Computing power leasing management method, device and storage medium

Abstract

The present application discloses a computing power leasing management method, device, and storage medium. The present application relates to the field of digital data processing technology. The method includes: receiving a computing power leasing form, determining the task type, required duration, and computing power specifications corresponding to the computing power leasing form; determining the user's computing power demand pattern based on a time series based on the historical computing power usage record of the user corresponding to the computing power leasing form; determining computing power analysis data associated with the task type based on platform historical data; calculating a first computing power requirement based on the task type, required duration, computing power specifications, and computing power demand pattern; and calculating a second computing power requirement based on the task type, required duration, computing power specifications, and computing power analysis data; and determining a computing power output in response to the computing power leasing form based on the first computing power requirement and the second computing power requirement, as well as a dynamic weight associated with the date. The present application can achieve the technical effect of reducing computing power usage costs.

Description

Technical Field

[0001] The present application relates to the field of digital data processing technology, and in particular to a computing power leasing management method, device and storage medium. Background Art

[0002] With the rapid development of artificial intelligence, large AI models have been deeply integrated into numerous business scenarios, including intelligent customer service, image recognition, and natural language processing. These large AI models place extremely high demands on computing power, both during model training and actual inference. However, the computing power requirements of many businesses fluctuate significantly during operation, and are not always consistently high. For example, during major promotional events on e-commerce platforms, the computing power requirements of their intelligent recommendation systems can surge, only to decline significantly during daily operations.

[0003] In order to meet the computing power requirements during business peaks, companies need to purchase a large number of high-end hardware graphics cards. The high hardware procurement costs and subsequent operation and maintenance and electricity consumption result in excessively high computing power costs. Summary of the Invention

[0004] The main purpose of this application is to provide a computing power leasing management method, device and storage medium, aiming to solve the technical problem of excessively high computing power costs caused by hardware procurement costs and subsequent operation and maintenance and power consumption.

[0005] To achieve the above objectives, the present application provides a computing power leasing management method, which is applied to a computing power leasing management platform. The computing power leasing management method includes:

[0006] Receive the computing power leasing form and determine the task type, required duration, and computing power specifications corresponding to the computing power leasing form;

[0007] Determine the time series-based computing power demand pattern of the user based on the historical computing power usage record of the user corresponding to the computing power rental form;

[0008] Determine computing power analysis data associated with the task type based on historical platform data;

[0009] Calculating a first computing power requirement based on the task type, the required duration, the computing power specification, and the computing power requirement rule; and calculating a second computing power requirement based on the task type, the required duration, the computing power specification, and the computing power analysis data;

[0010] Determine a computing power output in response to the computing power leasing form based on the first computing power requirement and the second computing power requirement, and a dynamic weight associated with a date.

[0011] In one embodiment, before the steps of receiving a computing power leasing form and determining the task type, required duration, and computing power specifications corresponding to the computing power leasing form, the following steps are included:

[0012] In response to an input operation of a form submission control, determining the task type;

[0013] Render the requirement details control associated with the task type, where different task types correspond to different requirement details controls;

[0014] In response to an input operation of the demand details control, parsing the demand duration and the computing power specification;

[0015] The computing power leasing form is generated according to the task type, the required duration, and the computing power specification.

[0016] In one embodiment, the step of determining a time-series-based computing power demand pattern of the user based on the historical computing power usage record of the user corresponding to the computing power rental form includes:

[0017] Obtain the user's historical computing power usage records, including the execution period, computing power consumption value, and business scenario label;

[0018] Aggregating the historical computing power usage records by buckets based on a preset time window to generate periodic computing power consumption characteristics;

[0019] Modeling the periodic computing power consumption characteristics to identify trend terms, seasonal terms, and residual terms of user computing power demand;

[0020] The prediction results of the trend item, seasonal item and residual item are corrected in combination with the business scenario label to generate the user's computing power demand pattern based on the time series.

[0021] In one embodiment, the step of determining computing power analysis data associated with the task type based on platform historical data includes:

[0022] Obtain all historical computing power request records matching the task type in the platform's historical data;

[0023] Perform three-dimensional classification aggregation according to the task type, execution period, and business scenario label;

[0024] Calculate the computing power consumption characteristic values under each category, including mean, variance, quantile, and resource utilization efficiency coefficient;

[0025] Correcting the computing power consumption characteristic value in combination with external data;

[0026] An association model based on a random forest algorithm is constructed to analyze the dependency between the task type and the computing power consumption characteristic value, and to generate computing power analysis data associated with the task type.

[0027] In one embodiment, the step of calculating the first computing power requirement based on the task type, the required duration, the computing power specification, and the computing power requirement rule includes:

[0028] Decomposing the computing power demand pattern into trend term, seasonal term and random term;

[0029] Establish a mapping relationship between task types and standard computing power units, and convert the computing power specifications into standardized computing power values;

[0030] A linear regression model is used to perform a weighted combination of the standardized computing power value, the demand duration, the trend term, and the seasonal term to determine a regression result;

[0031] The first computing power requirement is determined based on a time decay factor, a resource utilization efficiency coefficient of the user's historical computing power usage, and the regression result.

[0032] In one embodiment, the step of calculating the second computing power requirement based on the task type, the required duration, the computing power specification, and the computing power analysis data includes:

[0033] Extracting computing power consumption characteristic values matching the task type from the computing power analysis data, including historical average computing power consumption, time period fluctuation coefficient, and resource utilization efficiency benchmark value;

[0034] Based on the mapping relationship between the computing power specification and the hardware performance parameters, the computing power specification is converted into the standard computing power of the unit;

[0035] Determining the period correction performance based on the period fluctuation coefficient and the standard computing power of the unit;

[0036] Determine a duration efficiency product according to the required duration and the resource utilization efficiency benchmark value;

[0037] The second computing power requirement is determined based on the time period correction performance, the time efficiency product, and the historical average computing power consumption processed by the gradient boosting tree.

[0038] In one embodiment, the step of determining the computing power output in response to the computing power rental form based on the first computing power requirement and the second computing power requirement, and a dynamic weight associated with a date, includes:

[0039] Extracting date type labels from the platform's historical data and establishing a mapping relationship between date types and business fluctuation coefficients;

[0040] Determining the dynamic weight based on a mapping relationship between the date type and the business fluctuation coefficient, and a standard deviation of historical computing power requirements;

[0041] Determine the computing power output according to a weighted combination of the first computing power requirement, the second computing power requirement, and the dynamic weight.

[0042] In one embodiment, after the step of determining the computing power output in response to the computing power rental form based on the first computing power requirement and the second computing power requirement, and a dynamic weight associated with a date, the following steps are performed:

[0043] Collect actual computing power output data;

[0044] Outputting an alarm prompt message based on the computing power consumption threshold and the actual computing power output data.

