Computing power resource allocation method and device, equipment and storage medium

Abstract

The invention provides a computing power resource allocation method and device, equipment and a storage medium, and the method comprises the steps: obtaining an opening application of a tenant, determining the grade type of the tenant according to the property of the tenant of the opening application, and determining the number range of target adaptive nodes according to the grade type, the target adaptive node number range is limited based on the tenant property, and resource waste caused by the fact that the tenants occupy too many nodes is avoided; according to the target adaptive node number range, configuring the scaling strategy of the computing power cluster nodes of the computing power cluster, monitoring the function indexes of the computing power cluster in the process that the computing power cluster is in the use state, and dynamically adjusting the number of the computing power cluster nodes according to the function indexes and the scaling strategy, so that the number of the computing power cluster nodes can be automatically and dynamically adjusted; and effective allocation of computing power resources of the computing power cluster nodes is facilitated.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to a computing resource allocation method and device, equipment and storage medium. BACKGROUND

[0002] At present, due to the limited resources of computing chips, GPUs, CPUs and other key components, and the increasing demand for computing power in various industries, cloud management platforms have emerged, which provide computing resources for tenants by computing clusters of cloud management platforms. At present, the node allocation of the computing cluster of the cloud management platform is usually based on the cost of its own money investment, and the tenants with strong financial resources apply for more nodes, but in many cases, the actual node resources required by the tenants are much smaller than the allocated amount, so that the node resources are not effectively used, which may lead to the fact that other tenants with real needs cannot be allocated to node resources, and therefore a solution is needed. SUMMARY

[0003] The embodiments of the present application provide a computing resource allocation method, device, equipment and storage medium to solve at least one problem in the related art, and the technical solutions are as follows:

[0004] In a first aspect, the embodiments of the present application provide a computing resource allocation method, comprising:

[0005] Obtaining an opening application of a tenant, and determining a level type of the tenant according to a tenant property of the opening application;

[0006] According to the level type, determining a target adaptive node quantity range;

[0007] According to the target adaptive node quantity range, configuring a scaling strategy of a computing cluster node of a computing cluster;

[0008] In the process that the computing cluster is in a use state, monitoring a function index of the computing cluster, and dynamically adjusting the number of the computing cluster nodes according to the function index and the scaling strategy.

[0009] In an implementation manner, the determining the target adaptive node quantity range according to the level type comprises:

[0010] Matching the level type with a plurality of preset level types to determine a target level type, wherein each preset level type has a corresponding mapped level value range, and the adjustable amplitude of each level value range is different;

[0011] Taking the level value range mapped by the target level type as the target adaptive node quantity range.

[0012] In an embodiment, the scaling strategy of the computing power cluster nodes of the computing power cluster according to the target adaptive node quantity range comprises:

[0013] setting a minimum value of the number of computing power cluster nodes according to a minimum boundary value of the target adaptive node quantity range;

[0014] setting a guaranteed maximum value of the number of computing power cluster nodes according to a maximum boundary value of the target adaptive node quantity range;

[0015] setting an initial value of the number of computing power cluster nodes according to the target adaptive node quantity range, the initial value being between the minimum value and the guaranteed maximum value;

[0016] configuring a monitoring threshold of the function index according to the configuration instruction;

[0017] determining the scaling strategy of the computing power cluster nodes of the computing power cluster according to the minimum value, the guaranteed maximum value, the initial value, and the monitoring threshold.

[0018] In an embodiment, the monitoring of the function index of the computing power cluster and the dynamic adjustment of the number of computing power cluster nodes according to the function index and the scaling strategy in the process that the computing power cluster is in use comprises:

[0019] monitoring the function index of the computing power cluster in the process that the computing power cluster nodes of the initial value are in use;

[0020] executing a function index monitoring rule to compare the function index with the monitoring threshold;

[0021] when the function index is greater than or equal to the monitoring threshold, increasing the number of computing power cluster nodes in the range of the initial value and the guaranteed maximum value;

[0022] when the function index is less than the monitoring threshold and the function index continuously decreases, decreasing the number of computing power cluster nodes in the range of the minimum value and the initial value.

[0023] In an embodiment, the dynamic adjustment of the number of computing power cluster nodes according to the function index and the scaling strategy further comprises:

[0024] when the number of computing power cluster nodes is increased, and the number of computing power cluster nodes reaches the guaranteed maximum value and the function index continuously increases, performing capacity analysis on the computing power cluster to determine the available amount of the computing power cluster;

[0025] determining a sum value of the guarantee maximum value and the available quota amount, and increasing the number of the computing cluster nodes to the sum value.

