Computing power resource allocation method, storage medium and computer program product

Abstract

The invention discloses a computing power resource allocation method, a storage medium and a computer program product. Relates to the technical field of communication. The method comprises the following steps: acquiring a target task demand of a target account; determining target weight values respectively corresponding to the plurality of computing power factors based on the target task demand; based on computing power factor sets corresponding to the multiple computing power resources and target weight values corresponding to the multiple computing power factors, scoring results corresponding to the multiple computing power resources are determined, and the computing power factor sets comprise computing power factor measurement parameters corresponding to the corresponding computing power resources in the multiple computing power factors; and according to the scoring results corresponding to the plurality of computing power resources, determining a target computing power resource allocated to the target account from the plurality of computing power resources. According to the method and the device, the technical problems of low allocation efficiency and poor flexibility when a unified cloud platform is utilized to manage and allocate various computing power resources in related technologies are solved.

Description

Technical Field

[0001] The present invention relates to the field of communication technologies, and in particular to a computing resource allocation method, a storage medium, and a computer program product. Background Art

[0002] At present, the heterogeneous cloud management platform in related technologies provides unified cloud resource management capabilities, covering cloud resource orchestration, cloud resource configuration, unified monitoring and operation and maintenance of cloud resources and other basic cloud management functions. Some cloud vendors can support network interconnection within a single cloud, but to achieve network interconnection between heterogeneous cloud resources, users still need to use the traditional manual configuration routing method. It is impossible to truly achieve network interconnection between heterogeneous clouds through heterogeneous cloud management platforms, and it is also impossible to achieve true cloud-network integrated orchestration.

[0003] In addition, although the heterogeneous cloud management platform in the related technology provides users with a cloud-neutral management platform, users can implement basic configuration monitoring and other operations on cloud resources without paying attention to the differences between different cloud providers, but it only provides the role of the infrastructure management layer in the entire computing power scheduling platform, and is far from achieving the unified orchestration and scheduling of computing power, disaster recovery processing of resources during the application process, and dynamic adjustment of resources according to performance. This results in low computing power allocation efficiency and poor flexibility.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present invention provide a computing resource allocation method, storage medium and computer program product to at least solve the technical problems of low allocation efficiency and poor flexibility in the related art when using a unified cloud platform to manage and allocate various computing resources.

[0006] According to one aspect of an embodiment of the present invention, a computing power resource allocation method is provided, comprising: obtaining a target task requirement of a target account; determining target weight values ​​corresponding to a plurality of computing power factors respectively based on the target task requirement; determining scoring results corresponding to the plurality of computing power resources respectively based on a set of computing power factors corresponding to the plurality of computing power resources respectively, and the target weight values ​​corresponding to the plurality of computing power factors respectively, wherein the computing power factor set includes computing power factor measurement parameters corresponding to the plurality of computing power factors respectively of the corresponding computing power resources; determining a target computing power resource allocated to the target account from the plurality of computing power resources according to the scoring results corresponding to the plurality of computing power resources respectively.

[0007] According to another aspect of an embodiment of the present invention, a non-volatile storage medium is provided, wherein the non-volatile storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing any one of the computing power resource allocation methods.

[0008] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps of any one of the computing power resource allocation methods are implemented.

[0009] In an embodiment of the present invention, the target task requirements of the target account are obtained; based on the target task requirements, the target weight values ​​corresponding to the multiple computing power factors are determined; based on the computing power factor sets corresponding to the multiple computing power resources, and the target weight values ​​corresponding to the multiple computing power factors, the scoring results corresponding to the multiple computing power resources are determined, wherein the computing power factor set includes computing power factor measurement parameters corresponding to the corresponding computing power resources in the multiple computing power factors; according to the scoring results corresponding to the multiple computing power resources, the target computing power resources allocated to the target account are determined from the multiple computing power resources, so as to achieve the purpose of dynamically allocating computing power factor weights and calculating scoring results according to the user's task requirements, so as to flexibly allocate computing power resources, thereby achieving the technical effect of improving the computing power resource allocation efficiency and allocation flexibility when managing and allocating multiple computing power resources, thereby solving the technical problems of low allocation efficiency and poor flexibility in the related technology when using a unified cloud platform to manage and allocate multiple computing power resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0011] Figure 1 is a flow chart of a computing resource allocation method according to an embodiment of the present invention;

[0012] Figure 2 is a schematic diagram of an optional computing-network integration system framework according to an embodiment of the present invention;

[0013] Figure 3 It is an optional interactive logic diagram based on a computing-network integration system according to an embodiment of the present invention;

[0014] Figure 4 is a schematic diagram of an optional computing power scheduling process according to an embodiment of the present invention;

[0015] Figure 5 It is a structural diagram of a computing resource allocation device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0016] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0017] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0018] First, to facilitate understanding of the embodiments of the present invention, some terms or nouns involved in the present invention are explained below:

[0019] Heterogeneous cloud management platform is a management tool that can manage and integrate multiple cloud computing platforms of different types and manufacturers. This platform can help users uniformly manage and monitor multiple cloud computing resources, including public clouds, private clouds, hybrid clouds, etc., so that users can use cloud computing resources more efficiently and realize automatic allocation and optimization of resources.

[0020] With the rapid development of AI big model technology, the demand for computing resources has increased rapidly, but it faces challenges such as insufficient computing resources, imbalance between supply and demand, and waste of computing resources. When users purchase computing resources and services, they need to query on multiple computing platforms. There is a lack of a unified resource allocation interface, efficiency cannot be maximized, and there is a certain degree of waste of computing resources. At present, the best solution is to achieve unified scheduling of computing resources, but there are still many problems and bottlenecks in computing scheduling. The main reason is that cloud computing vendors have grown independently, with the main goal of quickly occupying the market in the early stage. They have not invested enough in the reliability of cloud services and the integration technology of cloud services, and the development of multi-cloud interconnection technology platforms and cloud services is seriously unbalanced. The key technologies and difficult problems are mainly the complexity of unified multi-cloud management, how to ensure that the performance of applications is consistent in multi-cloud scenarios when computing resources are scheduled, how to overcome the barriers and technical differences of multi-cloud heterogeneous interconnection networks with different computing resources; and how to achieve information communication between computing resources through the existing network to achieve the goal of a true computing network.

[0021] The current heterogeneous cloud management platform technology mainly solves the problem of unified management of multi-cloud environments, focusing on the Platform as a Service (Paas) layer to provide users with a unified heterogeneous cloud management and operation platform. By designing a unified adaptation layer in the background of the platform, connecting to different cloud platforms in the south, shielding the different interfaces of the underlying cloud platforms, and providing a unified interface call in the north. Using a unified cloud management interface to manage different cloud platforms, users can get a consistent experience and see the global resource usage.