[0045] In addition, to achieve the above-mentioned purpose, the present application also provides a computing power leasing management device, which includes: a memory, a processor, and a computer program stored on the memory and runnable on the processor, and the computer program is configured to implement the steps of the computing power leasing management method described above.

[0046] In addition, to achieve the above-mentioned purpose, the present application also provides a storage medium, which is a computer-readable storage medium, and the computer-readable storage medium stores a program for implementing the computing power leasing management method. The program for implementing the computing power leasing management method is executed by the processor to implement the steps of the computing power leasing management method as described above.

[0047] The present application provides a computing power leasing management method. The present application first determines the task type, required duration, and computing power specifications corresponding to the computing power leasing form by receiving the computing power leasing form; determines the computing power demand pattern of the user based on the time series based on the historical computing power usage record of the user corresponding to the computing power leasing form; determines the computing power analysis data associated with the task type based on the platform historical data; calculates the first computing power requirement based on the task type, the required duration, the computing power specification, and the computing power demand pattern; and calculates the second computing power requirement based on the task type, the required duration, the computing power specification, and the computing power analysis data; determines the computing power output in response to the computing power leasing form based on the first computing power requirement and the second computing power requirement, as well as the dynamic weight associated with the date. This solves the technical problem of high computing power costs caused by high hardware procurement costs and subsequent operation and maintenance and electricity consumption, and achieves the technical effect of reducing computing power usage costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0049] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0050] Figure 1 A flowchart illustrating the first embodiment of the computing power leasing management method of this application;

[0051] Figure 2 A flowchart of the third embodiment of the computing power leasing management method of this application is provided;

[0052] Figure 3 A flowchart illustrating the fourth embodiment of the computing power leasing management method of this application;

[0053] Figure 4 This is a schematic diagram of the hardware structure of the computing power leasing management equipment involved in this application.

[0054] The purpose, features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0055] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not intended to limit the present application.

[0056] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0057] Currently, with the rapid development of artificial intelligence, large AI models have been deeply integrated into numerous business scenarios, including intelligent customer service, image recognition, and natural language processing. Whether in model training or actual inference, large AI models place extremely high demands on computing power. However, the computing power requirements of many businesses fluctuate significantly during operation and are not always maintained at a high level. For example, during major promotional events on e-commerce platforms, the computing power requirements of the intelligent recommendation system will rise sharply, but during daily operations, computing power requirements will drop significantly. To meet the computing power requirements during peak business operations, companies need to purchase a large number of high-end graphics cards. The high hardware procurement costs, as well as the subsequent operation and maintenance and electricity consumption, lead to excessively high computing power costs.

[0058] Therefore, this application receives a computing power rental form, determines the task type, required duration and computing power specifications corresponding to the computing power rental form; determines the computing power demand pattern of the user based on the time series based on the historical computing power usage record of the user corresponding to the computing power rental form; determines the computing power analysis data associated with the task type based on the platform historical data; calculates the first computing power requirement based on the task type, the required duration, the computing power specifications and the computing power demand pattern; and calculates the second computing power requirement based on the task type, the required duration, the computing power specifications and the computing power analysis data; determines the computing power output in response to the computing power rental form based on the first computing power requirement and the second computing power requirement, as well as the dynamic weight associated with the date, thereby achieving the technical effect of reducing the computing power usage cost.

[0059] An energy storage system is a system that stores energy through a medium or device and releases it when needed. It typically consists of a medium or device for storing energy, equipment for converting energy, and equipment for managing the stored energy.

[0060] It should be noted that the execution entity of this embodiment can be a computing power leasing management platform, or a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, mobile phone, etc., or a computing power leasing management device capable of performing the above functions, etc. This embodiment does not specifically limit this. The following uses the computing power leasing management platform as an example to illustrate this embodiment and the following embodiments.

[0061] Based on this, this application proposes a computing power leasing management method of the first embodiment, please refer to Figure 1 The computing power leasing management method includes steps S10 to S50:

[0062] Step S10: Receive a computing power leasing form and determine the task type, required duration, and computing power specifications corresponding to the computing power leasing form.

[0063] In this example, the computing power leasing form is a document used by users to submit a computing power leasing application to the computing power leasing management platform. It contains key information such as the task type, required duration, and computing power specifications. The task type indicates the business scope for which the user is leasing computing power, such as image recognition in smart security; the required duration refers to the time span over which the user plans to use the computing power; and the computing power specifications specify the specific configuration parameters of the computing power equipment required, such as the model and number of GPUs.

[0064] As an optional implementation method, the platform sets up a rental application entrance on both the web and mobile terminals. Users fill out the form on the corresponding interface. The platform receives the form data in real time through the data interface and uses preset data parsing rules to extract the task type, required duration and computing power specifications.

[0065] Step S20: Determine the time series-based computing power demand pattern of the user according to the historical computing power usage record of the user corresponding to the computing power rental form.

[0066] In this embodiment, historical computing power usage records refer to detailed data on computing power users have rented on the platform in the past, including information such as usage time and amount. Time series computing power demand patterns are derived by performing time series analysis on these historical data to reveal the trends and characteristics of user computing power demand over time.

[0067] As an optional implementation, the platform uses Python's pandas and statsmodels libraries to preprocess and model users' historical computing power usage records, identify periodicity, seasonality, and long-term trends in the data, and thus determine the patterns of users' computing power demand.

[0068] Step S30: Determine computing power analysis data associated with the task type based on platform historical data.

[0069] In this embodiment, platform historical data includes computing power rental and usage data for all users during platform operation. Task-type-associated computing power analysis data refers to computing power usage characteristics for specific task types, derived through statistical analysis of platform historical data. This includes average computing power consumption for that task type and the frequency of peak computing power usage.

[0070] As an optional implementation, the platform uses big data analysis tools Hive and Spark to regularly summarize and analyze the platform's historical data, generate a mapping table between task types and computing power analysis data, and store it in the distributed database HBase for fast query and call.

[0071] Step S40, calculating the first computing power requirement based on the task type, the required duration, the computing power specification and the computing power requirement rule; and calculating the second computing power requirement based on the task type, the required duration, the computing power specification and the computing power analysis data.

[0072] In this embodiment, the first computing power requirement is an estimated computing power value by comprehensively considering the user's own task type, required duration, computing power specifications and computing power demand rules; the second computing power requirement is a value calculated based on the task type, required duration, computing power specifications and computing power analysis data associated with the task type.

[0073] As an optional implementation method, a first computing power requirement calculation model based on linear regression is constructed, and relevant data is used as model input to obtain the first computing power requirement; a second computing power requirement calculation model is constructed using an algorithm based on statistical average, and the corresponding data is input to calculate the second computing power requirement.

[0074] Step S50: Determine a computing power output in response to the computing power rental form based on the first computing power requirement and the second computing power requirement, and a dynamic weight associated with a date.

