Computing power resource transaction management method and system

Abstract

The invention provides a computing power resource transaction management method and system, and relates to the technical field of computing power resource transaction. The method comprises the following steps: acquiring a computing power transaction request sent by a computing power demand side and determining a plurality of reference computing power providers; acquiring historical computing power transaction data of a plurality of reference computing power providing ends, analyzing the plurality of groups of historical computing power transaction data, and determining a computing power transaction evaluation value of each reference computing power providing end; determining a target computing power providing end from the plurality of reference computing power providing ends based on a preset computing power transaction screening threshold, and recommending computing power transaction objects to the computing power demand side based on the target computing power providing end; and in response to the calculation power demand side selecting to perform calculation power transaction with the target calculation power providing side, providing calculation power service for the calculation power demand side based on the target calculation power providing side. According to the invention, the computing power transaction experience of the computing power demand side is improved.

Description

Technical Field

[0001] The present invention relates to the field of computing power resource trading technology, and in particular to a computing power resource trading management method and system. Background Art

[0002] With the rapid development of cloud computing and big data technologies, the demand for computing resources continues to grow. Effective management and efficient use of computing resources have become important factors in improving overall computing performance and cost-effectiveness.

[0003] In the prior art, the computing power provider is simply selected based on the computing power demand of the computing power demander and the computing power service is provided to the computing power demander, without considering the volatility of computing resource usage. When computing resources of different computing power providers are executing tasks, due to the influence of various factors such as hardware performance, task complexity, system load, etc., the actual execution time and resource utilization may fluctuate to a certain extent, which affects the computing power demander when experiencing computing power services. At the same time, the failures that may be encountered during the task execution will also affect the time to finally complete the computing goal. Under some pricing methods, the estimated cost of the computing power demander may deviate from the actual cost, affecting the accuracy of the decision-making of the computing power demander. Summary of the invention

[0004] The present invention proposes a computing power resource transaction management method and system, aiming to solve at least one technical problem existing in the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention provides a computing resource transaction management method in a first aspect, comprising: Obtain a computing power transaction request from a computing power demander, determine computing power demand information according to the computing power transaction request, and determine multiple reference computing power providers from the computing power transaction platform based on the computing power demand information; Obtain historical computing power transaction data of multiple reference computing power providers, analyze multiple groups of historical computing power transaction data, and determine the computing power transaction evaluation value of each reference computing power provider; Determine the target computing power provider from multiple reference computing power providers based on the preset computing power transaction screening threshold, and recommend computing power transaction objects to the computing power demander based on the target computing power provider; In response to the computing power demander selecting to conduct a computing power transaction with the target computing power provider, the computing power provider provides computing power services to the computing power demander based on the target computing power provider; Among them, multiple groups of historical computing power transaction data are analyzed to determine the computing power transaction evaluation value of each reference computing power provider, including: The task execution time data, resource utilization data and failure rate data corresponding to different computing tasks in each set of historical computing power transaction data are extracted, and the efficiency factor and failure rate factor are determined according to the task execution time data and failure rate data. The resource utilization factor is determined by analyzing the resource utilization data based on the computing power configuration data of each reference computing power provider. The computing power transaction evaluation value of each reference computing power provider is calculated using the following formula:

[0006] In the formula, is the computing power transaction evaluation value, is the efficiency factor, is the failure rate factor, is the resource utilization factor, , , are the weight parameters of efficiency factor, failure rate factor and resource utilization factor respectively.

[0007] Preferably, the efficiency factor, the failure rate factor and the resource utilization factor further include: For any reference computing power provider, determine the reference computing amount of each computing task in the historical computing power transaction data of the reference computing power provider, determine the first completion time and the second completion time of each computing task based on the reference computing amount of each computing task, and calculate the efficiency reference factor of each computing task based on the actual completion time of each computing task: ,in, is the first completion time, The second completion time, The actual completion time; Calculate the average of the efficiency reference factors of multiple computing tasks and record it as the efficiency factor of the reference computing power provider; Calculate the average failure rate of multiple computing tasks, and determine the failure rate factor of the reference computing power provider based on the average failure rate: ,in, is the mean failure rate; Multiple computing resource usage parameters of each computing task are calculated based on the computing power configuration data of the computing power provider and the resource utilization data of each computing task. The resource utilization reference factor of each computing task is determined based on the multiple computing resource usage parameters: ,in, Indicates the first Resource usage parameters, Indicates The standard parameters of resource usage parameters, Indicates the number of resource usage parameters; The average of the resource utilization reference factors of multiple computing tasks is calculated and recorded as the resource utilization factor of the reference computing power provider.

