A cloud computing resource prediction method, system, device and storage medium
By combining neural networks and autoregressive time series forecasting with control theory to dynamically adjust cloud computing resources, the problem of uneven allocation of cloud computing resources in rail transit is solved, thereby improving load utilization and system performance.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA RAILWAY SIYUAN SURVEY & DESIGN GRP CO LTD
- Filing Date
- 2022-11-08
- Publication Date
- 2026-07-07
AI Technical Summary
The existing cloud computing resource allocation method for rail transit cannot track changes in business load in real time, resulting in computing resources being unable to meet the normal operation of the load, affecting system performance and service quality, and the resource allocation is uneven.
Load changes are predicted using neural networks and autoregressive time series forecasting methods. Resource allocation is dynamically adjusted by combining control theory. A resource allocation system is constructed using a resource analyzer, predictor, adjuster, and filter. Load data is processed using Z-transform to achieve dynamic and adjustable resource allocation.
It improved the utilization of load resources, reduced latency, enhanced system security and virtual machine performance, and ensured the real-time performance and stability of the system.
Smart Images
Figure CN115686856B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of cloud computing technology, and more specifically, to a cloud computing resource prediction method, system, device, and storage medium. Background Technology
[0002] Currently, resource allocation methods in private cloud computing for rail transit can be broadly categorized into two types: non-resource prediction allocation and resource prediction allocation. Non-resource prediction allocation includes the most direct average allocation method, which fails to consider load differences between different virtual machines (VMs) and the varying load conditions of each VM at different times. The other type is threshold-based dynamic resource allocation. This method faces challenges in determining suitable upper and lower bounds for resource utilization for each VM; because resource allocation is only adjusted when significant load changes are detected, this method introduces a time delay and cannot effectively track real-time changes in rail transit service load.
[0003] Existing urban rail cloud platforms often suffer from insufficient computing resources to meet normal load conditions, impacting system performance and potentially failing to meet the Quality of Service (QoS) requirements of cloud computing users. Furthermore, the load balancing of urban rail cloud platforms is insufficient. This invention employs time-series forecasting methods, including neural networks and autoregressive methods, to predict load changes in urban rail cloud platform applications over a specific time period. It also applies control theory to the resource allocation system, making cloud resource allocation a dynamically adjustable process. This increases resource utilization, improves the performance of virtual machines in the urban rail transit cloud, reduces latency in urban rail cloud service execution, and enhances overall system security. Summary of the Invention
[0004] To address at least one deficiency or improvement need in the prior art, the present invention provides a cloud computing resource prediction method and system, which improves the prediction accuracy of load type and the utilization rate of cloud platform computing resources.
[0005] To achieve the above objectives, according to a first aspect of the present invention, a cloud computing resource prediction method is provided, the method comprising:
[0006] Obtain load data for different business types on the cloud platform, analyze the load data to obtain the predicted business type of the load data, and obtain the corresponding load resource allocation scheme based on the predicted business type.
[0007] A resource allocation model is constructed based on the predicted business type and the load resource allocation scheme of the load data. The predicted business type is input into the resource allocation model to obtain the corrected predicted business type.
[0008] Error processing is performed on the corrected predicted service types to obtain the final load resource allocation scheme.
[0009] Furthermore, the aforementioned cloud computing resource prediction method, wherein obtaining load data for different business types of the cloud platform, analyzing the load data to obtain the predicted business type of the load data, and obtaining the corresponding load resource allocation scheme based on the predicted business type, specifically includes:
[0010] Build and train a resource analysis model based on historical load data from the cloud platform;
[0011] Obtain the discrete load time series S of the cloud platform T The discrete load time series S T Input the discrete load time series S into the trained resource analysis model T Analysis was performed to obtain the time period T. p The predicted sequence ξ for this load service type i ;
[0012] Based on the time period T p Prediction sequence ξ of internal load service type i Calculate the time period T p Internal load resource allocation scheme sequence ξ o .
[0013] Furthermore, in the aforementioned cloud computing resource prediction method, the step of constructing a resource allocation model based on the predicted business type and load resource allocation scheme of the load data specifically includes:
[0014] The time period T is calculated according to the following formula (1). p Prediction sequence ξ of internal load service type i With time period T p Internal load resource allocation scheme sequence ξ o The difference,
[0015] ε l [T p ]=ξ i [T p ]-ξ o [T p (1)
[0016] The resource allocation model δC is constructed based on the linear function of equation (2). l [T p ],
[0017] δC l [T p ]=β0+β1(ε l [T p])+μ i (2)
[0018] Where δ is a constant, and 0≤δ≤1.