[0075] In this embodiment, the dynamic weight associated with a date is a pre-set weighting factor based on the degree to which different dates affect computing power demand. For example, the weights for different dates, such as weekdays, holidays, and e-commerce promotion days, may vary. Computing power output refers to the amount of computing power resources ultimately allocated to users by the platform.

[0076] As an optional implementation, the computing power configuration provided to the user is calculated using the formula "computing power output = first computing power requirement × dynamic weight of corresponding date + second computing power requirement × (1 - dynamic weight of corresponding date)".

[0077] For example, user A submits a computing power rental form to the computing power rental management platform. The form indicates that the task type is live streaming acceleration, the required duration is from the beginning to the end of summer vacation, a total of two months, and the computing power specifications require a 32-core CPU, 128GB of memory, and 8 high-performance GPUs.

[0078] After receiving the form, the platform extracts information about the task type, required duration, and computing power specifications. An analysis of the online education platform's historical computing power usage revealed that computing power demand surged during the holidays due to an increase in live courses, with a relatively low demand trend occurring on weekends. The platform also retrieved computing power analysis data for live streaming acceleration tasks, such as average holiday computing power consumption and peak computing power, from a mapping table of task types and computing power analysis data. Based on this data, the platform used a first computing power demand calculation model, combined with the online education platform's own computing power demand patterns, to calculate the first computing power requirement. A second computing power demand calculation model, combined with the general computing power requirements for live streaming acceleration tasks, calculated the second computing power requirement. Considering that summer is peak business season for online education platforms, the platform assigns a higher dynamic weight to this period. Using a formula, the platform determined that it would provide the online education platform with a computing power output of 28 CPU cores, 112GB of memory, and seven high-performance GPUs. This not only meets the platform's computing power needs for live streaming acceleration during the summer, but also avoids over-provisioning of computing resources and reduces costs.

[0079] This application first receives a computing power rental form, determines the task type, required duration, and computing power specifications corresponding to the computing power rental form; determines the computing power demand pattern of the user based on the time series based on the historical computing power usage record of the user corresponding to the computing power rental form; determines the computing power analysis data associated with the task type based on the platform's historical data; calculates the first computing power requirement based on the task type, the required duration, the computing power specification, and the computing power demand pattern; and calculates the second computing power requirement based on the task type, the required duration, the computing power specification, and the computing power analysis data; determines the computing power output in response to the computing power rental form based on the first computing power requirement and the second computing power requirement, as well as the dynamic weight associated with the date. This solves the technical problem of high computing power costs caused by high hardware procurement costs and subsequent operation and maintenance and electricity consumption, and achieves the technical effect of reducing computing power usage costs.

[0080] Based on any of the above embodiments, in the second embodiment of the present application, before the steps of receiving a computing power rental form and determining the task type, required duration, and computing power specifications corresponding to the computing power rental form, the following steps are included:

[0081] Step S1: In response to an input operation of a form submission control, determining the task type.

[0082] In this embodiment, the form submission control is an interactive element on the user interface used to initiate the computing power leasing form submission operation, such as a "Submit" button in a web page or application; the task type indicates the business category that the user needs to carry out when leasing computing power, such as intelligent customer service and big data analysis.

[0083] As an optional implementation method, the platform deploys a drop-down menu on the interactive page where users rent computing power. After the user clicks the menu, he or she selects an item from the pre-set task type options. When the user clicks the "Submit" button, the platform captures the operation and obtains the task type selected by the user.

[0084] Step S2: Rendering a requirement detail control associated with the task type, wherein different task types correspond to different requirement detail controls.

[0085] In this embodiment, the Demand Details control is a collection of interactive components used to collect users' specific requirements for computing power rental. Due to the varying computing power requirements of different task types, the corresponding Demand Details control also varies. For example, an image rendering task may require the user to enter information such as image resolution and rendering frame rate, while a natural language processing task may focus on information such as text processing capacity.

[0086] As an optional implementation, the platform pre-establishes a mapping relationship between task types and requirement details control templates. After obtaining the task type selected by the user, the corresponding requirement details control template is loaded from the server based on the mapping relationship and rendered on the page.

[0087] Step S3: In response to the input operation of the demand details control, the demand duration and the computing power specification are parsed.

[0088] In this embodiment, the required duration refers to the time range in which the user plans to use computing power, and the computing power specification is used to describe the performance parameters of the computing power device required by the user, such as the number of CPU cores and GPU memory capacity.

[0089] As an optional implementation, the platform binds an event catcher to each input item in the requirement details control. When the user enters data in these input boxes and completes the operation, the event catcher is triggered, and the platform obtains the input data and parses it according to pre-set parsing rules. For example, for the required duration, the platform recognizes the time unit and value entered by the user and converts it into a unified time format; for the computing power specification, the platform parses the corresponding hardware configuration information based on the hardware parameter description entered by the user.

[0090] Step S4: Generate the computing power leasing form according to the task type, the required duration, and the computing power specification.

[0091] In this embodiment, the computing power leasing form is a file that integrates key information about the user's computing power leasing for subsequent processing by the platform.

[0092] As an optional implementation, the platform generates a computing power rental form in JSON format. The information obtained in the previous step, such as task type, required duration, and computing power specifications, is organized according to a preset JSON structure to generate a complete computing power rental form data. This data is then stored in the server database for subsequent computing power rental management operations.

[0093] For example, a gaming company plans to lease computing power for game testing. On the platform's computing power rental page, a gaming company employee selects "Game Testing" from the Task Type drop-down menu and clicks "Submit." The platform captures this action and identifies the task type as "Game Testing." Based on the mapping between the task type and the Requirement Details control template, the platform renders the Requirement Details control corresponding to the game testing task on the page. This control includes input fields such as "Test Duration" and "Required Server Configuration (CPU, GPU, Memory, etc.)." The gaming company employee enters "30 days" in the "Test Duration" input field and "32-core CPU, 64GB of RAM, and 8 RTX 3080 GPUs" in the "Required Server Configuration" input field. The platform triggers the event capturer bound to these input fields, captures the input data, and, according to the parsing rules, interprets the required duration as 30 days and the computing power specifications as 32-core CPU, 64GB of RAM, and 8 RTX 3080 GPUs. Finally, the platform organizes the task type "game test", the required duration "30 days", and the computing power specifications "32-core CPU, 64GB memory, 8 RTX3080GPUs" in JSON format, generates a computing power rental form and stores it in the database for use in subsequent computing power rental management processes.

[0094] Since a demand detail control associated with the task type is used for rendering, different demand detail controls correspond to different task types, and the computing power requirements are quantified in different formats according to different task types, thereby improving the accuracy of computing power requirement assessment.

[0095] Based on any of the above embodiments, in the third embodiment of the present application, refer to Figure 2 The step of determining the time-series computing power demand pattern of the user based on the historical computing power usage record of the user corresponding to the computing power rental form includes:

[0096] Step S21, obtaining the user's historical computing power usage record, wherein the historical computing power usage record includes an execution period, computing power consumption value, and a business scenario tag.