[0008] Preferably, determining a target computing power provider from a plurality of reference computing power providers based on a preset computing power transaction screening threshold comprises: Select multiple reference computing power providers whose computing power transaction evaluation values ​​are greater than a preset computing power transaction screening threshold from multiple reference computing power providers and construct a target computing power set; Calculate the computing power stability factor of each reference computing power provider in the target computing power set based on the historical computing power transaction data corresponding to each reference computing power provider, calculate the computing power transaction reference price of each reference computing power provider in the target computing power set based on the computing power stability factor, and select the reference computing power provider with the lowest computing power transaction reference price as the target computing power provider; Among them, the reference price of computing power transaction of each reference computing power provider in the target computing power set is calculated according to the computing power stability factor, including:

[0009] In the formula, It is the reference price for computing power transactions. Estimated price for computing power transactions, It is the computing power stability factor.

[0010] Preferably, the computing power stability factor of each reference computing power provider in the target computing power set is calculated according to the historical computing power transaction data corresponding to each reference computing power provider, including: For any reference computing power provider, the execution time deviation of each computing task is calculated based on the mean value of the efficiency reference factor, and the resource usage deviation of each computing task is calculated based on the resource utilization reference factor; The computing power stability reference factor of each computing task is calculated using the following formula:

[0011] In the formula, It is a reference factor for computing power stability. is the execution time deviation, is resource usage bias, is the failure rate; Calculate the standard deviation of the computing power stability reference factors of multiple computing tasks under the reference computing power provider as the computing power stability factor of the reference computing power provider.

[0012] Preferably, the difference between the efficiency reference factor of the computing task and the average of multiple efficiency reference factors is the execution time deviation of the computing task, and the difference between the resource utilization reference factor of the computing task and the average of the resource utilization reference factors is the resource usage deviation of the computing task.

[0013] Preferably, the computing power demand information includes computing task type data, computing resource specification data, computing task volume data, time requirement data and budget data.

[0014] A second aspect of the present invention provides a computing power resource transaction management system, which is used to implement the above-mentioned computing power resource transaction management method, including: A computing power transaction request processing module is used to obtain a computing power transaction request issued by a computing power demander, determine computing power demand information according to the computing power transaction request, and determine multiple reference computing power providers from the computing power transaction platform based on the computing power demand information; The computing power transaction evaluation module is used to obtain the historical computing power transaction data of multiple reference computing power providers, analyze multiple groups of historical computing power transaction data, and determine the computing power transaction evaluation value of each reference computing power provider; A computing power transaction provider recommendation module is used to determine a target computing power provider from multiple reference computing power providers based on a preset computing power transaction screening threshold, and recommend a computing power transaction object to a computing power demander based on the target computing power provider; The computing power transaction execution module is used to respond to the computing power demander's choice to conduct computing power transactions with the target computing power provider, and provide computing power services to the computing power demander based on the target computing power provider.

[0015] Preferably, for the computing power transaction evaluation module, multiple groups of historical computing power transaction data are analyzed to determine the computing power transaction evaluation value of each reference computing power provider, including: The task execution time data, resource utilization data and failure rate data corresponding to different computing tasks in each set of historical computing power transaction data are extracted, and the efficiency factor and failure rate factor are determined according to the task execution time data and failure rate data. The resource utilization factor is determined by analyzing the resource utilization data based on the computing power configuration data of each reference computing power provider. The computing power transaction evaluation value of each reference computing power provider is calculated using the following formula:

[0016] In the formula, is the computing power transaction evaluation value, is the efficiency factor, is the failure rate factor, is the resource utilization factor, , , are the weight parameters of efficiency factor, failure rate factor and resource utilization factor respectively.