[0022] However, the heterogeneous cloud management platform in the related technology provides the unified management capability of cloud resources, and its functions cover the basic cloud management functions such as the orchestration of cloud resources, configuration of cloud resources, unified monitoring and operation and maintenance of cloud resources. Some cloud vendors can support network interconnection within a single cloud, but to achieve network interconnection between heterogeneous cloud resources, users still need to use the traditional manual configuration routing method. It is impossible to truly achieve network interconnection between heterogeneous clouds through the heterogeneous cloud management platform, and it is also impossible to achieve true cloud-network integrated orchestration. In addition, although the heterogeneous cloud management platform in the related technology provides users with a cloud-neutral management platform, users can ignore the differences between different cloud vendors and implement basic configuration monitoring and other operations on cloud resources, but it only provides the role of the infrastructure management layer in the entire computing power scheduling platform, and is far from meeting the requirements of unified orchestration and scheduling of computing power, disaster recovery processing of resources during the application process, and dynamic adjustment of resources according to performance.

[0023] In summary, the unified scheduling of computing resources in related technologies cannot achieve unified coordination and management of computing resources of different cloud providers, cannot compare resource utilization and other resource performance across cloud providers, and truly achieve reasonable allocation of computing resources across the entire network. The computing power allocation efficiency is low and the flexibility is poor.

[0024] In response to the above problems, an embodiment of the present invention provides an embodiment of a method for allocating computing power resources. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0025] Figure 1 is a flow chart of a computing resource allocation method according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0026] Step S102, obtaining the target task requirements of the target account;

[0027] In step S102 of the present application, by obtaining the task requirements of the target account, preparation is made for subsequently customizing the allocation strategy of computing resources according to the specific needs of each user.

[0028] As an optional embodiment, the execution subject of this embodiment can be a heterogeneous cloud management platform, which is used to access multiple different cloud platforms (i.e., cloud servers). Multiple different cloud platforms are used to provide multiple different heterogeneous cloud resources, which may include but are not limited to public cloud resources, private cloud resources, and edge device resources such as cloud boxes. Each heterogeneous cloud resource can be abstracted as a computing power resource with the same basic type and used as the computing power supply side. All resources can be measured by a unified computing power model. Through the heterogeneous cloud management platform, unified scheduling and allocation of multiple computing power resources can be performed, wherein multiple different cloud platforms correspond one-to-one to multiple computing power resources.

[0029] Optionally, the target task requirement is used to indicate the priority of the target account's requirements for different computing power factors. The target task requirement can be, but is not limited to, any of the following: computing efficiency first, computing cost first, network transmission efficiency first, and taking both computing power and network transmission into consideration.

[0030] In an optional embodiment, obtaining the target task requirements of the target account includes: obtaining the target business scenario requirements of the target account; determining the task requirements that match the target business scenario requirements among multiple task requirements as the target task requirements, wherein the multiple task requirements correspond one-to-one to the multiple business scenario requirements.

[0031] Optionally, the task requirements that match the target business scenario requirements are determined as the target task requirements, which means that the system can automatically filter out the most matching computing power configuration solution from the preset multiple task requirements based on the characteristics of the business scenario. This task requirement matching based on business scenarios can further improve the accuracy of resource allocation and ensure a high degree of fit between computing power resources and business needs. Multiple task requirements correspond one-to-one with multiple business scenario requirements, indicating the ability to dynamically map and adjust task requirements. This not only simplifies the expression of user requirements, but also allows the system to automatically adjust the task requirement configuration according to the changing business scenarios to adapt to real-time changes in the business.

[0032] Optionally, various business scenario requirements include: cloud computing (such as large-scale cloud computing), stable businesses such as online services, remote control and other businesses, and cloud-edge collaborative businesses. Different business scenario requirements correspond to different task requirements. The mapping relationship between business scenario requirements, task requirements and computing power characteristics is shown in Table 1 below.

[0033] Table 1

[0034]

[0035] It should be noted that, in view of the different demands for computing power in business scenarios, this embodiment analyzes the business scenario demands, maps them to the demands for computing power resources, and dynamically adjusts the computing power measurement scheme to achieve intelligent processing of scenario demands to computing power scheduling. For example, the business scenarios in Table 1 can be mapped to the corresponding task demands after analyzing the demand for resources of the tasks. According to the specific task demands, the weights of each computing power factor are dynamically adjusted (the basic adjustment principle is that the more attention the business scenario pays, the larger the adjustment coefficient of the corresponding factor) and combined with the weight calculation of other factors, the computing power factor measurement parameters are obtained in combination with the platform calculation to obtain the optimal computing power matching.

[0036] In an optional embodiment, a task requirement among multiple task requirements that matches the target business scenario requirement is determined as the target task requirement, including: querying whether there is a preset task requirement pre-configured for the target account; if no preset task requirement is found, determining the task requirement among multiple task requirements that matches the target business scenario requirement as the target task requirement.

[0037] It should be noted that, considering that users may have different business requirements in specific business scenarios, in order to better meet the business needs of users, it is set that users can customize the configuration of task requirements. That is, before determining the target task requirements according to the target business scenario, first check whether there are preset task requirements pre-configured for the target account, which means that users can customize the task requirement configuration according to the particularity of their own business or historical experience. This personalized setting can better meet the specific needs of users and improve the customization level of resource allocation. If the target account does not have preset task requirements, the most appropriate task requirements can be automatically matched from a variety of task requirements according to the requirements of the target business scenario. Automatic matching of task requirements based on business scenario requirements can reduce user intervention in resource selection, thereby improving the efficiency of task requirement configuration.

[0038] In an optional embodiment, the method further includes: in case a preset task requirement is queried, determining the preset task requirement as the target task requirement.

[0039] Optionally, if the system finds that the target account has pre-configured task requirements, these preset requirements will be directly used as the target task requirements to better meet the user's personalized task requirements and improve the user experience. The direct application of preset task requirements can avoid the system from repeatedly analyzing and matching task requirements every time resources are allocated, thereby improving the efficiency of resource allocation. Especially when user requirements are relatively fixed or have periodic changes, preset requirements can ensure that the system responds quickly, reduce decision-making time, and improve business processing speed.

[0040] Step S104, determining target weight values ​​corresponding to the multiple computing power factors respectively based on the target task requirements;

[0041] In step S104 of the present application, by assigning weights to different computing power factors, the focus of resource selection can be adjusted according to the specific task requirement priority of the target account. For example, if the target requirement emphasizes computing performance rather than cost, then the weights of factors such as CPU performance and GPU performance will be higher, so that options with excellent performance are more inclined in resource selection. Allocating computing power factor weights based on target task requirements can more accurately match user needs, dynamically adjust resource allocation, and improve user experience, especially in scenarios such as real-time processing, high concurrency, and large-scale data processing, ensuring that users obtain the best computing power support and service quality.