[0097] In this embodiment, historical computing power usage records are a collection of data related to a user's previous computing power usage on the computing power leasing management platform. The execution period refers to the specific time range in which a user uses computing power, reflecting the temporal distribution of computing power usage. The computing power consumption value represents the actual amount of computing power consumed by the user during the execution period and is a key indicator for measuring computing power demand. The business scenario tag identifies the specific business scenario for which the user's computing power is used, such as intelligent customer service or image rendering.

[0098] As an optional implementation, the platform establishes a storage table for users' historical computing power usage records in the database. This table collects users' computing power usage information from the log system in real time through a data interface and stores it in a structured manner based on fields such as execution period, computing power consumption value, and business scenario tags. When a user's historical computing power usage records are needed, relevant data is extracted from the database using SQL queries.

[0099] Step S22: bucket-aggregate the historical computing power usage records based on a preset time window to generate periodic computing power consumption characteristics.

[0100] In this embodiment, the preset time window is a fixed, pre-set duration, such as a day, week, or month, used to divide a continuous timeline into several equal time periods. Bucket aggregation groups historical computing power usage records according to the preset time window and performs statistical calculations, such as summation and average, on the computing power consumption values within each time period. Periodic computing power consumption characteristics are obtained through bucket aggregation and reflect the user's computing power consumption patterns over different time periods.

[0101] As an optional implementation, we use the Python pandas library to process the historical computing power usage records. First, we convert the execution period into a time index. Then, we use the resample() method to resample the data according to the preset time window. The computing power consumption values are summed to generate the total computing power consumption within each time window, thus obtaining the periodic computing power consumption characteristics.

[0102] Step S23: Model the periodic computing power consumption characteristics to identify the trend item, seasonal item, and residual item of the user computing power demand.

[0103] In this example, modeling involves fitting the periodic characteristics of computing power consumption using an appropriate time series analysis model to reveal underlying patterns in the data. The trend term represents the long-term trend in computing power demand over time, such as a gradual increase or decrease. The seasonal term reflects the recurring pattern of changes in computing power demand over a fixed time period, such as weekly or monthly cyclical fluctuations. The residual term represents the random error portion that the model cannot explain.

[0104] As an optional implementation, the SARIMAX (Seasonal Autoregressive Integrated Moving Average Exogenous Regression) model from the Python statsmodels library is used to model the periodic computing power consumption characteristics. This model can simultaneously consider the trend, seasonality, and randomness of the data. By fitting the model and estimating its parameters, it can separate the trend term, seasonal term, and residual term.

[0105] Step S24: Correct the prediction results of the trend item, seasonal item, and residual item in combination with the business scenario label to generate the computing power demand pattern of the user based on the time series.

[0106] In this example, the business scenario tag contains the specific business context of the user's computing power usage. Different business scenarios may have different impacts on computing power requirements. By combining business scenario tags, the model's prediction results can be adjusted and optimized to make them more consistent with actual conditions.

[0107] As an optional implementation, a mapping table is established between business scenario tags and computing power demand correction coefficients. For each business scenario tag, a corresponding correction coefficient is set based on historical data and business experience. After obtaining the forecast results for the trend, seasonal, and residual terms, the corresponding correction coefficient is searched based on the business scenario tag. The forecast results are weighted and corrected to ultimately generate a time-series-based computing power demand pattern for the user.

[0108] Through this embodiment, it is possible to more accurately grasp the patterns of users' computing power needs. Periodic features are mined from a large amount of historical computing power usage records, and then trend items, seasonal items, and residual items are analyzed using time series models. Finally, corrections are made based on business scenario labels to make the prediction results more consistent with the user's actual computing power needs. This helps the computing power leasing management platform to more reasonably allocate computing power resources and avoid resource waste or shortages caused by improper allocation of computing power resources. For example, during periods of low user computing power demand, unnecessary computing power supply can be reduced to reduce costs; during peak demand periods, computing power reserves can be made in advance to ensure service stability and reliability, thereby improving the platform's operational efficiency and user satisfaction.

[0109] Based on any of the above embodiments, in the fourth embodiment of the present application, refer to Figure 3 The step of determining computing power analysis data associated with the task type based on platform historical data includes:

[0110] Step S31: Obtain all historical computing power request records matching the task type in the platform's historical data.

[0111] In this embodiment, platform historical data refers to all data related to computing power usage accumulated during the operation of the computing power leasing management platform. Historical computing power request records contain detailed information about computing power requests initiated by users to the platform, including task type, execution period, business scenario tag, and computing power consumption values. Task type refers to the business category for which users use computing power, such as natural language processing or video encoding.

[0112] As an optional implementation method, the platform can set a corresponding index in the database and use SQL query statements to filter out all historical computing power request records whose task types are consistent with the current task type from the database.

[0113] Step S32: performing three-dimensional classification aggregation according to the task type, execution period, and business scenario label.

[0114] In this embodiment, three-dimensional classification aggregation groups historical computing power request records according to three dimensions: task type, execution period, and business scenario tag. The records within each group are then aggregated. The execution period refers to the specific time range in which a user uses computing power, and the business scenario tag identifies the specific business scenario in which the user uses computing power, such as e-commerce promotions or live online education courses.

[0115] As an optional implementation, use Python's pandas library to process historical computing power request records, group them by task type, execution period, and business scenario label using the groupby() method, and then perform aggregation operations such as summing and counting the computing power consumption values within the group.

[0116] Step S33: Calculate the computing power consumption characteristic values under each category, including mean, variance, quantile and resource utilization efficiency coefficient.

[0117] In this embodiment, the computing power consumption characteristic value is a statistical indicator used to describe the computing power consumption of each category. The mean represents the average level of computing power consumption within that category; the variance reflects the degree of dispersion of computing power consumption; the quantile provides the numerical value of computing power consumption at different locations; and the resource utilization efficiency coefficient measures the efficiency of computing power resource utilization within that category.

[0118] As an optional implementation, relevant functions of the pandas library are used to calculate the mean, variance, quantile, etc. of the computing power consumption values under each category. At the same time, the resource utilization efficiency coefficient is obtained by calculating the ratio of actual computing power consumption to the theoretical maximum computing power consumption.

[0119] Step S34: correct the computing power consumption characteristic value in combination with external data.

[0120] In this embodiment, external data refers to data related to computing power usage from outside the platform, such as market data, industry dynamics data, weather data, etc. This external data may have an impact on computing power consumption. By combining external data, the computing power consumption characteristic value can more accurately reflect the actual situation.

[0121] As an optional implementation, a correlation model is established between external data and the characteristic value of computing power consumption, and the characteristic value of computing power consumption is adjusted according to changes in external data. For example, if market data shows that a certain industry is in its peak season, the expected computing power consumption of task types related to this industry may increase, thereby increasing the corresponding characteristic value of computing power consumption.

