trend characteristics, data periodicity characteristics, and data change rate characteristics corresponding to the performance reference indicators based on actual performance data under the performance reference indicators and actual historical performance data corresponding to multiple historical moments before the current moment.

[0042] Among them, multiple historical moments can be pre-set, multiple moments before the current moment. The actual performance data corresponding to the historical moments are the actual historical performance data. The data trend feature can be a feature used to characterize the trend of changes in actual performance data. The data periodicity feature can be used to characterize the characteristics of periodic changes in actual performance data in the Dianhong operating system. The data change rate feature can be used to characterize the characteristics of instantaneous changes in actual performance data.

[0043] Specifically, for multiple performance reference indicators, the actual performance data under each performance reference indicator and the actual historical performance data corresponding to multiple historical moments before the current moment can be fitted by a linear regression method to obtain the data trend characteristics corresponding to each performance reference indicator. For multiple performance reference indicators, the actual performance data under each performance reference indicator and the actual historical performance data corresponding to multiple historical moments before the current moment can be processed by a fast Fourier transform method to obtain data periodic change characteristics. For multiple performance reference indicators, the data change rate characteristics corresponding to the current performance reference indicator can be determined by the actual performance data corresponding to the current performance reference indicator at the current moment, the actual historical performance data at a historical moment before the current moment, and the duration between the current moment and the historical moment. Based on this, the data change rate characteristics corresponding to each performance reference indicator are obtained.

[0044] In an embodiment of the present invention, the method for determining data trend characteristics can be: based on the amount of data corresponding to the sliding window, obtaining the actual historical performance data under the performance reference index for multiple historical moments before the current moment; based on the actual historical performance data, the actual performance data and a preset linear fitting function, determining the data trend characteristics corresponding to the current moment.

[0045] The sliding window may be a time window selected according to actual needs. The sliding window may be used to determine the number of performance data corresponding to each performance reference indicator at each historical moment, that is, the number of data. The preset linear fitting function may be a function for performing linear regression processing on actual performance data and actual historical performance data. The data trend feature may be understood as a linear fitting feature determined based on a preset linear fitting function.

[0046] The preset linear fitting function can be expressed as follows:

[0047] X(t)=a+bt+∈(t)

[0048] Among them, X(t) represents the performance data corresponding to time t, which can be the actual performance data corresponding to the current time t, or the actual historical performance data corresponding to the historical time t. a represents the intercept, b represents the data trend characteristics, and ∈(t) represents the error term.

[0049] Specifically, according to the amount of data corresponding to the sliding window, the actual historical performance data under the performance reference index at multiple historical moments before the current moment is determined. The actual historical performance data and the actual performance data are substituted into the preset linear fitting function to determine the data trend characteristics corresponding to the current moment. Optionally, the average of the actual historical performance data corresponding to the current sliding window can also be determined by obtaining the actual historical performance data and the amount of data at multiple historical moments. By sliding the sliding window multiple times, multiple averages of actual historical performance data are obtained. Multiple averages of actual historical performance data and the actual performance data are substituted into the preset linear fitting function to calculate the data trend characteristics corresponding to the current moment.

[0050] Exemplarily, a plurality of actual historical performance means are determined by a sliding average method, and the sliding average method can be represented by the following function.

[0051]

[0052] Among them, MA(t) represents the actual historical performance mean, N represents the size of the sliding window, i represents the historical moment, and X(i) represents the actual historical performance data at the historical moment i.

[0053] Substitute multiple actual historical performance averages MA(t) as X(t) into the preset linear fitting function, and substitute the actual performance data X(t) corresponding to the current moment t into the preset linear fitting function, and jointly calculate to obtain the data trend feature b corresponding to the current moment. When the data trend feature b>0, it means that the system load of the Dianhong operating system is gradually increasing, and when the data trend feature b<0, it means that the system load of the Dianhong operating system is gradually decreasing.

[0054] In an embodiment of the present invention, the method for determining the data periodicity characteristics can be: based on the actual performance data under the performance reference indicator, the actual historical performance data corresponding to multiple historical moments before the current moment, a pre-set amount of data, and a frequency domain processing function, determine the data periodicity characteristics corresponding to the current moment.

[0055] The preset data quantity may be the total data quantity of the actual historical performance data and the actual performance data. The frequency domain processing function may be a function for performing fast Fourier transform processing on the actual performance data and the actual historical performance data. The data amplitude corresponding to the collected performance data may be determined by the frequency domain processing function. Optionally, the frequency domain processing function may be as shown in the following formula.

[0056]

[0057] Wherein, X(f) represents the amplitude at frequency f, and frequency f is the frequency of collecting performance data. Frequency f can be determined by the time interval for collecting performance data and the number of collections corresponding to the time interval. T represents the preset number of data, and X(t) represents the performance data corresponding to the collection time t, which can be the actual performance data corresponding to the current time t, or the actual historical performance data corresponding to the historical time t. The periodicity characteristics of the data can be determined by the frequency and amplitude in the above function. Optionally, the frequency f is used as the main period to characterize the periodic changes of the data, and X(f) is used as the intensity to characterize the periodic changes of the data.