[0019] Furthermore, in the aforementioned cloud computing resource prediction method, the step of inputting the predicted service type into the resource allocation model to obtain the corrected predicted service type specifically includes:
[0020] By allocating model δC l [T p Based on the allocation result C output by the resource allocation model in the previous time period l [T p-1 For time period T p The allocation results within are corrected, specifically as shown in equation (3).
[0021] C l [T p ] = C l [T p-1 ]+δC l [T p (3).
[0022] Furthermore, in the aforementioned cloud computing resource prediction method, the step of performing error processing on the corrected predicted service type to obtain the final load resource allocation scheme specifically includes:
[0023] The corrected predicted business type is input into the smoothing processor for error processing according to the following formula (4).
[0024] Q l [T p ]=ωQ l [T p-1 ]+(1-ω)C l [T p (4)
[0025] Where ω is a constant, and 0≤ω≤1;
[0026] The output Q of the smoothing processor l [T p Summing these values yields the final load resource allocation scheme.
[0027] Furthermore, the aforementioned cloud computing resource prediction method also includes;
[0028] Define the discrete load time series S T Analysis was performed to obtain the time period T. p The predicted sequence ξ for this load service type iThe Z-domain function corresponding to the process is fitted as F(Z), and the Z-transform form of equation (1) is obtained according to the Z-transform rule as Z. -1 The Z-transform form of Equation 3 is as follows: The Z-transform form of Equation 4 is as follows: For the Q l [T p The Z-transform corresponding to the summation is:
[0029] Based on the Z-transform forms of Equation (1), Equation (3), and Equation (4), the function G(z) is defined as shown in Equation (5) below.
[0030]
[0031] The final load resource allocation scheme is obtained based on the functions F(Z) and G(z), as shown in equation (6) below.
[0032] ξ o [T p ]=ε l F(Z)G(Z) (6)
[0033] At this time, the discrete load time series S input to the system T [T p ] and the output load resource allocation scheme sequence ξ o [T p The relationship between them is shown in equation (7) below.
[0034]
[0035] According to a second aspect of the present invention, a cloud computing resource prediction system is also provided, comprising a resource analyzer, a resource predictor, a resource adjuster, a smoothing filter, and a resource allocator connected in sequence; wherein,
[0036] The resource analyzer is used to acquire load data for different business types on the cloud platform, analyze the load data, and obtain the predicted business type based on the load data.
[0037] The resource predictor is used to construct a resource allocation model based on the predicted service type and the load resource allocation scheme of the load data. The predicted service type is input into the resource allocation model to obtain the corrected predicted service type.
[0038] The resource adjuster and smoothing filter are used to perform error processing on the corrected predicted service type.
[0039] The resource allocator is used to obtain a load resource allocation scheme based on the predicted service type after error processing.
[0040] Furthermore, the aforementioned cloud computing resource prediction system also includes a Z-transform module. The resource analyzer, resource predictor, resource adjuster, smoothing filter, and resource allocator each include at least one Z-transform module. The Z-transform module is used to perform Z-transform processing on the processing results of the resource analyzer, the resource predictor, the resource adjuster, the smoothing filter, and the resource allocator, respectively.
[0041] According to a third aspect of the present invention, a cloud computing resource prediction device is also provided, comprising at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program that, when executed by the processing unit, causes the processing unit to perform the steps of any of the methods described above.
[0042] According to a fourth aspect of the invention, a storage medium is also provided, which stores a computer program executable by a cloud computing resource prediction device, which, when run on the cloud computing resource prediction device, causes the cloud computing resource prediction device to perform the steps of any of the methods described above.
[0043] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects:
[0044] (1) The cloud computing resource prediction method and system provided by the present invention analyzes the load data of different business types of the cloud platform to obtain the predicted business type of the load data, obtains the corresponding load resource allocation scheme according to the predicted business type, and corrects the predicted business type of the load data through the load resource allocation scheme. Through the dynamic adjustment of the feedback loop, the accuracy of load prediction and the utilization rate of cloud platform computing resources are improved.