[0122] Step S35: construct an association model based on the random forest algorithm, analyze the dependency between the task type and the computing power consumption characteristic value, and generate computing power analysis data associated with the task type.

[0123] In this embodiment, the random forest algorithm is an ensemble learning algorithm that effectively handles high-dimensional data and complex nonlinear relationships by constructing multiple decision trees and synthesizing their results. An association model is used to analyze the inherent connection between task types and computing power consumption characteristics. Task-type-associated computing power analysis data is generated by comprehensively considering task types and computing power consumption characteristics, and can support platform decision-making, such as providing recommendations for appropriate computing power configurations for different task types and business scenarios.

[0124] As an optional implementation, a random forest model is constructed using Python's scikit-learn library. Task type and computing power consumption characteristic values are used as input features. The dependency relationship between task type and computing power consumption characteristic values is obtained by training the model, and computing power analysis data associated with the task type is generated.

[0125] Through this embodiment, comprehensive and accurate computing power analysis data associated with task types is generated. Records matching the task type are screened out from a large amount of platform historical data, and multi-dimensional classification aggregation and eigenvalue calculation are performed. Then, corrections are made in combination with external data. Finally, a correlation model is established using the random forest algorithm, which can deeply explore the intrinsic connection between task type and computing power consumption. This helps the computing power leasing management platform to more accurately understand the computing power demand characteristics of different task types in various business scenarios, thereby providing users with more reasonable computing power leasing solutions, improving the allocation efficiency of computing power resources, and reducing the operating costs of the platform. At the same time, it can also improve the user experience and enhance the platform's competitiveness in the market.

[0126] Based on any of the above embodiments, in Embodiment 5 of the present application, the step of calculating the first computing power requirement based on the task type, the required duration, the computing power specification, and the computing power requirement rule includes:

[0127] Step S41: Decompose the computing power demand pattern into trend items, seasonal items, and random items.

[0128] In this example, the computing power demand pattern is derived from an analysis of historical user computing power usage records, reflecting the pattern of changes in computing power demand over time. The trend term represents an upward or downward trend in computing power demand over a long period of time; the seasonal term reflects the recurring fluctuation pattern of computing power demand within a fixed period (such as daily, weekly, or monthly); and the random term includes irregular, random changes that cannot be explained by the trend and seasonal terms.

[0129] As an optional implementation, the seasonal decomposition algorithm (such as STL decomposition) in Python's statsmodels library is used to decompose the computing power demand pattern data in the form of time series into trend terms, seasonal terms, and random terms.

[0130] Step S42: Establish a mapping relationship between the task type and the standard computing power unit, and convert the computing power specification into a standardized computing power value.

[0131] In this embodiment, the task type represents the business category for which the user leases computing power, such as video rendering or big data analysis. The standard computing power unit (SPU) is a benchmark unit for standardizing different computing power specifications. The standardized computing power value converts various computing power specifications into a value in SPUs, based on the mapping between task types and SPUs, to facilitate subsequent calculations.

[0132] As an optional implementation, the platform pre-builds a mapping table between task types and standard computing units. For example, for video rendering tasks, a specific GPU model is assigned a certain number of standard computing units. Once the computing power specifications are obtained, they are converted into standardized computing power values according to the mapping table.

[0133] Step S43: Using a linear regression model, weighted combination of the standardized computing power value, demand duration, trend item, and seasonal item is performed to determine a regression result.

[0134] In this example, a linear regression model is a statistical model used to analyze linear relationships between variables. It describes the relationship between independent and dependent variables by constructing a linear equation. Standardized computing power, demand duration, trend terms, and seasonal terms serve as independent variables, and the regression results serve as predicted values for the dependent variables. Weighted combination involves assigning different weights to each independent variable to reflect its influence on the dependent variable.

[0135] As an optional implementation, use Python's scikit-learn library to build a linear regression model, using the standardized computing power value, demand duration, trend term, and seasonal term as input features. Through model training, the weight of each feature is determined, and then the regression result is calculated.

[0136] Step S44: Determine the first computing power requirement based on the time decay factor, the resource utilization efficiency coefficient of the user's historical computing power usage, and the regression result.

[0137] In this embodiment, the time decay factor reflects the change in computing power demand over time and is typically used to adjust long-term forecast results to make them more realistic. The resource utilization efficiency coefficient for a user's historical computing power usage reflects the efficiency of the user's past computing power usage and is used to correct regression results. The first computing power requirement is the computing power value calculated by combining the above factors to meet the user's business needs.

[0138] As an optional implementation, the first computing power requirement is calculated using the formula "first computing power requirement = regression result × time decay factor × resource utilization efficiency coefficient of user's historical computing power usage".

[0139] For example, a short video production company plans to lease computing power for video rendering and submits a form to the computing power rental management platform. The task type is video rendering, the required duration is 10 days, and the computing power specification is four RTX 3090 GPUs. The platform first obtains the company's computing power demand pattern and decomposes it into trend, seasonal, and random terms using STL decomposition. Using a pre-built mapping table, the four RTX 3090 GPUs are converted into normalized computing power values. Next, a linear regression model is trained using the scikit-learn library using the normalized computing power values, the 10-day required duration, and the decomposed trend and seasonal terms as inputs to produce the regression results. Considering the time-sensitive nature of short video production, a time decay factor of 0.95 is set. Based on the company's historical computing power usage data, the resource utilization efficiency coefficient is calculated to be 0.85. Finally, a formula is used to calculate the first computing power requirement, providing a basis for the platform to rationally allocate computing power to the company.

[0140] This embodiment integrates multiple factors to accurately calculate the first computing power requirement. By decomposing the patterns of computing power demand, we can meticulously grasp the changing trends of user computing power requirements; by standardizing computing power specifications, we unify the calculation criteria; by using a linear regression model for weighted combination, we fully consider the impact of various factors on computing power demand; and by combining the time decay factor and resource utilization efficiency coefficient, we further optimize the calculation results. This enables the platform to more accurately allocate computing power resources to users, avoiding waste or insufficient computing power resources. This reduces users' computing power rental costs while improving the utilization rate of the platform's computing power resources, enhancing the platform's market competitiveness.

[0141] Based on any of the above embodiments, in Embodiment 6 of the present application, the step of calculating the second computing power requirement based on the task type, the required duration, the computing power specification, and the computing power analysis data includes:

[0142] Step S45: extract computing power consumption characteristic values matching the task type from the computing power analysis data, including historical average computing power consumption, time period fluctuation coefficient, and resource utilization efficiency benchmark value.