[0058] Specifically, the actual performance data under the performance reference index and the actual historical performance data corresponding to multiple historical moments before the current moment and the preset data quantity can be substituted into the frequency domain processing function to determine the amplitude corresponding to the frequency of the collected performance data. According to the frequency and amplitude of the collected performance data, the data periodicity characteristics corresponding to the current moment are determined.

[0059] Optionally, the actual historical performance data corresponding to different historical moments before the current moment are obtained multiple times, and the above process of calculating the amplitude is repeated to obtain multiple amplitudes. The multiple amplitudes are numerically sorted, and one or more amplitudes with the largest amplitude are used as the main amplitude. According to the main amplitude and the frequency of the corresponding collected performance data, the data periodicity characteristics corresponding to the current moment are obtained. That is, the frequency is used as the main period to characterize the periodic change of the data, and the main amplitude is used as the intensity to characterize the periodic change of the data.

[0060] In an embodiment of the present invention, the method for determining the data change rate characteristics can be: based on the actual performance data at the current moment under the performance reference index, the actual historical performance data corresponding to the previous historical moment of the current moment, and the instantaneous change rate determination function, determine the data change rate characteristics corresponding to the current moment.

[0061] The instantaneous change rate determination function may be a function used to determine the instantaneous change rate of data. The instantaneous change rate determination function may be expressed as:

[0062]

[0063] Among them, R(t) represents the instantaneous change rate of the data corresponding to the current moment, that is, the data change rate characteristic, t represents the current moment, t-1 represents the previous historical moment of the current moment, X(t) represents the actual performance data of the current moment, and X(t-1) represents the actual historical performance data corresponding to the previous historical moment of the current moment.

[0064] Specifically, the actual performance data at the current moment under the current performance reference index and the actual historical performance data corresponding to the previous historical moment of the current moment are substituted into the instantaneous change rate determination function to obtain the instantaneous change rate of the data corresponding to the current moment, that is, the data change rate characteristic.

[0065] S130. Analyze and process the data trend characteristics, data periodicity characteristics, and data change rate characteristics corresponding to each performance reference indicator to obtain target characteristics to be input into the target performance parameter prediction model.

[0066] The target feature may be a feature associated with the system load condition obtained by principal component analysis of data trend features, data periodicity features, and data change rate features. The target performance parameter prediction model may be the performance data required to determine different performance reference indicators at the prediction moment. The prediction moment may be a pre-set moment in the future of the current moment.

[0067] Specifically, principal component analysis is performed on the data trend characteristics, data periodicity characteristics, and data change rate characteristics corresponding to each performance reference indicator to obtain target characteristics, so as to input the target characteristics into the target performance parameter prediction model.

[0068] S140: Input the target features into the target performance parameter prediction model, and output the predicted performance data corresponding to the multiple performance reference indicators at the prediction time.

[0069] The target performance parameter prediction model may be pre-trained and used to determine the predicted performance data corresponding to multiple performance reference indicators at the prediction moment. Optionally, the target performance parameter prediction model may include multiple processing branches, each processing branch is used to determine the predicted performance data corresponding to the corresponding performance reference indicator. The predicted performance data can be used to characterize the performance data required to process the target business at the prediction moment. The prediction moment may be a moment in the future of the current moment.

[0070] Specifically, the target features are input into the target performance parameter prediction model, and based on multiple processing branches of the target performance parameter prediction model, predicted performance data corresponding to multiple performance reference indicators at the prediction moment are output.

[0071] Optionally, before processing the target features based on the target performance parameter prediction model, the target performance parameter prediction model can be trained first. The specific method can be: obtain multiple training samples, wherein the training samples include sample features and theoretical performance data corresponding to the sample features. Input the training samples into the target performance parameter prediction model to be trained to obtain the output performance data corresponding to each processing branch. Use a preset loss function to perform loss processing on the output performance data and theoretical performance data corresponding to each processing branch to obtain a loss value. Based on the loss value, the target performance parameter prediction model to be trained is corrected to obtain a trained target performance parameter prediction model.

[0072] The sample features may be target features determined based on actual historical performance data. The theoretical performance data may be pre-set performance data corresponding to each performance reference indicator at the prediction moment. The output performance data may be performance data output by the target performance parameter prediction model to be trained based on the training samples.

[0073] The preset loss function may be a pre-set function for loss processing. Optionally, in the case where the processing branch includes a demand branch corresponding to a plurality of performance reference indicators and a response time branch corresponding to a system average response time indicator, the preset loss function may be expressed as:

[0074] L total =αL resource +βL response

[0075] Wherein, α represents a pre-set, first weight coefficient. β represents a pre-set, second weight coefficient. L resource Indicates the loss value corresponding to the demand branch of multiple performance reference indicators. response Indicates the loss value corresponding to the response time branch of the system average response time indicator.

[0076]

[0077] in, Represents the loss value corresponding to the demand branch of multiple performance reference indicators, and N represents the number of training samples. represents the theoretical performance data corresponding to the nth performance reference indicator of the ith training sample, Indicates the output performance data corresponding to the nth performance reference indicator of the i-th training sample. For example, to determine the output performance data under the CPU instantaneous occupancy indicator, the memory instantaneous usage indicator, and the disk read / write indicator, then n = 3, It represents the output performance data corresponding to the CPU instantaneous occupancy index of the i-th training sample. Represents the output performance data corresponding to the instantaneous memory usage index of the i-th training sample, Represents the output performance data corresponding to the disk read and write indicators of the i-th training sample.