[0045] (2) By using the cloud computing resource prediction method and system provided by the present invention, control theory is applied to the resource allocation system, making cloud resource allocation a dynamic and adjustable process, which greatly increases the resource utilization rate of the load, improves the performance of virtual machines in the cloud, reduces the latency of cloud service execution, and improves the overall system security. Attached Figure Description
[0046] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0047] Figure 1A flowchart illustrating a cloud computing resource prediction method provided in an embodiment of this application;
[0048] Figure 2 This is a connection diagram of a cloud computing resource prediction system provided in an embodiment of this application;
[0049] Figure 3 This is a schematic diagram of Z-transform processing provided in an embodiment of this application;
[0050] Figure 4 This is a schematic diagram comparing the computing resource requirements and the predicted situation over a certain period of time, as provided in an embodiment of this application. Detailed Implementation
[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.
[0052] The terms "first," "second," "third," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.
[0053] On the one hand, this application provides a method for predicting cloud computing resources. Figure 1 Please refer to the flowchart of a cloud computing resource prediction method provided in this application embodiment. Figure 1 The cloud computing resource prediction method provided in this application can better address the problem of unreasonable resource allocation in urban rail transit cloud environments and improve the performance of systems within virtual machines on urban rail transit platforms. It can reduce application execution latency and, while ensuring the normal operation of the urban rail transit system, further improve the real-time performance and security of the system's services.
[0054] The method includes the following steps:
[0055] (1) Obtain load data for different business types on the cloud platform, analyze the load data, obtain the predicted business type of the load data, and obtain the corresponding load resource allocation scheme based on the predicted business type.
[0056] Specifically, a resource analysis model is built and trained based on the historical load data of the cloud platform;
[0057] Obtain the discrete load time series S of the cloud platform T The discrete load time series S T Input the discrete load time series S into the trained resource analysis model T Analysis was performed to obtain the time period T. p The predicted sequence ξ for this load service type i ;
[0058] Based on the time period T p Prediction sequence ξ of internal load service type i Calculate the time period T p Internal load resource allocation scheme sequence ξ o .
[0059] (2) Construct a resource allocation model based on the predicted business type and the load resource allocation scheme of the load data, input the predicted business type into the resource allocation model, and obtain the corrected predicted business type.
[0060] The time period T is calculated according to the following formula (1). p Prediction sequence ξ of internal load service type i With time period T p Internal load resource allocation scheme sequence ξ o The difference,
[0061] ε l [T p ]=ξ i [T p ]-ξ o [T p (1)
[0062] The resource allocation model δC is constructed based on the linear function of equation (2). l [T p ],
[0063] δC l [T p ]=β0+β1(ε l [T p ])+μ i (2)
[0064] Where δ is a constant, and 0≤δ≤1.
[0065] By allocating model δC l [T p Based on the allocation result C output by the resource allocation model in the previous time period l [T p-1For time period T p The allocation results within the scope are fed back and corrected, as shown in equation (3) below.
[0066] C l [T p ] = C l [T p-1 ]+δC l [T p (3).
[0067] (3) Perform error processing on the corrected predicted service types to obtain the final load resource allocation scheme.
[0068] Specifically, the corrected predicted service type is input into the smoothing processor for error processing according to the following formula (4).
[0069] Q l [T p ]=ωQ l [T p-1 ]+(1-ω)C l [T p (4)
[0070] Where ω is a constant, and 0≤ω≤1;
[0071] The output Q of the smoothing processor l [T p Summing these values yields the final load resource allocation scheme.
[0072] The Z-transform is a widely used classical technique. It's a mathematical transformation of discrete sequences, converting time-domain signals (i.e., discrete-time sequences) into expressions in the complex frequency domain. The Z-transform of discrete-time signals is an important tool for analyzing linear time-invariant discrete-time systems. It transforms the time-domain mathematical model of a linear time-invariant discrete system—the difference equation—into an algebraic equation in the Z-domain, simplifying the analysis of discrete systems. Furthermore, it allows the analysis of the system's time-domain characteristics, frequency response, and stability using system functions.
[0073] Further, a Z-transform is performed; specifically, the discrete load time series S is defined as follows. T Analysis was performed to obtain the time period T. p The predicted sequence ξ for this load service type i The Z-domain function corresponding to the process is fitted as F(Z), where F(Z) is a function obtained through neural network fitting and cannot be expressed as a specific function expression. According to the Z-transform rule, the Z-transform form of equation (1) is Z... -1The Z-transform form of Equation 3 is as follows: The Z-transform form of Equation 4 is as follows: For the Q l [T p The Z-transform corresponding to the summation is:
[0074] Based on the Z-transform forms of Equation (1), Equation (3), and Equation (4), the function G(z) is defined as shown in Equation (5) below.