[0017] The present invention has the following beneficial effects: The present invention conducts a detailed analysis of historical computing power transaction data, fully considers factors such as hardware performance fluctuations, task complexity, and system load changes that may occur in the calculation process, and generates a computing power transaction evaluation value for evaluating the stability and reliability of the computing power provider by analyzing the computing efficiency factor, failure rate factor, and resource utilization factor. It provides a comprehensive, accurate, standardized, and quantified computing power provider evaluation method, which not only improves the accuracy and comprehensiveness of computing resource selection, but also facilitates automated evaluation and decision-making, which helps computing power demanders to complete computing tasks more scientifically, economically, and efficiently when selecting and using computing resources, thereby improving the computing power transaction experience of computing power demanders. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flowchart of a computing resource transaction management method provided in an embodiment of the present invention.

[0019] Figure 2 A schematic diagram of the structure of a computing power resource transaction management system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to make those skilled in the art better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0021] See also Figure 1 The present invention provides a computing resource transaction management method, which specifically includes the following steps: S10, obtaining a computing power transaction request issued by a computing power demander, determining computing power demand information according to the computing power transaction request, and determining multiple reference computing power providers from the computing power transaction platform based on the computing power demand information; Specifically, the computing power trading platform is an intermediary platform that connects computing power providers and demanders, providing transaction infrastructure and services. The platform is responsible for the release, matching, transaction execution, payment and settlement of computing power resources. The computing power providers who have computing resources and are willing to rent or sell them can provide computing power services to computing power demanders through the computing power trading platform. For the computing power transaction requests issued by the computing power demanders, the computing power demand information is extracted from them, including but not limited to computing task type data: such as data processing, AI training, image rendering, etc.; computing resource specification data: such as the number of CPU cores, the number of GPUs, memory size, storage capacity, etc.; computing task volume data: such as the expected computing volume (GFLOPS, data processing volume, etc.); time requirement data: such as task start time, deadline, execution time range, etc.; budget data: the range of fees that the demander is willing to pay. Based on the computing power demand information, multiple computing power providers can be preliminarily screened to determine multiple reference computing power providers that meet the computing power demand information.

[0022] S20, obtaining historical computing power transaction data of multiple reference computing power providers, analyzing multiple groups of historical computing power transaction data, and determining a computing power transaction evaluation value of each reference computing power provider; Specifically, the historical computing power transaction data at least includes the transaction data involved in each computing power transaction process, as well as specific task execution data, such as time data, resource utilization data, computing volume data, fault data, etc. during the execution of computing tasks. When computing resources of different computing power providers execute tasks, due to the influence of various factors such as hardware performance, task complexity, and system load, the actual execution time and resource utilization may fluctuate greatly, resulting in differences in the actual execution of different reference computing power providers even if multiple reference computing power providers meet the computing power requirements. In order to improve the experience of computing power demanders, the historical computing power transaction data of each reference computing power provider is analyzed to calculate the computing power transaction evaluation value that represents the degree to which each reference computing power provider is worth trading, so as to assist computing power demanders in selecting computing power providers.

[0023] S30, determining a target computing power provider from multiple reference computing power providers based on a preset computing power transaction screening threshold, and recommending a computing power transaction object to a computing power demander based on the target computing power provider; S40, in response to the computing power demander selecting to conduct a computing power transaction with the target computing power provider, providing computing power services to the computing power demander based on the target computing power provider; Specifically, the target computing power provider is determined from multiple reference computing power providers in combination with the preset computing power transaction screening threshold. If the computing power transaction evaluation value corresponding to the target computing power provider is greater than the preset computing power transaction screening threshold, the target computing power provider is recommended to the computing power demander as the computing power transaction object. If the computing power demander chooses to conduct computing power transaction with the target computing power provider, computing power services can be provided to the computing power demander based on the target computing power provider to complete the computing power resource transaction process.