[0078]

[0079] in, Indicates the loss value corresponding to the response time branch of the system average response time indicator, and N represents the number of training samples. i response Represents the theoretical performance data under the system average response time indicator of the i-th training sample. Represents the output performance data under the system average response time indicator of the i-th training sample.

[0080] Specifically, multiple training samples are obtained. Each training sample is input into the target performance parameter prediction model to be trained to obtain the output performance data corresponding to each processing branch. Based on the above-mentioned preset loss function, the output performance data and theoretical performance data of each processing branch are subjected to loss processing to obtain a loss value. When the model parameters of the target performance parameter prediction model to be trained are corrected according to the loss value, the convergence of the loss function can be used as a training target, such as whether the training error is less than the preset error, or whether the error change tends to be stable, or whether the current number of iterations is equal to the preset number of iterations. If the detection reaches the convergence condition, such as the training error of the loss function is less than the preset error, or the error change trend tends to be stable, it indicates that the training of the target performance parameter prediction model to be trained is completed, and the iterative training can be stopped at this time. If it is detected that the convergence condition is not reached at present, other training samples can be further obtained to continue training the target performance parameter prediction model to be trained until the training error of the loss function is within the preset range. When the training error of the loss function reaches convergence, the trained target performance parameter prediction model can be obtained.

[0081] S150: Determine resource data allocated to the target service at the prediction time according to the predicted performance data and the actual performance data corresponding to each performance reference indicator.

[0082] The resource data may be understood as system resource data that needs to be allocated to the target service at the prediction time. The resource data may include system resource data corresponding to a plurality of performance parameter indicators.

[0083] Specifically, the resource demand difference data corresponding to each performance reference indicator is determined according to the predicted performance data and the actual performance data corresponding to each performance reference indicator. The resource data allocated to the target service at the prediction time is determined according to the resource demand difference data and the actual performance data.

[0084] In an embodiment of the present invention, the method for determining resource data can be: for multiple performance reference indicators, performance difference data is determined based on the predicted performance data and actual performance data corresponding to the performance reference indicators; the performance difference data is adjusted to determine the resource data allocated to the target business under the performance reference indicators based on the actual performance data and the adjusted performance difference data.

[0085] The performance difference data, i.e., the resource demand difference data mentioned above, is used to characterize the degree of difference in system resources between the predicted time and the current time. The performance difference data is adjusted by a preset adjustment coefficient to obtain adjusted performance difference data. The preset adjustment coefficient can be a pre-set coefficient used to adjust the performance difference data.

[0086] Specifically, for multiple performance reference indicators, the performance difference data can be determined by the predicted performance data and actual performance data corresponding to the performance reference indicators. The performance difference data is adjusted according to a preset adjustment coefficient to obtain the adjusted performance difference data. The actual performance data and the adjusted performance difference data are added together to obtain the resource data allocated to the target business under each performance reference indicator.

[0087] Exemplarily, the performance difference data is determined by the predicted performance data and the actual performance data corresponding to the performance reference indicator, as shown in the following formula.

[0088] ΔR n =R n,current -R n,predicted

[0089] Among them, R n,predicted Represents the predicted performance data corresponding to the nth performance reference indicator. n,current Indicates the actual performance data corresponding to the nth performance reference indicator. ΔR n Indicates performance difference data.

[0090] Adjustments to the performance difference data can be expressed by the following function:

[0091] ΔT n =α n ΔR n

[0092] Among them, α n Indicates the preset adjustment factor, ΔTn Represents adjusted performance difference data.

[0093] The resource data allocated to the target business under the performance reference index is determined based on the actual performance data and the adjusted performance difference data, as shown in the following formula.

[0094] R n,adjusted =ΔT n +R n,current

[0095] Among them, R n,adjusted Indicates the resource data allocated to the target service under the nth performance reference indicator.

[0096] The technical solution of this embodiment obtains the actual performance data of multiple performance reference indicators at the current moment during the execution of the target business, and determines the data trend characteristics, data periodicity characteristics and data change rate characteristics corresponding to the performance reference indicators based on the actual performance data under each performance reference indicator and the actual historical performance data corresponding to multiple historical moments before the current moment. By analyzing and processing the data trend characteristics, data periodicity characteristics and data change rate characteristics corresponding to each performance reference indicator, the target characteristics to be input into the target performance parameter prediction model are obtained, providing accurate input characteristics for subsequent prediction processing. The target characteristics are input into the target performance parameter prediction model, and the predicted performance data corresponding to multiple performance reference indicators at the prediction moment are output. According to the predicted performance data and actual performance data corresponding to each performance reference indicator, the resource data allocated to the target business at the prediction moment is determined, so as to realize the resource allocation processing of the target business through the resource data, and realize the intelligent scheduling of system resources. It solves the problems of low flexibility and ineffective utilization of system resources caused by static allocation of system resources, effectively improves system response speed, system resource utilization and system stability, especially in complex and high-load industrial environments, ensures that the system can operate stably and efficiently, reduces the risk of system failures caused by overload or insufficient resources, and enhances the reliability and scalability of the overall system.