[0075]
[0076] The final load resource allocation scheme is obtained based on the functions F(Z) and G(z), as shown in equation (6) below.
[0077] ξ o [T p ]=ε l F(Z)G(Z) (6)
[0078] At this time, the discrete load time series S input to the system T [T p ] and the output load resource allocation scheme sequence ξ o [T p The relationship between them is shown in equation (7) below.
[0079]
[0080] On the other hand, this application provides a cloud computing resource prediction system. Figure 2 This is a connection diagram of a cloud computing resource prediction system provided in an embodiment of this application. The system includes a resource analyzer, a resource predictor, a resource adjuster, a smoothing filter, and a resource allocator connected in sequence.
[0081] The resource analyzer identifies the type of each application load based on the characteristics of historical load data for different businesses, and selects appropriate resource analysis models for resource prediction. These models are trained using neural networks based on historical data for each type of business load. Since model matching may introduce some error, an error precision is set; matching errors less than the precision are considered successful business load type matches. Furthermore, during the predictor's matching process, multiple models similar to the current load may appear; in this case, the model with the highest probability is designated as the current business load mode. During cloud platform operation, the resource analyzer outputs the load type analysis results—the predicted business type of the load data—to the resource predictor in real time.
[0082] The cloud resource predictor includes a resource allocation model, which is constructed by linearly fitting the difference between the predicted business type of the load data and the load resource allocation scheme. This model is used to correct the prediction error based on the final output load resource allocation scheme.
[0083] The cloud resource adjuster is used to correct errors in the prediction results of the resource predictor. If the predicted resource allocation value is greater than the actual required resource allocation value, it will result in the waste of some resources; if the predicted resource allocation value is less than the actual required resource allocation value, it will lead to a decrease in system performance, an increase in service latency, and other consequences that reduce QoS. Therefore, a cloud resource adjuster is added to the system loop to correct the errors.
[0084] At this point, the adjusted value output by the cloud resource adjuster may still have an error compared to the expected allocation result. A low-pass smoothing filter is introduced into the system to eliminate the error.
[0085] The resource allocator allocates resources to different loads running in each virtual machine according to the corresponding matching load type in each cycle, and outputs the resource allocation scheme.
[0086] Furthermore, the cloud computing resource prediction system also includes a Z-transform module. Figure 3 The diagram illustrates Z-transform processing provided in this application embodiment. The resource analyzer, resource predictor, resource adjuster, smoothing filter, and resource allocator each include at least one Z-transform module. The Z-transform module is used to perform Z-transform processing on the processing results of the resource analyzer, resource predictor, resource adjuster, smoothing filter, and resource allocator, respectively.
[0087] Figure 4This illustration, provided as an embodiment of the present application, compares the computational resource demand with the predicted situation over a specific period. In one specific embodiment, taking an urban rail transit cloud platform as an example, multiple services typically run within the urban rail transit cloud. For the CBTC (Combined Train Control) system, ATS (Automatic Train Service) and intelligent signal maintenance and monitoring services are placed in the cloud. For the internal management network domain of the urban rail transit system, human resources systems, integrated safety systems, and document management systems can all be migrated to the cloud. For the external service domain of the urban rail transit system, the urban rail transit enterprise portal system, internet ticketing system, and video surveillance system are typically migrated to the cloud. For the aforementioned series of urban rail transit services, the application load of each service exhibits different patterns of change in time and space. Temporally, it is usually divided into peak and off-peak periods; spatially, it generally manifests as different application load conditions in different sections or stations. Directly predicting the load for the next time period and allocating resources for the loads of these different services is relatively difficult. Therefore, this application first analyzes the types of loads. Different service loads have different patterns of change. Therefore, based on the historical change patterns of different service loads, the specific types of loads are analyzed, and the load types and load change sequences are passed to the resource predictor.
[0088] Specifically, the system predicts resource usage for the next stage under different load models using a trained model, generating a resource allocation prediction sequence. For computationally intensive, resource-intensive, and relatively stable services, such as ATS (Automatic Train Service) and ticketing / sales services, more resources need to be allocated. In the next stage, a resource adjuster and smoothing filter adjust the resource allocation prediction sequence and feed the output resource allocation scheme sequence back to the predictor through a control loop, achieving real-time adjustment of the resource allocation sequence. This system predicts resource usage for the next time period based on the load situation of the current time period, thus ensuring real-time performance and significantly improving overall system performance. Figure 4 This illustration shows a comparison between computing resource demand and prediction over a specific period, as provided in an embodiment of this application. The computing resource demand curve for a cloud service system is a solid curve, while the dashed curve represents the computing resource allocation curve predicted using the resource prediction system provided in this application. The curves demonstrate that the predictor, adjuster, filter, and allocator work together in the computing resource allocation process to accurately track the resource demand for this service over a given period, with minimal error between the allocation curve and the demand curve. Furthermore, both the allocation curve and the demand curve exhibit dynamic changes. The dynamic adjustment following the computing resource prediction is achieved through a control loop, realizing the desired effect.