[0024] In an optional implementation process, for step S20, multiple groups of historical computing power transaction data are analyzed to determine the computing power transaction evaluation value of each reference computing power provider, specifically including: The task execution time data, resource utilization data and failure rate data corresponding to different computing tasks in each set of historical computing power transaction data are extracted, the efficiency factor and failure rate factor are determined according to the task execution time data and failure rate data, and the resource utilization factor is determined by analyzing the resource utilization data based on the computing power configuration data of each reference computing power provider.

[0025] Taking any reference computing power provider as an example, for the calculation of the efficiency factor, the reference computing amount of each computing task in the historical computing power transaction data of the reference computing power provider is determined, and the first completion time and second completion time of each computing task are determined based on the reference computing amount of each computing task.

[0026] Specifically, for multiple computing tasks involving multiple computing power transactions in the historical computing power transaction data, the reference computing amount corresponding to the multiple computing tasks involved can be determined based on the historical computing power transaction data, and then the first completion time and the second completion time required to complete the reference computing amount can be estimated based on the computing power configuration data of the reference computing power provider. Specifically, the estimation is based on the theoretical performance of the computing power provider, and the first completion time and the second completion time correspond to the longest completion time and the shortest completion time of the reference computing power provider, respectively. Exemplarily, the theoretical computing amount of task X is 1000 GFLOPS. The maximum computing power of the computing power provider under high performance is 200 GFLOPS, and the estimated shortest completion time is 5 seconds. The minimum computing power under low performance is 50 GFLOPS, and the estimated longest completion time is 20 seconds.

[0027] It is worth noting that the reference computing amount can be estimated based on historical data. For example, based on historical data, it is determined that the actual computing amount of processing 1,000 images based on a certain AI model is 200 GFLOPS (Giga Floating Point Operations Per Second). For a new computing task, for example, 5,000 images need to be processed by the corresponding model, and the estimated reference computing amount is 1000 GFLOPS. Optionally, it can also be determined based on the actual execution data of each computing task, for example, according to the task execution time (the time from the start to the end of the task), resource utilization (the utilization of resources such as CPU, GPU, memory, etc. during the task execution process) and hardware performance (performance parameters of computing resources, such as GFLOPS of CPU, TFLOPS of GPU, etc.) and other data, to determine the actual computing amount, and then consider the computing amount wasted in the fault process, so as to obtain the computing amount required to complete the corresponding computing task in theory, that is, the reference computing amount. The method for determining the reference computing amount is a technical means well known to those skilled in the art, and will not be repeated in this embodiment.

[0028] After determining the first completion time and the second completion time of each computing task, the efficiency reference factor of each computing task is calculated according to the actual completion time of each computing task, specifically: Efficiency reference factor = ,in, is the first completion time, The second completion time, The actual completion time.

[0029] After the efficiency reference factor of each computing task is calculated in the above manner, the average of the efficiency reference factors of multiple computing tasks is calculated and recorded as the efficiency factor of the reference computing power provider.

[0030] For the calculation of the failure rate factor, the average failure rate of multiple computing tasks is calculated, and the failure rate factor of the reference computing power provider is determined based on the average failure rate. Specifically: Failure rate factor = ,in, is the mean failure rate.

[0031] For the calculation of the resource utilization factor, multiple computing resource usage parameters of each computing task are calculated based on the computing power configuration data of the reference computing power provider and the resource utilization data of each computing task, wherein the computing resource usage parameter is used to characterize the usage of a certain computing resource per unit time. For example, assuming that the CPU parameter of the reference computing power provider is 100 GFLOPS and the CPU utilization rate when executing a certain computing task is 70%, the computing resource usage parameter corresponding to the CPU is 70GFLOPS, which is used to characterize the actual CPU computing power used.

[0032] After calculating multiple computing resource usage parameters for each computing task, the resource utilization reference factor for each computing task is determined using the following formula: Resource utilization reference factor = ,in, Indicates the first Resource usage parameters, Indicates The standard parameters of resource usage parameters, Indicates the number of resource usage parameters.