[0143] In this embodiment, the computing power analysis data is a data set related to computing power consumption, which is obtained by analyzing different task types based on the platform's historical data. The historical average computing power consumption refers to the average amount of computing power consumed when executing this task type in the past, reflecting the basic level of computing power required by this task type. The time period fluctuation coefficient reflects the fluctuation of computing power consumption of this task type in different time periods. For example, the computing power consumption of certain tasks will increase significantly during specific time periods (such as peak hours on weekdays). The resource utilization efficiency baseline value represents the theoretically reasonable computing power resource utilization efficiency when executing this task type. As an optional implementation method, the platform stores computing power analysis data in a database and creates an index for each task type. By querying the index, the computing power consumption feature value that matches the current task type can be quickly extracted.

[0144] Step S46: Based on the mapping relationship between the computing power specification and the hardware performance parameter, the computing power specification is converted into a unit of standard computing power.

[0145] In this embodiment, the computing power specification describes the hardware configuration information for the computing power requested by the user, such as the number of CPU cores and the model and number of GPUs. Hardware performance parameters are quantitative indicators of the computing power of different hardware devices. The unit's standard computing power is a unified measurement standard used to convert different computing power specifications into comparable values. For example, a specific CPU model is used as the standard, and other hardware configurations are converted to the standard computing power based on the performance ratio of the CPU to the standard CPU.

[0146] As an optional implementation method, the platform pre-establishes a mapping table between computing power specifications and hardware performance parameters. After obtaining the user's computing power specifications, it converts them into the unit's standard computing power by looking up the mapping table and performing corresponding calculations.

[0147] Step S47: determining the time period correction performance according to the time period fluctuation coefficient and the standard computing power of the unit.

[0148] In this embodiment, the time-period-adjusted performance is calculated by adjusting the unit standard computing power to account for the impact of time-period fluctuations on computing power requirements. Because the computing power requirements of tasks may vary in different time periods, adjusting the unit standard computing power using the time-period fluctuation coefficient can more accurately reflect the computing power required to complete tasks within a specific time period.

[0149] As an optional implementation, the unit's standard computing power is multiplied by the period fluctuation coefficient to obtain the period correction performance.

[0150] Step S48: determining a duration-efficiency product according to the required duration and the resource utilization efficiency benchmark value.

[0151] In this embodiment, the required duration is the length of time the user plans to use computing power. The duration-efficiency product is the result of multiplying the required duration by the resource utilization efficiency benchmark. It comprehensively considers the impact of task execution time and resource utilization efficiency on computing power requirements. For example, if the required duration is long and resource utilization efficiency is low, the required computing power may increase accordingly.

[0152] As an optional implementation, the required duration is directly multiplied by the resource utilization efficiency benchmark value to obtain the duration efficiency product.

[0153] Step S49: Determine the second computing power requirement based on the time period correction performance, the time efficiency product, and the historical average computing power consumption of the gradient boosting tree processing.

[0154] In this embodiment, the gradient boosting tree is an ensemble learning algorithm that iteratively trains multiple weak learners (decision trees) and combines them into a strong learner. It effectively handles nonlinear relationships and complex data. Using the time period correction performance, the product of duration efficiency, and historical average computing power consumption as input features, the gradient boosting tree algorithm is used for training and prediction to obtain the final secondary computing power requirement.

[0155] As an optional implementation, use the GradientBoostingRegressor class in Python's scikit-learn library to build a gradient boosting tree model, use the above three features as input for training, and then use the trained model to predict the second computing power requirement.

[0156] For example, an e-commerce platform plans to lease computing power during the "Double 11" (Singles' Day) promotional event to optimize its product recommendation system. The task type is to run a product recommendation algorithm. The demand period is eight days, from the week before "Double 11" to the actual day of "Double 11." The computing power specifications are a 32-core CPU and eight NVIDIA Tesla V100 GPUs. The platform extracts computing power analysis data and finds that the historical average computing power consumption for the product recommendation algorithm task type is 500 standard computing power units. The period fluctuation coefficient during "Double 11" is 1.5 (indicating that computing power demand during this period is 50% higher than usual), and the resource utilization efficiency benchmark is 0.8. Using a pre-established mapping table, the 32-core CPU and eight NVIDIA Tesla V100 GPUs are converted to a unit of standard computing power of 1200 standard computing power units. Based on the period fluctuation coefficient and the unit standard computing power, the period-corrected performance is calculated as 1200 × 1.5 = 1800 standard computing power units. Based on the demand period and the resource utilization efficiency benchmark, the period efficiency product is calculated as 8 × 0.8 = 6.4. Using the time-period-corrected performance of 1800, the time-efficiency product of 6.4, and the historical average computing power consumption of 500 as input, the gradient boosting tree model was trained and predicted, ultimately resulting in a second computing power requirement of 2000 standard computing power units. Based on this result, the platform rationally allocates computing power resources to e-commerce platforms.

[0157] This embodiment calculates the second computing power requirement by fully considering multiple factors, including the historical characteristics of the task type, time period fluctuations, resource utilization efficiency, and computing power specifications. The gradient boosting tree algorithm is used to handle these complex relationships, making the calculation results more accurate and reliable. This helps the computing power rental management platform provide users with more precise computing power configuration solutions, avoiding resource waste or shortages caused by unreasonable computing power allocation. This reduces users' computing power rental costs, improves user satisfaction, and increases the utilization rate of the platform's computing power resources.

[0158] Based on any of the above embodiments, in Embodiment 7 of the present application, the step of determining the computing power output in response to the computing power rental form based on the first computing power requirement and the second computing power requirement, and the dynamic weight associated with the date, includes:

[0159] Step S51: extract date type tags from the platform historical data, and establish a mapping relationship between date types and business fluctuation coefficients.

[0160] In this embodiment, platform historical data refers to all data related to the computing power leasing business accumulated by the computing power leasing management platform during its past operations. This data covers information such as computing power demand and business execution status on different dates. Date type labels categorize dates, such as weekdays, weekends, holidays, and promotional days. Different date types often correspond to different business busyness and computing power demand characteristics. The business volatility coefficient measures the degree of fluctuation in business computing power demand relative to the average level on a specific date type. A larger volatility coefficient indicates a more drastic change in business computing power demand on that date type.

[0161] As an optional implementation, the platform can use data mining techniques to analyze and categorize date information in historical data, extracting different date type labels. It then collects data on business computing power requirements for each date type and calculates its ratio to the average computing power requirement to obtain the corresponding business volatility coefficient. A mapping relationship between the two is then established and stored in the database.

[0162] Step S52: determining the dynamic weight according to the mapping relationship between the date type and the business fluctuation coefficient, and the standard deviation of the historical computing power requirements.

[0163] In this embodiment, the standard deviation of historical computing power requirements measures the degree of dispersion of historical computing power requirements data, reflecting the fluctuations in computing power requirements at different time points. Dynamic weights are dynamically adjusted based on the date type and historical computing power demand fluctuations. They are used to appropriately allocate the influence of the primary and secondary computing power requirements when calculating computing power output.