[0026] In an implementation, the scaling strategy further comprises a cost threshold value and a priority; and the dynamically adjusting the number of the computing cluster nodes according to the function index and the scaling strategy further comprises:

[0027] calculating a total cost according to a pre-designed cost rule, the number of the computing cluster nodes, and a usage time;

[0028] if the total cost is greater than the cost threshold value, when the priority is that the cost rule is higher than the function index monitoring rule, decreasing the number of the computing cluster nodes so that the total cost is less than or equal to the cost threshold value;

[0029] if the total cost is greater than the cost threshold value, when the priority is that the cost rule is lower than the function index monitoring rule, performing the step of comparing the function index with the monitoring threshold value.

[0030] In an implementation, the dynamically adjusting the number of the computing cluster nodes according to the function index and the scaling strategy comprises:

[0031] determining a preferential period and displaying the preferential period;

[0032] in response to a triggering instruction, setting a triggering condition of the monitoring threshold value and an adjustment value;

[0033] when an actual time is the preferential period, triggering the triggering condition and reducing a value of a current monitoring threshold value to the adjustment value.

[0034] In a second aspect, an embodiment of the present application provides a computing resource allocation apparatus, comprising:

[0035] a first determining module configured to acquire an opening application of a tenant and determine a level type of the tenant according to a tenant property of the opening application;

[0036] a second determining module configured to determine a target adaptive node number range according to the level type;

[0037] a configuration module configured to configure a scaling strategy of computing cluster nodes of a computing cluster according to the target adaptive node number range;

[0038] an adjustment module configured to monitor a function index of the computing cluster in a process in which the computing cluster is in a usage state, and dynamically adjust the number of the computing cluster nodes according to the function index and the scaling strategy.

[0039] In a third aspect, an electronic device is provided, including a processor and a memory, the memory storing instructions which are loaded and executed by the processor to implement the method in any of the embodiments of the above aspects.

[0040] In a fourth aspect, a computer readable storage medium is provided, storing a computer program which, when executed, implements the method in any of the embodiments of the above aspects.

[0041] The beneficial effects of the above technical solutions at least include:

[0042] By obtaining the opening application of the tenant, and determining the level type of the tenant according to the tenant property of the opening application, the target adaptive node quantity range is determined according to the level type, and the target adaptive node quantity range is limited based on the tenant property, which is beneficial to avoid resource waste caused by too many nodes occupied by the tenant; according to the target adaptive node quantity range, the scaling strategy of the computing power cluster node of the computing power cluster is configured, the function index of the computing power cluster is monitored in the process that the computing power cluster is in a use state, and the number of the computing power cluster nodes is dynamically adjusted according to the function index and the scaling strategy, which can automatically and dynamically adjust the number of the computing power cluster nodes, and is more beneficial to the effective allocation of the computing power resource of the computing power cluster node.

[0043] The above summary is merely intended to illustrate the present description and is not intended to limit in any way. In addition to the illustrative aspects, embodiments and features described above, further aspects, embodiments and features of the present application will be apparent from the drawings and the following detailed description. BRIEF DESCRIPTION OF DRAWINGS

[0044] In the drawings, like numerals refer to like elements throughout the various drawings. The drawings are not necessarily to scale, the emphasis instead being placed on the relations between various elements. It should be understood that the drawings only depict some embodiments in accordance with the present disclosure and should not be considered limiting of the scope of the disclosure.

[0045] Figure 1 A step flowchart of a computing power resource allocation method according to an embodiment of the present disclosure;

[0046] Figure 2 A schematic diagram of a cloud management platform according to an embodiment of the present disclosure;

[0047] Figure 3 An effect diagram of a cluster node resource pool usage according to an embodiment of the present disclosure;

[0048] Figure 4 A structural block diagram of a computing power resource allocation device according to an embodiment of the present disclosure;

[0049] Figure 5 Structure block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0050] In the following, only certain exemplary embodiments are simply described. As those skilled in the art can recognize, the described embodiments can be modified in various different ways without departing from the spirit or scope of the present application. Therefore, the drawings and the description are considered to be exemplary in nature rather than limiting.