[0042] Optionally, multiple computing power factors may include, but are not limited to, computing power distribution (regional / area information), computing power resource type (central processing unit (CPU), graphics processing unit (GPU), tensor processing unit (TPU)), computing power resource device information (CPU core number / memory capacity / system disk capacity / hard disk capacity, etc.), computing power resource real-time utilization information, inbound and outbound network bandwidth, computing power resource unit cost, and one or more of network resource load.

[0043] In an optional embodiment, based on the target task requirements, target weight values ​​corresponding to multiple computing power factors are determined, including: querying whether there are preset weight values ​​corresponding to multiple computing power factors, wherein the preset weight values ​​are pre-configured for the target account; if the preset weight values ​​corresponding to multiple computing power factors are not queried, based on the target task requirements, the target weight values ​​corresponding to the multiple computing power factors are determined.

[0044] It should be noted that, considering that users may have different weight requirements for different computing power factors in specific business scenarios, in order to better meet the personalized task requirements of users and improve user experience. Allowing users to pre-configure the weight values ​​of computing power factors according to the characteristics and priorities of their own business means that users can customize the consideration criteria for computing power resource allocation, such as giving priority to cost, performance, geographical location or network latency, etc., and these priorities set by users can be respected first when allocating resources. In the case where the user does not pre-configure the weight value, the weight of the computing power factor will be automatically determined based on the target task requirements. This ensures that even without user intervention, resource allocation can be reasonably optimized according to the characteristics of the business scenario to ensure the best use of resources. The automatic weight allocation mechanism enables dynamic adaptation to different business scenarios and demand changes. Even if the user does not foresee a specific task demand scenario, the weight of the computing power factor can be adjusted in real time according to the characteristics and priorities of the current demand to provide the most appropriate resources.

[0045] In an optional embodiment, the method further includes: when preset weight values ​​corresponding to multiple computing power factors are queried, the preset weight values ​​corresponding to the multiple computing power factors are determined as target weight values ​​corresponding to the multiple computing power factors.

[0046] Optionally, when the preset weight values ​​corresponding to multiple computing power factors are queried, these weight values ​​can be directly used as the target weight values, that is, the user can set the weights of the computing power factors in advance according to business needs, cost considerations, performance priorities and other factors, and these user-customized weights will be applied preferentially during resource allocation to provide services that are closer to user needs. In addition, directly using the user-preset weight values ​​as the target weight values ​​can avoid the complex process of recalculating the weight values ​​each time resources are allocated, thereby improving the efficiency of resource allocation. Especially when user needs are relatively fixed or have periodic changes, the application of preset weight values ​​can ensure that the system responds quickly and reduce decision-making time. The use of preset weight values ​​can also simplify the decision-making process in resource allocation, especially for business scenarios where the weights of computing power factors do not change much or have clear preferences, and the priority and allocation strategy of computing power resources can be quickly determined, thereby avoiding unnecessary optimization calculations and reducing decision-making costs.

[0047] In an optional embodiment, based on the target task requirements, target weight values ​​corresponding to multiple computing power factors are determined, including: based on the target task requirements, demand scores corresponding to multiple computing power factors are determined, wherein the demand scores are used to indicate the target account's preference for the corresponding computing power factors; based on the demand scores corresponding to the multiple computing power factors, target weight values ​​corresponding to the multiple computing power factors are determined.

[0048] Optionally, by determining the demand score, you can understand the preference of different accounts or tasks for specific computing power factors. For example, a data processing task may prefer a high-speed central processing unit (CPU), while another task may require a large amount of storage space. Based on this personalized demand, resources can be allocated more accurately to meet the specific requirements of different tasks. The determination of the target weight value helps to optimize the efficiency of resource use. When resources are limited, computing power factors with high weight values ​​will be given priority, thereby ensuring that the most critical or urgent computing power needs are met, thereby improving the efficiency and response speed of overall task processing. The dynamic adjustment mechanism of the weight value enables better adaptation to changing task requirements. For example, if a task suddenly requires more memory, by adjusting the weight value, it can respond quickly and allocate more memory resources to the task. By accurately matching task requirements and computing power factors, resource waste can be reduced. For example, if a task does not require a high-performance CPU, then the reasonable setting of the weight value can ensure that CPU resources are allocated to tasks that need it more, thereby avoiding idleness and waste of resources. For users or accounts, resource allocation based on task requirements can significantly improve their experience. Users can obtain the most optimized resource support based on their task requirements, thereby speeding up task completion.

[0049] Optionally, if the user does not manually specify the weights of different types of factors, that is, the preset weight values ​​set for multiple computing power factors by the target account are not queried, then the target weight values ​​of different computing power factors are calculated according to the task requirements. The target weight value corresponding to any computing power factor among the multiple computing power factors can be obtained in the following way based on, but not limited to, the demand scores corresponding to the multiple computing power factors:

[0050] fi weight =weight adj *weight average / S weight

[0051] Among them, weight Indicates the target weight value of any computing power factor, weight adj Indicates the demand score of any computing power factor, weight average Represents the average value of the demand scores corresponding to multiple computing power factors; S weight Represents the sum of the demand scores corresponding to multiple computing power factors.

[0052] For example, multiple computing power factors include computing power resource type, computing power distribution, computing power resource real-time utilization information, network bandwidth, and unit price cost, and the corresponding demand scores are 10, 9, 8, 7, and 6, respectively. Then:

[0053] S weight =6+7+8+9+10=40;

[0054] weight average =(10+9+8+7+6) / 5=8;

[0055] The target weight value of the computing resource type is: 10*8 / 40=2;

[0056] The target weight value of the computing power distribution factor is: 9*8 / 40=1.8;

[0057] The target weight value of the real-time utilization of computing resources is: 8*8 / 40=1.6;

[0058] The target weight value of network bandwidth is: 7*8 / 40=1.4;

[0059] The target weight value of unit cost is: 6*8 / 40=1.2.

[0060] As an optional embodiment, based on the target task requirements, target weight values ​​corresponding to multiple computing power factors are determined, including: based on the target task requirements, from multiple indicator weight sets, a target indicator weight set matching the target task requirements is determined, wherein the multiple indicator weight sets respectively include a group of candidate weight values ​​corresponding to multiple computing power factors, and the multiple indicator weight sets correspond one-to-one to the multiple task requirements; based on the target indicator weight set, the target weight values ​​corresponding to the multiple computing power factors are determined.