[0033] Specifically, the standard parameter of the resource usage parameter is a pre-set value, which is used to convert multiple resource usage parameters into dimensionless parameters to achieve data standardization. After calculating the resource utilization reference factor of each computing task, the average of the resource utilization reference factors of multiple computing tasks is calculated and recorded as the resource utilization factor of the reference computing power provider.

[0034] Finally, based on the efficiency factor, failure rate factor and resource utilization factor, the following formula is used to calculate the computing power transaction evaluation value of each reference computing power provider:

[0035] In the formula, is the computing power transaction evaluation value, is the efficiency factor, is the failure rate factor, is the resource utilization factor, , , are the weight parameters of efficiency factor, failure rate factor and resource utilization factor respectively.

[0036] The above-mentioned historical computing power transaction data analysis and computing power transaction evaluation value calculation process, through detailed analysis of historical computing power transaction data, the computing power transaction evaluation value calculated can accurately evaluate the computing power and reliability of each computing power provider, taking into account multiple key factors such as task execution time, resource utilization and failure rate, and comprehensively reflecting the actual performance of the computing power provider, which can help computing power demanders to more accurately select computing power resources that meet their needs, avoid performance problems and cost increases caused by improper selection, and evaluate the comprehensive performance of computing power providers from multiple angles to avoid the bias that may be caused by a single indicator, making the evaluation results more comprehensive and reliable, and the resource utilization of different computing power providers can be fairly compared under the same standard. The solution for quantitative analysis of computing power transaction evaluation values ​​can be directly used in automated evaluation and decision-making processes, is convenient for computer system processing, and can be integrated into the computing power trading platform to realize the automation and intelligence of computing power resource selection.

[0037] In an optional implementation process, for step S30, determining a target computing power provider from multiple reference computing power providers based on a preset computing power transaction screening threshold specifically includes: Select multiple reference computing power providers whose computing power transaction evaluation values ​​are greater than a preset computing power transaction screening threshold from multiple reference computing power providers and construct a target computing power set; Specifically, taking the preset computing power transaction screening threshold as the reference standard, a preliminary screening is first performed on multiple reference computing power providers, and the multiple reference computing power providers obtained by screening are stored through the target computing power set.

[0038] Calculate the computing power stability factor of each reference computing power provider in the target computing power set based on the historical computing power transaction data corresponding to each reference computing power provider, calculate the computing power transaction reference price of each reference computing power provider in the target computing power set based on the computing power stability factor, and select the reference computing power provider with the lowest computing power transaction reference price as the target computing power provider; Among them, the computing power stability factor of each reference computing power provider in the target computing power set is calculated according to the historical computing power transaction data corresponding to each reference computing power provider, specifically including: For any reference computing power provider, the execution time deviation of each computing task is calculated according to the mean of the efficiency reference factor, and the resource usage deviation of each computing task is calculated according to the resource utilization reference factor, wherein the execution time deviation of the computing task is specifically the difference between the efficiency reference factor of the computing task and the mean of multiple efficiency reference factors, and the resource usage deviation of the computing task is specifically the difference between the resource utilization reference factor of the computing task and the mean of the resource utilization reference factors.

[0039] After calculating the execution time deviation and resource usage deviation of each computing task, the computing power stability reference factor of each computing task is calculated using the following formula:

[0040] In the formula, It is a reference factor for computing power stability. is the execution time deviation, is resource usage bias, is the failure rate; After calculating the computing power stability reference factor of each computing task corresponding to the reference computing power provider in the above manner, the standard deviation of the computing power stability reference factors of multiple computing tasks under the reference computing power provider is calculated as the computing power stability factor of the reference computing power provider.

[0041] It is worth noting that, considering the computing power transaction process, some pricing methods use time-based pricing to determine the computing power transaction price. These methods do not take into account the fluctuations in actual execution time and resource utilization caused by various factors such as hardware performance, task complexity, and system load. As a result, in the process of providing computing power services to computing power demanders, the actual time spent is greater than the pre-estimated price, which may cause cost overruns for computing power demanders when experiencing computing power services.