[0051] Noun introduction:

[0052] IaaS (Infrastructure as a Service): a cloud computing service model that provides users with a set of basic computing resources and virtualization technology platform. Through IaaS, users can rent necessary computing resources such as virtual machines, data storage and network services from cloud service providers. This service model allows users to more flexibly and economically meet their computing needs without having to purchase and maintain expensive physical hardware facilities.

[0053] Computing power: the computing power of a device to process data and achieve a specific result output. The core of computing power is CPU, GPU, FPGA, ASIC and other types of computing chips, and it is carried by computers, servers, high-performance computing clusters and various intelligent terminals. Mass data processing and various digital applications cannot be separated from the processing and calculation of computing power. The larger the computing power value represents the stronger the comprehensive computing power. Computing power can be divided into basic computing power, intelligent computing power and supercomputing power, which provide basic general computing (based on CPU), artificial intelligence computing (based on GPU, FPGA, ASIC, etc.) and scientific engineering computing (based on supercomputers). The computing power referred to in the present application specifically refers to the computing power based on CPU, GPU and other chips.

[0054] GPU: GPU (Graphics Processing Unit) is a microprocessor specially designed for processing graphics and image calculations. Compared with traditional CPUs, GPUs have higher parallel processing capabilities, making them perform well in graphics rendering, numerical analysis, financial analysis, password cracking and other mathematical calculations and geometric operation tasks.

[0055] Reference Figure 1 , a flowchart of a computing power resource allocation method according to an embodiment of the present application is shown, which can at least include steps S100-S400:

[0056] S100, obtain the opening application of the tenant, and determine the level type of the tenant according to the nature of the tenant of the opening application.

[0057] S200, determine a target adaptive node quantity range according to the level type.

[0058] S300, configure a scaling strategy of the computing power cluster node of the computing power cluster according to the target adaptive node quantity range.

[0059] S400, monitor a function index of the computing power cluster in a process in which the computing power cluster is in a use state, and dynamically adjust the quantity of the computing power cluster node according to the function index and the scaling strategy.

[0060] The technical scheme of the embodiment of the application, by obtaining the opening application of the tenant, and determining the level type of the tenant according to the tenant property of the opening application, determining the target adaptive node quantity range according to the level type, limiting the target adaptive node quantity range based on the tenant property, is beneficial to avoid resource waste caused by too many nodes occupied by the tenant; configuring the scaling strategy of the computing power cluster node of the computing power cluster according to the target adaptive node quantity range, monitoring the function index of the computing power cluster in the process in which the computing power cluster is in the use state, and dynamically adjusting the quantity of the computing power cluster node according to the function index and the scaling strategy, can automatically and dynamically adjust the quantity of the computing power cluster node, and is more beneficial to effective allocation of the computing power resource of the computing power cluster node.

[0061] As shown in the figure, in an embodiment, each tenant (also referred to as a user) can initiate an opening application to the cloud management self-service portal of the cloud management platform, and the cloud management platform can determine the level type of the tenant based on the tenant property carried in the opening application after obtaining the opening application of the tenant. Figure 2

[0062] In an embodiment, step S200 includes steps S210-S220:

[0063] S210, match the level type with a plurality of preset level types to determine a target level type.

[0064] Optionally, a plurality of preset level types can be preset in the cloud management platform, and each preset level type has a corresponding mapped level value range, and the adjustable amplitude of each level value range is different, for example, the level value range of the level type of the first level can be 0-100, for example, the level value range of the level type of the second level can be 0-200, and the higher the level, the higher the adjustable amplitude. Therefore, matching the level type with the plurality of preset level types determines the matching preset level type as the target level type.

[0065] ​In an embodiment, the cloud management platform can set the preset level type corresponding to the different tenant properties as the level type of the tenant, or the administrator can configure the level type of the tenant based on the user tenant properties. For example, the tenant properties can be personal, small and medium-sized enterprise, large enterprise, etc. When the number of people and the corresponding resource demand are more, the tenant can be configured with a higher level type. In other embodiments, the level type can be configured based on other dimensions. In general, the level type can be configured as ordinary, intermediate, high, VIP, and each level type can be defined with specific service terms and resource access permissions.

[0066] S220, mapping the level value range of the target level type as the target adaptive node quantity range.