[0061] It can be understood that from multiple preset indicator weight sets, the most suitable indicator weight set is selected according to the target task requirements. This means that the user's specific business scenario requirements can be understood more accurately and meticulously, and the most appropriate computing power resource allocation strategy can be selected based on these requirements, so as to achieve refined matching of task requirements. By finding an indicator weight set that matches the target task requirements, the target weight values ​​of multiple computing power factors can be quickly determined, avoiding the separate calculation of each factor weight, thereby improving the overall efficiency of resource allocation. In addition, the above method does not require an in-depth understanding of the configuration details of the underlying computing power resources. Only business requirements need to be provided, and the system can automatically match the most appropriate resource allocation strategy, which can simplify user operations and improve user experience. The indicator weight set corresponds one-to-one to a variety of task requirements, and can dynamically adjust the resource allocation strategy according to different business scenarios. Even when business requirements change rapidly, it can respond quickly and provide the most appropriate resource allocation solution.

[0062] Optionally, each indicator weight set in the multiple indicator weight sets may include candidate weight values ​​corresponding to multiple candidate factors, and the multiple computing power factors are screened out from the multiple candidate factors. It can be, but is not limited to, based on screening out candidate factors that match the target business scenario requirements from the multiple candidate factors as multiple computing power factors. The above method can flexibly screen and set computing power factors according to business scenario needs. On the one hand, it can meet the needs of different business scenarios for computing power factors, screen out computing power factors with high relevance to business scenario needs to participate in the scoring of computing power resources, and exclude the interference of irrelevant or less relevant factors on the scoring results; on the other hand, through the screening of computing power factors, it can reduce the calculation pressure of subsequent computing power resource scoring and improve the efficiency of computing power resource scoring.

[0063] For example, multiple candidate factors are set including computing power distribution (region / area information), computing power resource type (CPU / GPU / TPU), computing power resource device information (CPU core number / memory capacity / system disk capacity / hard disk capacity, etc.), computing power resource real-time utilization information, inbound and outbound network bandwidth, computing power resource unit price cost, network resource load. If the target task requirement corresponding to the target account prioritizes computing efficiency, the corresponding multiple computing power factors can be computing power resource type, computing power distribution, computing power resource real-time utilization information, network bandwidth, and unit price cost. The weight values ​​of the above computing power factors are the corresponding preset values, such as 1.0, 0.9, 0.8, 0.7, and 0.6, respectively.

[0064] Step S106, based on the computing power factor sets corresponding to the multiple computing power resources and the target weight values ​​corresponding to the multiple computing power factors, determining the scoring results corresponding to the multiple computing power resources, wherein the computing power factor set includes the computing power factor measurement parameters corresponding to the corresponding computing power resources in the multiple computing power factors;

[0065] In step S106 of the present application, various characteristics of computing power resources (such as CPU performance, memory size, network bandwidth, cost, geographical location, etc.) are converted into specific numerical indicators through computing power factor measurement parameters, which can realize the quantitative evaluation of computing power resources and provide a solid data foundation for subsequent intelligent decision-making and resource scheduling. Combined with the target weight value, that is, different levels of importance are given to computing power factors according to the specific needs of users or business scenarios, each computing power resource can be scored individually. This means that the scoring of resources is no longer static or universal, but is dynamically adjusted according to current needs, which is more in line with actual usage scenarios, thereby improving the accuracy of resource allocation. The determination of the scoring results can intelligently screen out the computing power resources that best meet current needs. Resources with higher scores will be considered better matching options, which helps to quickly and accurately select resources in multi-cloud, heterogeneous resource environments, avoiding the complexity of manual configuration and comparison.

[0066] Optionally, the computing power factor measurement parameter refers to a specific value used to quantify and describe each key characteristic or performance indicator of computing power resources in the scheduling and management of computing power resources. The scoring result of any computing power resource among multiple computing power resources is obtained in the following manner: based on the computing power factor measurement parameters corresponding to multiple computing power factors included in the computing power factor set corresponding to any computing power resource, and the weight values ​​corresponding to the multiple computing power factors, a weighted calculation is performed to obtain the scoring result of any computing power resource, wherein the formula for the weighted calculation can be expressed as:

[0067]

[0068] Among them, R rIndicates the scoring result of any computing resource (i.e., computing resource measurement adjustment parameter); fi seq Represents the power factor measurement parameter of any power factor among multiple power factors; fi weight Represents the target weight value of any computing power factor, i=1, 2, ..., N represents any computing power factor among multiple computing power factors; N represents the total number of multiple computing power factors.

[0069] For example, take multiple computing resources including computing resource 1 and computing resource 2 as an example, where the relevant information of computing resource 1 and computing resource 2 is as follows: computing resource 1: cloud platform 1-region 1, CPU: 4vCPU 16GiB, bandwidth: 100M, latency: 18ms, unit price 0.25 yuan / hour;

[0070] Computing resource 2: Cloud platform 2-Region 2, GPU: 4vCPU 16GiB, bandwidth: 100M, latency: 18ms, unit price 1 yuan / hour;

[0071] In order to show the impact of factors, the two computing resources only set the computing power distribution, and the computing power resource type and computing power resource unit cost are different. Specifically: In order to show the impact of factors, the two computing power resources only set the computing power address, and the computing power type and price are different. Specifically:

[0072] If the business address of the target account is in region 3, then based on the computing power distribution factor, the calculation is based on the region:

[0073] f1-1seq=0.54, f2-1seq=0.46;

[0074] Among them, f1-1seq represents the measurement parameter of the computing power distribution of computing power resource 1; f2-1seq represents the measurement parameter of the computing power distribution of computing power resource 2;

[0075] Metric parameters of computing resource types. The default parameters for the specified type are as follows:

[0076] f1-2seq=1, f2-2seq=2;

[0077] Among them, f1-2seq represents the measurement parameters of the computing resource type of computing resource 1; f2-2seq represents the measurement parameters of the computing resource type of computing resource 2;

[0078] Computing power resources are ranked by unit price cost, and the corresponding metric parameters are calculated based on the unit price:

[0079] f1-3seq=4, f2-3seq=1;

[0080] Among them, f1-3seq represents the measurement parameters of the unit price cost of computing power resources of computing power resource 1; f2-3seq represents the measurement parameters of the unit price cost of computing power resources of computing power resource 1.

[0081] The calculation formula for the scoring result of the corresponding computing power resources can be expressed as: Rr = computing power distribution ranking * target weight value of computing power distribution + computing power resource type * target weight value of computing power type + computing power resource unit price cost ranking * target weight value of unit price cost;

[0082] Combining the target weight values ​​of each computing power factor determined above, we can get:

[0083] The score result of computing resource 1 is Rr1=0.54*1.8+1*2+4*1.2=7.772;

[0084] The scoring result of computing resource 2 is Rr2 = 0.46*1.8 + 4*2 + 1*1.2 = 10.028;

[0085] According to the calculation resource measurement adjustment parameters (i.e., the scoring results), the optimal calculation power is calculated as calculation resource 2.