[0097] Embodiment 2

[0098] Figure 2 1 is a flowchart of a data processing method provided by Embodiment 2 of the present invention. This embodiment is a preferred embodiment of the above embodiment. The specific implementation method can refer to the technical solution of this embodiment. Among them, the technical terms that are the same as or corresponding to the above embodiment are not repeated here. Figure 2 As shown, the method includes:

[0099] S210: During the execution of the target service, actual performance data of multiple performance reference indicators at the current moment are obtained.

[0100] Among them, the multiple performance reference indicators include at least two of the CPU instantaneous occupancy indicator, memory instantaneous usage indicator, disk read and write indicator, network average traffic indicator, system average response time indicator, process indicator, temperature and energy consumption indicator.

[0101] S220. For multiple performance reference indicators, determine data trend characteristics, data periodicity characteristics, and data change rate characteristics corresponding to the performance reference indicators based on actual performance data under the performance reference indicators and actual historical performance data corresponding to multiple historical moments before the current moment.

[0102] S230. Use principal component analysis to process data trend characteristics, data periodicity characteristics, and data change rate characteristics to obtain a covariance matrix.

[0103] Among them, principal component analysis is a data dimensionality reduction technology, which is used to convert data trend characteristics, data periodicity characteristics and data change rate characteristics into characteristic vectors in the principal component direction through linear transformation. The covariance matrix can be a characteristic matrix composed of data trend characteristics, data periodicity characteristics and data change rate characteristics under each performance reference indicator.

[0104] Specifically, the data trend characteristics, data periodicity characteristics and data change rate characteristics under each performance reference indicator are combined into a covariance matrix to be adjusted. Each element in the covariance matrix to be adjusted is adjusted according to a preset weight coefficient, and a covariance matrix is ​​obtained based on the adjusted elements.

[0105] For example, a weight coefficient may be assigned to each feature corresponding to each performance reference indicator. The weight coefficient may be calculated as follows:

[0106]

[0107] Among them, σ j (t) represents the standard deviation of the jth feature in the preset time t. The preset time t can be the time length between the current moment and the preset historical moment. D represents the number of performance reference indicators corresponding to the same feature. For example, the number of performance reference indicators corresponding to the same data trend feature.

[0108] The covariance matrix to be adjusted is determined using a rolling window. That is,

[0109]

[0110] Among them, Y(t) represents the standardized feature corresponding to the current time t, and Y(t) can be the standardized data trend feature, data periodicity feature, and data change rate feature. Y(t) Trepresents the transpose of Y(t). w Indicates the size of the rolling window. According to the weight coefficient w j (t) Perform weighted adjustment on the covariance matrix to be adjusted to obtain the adjusted covariance matrix Σ t '.

[0111] S240, by performing eigenvalue decomposition on the covariance matrix, eigenvalues ​​and eigenvectors are obtained, and data trend characteristics, data periodicity characteristics, and data change rate characteristics are projected into the principal component space based on the eigenvalues ​​and eigenvectors to obtain projected feature data.

[0112] The eigenvalue may represent the variance in the feature direction. The eigenvector is used to represent the feature direction. The principal component space may be the space corresponding to the principal component direction vector. The projected feature data may be the feature obtained by projecting the data trend feature, the data periodicity feature, and the data change rate feature into the principal component space.

[0113] Specifically, the covariance matrix is ​​subjected to eigenvalue decomposition to obtain eigenvalues ​​and eigenvectors. The eigenvector with the largest eigenvalue is selected as the principal component direction vector. The data trend feature, data periodicity feature, and data change rate feature are projected into the principal component space to obtain the projected feature data.

[0114] Exemplarily, in combination with the above example, the covariance matrix Σ t 'Perform eigenvalue decomposition to obtain eigenvalue λ i (t) and the eigenvector v i (t). That is,

[0115] Σ t 'v i (t) = λ i (t)v i (t)

[0116] Among them, v i (t) represents the i-th eigenvector at the current time t, λ i (t) is the i-th eigenvalue at the current time t.

[0117] The directions corresponding to the k eigenvectors with the largest eigenvalues ​​are taken as the principal component directions, and the data trend characteristics, data periodicity characteristics and data change rate characteristics are projected onto the principal component directions to obtain the projected feature data Z(t).

[0118] Z(t)=Y(t)v i '(t)

[0119] Among them, v i'(t) represents the principal component direction vector at the current time t, Y(t) represents the standardized feature corresponding to the current time t, and Z(t) represents the feature data after dimensionality reduction, that is, the feature data after projection.

[0120] By calculating the covariance matrix through a rolling window, the covariance structure of the covariance matrix can be updated in real time, thereby promptly reflecting the fluctuations of the load changes and equipment status of the power grid, thereby improving the universality and accuracy of the Dianhong operating system in allocating system resources.

[0121] S250, after nonlinear mapping of the projected feature data, a kernel matrix is ​​constructed, and the target feature is obtained by principal component analysis of the kernel matrix.

[0122] The elements in the kernel matrix are used to characterize the linear relationship between the projected feature data corresponding to the current moment and the projected feature data corresponding to the preset historical moment under each performance parameter indicator.

[0123] The preset historical moment may be a historical moment that is preset relative to the current moment.