[0067] Optionally, the level value range mapped by the target level type is used as the target adaptive node quantity range. It should be noted that the level type of each tenant can be bound to priority mapping, priority driving, and other settings in addition to the binding of the corresponding target adaptive node quantity range. Through the binding of the target adaptive node quantity range, even if the tenant has enough fees, he cannot purchase more resources than his level under limited resources. In this way, the computing power cluster resource capacity applied by the tenant can be ensured to match the set level, and node resources can be avoided from being wasted. The computing power cluster of the present application adopts an auto-scaling Web service computing power cluster.

[0068] In an embodiment, the user can also select the node type, node virtual machine type, and node specification of the computing power cluster node when applying. The node type includes but is not limited to a central node, a computing node, a special task node, a log storage node, etc. The node virtual machine type needs to be determined according to the type of the job based on the selected virtual machine specification, mainly including GPU type, CPU type, GPU memory enhanced type, CPU memory enhanced type, etc. The node specification can include instance type, operating system, etc.

[0069] In an embodiment, the tenant can propose a change process of the level type to the cloud management platform, and the administrator can change the latest tenant level type after approval. If the level after the change is lower than before, the original computing power resource allocation will be automatically recovered gradually according to the scaling condition strategy set by the tenant.

[0070] In an embodiment, the tenant can select or configure the node specification of the cluster node resource pool in the cloud management platform, including instance type, operating system, etc. The user can create a new node pool or edit the configuration of an existing node pool, select different performance instance types as needed, usually based on different combinations of GPU, CPU, memory, storage and network performance (some combinations are only set for higher tenant level types), such as general-purpose: balanced CPU and memory configuration, suitable for a variety of applications; compute-optimized: stronger CPU performance, suitable for compute-intensive tasks; memory-optimized: larger memory capacity, suitable for memory-intensive applications; high I / O: high-speed disk I / O performance, suitable for applications that require fast storage access; GPU-accelerated: integrated GPU, suitable for graphics processing and deep learning, etc.

[0071] In an embodiment, step S300 includes steps S310-S350:

[0072] S310, according to the minimum boundary value of the target adaptive node quantity range, set the minimum value of the number of nodes of the computing power cluster.

[0073] Optionally, the tenant can customize the scaling strategy through the cloud management platform. For example, assuming that the determined target adaptive node quantity range is 0-100, at this time the user can set the minimum value of the number of nodes of the computing power cluster to 0 according to the minimum boundary value 0.

[0074] S320, according to the maximum boundary value of the target adaptive node quantity range, set the guaranteed maximum value of the number of nodes of the computing power cluster.

[0075] Optionally, assuming that the determined target adaptive node quantity range is 0-100, the tenant can set the guaranteed maximum value of the number of nodes of the computing power cluster to 100 according to the maximum boundary value 100 of the target adaptive node quantity range.

[0076] S330, according to the target adaptive node quantity range, set the initial value of the number of nodes of the computing power cluster, the initial value being between the minimum value and the guaranteed maximum value.

[0077] Optionally, assuming that the determined target adaptive node quantity range is 0-100, at this time the initial value between the minimum value 0 and the guaranteed maximum value 100 can be set, for example, 5.

[0078] S340, according to the configuration instruction, configure the monitoring threshold of the function index.

[0079] Optionally, the tenant sets the above-mentioned various values based on its actual needs, and can set the monitoring threshold of the function index. When the function index has multiple, different monitoring thresholds corresponding to different function indexes can be set.

[0080] S350, determine the scaling strategy of the computing power cluster node of the computing power cluster according to the minimum value, the guarantee maximum value, the initial value and the monitoring threshold.

[0081] Optionally, according to the minimum value, the guarantee maximum value, the initial value and the monitoring threshold, the adaptive range and the monitoring condition of the computing power cluster node of the computing power cluster can be determined as part of the scaling strategy of the computing power cluster node of the computing power cluster, realizing the self-defined scaling strategy. The scaling strategy is used for the control center of the cloud management platform to analyze the function indicators collected by the collection monitoring center, and then through the resource scheduling and scaling controller, to adaptively extend (increase the computing power cluster node) and shrink (reduce / remove the computing power cluster node).

[0082] It should be noted that the collection monitoring center is responsible for real-time collection and monitoring of performance data of the computing power cluster to obtain function indicators; the collection monitoring center is composed of components such as monitoring agents, monitoring services and index collectors, which work together to ensure comprehensive monitoring of the running status of the computing power cluster. The monitoring agent is deployed on each node of the cluster and is responsible for collecting function indicators. The collected data is periodically aggregated and transmitted to the control center for further analysis and decision-making. Through this process, the collection monitoring center provides the cloud management platform with in-depth insights into the health status of the computing power cluster, supporting intelligent decision-making based on data-driven, thereby optimizing resource allocation and improving the overall performance and efficiency of the cluster.