[0164] As an optional implementation, first, based on the date corresponding to the current computing power rental form, the business fluctuation coefficient of the date type is obtained from the mapping relationship. Then, combined with the historical computing power demand standard deviation, a preset calculation formula is used to determine the dynamic weight. For example, the formula: dynamic weight = business fluctuation coefficient / (business fluctuation coefficient + historical computing power demand standard deviation) can be used. This allows for dates with large business fluctuations to place greater emphasis on the first computing power requirement calculated based on historical patterns; while on dates with small business fluctuations, relatively more emphasis is placed on the second computing power requirement calculated based on the general characteristics of the task type.

[0165] Step S53: Determine the computing power output according to the first computing power requirement, the second computing power requirement, and the weighted combination of the dynamic weight.

[0166] In this embodiment, the first computing power requirement is a value calculated by comprehensively considering the user's task type, required duration, computing power specifications, and computing power demand patterns. The second computing power requirement is a value calculated based on task type, required duration, computing power specifications, and computing power analysis data associated with the task type. Weighted combination refers to multiplying the first and second computing power requirements by their corresponding dynamic weights and then adding the two results to obtain the final computing power output.

[0167] As an optional implementation, the formula is used: Computing power output = first computing power requirement × dynamic weight + second computing power requirement × (1-dynamic weight). In this way, the proportion of the two computing power requirements can be flexibly adjusted according to the business characteristics of different days, thereby more accurately determining the computing power resources provided to users.

[0168] For example, User B plans to lease computing power during the May Day holiday to cope with a potential increase in player traffic. The computing power lease form submitted to the platform shows a first computing power requirement of 8,000 units and a second computing power requirement of 7,500 units. The computing power lease management platform analyzes historical platform data and determines that May Day is a holiday, with a corresponding business fluctuation coefficient of 1.2 (indicating that computing power demand on this date is 20% higher than the average). Furthermore, by calculating historical computing power demand data, the standard deviation of historical computing power demand is 0.3. Based on this data, the dynamic weight is calculated using the formula: Dynamic Weight = 1.2 / (1.2 + 0.3) = 0.8. Finally, the computing power output is calculated using the weighted combination formula: Computing Power Output = 8,000 × 0.8 + 7,500 × (1 - 0.8) = 6,400 + 1,500 = 7,900 units. The platform allocates the corresponding computing power resources to the online gaming platform based on this result.

[0169] This embodiment determines computing power output, fully accounting for business fluctuations and the stability of historical computing power requirements across different date types. By utilizing dynamic weights to weight the first and second computing power requirements, the respective weights can be flexibly adjusted based on actual circumstances, ensuring that computing power output more closely matches actual business needs. This helps avoid insufficient computing power on dates with significant business fluctuations, ensuring service stability and user experience. It also prevents overallocation of computing power on dates with less volatility, thereby reducing computing power rental costs and improving computing resource utilization efficiency.

[0170] Based on any of the above embodiments, in Embodiment 8 of the present application, after the step of determining the computing power output in response to the computing power rental form based on the first computing power requirement and the second computing power requirement, and the dynamic weight associated with the date, the following steps are included:

[0171] Step B10: Collect actual computing power output data.

[0172] In this embodiment, actual computing power output data refers to the actual computing power consumed or used in the process of providing computing power services to users. This data can reflect the actual situation of users' computing power usage, including real-time computing power consumption values and computing power usage trends in different time periods.

[0173] As an optional implementation, the platform can deploy sensors or monitoring software at computing power output nodes to collect real-time operational data from computing power devices, such as CPU usage, GPU load, memory usage, and other information. This data is collected and transmitted to the platform's data center at regular intervals (e.g., every second or every minute), where it is stored in a database for subsequent analysis and processing.

[0174] Step B20: outputting an alarm prompt message based on the computing power consumption threshold and the actual computing power output data.

[0175] In this embodiment, the computing power consumption threshold is one or more computing power consumption limit values pre-set by the platform, which are used to determine whether the current computing power usage is within the normal range. These thresholds can be set according to different task types, user needs, and the platform's resource management policies. For example, for tasks with relatively stable computing power requirements, a fixed upper threshold can be set; for tasks with fluctuating characteristics, multiple thresholds of different levels can be set to cope with varying degrees of computing power fluctuations. Alarm messages are prompt messages sent by the platform to administrators or users when actual computing power output data exceeds or falls below the set computing power consumption threshold. Their purpose is to remind relevant personnel to pay attention to computing power usage and take timely measures to adjust it.

[0176] As an optional implementation, the platform can write a monitoring program that regularly reads actual computing power output data from the database and compares it with a preset computing power consumption threshold. If the actual computing power output data exceeds the threshold range, the program will trigger an alarm mechanism, sending an alarm prompt to relevant personnel via email, SMS, system pop-up window, etc., and recording detailed information about the alarm event, such as the alarm time, alarm type, and the specific value that exceeded the threshold.

[0177] By collecting actual computing power output data and generating alerts based on computing power consumption thresholds, the platform can monitor computing power resource usage in real time and promptly identify abnormal computing power consumption behavior. This helps platform administrators respond quickly and take appropriate measures to ensure the stability and reliability of computing power services. For example, when actual computing power output exceeds the threshold, administrators can promptly check for resource abuse, system failures, and other issues and take appropriate measures to avoid service interruptions or performance degradation caused by insufficient computing power. At the same time, for users, alerts can help them understand their computing power usage, adjust their business needs appropriately, and avoid unnecessary costs. Furthermore, by analyzing alert data, the platform can further optimize the setting of computing power consumption thresholds, improving the efficiency and accuracy of resource management.

[0178] The present application provides a computing power leasing management device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the computing power leasing management method in the above-mentioned embodiment one.

[0179] Reference below Figure 4 , which shows a schematic diagram of the structure of a computing power leasing management device suitable for implementing embodiments of the present application. The computing power leasing management device in the embodiments of the present application may include, but is not limited to, mobile terminals such as laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 4 The computing power leasing management device shown is merely an example and should not impose any limitations on the functions and scope of use of the embodiments of the present application.

[0180] like Figure 4As shown, the computing power leasing management device may include a processing device 1001 (e.g., a central processing unit, graphics processing unit, etc.), which can perform various appropriate actions and processes based on programs stored in a read-only memory (ROM) 1002 or programs loaded from a storage device 1003 into a random access memory (RAM) 1004. RAM 1004 also stores various programs and data required for the operation of the computing power leasing management device. Processing device 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication devices 1009 can allow the computing power leasing management device to communicate with other devices wirelessly or wired to exchange data. Although the figure shows a computing power leasing management device with various systems, it should be understood that implementation or presence of all the illustrated systems is not required. More or fewer systems may alternatively be implemented or present.

[0181] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0182] The computing power leasing management device provided in this application, which adopts the computing power leasing management method in the above-mentioned embodiment, can solve the technical problem of excessively high computing power costs caused by hardware procurement costs, subsequent operation and maintenance, and power consumption. Compared with the existing technology, the beneficial effects of the computing power leasing management device provided in this application are the same as those of the computing power leasing management device provided in the above-mentioned embodiment, and the other technical features of the computing power leasing management device are the same as those disclosed in the method of the previous embodiment, and are not further described here.