[0086] It should be noted that the above weighted calculation formula can be pre-set as a heterogeneous cloud unified computing power model and deployed to a heterogeneous cloud management platform. In actual application scenarios, when the target weight values ​​corresponding to multiple computing power factors and the corresponding computing power factor measurement parameters are determined, efficient calculation of the corresponding scoring results of each computing power resource can be achieved based on the heterogeneous cloud unified computing power model.

[0087] Step S108, determining the target computing resources to be allocated to the target account from the multiple computing resources according to the scoring results corresponding to the multiple computing resources.

[0088] In step S108 of the present application, the scoring result is obtained after a comprehensive evaluation of various computing resources based on the target task requirements and the computing factor weight values. Therefore, selecting the target computing resources based on the scoring results can ensure that the selected resources are the most in line with user needs and business scenarios, thereby achieving optimal resource allocation.

[0089] In an optional embodiment, according to the scoring results corresponding to the multiple computing power resources, the target computing power resources allocated to the target account are determined from the multiple computing power resources, including: according to the scoring results corresponding to the multiple computing power resources, the multiple computing power resources are sorted in descending order to obtain the sorting results; and the computing power resources ranked in the front predetermined position in the sorting results are determined as the target computing power resources.

[0090] Optionally, descending sorting and a pre-determined number of digits can quickly identify the resources that best meet user needs, avoiding detailed comparison and analysis of all resources, thereby improving the overall efficiency of resource allocation, especially in scenarios with abundant computing resources and a wide range of choices. The efficiency improvement is more significant. The scoring results include a comprehensive consideration of multiple factors such as cost, performance, geographic location, latency, etc. Through the sorting and optimization mechanism, the best balance between cost control and performance optimization can be found, achieving dual optimization of economic benefits and business effectiveness of resource allocation.

[0091] Optionally, the predetermined number of digits may be one, i.e., the computing power resource with the largest score result value is determined as the target computing power resource; the predetermined number of digits may also be greater than one, for example, the computing power resources ranked in the top two scores are determined as the target computing power resources. During the execution of the target task requirement, the target task requirement can be deployed on two target computing power resources at the same time.

[0092] Optionally, when the number of predetermined digits is greater than one, the corresponding predetermined number of target computing resources can be enabled for on-demand sharing scheduling; and the target task requirements are jointly executed based on the predetermined number of target computing resources.

[0093] Optionally, based on the unified management of computing resources and the network interconnection of computing resources, it can also support the deployment of applications on different heterogeneous cloud resources to achieve on-demand sharing of computing power. For example, deploy image recognition applications to two heterogeneous clouds (such as HW cloud in A and AL cloud computing power in B) as target computing power resources, and enable computing power on-demand sharing scheduling. The default A computing power shares 50% of the computing power tasks and B computing power shares 50% of the computing power tasks. The corresponding image recognition speed is: X images / second. Through deep computing power perception, the computing power sharing application adjusts the threshold value based on the computing power resource performance and the pre-configured computing power. When it is found that the difference in CPU utilization of the two computing power resources exceeds the adjustment threshold a, such as (CPU utilization of B computing power - CPU utilization of A computing power) / (B computing power CPU + A computing power CPU)>a, the adjustment of computing power task sharing is triggered, and the change is made in real time. A computing power shares 75% of the computing power tasks and B computing power shares 25% of the computing power tasks. The corresponding image recognition speed is: Y images / second, where Y>X.

[0094] In an optional embodiment, when the target computing power resources include main computing power resources and backup computing power resources, the computing power resources ranked in the front predetermined number of digits in the sorting result are determined as the target computing power resources, including: determining the computing power resources ranked in the front first digit in the sorting result as the main computing power resources; determining the computing power resources ranked in the front second digit among other computing power resources as the backup computing power resources, wherein the predetermined number of digits is the sum of the first digit and the second digit.

[0095] Optionally, by setting up both the main computing resources and the backup computing resources, you can ensure the high performance and high availability of the main resources, while configuring the backup resources to deal with possible failures or performance fluctuations, to ensure business continuity and data security. The division of primary and backup resources can improve the flexibility of the resource allocation plan and the reliability of the entire system architecture. The selection of primary resources focuses on the optimal performance match, while the backup resources provide high availability for the system, ensuring that when the main computing resources fail or the performance degrades, they can quickly switch to the backup resources to maintain the continuity of business operations. The division of primary and backup resources can ensure rapid recovery in the face of emergencies, reduce business interruption time, and improve service stability and user experience.

[0096] It should be noted that when the related technologies manage and allocate multiple computing resources based on a unified cloud platform, they are also unable to realize heterogeneous cloud network communication, computing network integration, data sharing, data disaster recovery and backup, load balancing and other functions. Based on the primary and backup resource division mechanism in this embodiment, not only disaster recovery switching is supported, but also load balancing can be achieved. When the primary resource load is too high, some tasks can be automatically migrated to the backup resource to ensure the maximum utilization of computing power while avoiding resource overload.

[0097] For example, the computing power resources ranked first and second in the sorting results, the HW cloud at A and the AL cloud computing power at B, can be used as the main computing power resources, and the computing power resource ranked third, a private cloud at C, can be used as the backup computing power resource. When the target task requirement is image recognition, the corresponding image recognition application can be deployed to three different heterogeneous cloud resources, among which the HW cloud at A and the AL cloud computing power at B are used as the main computing power resources to form active-active computing power, and a private cloud at C is used for off-site disaster recovery.

[0098] Optionally, based on the unified management of computing resources, it is possible to support the deployment of applications on different heterogeneous cloud resources to achieve off-site disaster recovery of computing resources. In case of emergencies, such as failure or performance degradation of the main computing resources, the computing power of the HW cloud in A and the AL cloud in B will be shut down, and the computing power will be automatically switched to a private cloud in C. The recognition speed of the private cloud in C is: Z images / second.

[0099] Through the above steps S102 to S108, it is possible to dynamically allocate computing power factor weights and calculate scoring results according to the user's task requirements, so as to flexibly allocate computing power resources, thereby achieving the technical effect of improving the efficiency and flexibility of computing power resource allocation when managing and allocating various computing power resources, and further solving the technical problems of low allocation efficiency and poor flexibility when using a unified cloud platform to manage and allocate various computing power resources in related technologies.