[0083] The control center is the core decision-making unit of the cloud management platform, responsible for executing automatic scaling operations of the cluster according to the function indicators and scaling strategies provided by the collection monitoring center. The center is composed of key components such as automatic scaling strategy managers and decision algorithm modules, which work together to achieve accurate decision-making. The decision engine of the control center analyzes the monitoring data comprehensively, evaluates the actual running state of the cluster, and matches with the scaling strategy to calculate and determine the necessary scaling actions, such as increasing / decreasing nodes or adjusting node configurations. Subsequently, the decision engine issues these operation instructions to the resource scheduling and scaling scheduler, which is responsible for specific execution work, ensuring that the size and configuration of the computing power cluster can flexibly adapt to changes in business demand, thereby optimizing resource utilization and system performance.

[0084] The resource scheduling and scaling controller is responsible for executing automatic scaling operation instructions, such as creating new computing power cluster nodes, destroying idle computing power cluster nodes, adjusting computing power cluster node configurations, etc. It is composed of scaling schedulers, resource managers, etc. The controller manages and schedules the computing power cluster nodes in the computing power cluster according to the instructions of the decision engine of the control center to meet the load requirements of the application.

[0085] The computing power cluster node is responsible for storage and processing of computing power cluster data, including various node types, for example, the center node is responsible for managing metadata information of the computing power cluster and namespace of the file system, and is the core node of the entire computing power cluster. The center node regularly obtains computing power cluster node distribution information from the acquisition and monitoring center unit to adjust task scheduling. In addition, the computing node is responsible for computing tasks and data storage of the computing power cluster. Special task node: used as a special task computing node (optional); log storage node: used for storing input and output data sets and log information (such as Mysql database, optional). When adding, these nodes must be registered in the acquisition and monitoring center, so that the control center node can master the situation of each node of the computing power cluster in real time. All operations of the scalable cluster are recorded in the database.

[0086] In the embodiments of the application, the cluster is initialized and started before the computing power cluster is used, to ensure that the automatic scaling effect is compatible with the computing power cluster, and to avoid the scaling effect exceeding the upper and lower limits of the cluster.

[0087] In one embodiment, step S400 can include steps S410-S440:

[0088] S410, in the process of the initial value of the computing power cluster node being in a use state, monitoring the function indicators of the computing power cluster.

[0089] Optionally, in the computing power cluster corresponding to the tenant, for example, the initial value is 5, at this time, the use of 5 computing power cluster nodes is taken as the initial state, and then in the process of the computing power cluster node being in a use state, the function indicators of the computing power cluster are monitored.

[0090] S420, executing a function indicator monitoring rule to compare the function indicators with monitoring thresholds.

[0091] Optionally, the function indicators include but are not limited to load indicators and performance indicators, the load indicators include but are not limited to, and the performance indicators include but are not limited to CPU, GPU, memory utilization, throughput, response time, etc. In the embodiments of the application, based on the function indicator monitoring rule of the cloud management platform, each function indicator is compared with the corresponding monitoring threshold.

[0092] S430, when the function indicators are greater than or equal to the monitoring thresholds, the number of computing power cluster nodes is increased within the range of the initial value and the guaranteed maximum value.

[0093] Optionally, if the function indicators are greater than or equal to the corresponding monitoring thresholds, at this time, the number of computing power cluster nodes is increased within the range of the initial value and the guaranteed maximum value based on the previously set initial value and guaranteed maximum value. It should be noted that the amplitude and frequency of the increase can be determined based on actual conditions, and are not specifically limited.

[0094] S440, when the function index is less than the monitoring threshold value and the function index continues to decrease, reducing the number of computing cluster nodes in the range between the minimum value and the initial value.

[0095] Optionally, when the function index is less than the corresponding monitoring threshold value and the function index continues to decrease, the number of computing cluster nodes is reduced in the range between the minimum value and the initial value based on the previously set minimum value, so as to realize dynamic automatic adjustment of the computing cluster nodes.

[0096] In an embodiment, step S400 can further include steps S450-S460:

[0097] S450, when the number of computing cluster nodes is increased, if the number of computing cluster nodes reaches the guarantee maximum value and the function index continues to increase, performing capacity analysis on the computing cluster to determine the available amount of the computing cluster.