[0124] Specifically, the projected feature data is mapped to a high-dimensional feature space through a nonlinear function to obtain the mapped feature data. A kernel matrix is ​​constructed based on the mapped feature data corresponding to the current moment and the mapped feature data corresponding to the preset historical moment. The kernel matrix is ​​subjected to principal component analysis to obtain the eigenvalues ​​and eigenvectors corresponding to the kernel matrix. According to the eigenvalues ​​and eigenvectors corresponding to the kernel matrix, the principal component direction corresponding to the kernel matrix is ​​determined to project the mapped feature data onto the principal component direction corresponding to the kernel matrix to obtain the target feature.

[0125] Exemplarily, the projected feature data is mapped to a high-dimensional feature space by the following nonlinear function to obtain the mapped feature data. That is,

[0126] φ(Z(t))=exp(-|(|Z(t)-z i |)| 2 / σ 2 )

[0127] Among them, Z(t) represents the feature data after projection, z i represents the sample point or center point in the pre-selected, projected feature data. σ represents the preset parameter used to control the distance metric in the high-dimensional feature space. φ(Z(t)) represents the mapped feature data.

[0128] The kernel matrix is ​​constructed by the following formula:

[0129] K ij =φ(Z(t i))·φ(Z(t j ))

[0130] Among them, φ(Z(t i )) represents the current time t i The corresponding mapped feature data, φ(Z(t j )) represents the preset historical moment t j Represents the mapped feature data corresponding to the preset historical moment, K ij Represents the elements in the kernel matrix, used to characterize Z(t i ) and Z(t j ) feature similarity.

[0131] The kernel matrix is ​​processed by principal component analysis using the following formula:

[0132] Kv i =λ i v i

[0133] Among them, K represents the kernel matrix, v i represents the eigenvector corresponding to the kernel matrix, λ i Represents the eigenvalue corresponding to the kernel matrix. The larger the eigenvalue, the higher the importance of the principal component. According to all the eigenvalues, the k eigenvectors with the largest eigenvalues ​​are selected as the principal component direction vectors to project the mapped feature data onto the principal component direction corresponding to the kernel matrix to obtain the target feature. By using nonlinear mapping to project the projected feature data into a high-dimensional feature space, it is possible to capture complex nonlinear relationships that cannot be discovered by traditional linear methods. By determining the eigenvalues ​​and eigenvectors corresponding to the kernel matrix, a more accurate principal component can be obtained, thereby optimizing the dimensionality reduction effect of the feature. The above has obvious effects on complex dynamic characteristics in the power grid (such as periodic load fluctuations, emergencies, etc.), and can improve the expression ability and prediction accuracy of the target performance parameter prediction model for complex system states.

[0134] S260: Based on the feature extraction layer in the target performance parameter prediction model, extract the features to be processed corresponding to the target features.

[0135] Among them, the feature extraction layer can be used to extract features from the target features. The extracted features are the features to be processed. Optionally, the feature extraction layer can be a shared feature extraction layer, which is used to extract common features between different performance reference indicators. The shared feature extraction layer can reduce the training time and computational overhead of the target performance parameter prediction model, and improve the generalization ability and prediction accuracy of the target performance parameter prediction model.

[0136] Specifically, the target features are processed by at least one convolution module and a fully connected module of the feature extraction layer in the target performance parameter prediction model to obtain features to be processed.

[0137] Optionally, the feature extraction layer may include a temporal convolution module, a fully connected module, and a regularization module. The target feature is convolved by the temporal convolution module. The feature output by the temporal convolution module is fully connected by the fully connected module. The feature output by the fully connected module is regularized by the regularization module to prevent overfitting of the target performance parameter prediction model.

[0138] Among them, the time series convolution module can include 64 nodes, which are used to receive the output of the previous neural network layer and generate the output of the current neural network layer. The time series convolution module can adopt the ReLU activation function to improve the nonlinear fitting ability of the target performance parameter prediction model and reduce overfitting. The fully connected module can include 128 nodes, each of which is connected to the node of the previous neural network layer, and is used to receive input features and generate output features. The fully connected module can adopt the ReLU activation function to alleviate the gradient vanishing problem of the target performance parameter prediction model and accelerate the training process of the target performance parameter prediction model. In the regularization module, that is, the Dropout module, the Dropout rate is set to 0.5, so that each node has a 50% probability of being discarded in each training iteration to improve the robustness of the target performance parameter prediction model.

[0139] S270 , processing the feature to be processed based on the processing branch corresponding to each performance reference indicator in the target performance parameter prediction model, and outputting predicted performance data corresponding to the corresponding performance reference indicator.

[0140] The processing branch may be a branch for predicting predicted performance data corresponding to the current performance reference indicator. The number of processing branches is consistent with the number of performance reference indicators. Optionally, the processing branch may include at least one fully connected layer and an output layer. The fully connected layer may include 64 nodes and use a ReLU activation function. The output layer may include one node and use a linear activation function.

[0141] Specifically, the features to be processed are input into the processing branches corresponding to each performance reference indicator in the target performance parameter prediction model, and the features to be processed are processed in sequence based on the fully connected layer in the processing branch and the activation function in the output layer to obtain the predicted performance data of the performance reference indicator corresponding to the processing branch.

[0142] Optionally, in the target performance parameter prediction model, the processing branch may include a processing branch corresponding to the system average response time indicator and a processing branch corresponding to other performance parameter indicators. Different processing branches contain different modules or neural network layers.