[0100] Based on the above embodiments and optional embodiments, the present invention proposes an optional implementation method of a computing power resource allocation method. Figure 2 is a schematic diagram of an optional computing-network integration system framework according to an embodiment of the present invention. The method can be applied to Figure 2 In the computing-network integration system shown, the computing-network integration system uses multiple different cloud platforms (such as public clouds, private clouds, and edge clouds) as computing resource providers, and accesses the cloud backbone network through the supply-side computing gateway, where the private cloud can include but is not limited to bare metal servers (Bare Metal, BM), virtual machines (Virtual Machine, VM), containers (Container, C), applications (Application, APP), etc.; user-end devices (such as enterprise-end and personal-end) are computing resource demanders and access the cloud backbone network through the demand-side computing gateway. All computing resources in the computing-network integration system are scheduled and allocated with the same computing power through a heterogeneous cloud management platform. In the cloud backbone network, a routing transmission mechanism between a software-defined network (Software-Defined Networking, SDN) controller and a dedicated network edge (Provider Edge, PE) is established through a route reflector (RouteReflector, RR), and the dedicated network PE devices are managed by the SDN controller.

[0101] In response to the demand for cross-domain heterogeneous cloud network intercommunication in the integrated computing and network system, a virtual private cloud (VPC) is created in the same region as the user's cloud resources, and a cloud gateway is created using a system gateway image. A tunnel is established between cloud gateways through a computer network system or an operator's backbone network to achieve network cloud interconnection, and a system customer premises equipment (CPE) image is used to create an access cloud gateway, thereby enabling user VPCs to access multi-cloud cloud gateways, ultimately achieving cross-domain heterogeneous cloud resource network intercommunication and cloud-network integrated management. Specifically, Figure 3 is an optional interactive logic diagram based on a computing-network integration system according to an embodiment of the present invention, such as Figure 3As shown, the computing management platform is a layer of the heterogeneous cloud management platform, which is a hierarchical management of the business. Multiple cloud management platforms can include but are not limited to public cloud management platforms, private cloud management platforms, central cloud management platforms, edge cloud management platforms, etc., which are used to manage cloud resources such as central clouds, edge clouds, public clouds, and private clouds. The computing management platform is used to manage different heterogeneous cloud platforms and unify the management of different cloud vendor resources as a computing resource pool for the entire computing power scheduling. The entire computing power scheduling process is: through the registration and management of different cloud vendors, the perception and registration of computing power are realized. After this step, all computing power resources of heterogeneous cloud resources are collected to the multi-cloud management platform as resources of the available computing power resource pool; when a user has a computing power request, different computing power requirements can be input through demand or business conversion, user specification, etc., and the computing power orchestration and scheduling of the multi-cloud platform is applied. The computing power resource allocation method in this embodiment is matched to the optimal computing power resources to achieve resource allocation. If the user still has network needs, Software-Defined Wide Area Network (SD-WAN) technology can be further used to achieve network connectivity.

[0102] Specifically, Figure 3 As shown in the figure, in the computing power perception and computing power registration stage, the first thing to do is computing power collection. This process will monitor and collect the status, performance and other information of computing power resources in real time through certain technical means (such as sensors, monitoring software, etc.). Computing power modeling will convert this information into understandable models. These models can reflect the characteristics, advantages and limitations of computing power resources and provide a basis for subsequent management and application. Computing power registration is to register the modeled computing power resource information into the cloud management platform. This process involves storing the model information, status information, performance information, etc. of computing power resources in a database or registry so that subsequent management and applications can easily access and utilize these resources. In the process of computing power perception and computing power registration, the cloud management platform may play an important role. The cloud management platform is a management platform for cloud computing resources, which is responsible for the configuration, scheduling, monitoring and optimization of cloud computing resources. In the perception and registration of computing power resources, the cloud management platform can provide necessary support and guarantees to ensure the rational use and efficient operation of cloud computing resources. The SDN controller is one of the core components of the software-defined network. In the perception and registration of computing resources, the SDN controller can ensure the flexibility and scalability of network resources. It can dynamically configure and optimize network resources according to the demand for computing resources and the status of the network to meet the needs of different tasks and applications.

[0103] In the computing power orchestration and computing power scheduling interaction stage, when a system or application needs computing power resources, a computing power discovery request will be initiated first. This request may come from individual users, mobile terminals or other devices and sent to the computing power gateway through the access network. As a bridge connecting users and computing power resources, the computing power gateway will receive and process these computing power discovery requests. It will select a suitable network channel (such as 5G / Internet, SRv6 large intranet, dedicated line, etc.) according to the characteristics and requirements of the request to forward the request to the corresponding computing power resources or services. In the computing power scheduling stage, computing power orchestration is an important link. It will select the most suitable resources from the available computing power resources for combination and optimization according to the specific requirements of the computing power discovery request. This process is implemented based on the computing power resource allocation method in this embodiment, and the computing power routing is responsible for determining the transmission path between computing power resources. After the computing power orchestration is completed, the computing power routing will select the optimal transmission path to allocate computing power resources to specific tasks or applications based on factors such as network topology, bandwidth, and delay. The computing management platform is a comprehensive management platform that is responsible for the management, monitoring and optimization of the entire computing power resources. During the process of computing power orchestration and scheduling, the computing management platform will provide necessary support and guarantees to ensure the rational use and efficient operation of computing power resources.

[0104] Figure 4 is a schematic diagram of an optional computing power scheduling process according to an embodiment of the present invention, such as Figure 4 As shown, the method includes:

[0105] S1, the computing power supply side initiates a computing power registration request to the heterogeneous cloud management platform, registers the computing power resources to the heterogeneous cloud management platform, manages the computing power resources through the heterogeneous cloud management platform, and realizes deep perception of the computing power resources.

[0106] S2, the heterogeneous cloud management platform forwards the computing power registration request initiated by the heterogeneous cloud management platform to the computing power scheduling system, notifies the computing power resources that have been included in the platform management to the computing power scheduling system, and stores the computing power resources in the computing power resource database.

[0107] S3: The network PE device initiates a computing power registration request to the SDN network controller, and the network resources of the dedicated network PE device are managed by the SDN network controller.

[0108] S4, the network PE device forwards the computing power registration request initiated by the network PE device to the computing power scheduling system, notifies the network resources to the computing power scheduling system, and stores the network resources in the network resource database.

[0109] S5, register the demand-side computing power gateway and the supply-side computing power gateway to the computing power scheduling system and store them through the gateway database.

[0110] S6: The administrator accesses the computing resource scheduling system to supplement resource data and manage computing strategies, make real-time strategy adjustments, and develop measurement plans.

[0111] S7, the computing power resource scheduling system performs computing power routing control on the computing power demand side gateway according to the information change in step S7, and realizes the synchronous update of the computing power routing table and computing power strategy table corresponding to the computing power demand side gateway.

[0112] S8, the heterogeneous cloud management platform periodically initiates computing power registration to the computing power scheduling system to synchronize computing power resources.