[0098] Optionally, when the number of computing cluster nodes is increased, if the number of computing cluster nodes reaches the guarantee maximum value and the function index continues to increase, for example, the load index still continues to increase, the cloud management platform performs capacity analysis on the computing cluster to determine the available amount of the current computing cluster, that is, the number of computing cluster nodes that can be used for allocation in the cluster node resource pool of the current computing cluster.

[0099] S460, determining the sum of the guarantee maximum value and the available amount, and increasing the number of computing cluster nodes to the sum.

[0100] Optionally, the sum of the guarantee maximum value and the available amount is determined, and then the number of computing cluster nodes is increased to the sum. It should be noted that the sum is equivalent to the limit value of the number of computing cluster nodes that the current computing cluster can provide. By analyzing and determining the limit value, the demand for sudden surge of node resources required by the tenant in some cases can be met, so as to meet the actual demand of the tenant as much as possible, and be more intelligent and humanized. For example, Figure 3 as shown, it is a cluster node resource pool usage effect diagram in some cases.

[0101] In an embodiment, when forming the scaling strategy, the tenant can also set a cost threshold and a priority in the cloud management platform as part of the formed scaling strategy; step S400 of the embodiment of the application can further include steps S470-S490:

[0102] S470, calculating the total cost according to the pre-designed cost rule, the number of computing cluster nodes and the use time.

[0103] Optionally, a pre-design fee rule is set in the cloud management platform, and the pre-design fee rule includes the node type of different computing power cluster nodes, the node virtual machine type corresponding to the computing power cluster nodes, and the use fee corresponding to the node specification, so that based on the pre-design fee rule, the number of computing power cluster nodes used by the tenant and the use time, the total fee can be calculated. It should be noted that when charging, the charging period and the scaling period are set as needed, the charging period can be by minute or by hour, and the scaling period should be set according to the specific business demand and performance target. The platform needs to be carefully planned and tested to find the most suitable scaling period and scaling strategy.

[0104] S480, if the total fee is greater than the cost threshold, when the priority of the cost rule is higher than that of the function index monitoring rule, the number of computing power cluster nodes is reduced so that the total fee is less than or equal to the cost threshold.

[0105] Optionally, if the total fee is greater than the cost threshold, if the priority set by the tenant is that the cost rule is higher than the function index monitoring rule, the number of current computing power cluster nodes, specifically the number of computing power cluster nodes in use, is preferentially reduced, so that the total fee is less than or equal to the cost threshold.

[0106] S490, if the total fee is greater than the cost threshold, when the priority of the cost rule is lower than that of the function index monitoring rule, the step of comparing the function index with the monitoring threshold is performed.

[0107] Optionally, if the total fee is greater than the cost threshold, when the priority of the cost rule is lower than that of the function index monitoring rule, the function is preferentially guaranteed, and the function index monitoring rule is preferentially considered, and the step of comparing the function index with the monitoring threshold is performed, that is, S420 is performed. It should be noted that the cost threshold can be set based on design, for example, 50%, 75%, 90% and 100% of the budget, and when the actual total fee reaches these thresholds, the cloud platform can also trigger a warning notification and send it to the user through email, SMS or the message system of the cloud platform.

[0108] In an embodiment, the number of nodes can also be automatically reduced in a specific period of time, such as at night or on weekends, and the cloud platform can automatically increase the nodes during the business peak period, such as during the day on weekdays.

[0109] Optionally, step S400 can further include steps S401-S403:

[0110] S401, determine the preferential period and display it.

[0111] Optionally, when the provider of the computing cluster offers preferential interactions, or the platform administrator sets preferential pre-processing for certain date segments or time segments in advance based on experience and actual conditions, preferential charging processing is adopted for these time segments. For example, a certain time segment is a preferential time segment, and related preferential time segment information can be displayed on the cloud management platform for tenants to understand.

[0112] At the same time, the cloud management platform has a charging preferential operation and maintenance calendar to realize the following functions:

[0113] Calendar information integration: The cloud management platform integrates and imports a public holiday and special time segment calendar information, which is used to determine the charging preferential time segment;

[0114] Pre-processing of preferential time segment: The cloud management platform administrator can set the preferential time segment of a specific date or time segment based on historical data, business trends and public holidays;

[0115] User interface display: The preferential time segment is clearly displayed in the user interface, including calendar view and list view, to facilitate user planning of job scheduling;

[0116] Preferential charging model: Different charging models are designed for preferential time segments, such as reducing usage rates, providing additional free quotas, or implementing discounts.