[0143] Optionally, when processing the feature to be processed according to the processing branch, the output result corresponding to the previous level is subjected to attention weighting processing based on the attention module in the processing branch to output predicted performance data.

[0144] Among them, the attention module is used to perform weighted processing on the output results of the previous level in the processing branch. The weight coefficients corresponding to the features of different performance reference indicators are different. For example, if the current processing branch is the processing branch corresponding to the CPU instantaneous occupancy indicator, the weight coefficient of the feature corresponding to the CPU instantaneous occupancy indicator can be adaptively increased. The output result corresponding to the previous level can be understood as the feature output by the previous module or the previous level in the attention module in the processing branch.

[0145] Specifically, based on the lightweight attention mechanism corresponding to the attention module in the processing branch, the output results corresponding to the previous layer are weighted to output the predicted performance data. Through the lightweight attention mechanism, weighted processing can be performed according to the relationship between different performance reference indicators. Especially in an environment with large resource competition and load fluctuations, the target performance parameter prediction model can dynamically adjust the weights according to the demand relationship of each performance reference indicator, thereby improving the modeling and processing capabilities in multiple performance reference indicator scenarios. The introduction of the attention mechanism allows the target performance parameter prediction model to automatically focus on the most important features for the current processing branch, improving the accuracy of the predicted performance data.

[0146] S280: Determine resource data allocated to the target service at the prediction time according to the predicted performance data and the actual performance data corresponding to each performance reference indicator.

[0147] The technical solution of this embodiment, in the process of executing the target business, obtains the actual performance data of multiple performance reference indicators at the current moment, and determines the data trend characteristics, data periodicity characteristics and data change rate characteristics corresponding to the performance reference indicators according to the actual performance data under each performance reference indicator and the actual historical performance data corresponding to multiple historical moments before the current moment. The principal component analysis method is used to process the data trend characteristics, data periodicity characteristics and data change rate characteristics to obtain a covariance matrix. By performing eigenvalue decomposition on the covariance matrix, eigenvalues ​​and eigenvectors are obtained, and the data trend characteristics, data periodicity characteristics and data change rate characteristics are projected into the principal component space based on the eigenvalues ​​and eigenvectors to obtain the projected feature data. After nonlinear mapping of the projected feature data, a kernel matrix is ​​constructed, and the target feature is obtained by principal component analysis of the kernel matrix, which provides accurate input features for subsequent prediction processing. Based on the feature extraction layer in the target performance parameter prediction model, the to-be-processed features corresponding to the target feature are extracted. Based on the processing branch corresponding to each performance reference indicator in the target performance parameter prediction model, the to-be-processed features are processed, and the predicted performance data corresponding to the corresponding performance reference indicator is output. According to the predicted performance data and actual performance data corresponding to each performance reference indicator, the resource data allocated to the target business at the predicted time is determined, so as to realize the resource allocation processing of the target business through the resource data, and realize the intelligent scheduling of system resources. It solves the problem of low flexibility and ineffective utilization of system resources caused by static allocation of system resources, effectively improves the system response speed, system resource utilization and system stability, especially in complex and high-load industrial environments, ensures that the Dianhong operating system can run stably and efficiently, reduces the risk of system failures caused by overload or insufficient resources, and enhances the reliability and scalability of the overall system.

[0148] Embodiment 3

[0149] Figure 3 Schematic diagram of the structure of a data processing device provided by Embodiment 3 of the present invention. Figure 3 As shown, the device includes: a data acquisition module 310, a feature determination module 320, a feature analysis module 330, a performance data prediction module 340 and a resource data determination module 350.

[0150] The data acquisition module 310 is used to acquire the actual performance data of multiple performance reference indicators at the current moment during the execution of the target business, wherein the multiple performance reference indicators include at least two of the CPU instantaneous occupancy index, the memory instantaneous usage index, the disk read / write index, the network average traffic index, the system average response time index, the process index, and the temperature and energy consumption index; the feature determination module 320 is used to determine the data trend features and data cycle corresponding to the performance reference indicators based on the actual performance data under the performance reference indicators and the actual historical performance data corresponding to multiple historical moments before the current moment. The module 330 is used to analyze and process the data trend characteristics, data periodicity characteristics and data change rate characteristics corresponding to each performance reference indicator to obtain the target characteristics to be input into the target performance parameter prediction model; the performance data prediction module 340 is used to input the target characteristics into the target performance parameter prediction model, and output the predicted performance data corresponding to multiple performance reference indicators at the prediction time; the resource data determination module 350 is used to determine the resource data allocated to the target business at the prediction time according to the predicted performance data and actual performance data corresponding to each performance reference indicator.

[0151] The technical solution of this embodiment obtains the actual performance data of multiple performance reference indicators at the current moment during the execution of the target business, and determines the data trend characteristics, data periodicity characteristics and data change rate characteristics corresponding to the performance reference indicators based on the actual performance data under each performance reference indicator and the actual historical performance data corresponding to multiple historical moments before the current moment. By analyzing and processing the data trend characteristics, data periodicity characteristics and data change rate characteristics corresponding to each performance reference indicator, the target characteristics to be input into the target performance parameter prediction model are obtained, providing accurate input characteristics for subsequent prediction processing. The target characteristics are input into the target performance parameter prediction model, and the predicted performance data corresponding to multiple performance reference indicators at the prediction moment are output. According to the predicted performance data and actual performance data corresponding to each performance reference indicator, the resource data allocated to the target business at the prediction moment is determined, so as to realize the resource allocation processing of the target business through the resource data, and realize the intelligent scheduling of system resources. It solves the problems of low flexibility and ineffective utilization of system resources caused by static allocation of system resources, effectively improves system response speed, system resource utilization and system stability, especially in complex and high-load industrial environments, ensures that the system can operate stably and efficiently, reduces the risk of system failures caused by overload or insufficient resources, and enhances the reliability and scalability of the overall system.