[0113] S9, the SDN network controller periodically initiates computing power registration to the computing power scheduling system to synchronize network resources.

[0114] S10: The user makes a computing power request, which is forwarded to the computing power scheduling system via the computing power gateway on the demand side.

[0115] S11, the computing power scheduling system orders solutions according to user demand scenarios, determines the computing power resource allocation strategy, and generates the corresponding computing power strategy table and computing power service table based on the computing power resource allocation strategy. The specific determination process of the computing power resource allocation strategy is the same as above and will not be repeated here.

[0116] S12, based on the computing power strategy table and the computing power service table, the computing power scheduling system arranges and configures computing power resources, and performs computing power routing control and computing power resource allocation through the demand-side computing power gateway.

[0117] S13, the computing power scheduling system responds to the computing power request initiated by the user and feeds back the computing power resource allocation result corresponding to the user to the user side via the demand side gateway.

[0118] S14, the computing power gateway (computing power demand-side gateway, computing power supply-side gateway) integrates, distributes and announces computing power resource information through computing power unloading / computing power loading, so as to realize the flexible deployment of cloud network computing power resources.

[0119] The registration of computing power is achieved through the entire process, and the computing power information and network resource information are published to the network for computing power consumption, thereby realizing the activation of computing power routing forwarding and related services.

[0120] This embodiment integrates deep perception of computing resources, resource processing and scheduling functions on the basis of unified management of computing resources across cloud vendors, regions, public clouds and private clouds by a large-scale heterogeneous cloud management platform, so as to enhance the perception ability of computing resources, improve the resource utilization rate of cloud resource pool, and realize reasonable allocation and intelligent scheduling of resources. Specifically, through this embodiment, the difference in computing resource information feedback between different cloud vendors, different regions, and different cloud platforms such as public clouds and private clouds can be shielded, and unified management and data monitoring can be achieved, so that the computing resources of the entire network can be regarded as a unified resource pool. A unified heterogeneous cloud unified computing model is established, based on a unified measurement system, and a computing capacity template is formed by uniformly abstractly describing different computing resources, so as to provide standard computing power measurement rules for computing power equipment management, computing power billing, etc. It should be noted that the ultimate goal of computing power resource perception is to enable flexible scheduling and use to achieve reasonable allocation of resources. This embodiment, based on unified measurement of computing power and network interconnection of computing power resources, realizes distributed computing of computing power, on-demand sharing of computing power resources, and automatic switching of computing power disaster recovery through real-time monitoring of the operating status of cloud resources and unified orchestration and scheduling.

[0121] It still needs to be explained that this embodiment can break down the barriers of different cloud platforms and different types of computing resources, virtualize the heterogeneous cloud computing power of the entire network, and use it as a unified computing resource pool for computing network scheduling. Based on the unified computing power model provided by this embodiment, the abstract computing power resources can be used as input for adaptation processing on the basis of deep perception of computing power, accurate model matching, model superposition, merged scene customization, and flexible adaptation to various computing power demand scenarios. Through intelligent scheduling of computing power resources across the entire network, applications can be dynamically adjusted according to the computing power status to achieve optimal computing power utilization and high application reliability; support off-site disaster recovery of computing power, real-time switching, and ensure high resistance to destruction of computing power.

[0122] In this embodiment, a computing resource allocation device is also provided, which is used to implement the above embodiments and preferred implementation modes, and will not be repeated here. As used below, the terms "module" and "device" can implement a combination of software and / or hardware for a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0123] According to an embodiment of the present invention, there is also provided an apparatus embodiment for implementing the above computing resource allocation method. Figure 5 is a schematic diagram of the structure of a computing resource allocation device according to an embodiment of the present invention. Figure 5 As shown, the computing resource allocation device includes: a task requirement acquisition module 500, a weight determination module 502, a scoring module 504, and a resource allocation module 506, wherein:

[0124] A task requirement acquisition module 500 is used to acquire the target task requirement of the target account;

[0125] The weight determination module 502 is connected to the task requirement acquisition module 500, and is used to determine the target weight values ​​corresponding to the multiple computing power factors respectively based on the target task requirements;

[0126] Scoring module 504, connected to weight determination module 502, is used to determine scoring results corresponding to the plurality of computing power resources respectively based on the computing power factor sets corresponding to the plurality of computing power resources respectively and the target weight values ​​corresponding to the plurality of computing power factors respectively, wherein the computing power factor set includes computing power factor measurement parameters corresponding to the plurality of computing power factors respectively of the corresponding computing power resources;

[0127] The resource allocation module 506 is connected to the scoring module 504 and is used to determine the target computing resources to be allocated to the target account from the multiple computing resources according to the scoring results corresponding to the multiple computing resources.

[0128] A task requirement acquisition module 500 is set up to acquire the target task requirement of the target account; a weight determination module 502 is connected to the task requirement acquisition module 500, and is used to determine the target weight values ​​corresponding to the multiple computing power factors based on the target task requirement; a scoring module 504 is connected to the weight determination module 502, and is used to determine the scoring results corresponding to the multiple computing power resources based on the computing power factor sets corresponding to the multiple computing power resources and the target weight values ​​corresponding to the multiple computing power factors, wherein the computing power factor set includes the computing power factor measurement parameters corresponding to the corresponding computing power resources in the multiple computing power factors; resource The allocation module 506 is connected to the scoring module 504 and is used to determine the target computing resources to be allocated to the target account from the multiple computing resources according to the scoring results corresponding to the multiple computing resources. It can achieve the purpose of dynamically allocating computing factor weights and calculating scoring results according to the user's task requirements, so as to flexibly allocate computing resources, thereby achieving the technical effect of improving the efficiency and flexibility of computing resource allocation when managing and allocating multiple computing resources, and further solving the technical problems of low allocation efficiency and poor flexibility when using a unified cloud platform to manage and allocate multiple computing resources in related technologies.

[0129] It should be noted that the above modules can be implemented by software or hardware. For example, for the latter, it can be implemented in the following ways: the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0130] It should be noted that the task requirement acquisition module 500, the weight determination module 502, the scoring module 504, and the resource allocation module 506 correspond to steps S102 to S108 in the embodiment, and the examples and application scenarios implemented by the modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above embodiment. It should be noted that the modules as part of the device can be run in a computer terminal.

[0131] It should be noted that the optional or preferred implementation of this embodiment can refer to the relevant description in the embodiment, which will not be repeated here.

[0132] The above-mentioned computing power resource allocation device may also include a processor and a memory. The above-mentioned task requirement acquisition module 500, weight determination module 502, scoring module 504, resource allocation module 506, etc. are all stored in the memory as program modules, and the processor executes the above-mentioned program modules stored in the memory to realize the corresponding functions.