[0183] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0184] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0185] The present application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, and the computer-readable program instructions are used to execute the computing power leasing management method in the above-mentioned embodiment.

[0186] The computer-readable storage medium provided herein may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems, or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including, but not limited to, wires, optical cables, radio frequency (RF), etc., or any suitable combination thereof.

[0187] The above-mentioned computer-readable storage medium may be included in the computing power leasing management device; or it may exist independently without being assembled into the computing power leasing management device.

[0188] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the computing power leasing management device, the computing power leasing management device is enabled to: receive a computing power leasing form, determine the task type, required duration and computing power specifications corresponding to the computing power leasing form; determine the computing power demand pattern of the user based on the time series according to the historical computing power usage record of the user corresponding to the computing power leasing form; determine the computing power analysis data associated with the task type according to the platform historical data; calculate the first computing power requirement based on the task type, the required duration, the computing power specifications and the computing power demand pattern; and calculate the second computing power requirement based on the task type, the required duration, the computing power specifications and the computing power analysis data; determine the computing power output in response to the computing power leasing form according to the first computing power requirement and the second computing power requirement, and the dynamic weight associated with the date.

[0189] Computer program code for performing the operations of the present application may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0190] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0191] The modules described in the embodiments of the present application may be implemented in software or hardware, wherein the name of a module does not necessarily limit the unit itself.

[0192] The computer-readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the aforementioned computing power leasing management method. This computer-readable storage medium can address the technical issue of excessively high computing power costs due to hardware procurement costs, subsequent operation and maintenance, and power consumption. Compared to the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the computing power leasing management method provided in the aforementioned embodiments, and are not further elaborated here.

[0193] An embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the computing power leasing management method as described above.

[0194] The computer program product provided in this application can address the technical issue of excessively high computing power costs due to hardware procurement costs, subsequent maintenance, and power consumption. Compared to the prior art, the beneficial effects of the computer program product provided in this embodiment are the same as those of the computing power leasing management method provided in the above-mentioned embodiment, and will not be elaborated on here.

[0195] The above are only preferred embodiments of the present application and do not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent processing scope of the present application.

Claims

1. A computing power leasing management method, characterized in that: Applied to a computing power leasing management platform, the computing power leasing management method includes: Receive the computing power leasing form and determine the task type, required duration, and computing power specifications corresponding to the computing power leasing form; Determine the time series-based computing power demand pattern of the user based on the historical computing power usage record of the user corresponding to the computing power rental form; Determine computing power analysis data associated with the task type based on historical platform data; Decomposing the computing power demand pattern into trend term, seasonal term and random term; Establish a mapping relationship between task types and standard computing power units, and convert the computing power specifications into standardized computing power values; A linear regression model is used to perform a weighted combination of the standardized computing power value, the demand duration, the trend term, and the seasonal term to determine a regression result; Determining a first computing power requirement based on a time decay factor, a resource utilization efficiency coefficient of the user's historical computing power usage, and the regression result; and calculating a second computing power requirement based on the task type, the required duration, the computing power specification, and the computing power analysis data; The step of calculating the second computing power requirement based on the task type, the required duration, the computing power specification, and the computing power analysis data includes: Extracting computing power consumption characteristic values matching the task type from the computing power analysis data, including historical average computing power consumption, time period fluctuation coefficient, and resource utilization efficiency benchmark value; Based on the mapping relationship between the computing power specification and the hardware performance parameters, the computing power specification is converted into the standard computing power of the unit; Determining the period correction performance based on the period fluctuation coefficient and the standard computing power of the unit; Determine a duration efficiency product according to the required duration and the resource utilization efficiency benchmark value; Determining the second computing power requirement based on the time period correction performance, the time efficiency product, and the historical average computing power consumption by processing the gradient boosting tree; Determine a computing power output in response to the computing power leasing form based on the first computing power requirement and the second computing power requirement, and a dynamic weight associated with a date.

2. The method according to claim 1, wherein Before the step of receiving the computing power leasing form and determining the task type, required duration, and computing power specifications corresponding to the computing power leasing form, the following steps are included: In response to an input operation of a form submission control, determining the task type; Render the requirement details control associated with the task type, where different task types correspond to different requirement details controls; In response to an input operation of the demand details control, parsing the demand duration and the computing power specification; The computing power leasing form is generated according to the task type, the required duration, and the computing power specification.

3. The method according to claim 1, wherein The step of determining the time-series-based computing power demand pattern of the user according to the historical computing power usage record of the user corresponding to the computing power rental form includes: Obtain the user's historical computing power usage records, including the execution period, computing power consumption value, and business scenario label; Aggregating the historical computing power usage records by buckets based on a preset time window to generate periodic computing power consumption characteristics; Modeling the periodic computing power consumption characteristics to identify trend terms, seasonal terms, and residual terms of user computing power demand; The prediction results of the trend item, seasonal item and residual item are corrected in combination with the business scenario label to generate the user's computing power demand pattern based on the time series.

4. The method according to claim 1, wherein The step of determining the computing power analysis data associated with the task type based on the platform historical data includes: Obtain all historical computing power request records matching the task type in the platform's historical data; Perform three-dimensional classification aggregation according to the task type, execution period, and business scenario label; Calculate the computing power consumption characteristic values under each category, including mean, variance, quantile, and resource utilization efficiency coefficient; Correcting the computing power consumption characteristic value in combination with external data; An association model based on a random forest algorithm is constructed to analyze the dependency between the task type and the computing power consumption characteristic value, and to generate computing power analysis data associated with the task type.

5. The method according to claim 1, wherein The step of determining the computing power output in response to the computing power leasing form based on the first computing power requirement and the second computing power requirement, and a dynamic weight associated with a date, includes: Extracting date type labels from the platform's historical data and establishing a mapping relationship between date types and business fluctuation coefficients; Determining the dynamic weight based on a mapping relationship between the date type and the business fluctuation coefficient, and a historical computing power demand standard deviation; Determine the computing power output according to a weighted combination of the first computing power requirement, the second computing power requirement, and the dynamic weight.

6. The method according to claim 1, wherein After the step of determining the computing power output in response to the computing power leasing form based on the first computing power requirement and the second computing power requirement, and the dynamic weight associated with the date, the method further includes: Collect actual computing power output data; Outputting an alarm prompt message based on the computing power consumption threshold and the actual computing power output data.

7. A computing power leasing management device, characterized in that: The computing power leasing management device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the computing power leasing management method according to any one of claims 1 to 6.

8. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the computer-readable storage medium. When the computer program is executed by a processor, the steps of the computing power leasing management method according to any one of claims 1 to 6 are implemented.

Images