[0152] On the basis of the above embodiment, optionally, the feature determination module includes a trend feature determination unit, which is used to obtain the actual historical performance data under the performance reference index for multiple historical moments before the current moment based on the number of data corresponding to the sliding window; based on the actual historical performance data, the actual performance data and a preset linear fitting function, determine the data trend characteristics corresponding to the current moment.

[0153] Optionally, the feature determination module includes a periodic feature determination unit, which is used to determine the data periodic characteristics corresponding to the current moment based on the actual performance data under the performance reference indicator, the actual historical performance data corresponding to multiple historical moments before the current moment, a pre-set data quantity and a frequency domain processing function.

[0154] Optionally, the feature determination module includes a change rate feature determination unit, which is used to determine the data change rate feature corresponding to the current moment based on the actual performance data at the current moment under the performance reference indicator, the actual historical performance data corresponding to the previous historical moment of the current moment, and the instantaneous change rate determination function.

[0155] Optionally, a feature analysis module is used to process data trend features, data periodicity features and data change rate features using principal component analysis to obtain a covariance matrix; obtain eigenvalues ​​and eigenvectors by performing eigenvalue decomposition on the covariance matrix, and project the data trend features, data periodicity features and data change rate features into the principal component space based on the eigenvalues ​​and eigenvectors to obtain projected feature data; construct a kernel matrix after nonlinear mapping of the projected feature data, and obtain target features through principal component analysis of the kernel matrix, wherein the elements in the kernel matrix are used to characterize the linear relationship between the projected feature data corresponding to the current moment under each performance parameter indicator and the projected feature data corresponding to a preset historical moment.

[0156] Optionally, the performance data prediction module includes: a unit for determining features to be processed, which is used to extract features to be processed corresponding to the target features based on the feature extraction layer in the target performance parameter prediction model; a performance data output unit, which is used to process the features to be processed based on the processing branches corresponding to each performance reference indicator in the target performance parameter prediction model, and output predicted performance data corresponding to the corresponding performance reference indicator.

[0157] Optionally, a performance data output unit is used to perform attention-weighted processing on the output result corresponding to the previous level based on the attention module in the processing branch to output predicted performance data.

[0158] Optionally, a resource data determination module is used to determine performance difference data for multiple performance reference indicators based on predicted performance data and actual performance data corresponding to the performance reference indicators; and adjust the performance difference data to determine the resource data allocated to the target business under the performance reference indicators based on the actual performance data and the adjusted performance difference data.

[0159] The data processing device provided in the embodiment of the present invention can execute the data processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0160] Embodiment 4

[0161] Figure 4 1 is a schematic diagram of the structure of an electronic device provided in Embodiment 4 of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0162] like Figure 4 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0163] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0164] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 performs the various methods and processes described above, such as a data processing method.

[0165] In some embodiments, the data processing method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the data processing method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the data processing method in any other appropriate manner (e.g., by means of firmware).

[0166] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0167] The computer program for implementing the data processing method of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program can be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0168] Embodiment 5

[0169] Embodiment 5 of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to cause a processor to execute a data processing method, the method comprising:

[0170] During the execution of the target business, actual performance data of multiple performance reference indicators at the current moment are obtained, wherein the multiple performance reference indicators include at least two of a CPU instantaneous occupancy indicator, a memory instantaneous usage indicator, a disk read / write indicator, a network average traffic indicator, a system average response time indicator, a process indicator, and a temperature and energy consumption indicator; for multiple performance reference indicators, data trend characteristics, data periodicity characteristics, and data change rate characteristics corresponding to the performance reference indicators are determined based on the actual performance data under the performance reference indicators and the actual historical performance data corresponding to multiple historical moments before the current moment; target characteristics to be input into a target performance parameter prediction model are obtained by analyzing and processing the data trend characteristics, data periodicity characteristics, and data change rate characteristics corresponding to each performance reference indicator; the target characteristics are input into the target performance parameter prediction model, and predicted performance data corresponding to multiple performance reference indicators at the prediction moment are output; and resource data allocated to the target business at the prediction moment is determined based on the predicted performance data and actual performance data corresponding to each performance reference indicator.