[0089] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method. The computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, microdrives, as well as magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic cards or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.
[0090] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.
[0091] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.
[0092] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some service interface; the indirect coupling or communication connection between devices or units may be electrical or other forms.
[0093] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0094] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0095] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.
[0096] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.
[0097] The foregoing description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Those skilled in the art will readily conceive of embodiments of this disclosure upon considering the specification and practicing the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described herein. The specification and embodiments are to be considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.
[0098] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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 specification.
[0099] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A cloud computing resource prediction method, characterized in that, include: The process involves acquiring load data for different business types on the cloud platform, analyzing the load data to determine the predicted business types, and then developing corresponding load resource allocation schemes based on these predicted business types. Specifically, this includes: building and training a resource analysis model based on historical load data from the cloud platform; and obtaining discrete load time series data from the cloud platform. Discrete load time series Input the discrete load time series into the trained resource analysis model Analysis was performed to obtain the time period. Predicted sequence of this load service type According to the time period Prediction sequence of internal load service types Calculation time period Internal load resource allocation scheme sequence Define the discrete load time series. Analysis was performed to obtain the time period. Predicted sequence of this load service type The process corresponding to Domain function fitting to function ; Based on the predicted business type and load resource allocation scheme of the load data, a resource allocation model is constructed. The predicted business type is input into the resource allocation model to obtain the corrected predicted business type, which specifically includes: according to formula (1) Calculation time period Prediction sequence of internal load service types With load resource allocation scheme sequence The difference; according to equation (2) Constructing a resource allocation model using linear functions Through resource allocation model Based on the allocation results output by the resource allocation model in the previous time period For time period The allocation results within are corrected, as shown in equation (3). ;in, It is a constant, and 0 ≤ ≤1; According to equation (4) The corrected predicted business type is input into the smoothing processor for error processing, and the output of the smoothing processor is... Summation; where, It is a constant, and 0 ≤ ≤1; according to The transformation rule yields equation (1). Transformation form is Equation (3) Transformation form is Equation (4) Transformation form is ,right Summation corresponding Transform into Based on equation (1) Transformation of form, equation (3) Transformation of form and equation (4) Functions defined in different forms Thus, we obtain equation (5). ; Based on functions AND function The final load resource allocation scheme is obtained as equation (6). At this time, the discrete load time series input to the discrete system With the output load resource allocation scheme sequence The relationship is given by equation (7). .
2. A cloud computing resource prediction system, used to execute the steps of the cloud computing resource prediction method as described in claim 1, characterized in that, It includes a resource analyzer, a resource predictor, a resource adjuster, a smoothing filter, and a resource allocator connected in sequence; among them, The resource analyzer is used to acquire load data for different business types on the cloud platform, analyze the load data, and obtain the predicted business type based on the load data. The resource predictor is used to construct a resource allocation model based on the predicted service type and the load resource allocation scheme of the load data. The predicted service type is input into the resource allocation model to obtain the corrected predicted service type. The resource adjuster and smoothing filter are used to process errors in the corrected predicted service type. The resource allocator is used to obtain a load resource allocation scheme based on the predicted service type after error processing.
3. The cloud computing resource prediction system as described in claim 2, wherein, Also includes The transformation module includes at least one of the following: resource analyzer, resource predictor, resource adjuster, smoothing filter, and resource allocator. The transformation module, the The transformation module is used to perform transformations on the processing results of the resource analyzer, the resource predictor, the resource adjuster, the smoothing filter, and the resource allocator. Transformation processing.
4. A cloud computing resource prediction device, characterized in that, It includes at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program that, when executed by the processing unit, causes the processing unit to perform the steps of the method of claim 1.
5. A storage medium, characterized in that, It stores a computer program that can be executed by a cloud computing resource prediction device, which, when run on the cloud computing resource prediction device, causes the cloud computing resource prediction device to perform the steps of the method of claim 1.