[0042] In this embodiment, based on the historical computing power transaction data corresponding to each reference computing power provider, each reference computing power provider in the target computing power set is analyzed from the perspective of computing power stability, and the computing power stability factor of each reference computing power provider is calculated. The computing power transaction estimated price is corrected by the computing power stability factor, and the computing power transaction reference price of each reference computing power provider is calculated. The reference computing power provider with the lowest computing power transaction reference price is selected and recorded as the target computing power provider.

[0043] Specifically, the reference price of computing power transaction of each reference computing power provider in the target computing power set is calculated based on the computing power stability factor, including:

[0044] In the formula, It is the reference price for computing power transactions. Estimated price for computing power transactions, It is the computing power stability factor.

[0045] It is worth noting that the estimated price of computing power trading is specifically the computing amount estimated according to the computing task that needs to be performed and the estimated time required for the computing power provider to complete the computing amount. The computing power service price to be paid under the estimated time is then determined according to the charging rules. On this basis, taking into account the computing power stability factor, the highest estimated price under fluctuations is calculated, which is the computing power trading reference price.

[0046] The calculation process of the above-mentioned reference price of computing power trading, from the perspective that the traditional pricing method based on time fails to fully consider the factors such as hardware performance fluctuations, task complexity and system load changes that may occur during the calculation process, which easily leads to the actual execution time exceeding expectations, thereby causing the problem of cost overruns, introduces the computing power stability factor to more accurately estimate the reference price of computing power trading, improve the accuracy of cost prediction, ensure that the computing power demander can complete the computing task within the expected cost, and help the computing power demander better understand and select the computing power resources that best suit their needs. The analysis of the computing power stability factor can assist in identifying those computing power providers that are stable and reliable under different computing loads, thereby improving the reliability of computing power resource selection. For the computing power demander, it can avoid task interruptions or calculation errors caused by unstable computing resources and ensure the smooth completion of the task. By selecting the computing power provider with the lowest reference price of computing power trading, the computing power demander can further optimize costs and improve economic benefits on the premise of ensuring the smooth completion of the computing task.

[0047] See also Figure 2 , a computing power resource transaction management system provided by an embodiment of the present invention includes: A computing power transaction request processing module is used to obtain a computing power transaction request issued by a computing power demander, determine computing power demand information according to the computing power transaction request, and determine multiple reference computing power providers from the computing power transaction platform based on the computing power demand information; The computing power transaction evaluation module is used to obtain the historical computing power transaction data of multiple reference computing power providers, analyze multiple groups of historical computing power transaction data, and determine the computing power transaction evaluation value of each reference computing power provider; Among them, multiple groups of historical computing power transaction data are analyzed to determine the computing power transaction evaluation value of each reference computing power provider, including: The task execution time data, resource utilization data and failure rate data corresponding to different computing tasks in each set of historical computing power transaction data are extracted, and the efficiency factor and failure rate factor are determined according to the task execution time data and failure rate data. The resource utilization factor is determined by analyzing the resource utilization data based on the computing power configuration data of each reference computing power provider. The computing power transaction evaluation value of each reference computing power provider is calculated using the following formula:

[0048] In the formula, is the computing power transaction evaluation value, is the efficiency factor, is the failure rate factor, is the resource utilization factor, , , are the weight parameters of efficiency factor, failure rate factor and resource utilization factor respectively.

[0049] For efficiency factors, failure rate factors, and resource utilization factors, it also includes: For any reference computing power provider, determine the reference computing amount of each computing task in the historical computing power transaction data of the reference computing power provider, determine the first completion time and the second completion time of each computing task based on the reference computing amount of each computing task, and calculate the efficiency reference factor of each computing task based on the actual completion time of each computing task: ,in, is the first completion time, The second completion time, The actual completion time; Calculate the average of the efficiency reference factors of multiple computing tasks and record it as the efficiency factor of the reference computing power provider; Calculate the average failure rate of multiple computing tasks, and determine the failure rate factor of the reference computing power provider based on the average failure rate: ,in, is the mean failure rate; Multiple computing resource usage parameters of each computing task are calculated based on the computing power configuration data of the computing power provider and the resource utilization data of each computing task. The resource utilization reference factor of each computing task is determined based on the multiple computing resource usage parameters: ,in, Indicates the first Resource usage parameters, Indicates The standard parameters of resource usage parameters, Indicates the number of resource usage parameters; The average of the resource utilization reference factors of multiple computing tasks is calculated and recorded as the resource utilization factor of the reference computing power provider.