[0133] The processor includes a kernel, which retrieves the corresponding program module from the memory. The kernel may be one or more. The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. The memory includes at least one memory chip.

[0134] According to an embodiment of the present application, an embodiment of a non-volatile storage medium is also provided. Optionally, in this embodiment, the non-volatile storage medium includes a stored program, wherein when the program is executed, the device where the non-volatile storage medium is located is controlled to execute any one of the computing resource allocation methods.

[0135] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group, and the non-volatile storage medium includes a stored program.

[0136] Optionally, when the program is running, the device where the non-volatile storage medium is located is controlled to perform the following functions: obtain the target task requirements of the target account; based on the target task requirements, determine the target weight values ​​corresponding to multiple computing power factors; based on the computing power factor sets corresponding to the multiple computing power resources, and the target weight values ​​corresponding to the multiple computing power factors, determine the scoring results corresponding to the multiple computing power resources, wherein the computing power factor set includes the computing power factor measurement parameters corresponding to the corresponding computing power resources in the multiple computing power factors; according to the scoring results corresponding to the multiple computing power resources, determine the target computing power resources allocated to the target account from the multiple computing power resources.

[0137] According to an embodiment of the present application, an embodiment of a processor is also provided. Optionally, in this embodiment, the processor is used to run a program, wherein the program executes any one of the computing resource allocation methods when it is run.

[0138] According to an embodiment of the present application, an embodiment of a computer program product is also provided, which, when executed on a data processing device, is suitable for executing a program that initializes any of the above-mentioned computing power resource allocation method steps.

[0139] Optionally, the above-mentioned computer program product, when executed on a data processing device, is suitable for executing a program that initializes the following method steps: obtaining the target task requirements of the target account; determining the target weight values ​​corresponding to multiple computing power factors based on the target task requirements; determining the scoring results corresponding to multiple computing power resources based on the computing power factor sets corresponding to the multiple computing power resources and the target weight values ​​corresponding to the multiple computing power factors, wherein the computing power factor set includes computing power factor measurement parameters corresponding to the corresponding computing power resources in the multiple computing power factors; determining the target computing power resources allocated to the target account from the multiple computing power resources according to the scoring results corresponding to the multiple computing power resources.

[0140] An embodiment of the present invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are implemented: obtaining the target task requirements of a target account; determining the target weight values ​​corresponding to a plurality of computing power factors based on the target task requirements; determining the scoring results corresponding to the plurality of computing power resources based on the computing power factor sets corresponding to the plurality of computing power resources and the target weight values ​​corresponding to the plurality of computing power factors, wherein the computing power factor set includes computing power factor measurement parameters corresponding to the corresponding computing power resources in the plurality of computing power factors; determining the target computing power resources allocated to the target account from the plurality of computing power resources according to the scoring results corresponding to the plurality of computing power resources.

[0141] The above sequence of the embodiments of the present invention is for description only and does not represent the superiority or inferiority of the embodiments.

[0142] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0143] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the above modules can be a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, modules or indirect coupling or communication connection of modules, which can be electrical or other forms.

[0144] The modules described above as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment.

[0145] In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of software functional modules.

[0146] If the above-mentioned integrated module is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable non-volatile storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a non-volatile storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the various embodiments of the present invention. The aforementioned non-volatile storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0147] The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A computing resource allocation method, characterized in that: include: Obtain the target task requirements of the target account; Based on the target task requirements, determine target weight values ​​corresponding to multiple computing power factors respectively; Based on the computing power factor sets corresponding to the multiple computing power resources, and the target weight values ​​corresponding to the multiple computing power factors, respectively, determining the scoring results corresponding to the multiple computing power resources, wherein the computing power factor set includes the computing power factor measurement parameters corresponding to the multiple computing power factors of the corresponding computing power resources; According to the scoring results respectively corresponding to the multiple computing resources, a target computing resource allocated to the target account is determined from the multiple computing resources.

2. The method according to claim 1, characterized in that The target task requirements of obtaining the target account include: Obtain target business scenario requirements of the target account; A task requirement among multiple task requirements that matches the target business scenario requirement is determined as the target task requirement, wherein the multiple task requirements correspond to the multiple business scenario requirements one by one.

3. The method according to claim 2, characterized in that The step of determining, among the multiple task requirements, a task requirement that matches the target business scenario requirement as the target task requirement includes: Check whether there is a preset task requirement pre-configured for the target account; In the case where the preset task requirement is not found, a task requirement among the multiple task requirements that matches the target business scenario requirement is determined as the target task requirement.

4. The method according to claim 3, characterized in that The method further comprises: When the preset task requirement is found, the preset task requirement is determined as the target task requirement.

5. The method according to claim 1, characterized in that The determining, based on the target task requirements, target weight values ​​corresponding to the multiple computing power factors respectively includes: Query whether there are preset weight values ​​corresponding to the multiple computing power factors respectively, wherein the preset weight values ​​are pre-configured for the target account; If the preset weight values ​​corresponding to the plurality of computing power factors are not found, determining the target weight values ​​corresponding to the plurality of computing power factors based on the target task requirements; or When the preset weight values ​​respectively corresponding to the multiple computing power factors are queried, the preset weight values ​​respectively corresponding to the multiple computing power factors are determined as the target weight values ​​respectively corresponding to the multiple computing power factors.

6. The method according to claim 1, characterized in that The determining, based on the target task requirements, target weight values ​​corresponding to the multiple computing power factors respectively includes: Based on the target task requirements, determining the demand scores corresponding to the multiple computing power factors respectively, wherein the demand scores are used to indicate the preference of the target account for the corresponding computing power factors; Based on the demand scores respectively corresponding to the multiple computing power factors, target weight values ​​respectively corresponding to the multiple computing power factors are determined.

7. The method according to any one of claims 1 to 6, characterized in that The determining, according to the scoring results respectively corresponding to the multiple computing resources, a target computing resource to be allocated to the target account from the multiple computing resources comprises: According to the scoring results respectively corresponding to the multiple computing resources, the multiple computing resources are sorted in descending order to obtain a sorting result; The computing power resource ranked in the front predetermined position in the sorting result is determined as the target computing power resource.

8. The method according to claim 7, characterized in that In the case where the target computing power resource includes a main computing power resource and a backup computing power resource, determining the computing power resource ranked in the first predetermined number of places in the sorting result as the target computing power resource includes: Determine the computing power resource ranked first in the sorting result as the main computing power resource; The computing power resource ranked second from the top among other computing power resources is determined as the backup computing power resource, wherein the predetermined number of digits is the sum of the first number of digits and the second number of digits.

9. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the computing power resource allocation method described in any one of claims 1 to 8.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the computing resource allocation method described in any one of claims 1 to 8 are implemented.

Images