[0171] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, device, or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0172] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0173] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0174] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0175] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0176] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A data processing method, characterized in that: include: During the execution of the target service, actual performance data of multiple performance reference indicators at the current moment are obtained, wherein the multiple performance reference indicators include at least two of a CPU instantaneous occupancy indicator, a memory instantaneous usage indicator, a disk read / write indicator, a network average flow indicator, a system average response time indicator, a process indicator, and a temperature and energy consumption indicator; For the multiple performance reference indicators, determine the data trend characteristics, data periodicity characteristics, and data change rate characteristics corresponding to the performance reference indicators based on the actual performance data under the performance reference indicators and the actual historical performance data corresponding to multiple historical moments before the current moment; By analyzing and processing the data trend characteristics, data periodicity characteristics and data change rate characteristics corresponding to each performance reference indicator, the target characteristics to be input into the target performance parameter prediction model are obtained; Inputting the target feature into the target performance parameter prediction model, and outputting the predicted performance data corresponding to the multiple performance reference indicators at the prediction time; The resource data allocated to the target service at the prediction moment is determined according to the predicted performance data and the actual performance data corresponding to each performance reference indicator.

2. The method according to claim 1, characterized in that: The determining of the data trend feature corresponding to the performance reference indicator based on the actual performance data under the performance reference indicator and the actual historical performance data corresponding to multiple historical moments before the current moment includes: According to the amount of data corresponding to the sliding window, obtain actual historical performance data under the performance reference indicator at multiple historical moments before the current moment; Based on the actual historical performance data, the actual performance data and a preset linear fitting function, a data trend feature corresponding to the current moment is determined.

3. The method according to claim 1, characterized in that The determining of the data periodicity characteristics corresponding to the performance reference indicator according to the actual performance data under the performance reference indicator and the actual historical performance data corresponding to a plurality of historical moments before the current moment includes: The data periodicity characteristics corresponding to the current moment are determined based on the actual performance data under the performance reference index, the actual historical performance data corresponding to multiple historical moments before the current moment, a preset data quantity and a frequency domain processing function.

4. The method according to claim 1, characterized in that: The determining of the data change rate characteristic corresponding to the current moment according to the actual performance data under the performance reference indicator and the actual historical performance data corresponding to multiple historical moments before the current moment includes: The data change rate characteristics corresponding to the current moment are determined based on the actual performance data at the current moment under the performance reference index, the actual historical performance data corresponding to the previous historical moment of the current moment, and the instantaneous change rate determination function.

5. The method according to claim 1, characterized in that The target features to be input into the target performance parameter prediction model are obtained by analyzing and processing the data trend features, data periodicity features, and data change rate features corresponding to each performance reference indicator, including: The data trend characteristics, the data periodicity characteristics and the data change rate characteristics are processed by principal component analysis to obtain a covariance matrix; By performing eigenvalue decomposition on the covariance matrix, eigenvalues ​​and eigenvectors are obtained, and based on the eigenvalues ​​and eigenvectors, the data trend feature, the data periodicity feature, and the data change rate feature are projected into a principal component space to obtain projected feature data; After nonlinear mapping of the projected feature data, a kernel matrix is ​​constructed, and the target feature is obtained by principal component analysis of the kernel matrix, wherein the elements in the kernel matrix are used to characterize the linear relationship between the projected feature data corresponding to the current moment under each performance parameter indicator and the projected feature data corresponding to the preset historical moment.

6. The method according to claim 1, characterized in that The step of inputting the target feature into the target performance parameter prediction model and outputting the predicted performance data corresponding to the multiple performance reference indicators at the prediction moment includes: Extracting the to-be-processed features corresponding to the target features based on the feature extraction layer in the target performance parameter prediction model; The feature to be processed is processed based on the processing branch corresponding to each performance reference indicator in the target performance parameter prediction model, and the predicted performance data corresponding to the corresponding performance reference indicator is output.

7. The method according to claim 6, characterized in that The feature to be processed is processed based on the processing branch corresponding to each performance reference indicator, and the predicted performance data corresponding to the corresponding performance reference indicator is output, including: Based on the attention module in the processing branch, the output result corresponding to the previous level is subjected to attention weighting processing to output the predicted performance data.

8. The method according to claim 1, characterized in that: The determining, according to the predicted performance data and the actual performance data corresponding to each performance reference indicator, the resource data allocated to the target service at the predicted time comprises: For the multiple performance reference indicators, determining performance difference data based on the predicted performance data and the actual performance data corresponding to the performance reference indicators; The performance difference data is adjusted to determine the resource data allocated to the target service under the performance reference indicator based on the actual performance data and the adjusted performance difference data.

9. A data processing device, characterized in that: include: A data acquisition module, used to acquire actual performance data of multiple performance reference indicators at the current moment during the execution of the target business, wherein the multiple performance reference indicators include at least two of a CPU instantaneous occupancy indicator, a memory instantaneous usage indicator, a disk read / write indicator, a network average flow indicator, a system average response time indicator, a process indicator, and a temperature and energy consumption indicator; A feature determination module, for determining, for the multiple performance reference indicators, data trend features, data periodicity features, and data change rate features corresponding to the performance reference indicators based on actual performance data under the performance reference indicators and actual historical performance data corresponding to multiple historical moments before the current moment; A feature analysis module is used to obtain target features to be input into the target performance parameter prediction model by analyzing and processing the data trend features, data periodicity features and data change rate features corresponding to each performance reference indicator; A performance data prediction module, used to input the target feature into the target performance parameter prediction model, and output the predicted performance data corresponding to the multiple performance reference indicators at the prediction time; The resource data determination module is used to determine the resource data allocated to the target business at the prediction moment according to the predicted performance data and the actual performance data corresponding to each performance reference indicator.

10. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the data processing method according to any one of claims 1 to 8.

Images