[0050] A computing power transaction provider recommendation module is used to determine a target computing power provider from multiple reference computing power providers based on a preset computing power transaction screening threshold, and recommend a computing power transaction object to a computing power demander based on the target computing power provider; Wherein, determining a target computing power provider from multiple reference computing power providers based on a preset computing power transaction screening threshold includes: Select multiple reference computing power providers whose computing power transaction evaluation values ​​are greater than a preset computing power transaction screening threshold from multiple reference computing power providers and construct a target computing power set; Calculate the computing power stability factor of each reference computing power provider in the target computing power set based on the historical computing power transaction data corresponding to each reference computing power provider, calculate the computing power transaction reference price of each reference computing power provider in the target computing power set based on the computing power stability factor, and select the reference computing power provider with the lowest computing power transaction reference price as the target computing power provider; Among them, the reference price of computing power transaction of each reference computing power provider in the target computing power set is calculated according to the computing power stability factor, including:

[0051] In the formula, It is the reference price for computing power transactions. Estimated price for computing power transactions, It is the computing power stability factor.

[0052] The computing power transaction execution module is used to respond to the computing power demander's choice to conduct computing power transactions with the target computing power provider, and provide computing power services to the computing power demander based on the target computing power provider.

[0053] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A computing resource transaction management method, characterized in that: include: Obtain a computing power transaction request from a computing power demander, determine computing power demand information according to the computing power transaction request, and determine multiple reference computing power providers from the computing power transaction platform based on the computing power demand information; Obtain historical computing power transaction data of multiple reference computing power providers, analyze multiple groups of historical computing power transaction data, and determine the computing power transaction evaluation value of each reference computing power provider; Determine the target computing power provider from multiple reference computing power providers based on the preset computing power transaction screening threshold, and recommend computing power transaction objects to the computing power demander based on the target computing power provider; In response to the computing power demander selecting to conduct a computing power transaction with the target computing power provider, the computing power provider provides computing power services to the computing power demander based on the target computing power provider; Among them, multiple groups of historical computing power transaction data are analyzed to determine the computing power transaction evaluation value of each reference computing power provider, including: The task execution time data, resource utilization data and failure rate data corresponding to different computing tasks in each set of historical computing power transaction data are extracted, and the efficiency factor and failure rate factor are determined according to the task execution time data and failure rate data. The resource utilization factor is determined by analyzing the resource utilization data based on the computing power configuration data of each reference computing power provider. The computing power transaction evaluation value of each reference computing power provider is calculated using the following formula: In the formula, is the computing power transaction evaluation value, is the efficiency factor, is the failure rate factor, is the resource utilization factor, , , are the weight parameters of efficiency factor, failure rate factor and resource utilization factor respectively.

2. A computing resource transaction management method according to claim 1, characterized in that: For efficiency factors, failure rate factors, and resource utilization factors, it also includes: For any reference computing power provider, determine the reference computing amount of each computing task in the historical computing power transaction data of the reference computing power provider, determine the first completion time and the second completion time of each computing task based on the reference computing amount of each computing task, and calculate the efficiency reference factor of each computing task based on the actual completion time of each computing task: ,in, is the first completion time, The second completion time, The actual completion time; Calculate the average of the efficiency reference factors of multiple computing tasks and record it as the efficiency factor of the reference computing power provider; Calculate the average failure rate of multiple computing tasks, and determine the failure rate factor of the reference computing power provider based on the average failure rate: ,in, is the mean failure rate; Multiple computing resource usage parameters of each computing task are calculated based on the computing power configuration data of the computing power provider and the resource utilization data of each computing task. The resource utilization reference factor of each computing task is determined based on the multiple computing resource usage parameters: ,in, Indicates the first Resource usage parameters, Indicates The standard parameters of resource usage parameters, Indicates the number of resource usage parameters; The average of the resource utilization reference factors of multiple computing tasks is calculated and recorded as the resource utilization factor of the reference computing power provider.

3. A computing resource transaction management method according to claim 2, characterized in that: The target computing power provider is determined from multiple reference computing power providers based on a preset computing power transaction screening threshold, including: Select multiple reference computing power providers whose computing power transaction evaluation values ​​are greater than a preset computing power transaction screening threshold from multiple reference computing power providers and construct a target computing power set; Calculate the computing power stability factor of each reference computing power provider in the target computing power set based on the historical computing power transaction data corresponding to each reference computing power provider, calculate the computing power transaction reference price of each reference computing power provider in the target computing power set based on the computing power stability factor, and select the reference computing power provider with the lowest computing power transaction reference price as the target computing power provider; Among them, the reference price of computing power transaction of each reference computing power provider in the target computing power set is calculated according to the computing power stability factor, including: In the formula, It is the reference price for computing power transactions. Estimated price for computing power transactions, It is the computing power stability factor.

4. A computing power resource transaction management method according to claim 3, characterized in that: The computing power stability factor of each reference computing power provider in the target computing power set is calculated based on the historical computing power transaction data corresponding to each reference computing power provider, including: For any reference computing power provider, the execution time deviation of each computing task is calculated based on the mean value of the efficiency reference factor, and the resource usage deviation of each computing task is calculated based on the resource utilization reference factor; The computing power stability reference factor of each computing task is calculated using the following formula: In the formula, It is a reference factor for computing power stability. is the execution time deviation, is resource usage bias, is the failure rate; Calculate the standard deviation of the computing power stability reference factors of multiple computing tasks under the reference computing power provider as the computing power stability factor of the reference computing power provider.

5. A computing resource transaction management method according to claim 4, characterized in that: The difference between the efficiency reference factor of the computing task and the average of multiple efficiency reference factors is the execution time deviation of the computing task, and the difference between the resource utilization reference factor of the computing task and the average of the resource utilization reference factors is the resource usage deviation of the computing task.

6. A computing power resource transaction management method according to claim 5, characterized in that: The computing power demand information includes computing task type data, computing resource specification data, computing task volume data, time requirement data and budget data.

7. A computing power resource transaction management system, used to implement a computing power resource transaction management method as described in any one of claims 1 to 6, characterized in that: include: A computing power transaction request processing module is used to obtain a computing power transaction request issued by a computing power demander, determine computing power demand information according to the computing power transaction request, and determine multiple reference computing power providers from the computing power transaction platform based on the computing power demand information; The computing power transaction evaluation module is used to obtain the historical computing power transaction data of multiple reference computing power providers, analyze multiple groups of historical computing power transaction data, and determine the computing power transaction evaluation value of each reference computing power provider; A computing power transaction provider recommendation module is used to determine a target computing power provider from multiple reference computing power providers based on a preset computing power transaction screening threshold, and recommend a computing power transaction object to a computing power demander based on the target computing power provider; The computing power transaction execution module is used to respond to the computing power demander's choice to conduct computing power transactions with the target computing power provider, and provide computing power services to the computing power demander based on the target computing power provider.

8. A computing power resource transaction management system according to claim 7, characterized in that: For the computing power transaction evaluation module, multiple groups of historical computing power transaction data are analyzed to determine the computing power transaction evaluation value of each reference computing power provider, including: The task execution time data, resource utilization data and failure rate data corresponding to different computing tasks in each set of historical computing power transaction data are extracted, and the efficiency factor and failure rate factor are determined according to the task execution time data and failure rate data. The resource utilization factor is determined by analyzing the resource utilization data based on the computing power configuration data of each reference computing power provider. The computing power transaction evaluation value of each reference computing power provider is calculated using the following formula: In the formula, is the computing power transaction evaluation value, is the efficiency factor, is the failure rate factor, is the resource utilization factor, , , are the weight parameters of efficiency factor, failure rate factor and resource utilization factor respectively.

Images