[0117] Notification and reminder mechanism: The user is notified in advance of the upcoming preferential time segment through email and SMS;

[0118] API support: Allows advanced users or third-party applications to automatically obtain preferential time segment information and adjust job scheduling.

[0119] Historical data analysis: Provide historical data analysis tools to help users understand past preferential time segment usage and cost-effectiveness.

[0120] S402, in response to the trigger instruction, set the trigger condition of the monitoring threshold and the adjustment value.

[0121] Optionally, when the tenant is interested in the preferential time segment, the trigger instruction can be generated through the configuration operation, and the cloud management platform sets the trigger condition of the monitoring threshold and the adjustment value based on the trigger instruction.

[0122] S403, when the actual time is a preferential time segment, trigger the trigger condition, and reduce the value of the current monitoring threshold to the adjustment value.

[0123] Optionally, when the actual time is a preferential time segment, trigger the trigger condition, and reduce the value of the current monitoring threshold to the adjustment value, so that the user can set a more relaxed monitoring threshold in the case of controlling costs, and use more computing cluster nodes when needed to meet the actual needs of the user.

[0124] It should be noted that the function indicators collected by the collection monitoring center and the monitoring results of the control center can form an operation statistical report and be saved to the database. In addition, the control center can configure algorithms such as Q-Learning, so as to summarize the corresponding rules of the function indicators and predict the best scaling scheme.

[0125] Through the method of the embodiment of the present application, the power resource allocation and billing self-service of the automatic scaling cluster can dynamically adjust the allocation of power resources according to user business needs and scaling strategies, so that the power cluster can quickly respond to the surge or reduction of traffic, not only enhancing the availability of services, but also significantly reducing the economic burden of users, allowing users to pay fees according to actual use; therefore, the power node resource pool efficiency can be maximized, and the cost-effectiveness optimization is realized by saving the increasingly scarce and expensive power resources.

[0126] The target of the embodiment of the present application is to solve the previous pain points, and integrate the power resource allocation and billing self-service function on the cloud platform to enhance the power resource management capability of the automatic scaling cluster and simplify the billing process. Through a series of refined power resource allocation settings and billing settings, users can not only determine the scale of the cluster and the cost budget independently, but also the platform can guide users to use resources during specific preferential periods to reduce costs. In addition, the platform has a cost warning function, which can notify users in time when the user consumption approaches or exceeds the budget, and automatically adjust the number of nodes when necessary.

[0127] In the case of tight power resource pool, the cloud platform can also reasonably allocate power resources according to different tenant level types of users to ensure fair and efficient use of resources. The introduction of this power resource allocation and billing self-service management function aims to achieve the optimal balance between user utilization rate, cost and power cluster utilization rate, thereby providing a cost-effective and flexible and reliable power service environment for users. In summary, the present application provides an innovative solution to the common market problem of limited resources and surging demand, limited budget and rising costs, especially for scenes such as scientific research projects and stress testing with limited funds, which brings significant economic advantages and operational convenience to users through flexible resource management and cost control.

[0128] Referring to Figure 4 , a structural block diagram of a power resource allocation device of an embodiment of the present application is shown, which can include:

[0129] The first determination module is configured to obtain an opening application of a tenant, and determine the level type of the tenant according to the tenant property of the opening application;

[0130] The second determining module is configured to determine a target adaptive node quantity range according to the level type.

[0131] The configuration module is configured to configure a scaling strategy of the computing power cluster nodes of the computing power cluster according to the target adaptive node quantity range.

[0132] The adjustment module is configured to monitor a function index of the computing power cluster in a process in which the computing power cluster is in a use state, and dynamically adjust the quantity of the computing power cluster nodes according to the function index and the scaling strategy.

[0133] The functions of each module in each device of the embodiments of the present application can be referred to the corresponding description in the above method, which will not be repeated here.

[0134] Referring to Figure 5 , a structural block diagram of an electronic device of an embodiment of the present application is shown, which includes a memory 310 and a processor 320, the memory 310 stores instructions executable on the processor 320, and the processor 320 loads and executes the instructions to implement the computing power resource allocation method in the above embodiments. The number of the memory 310 and the processor 320 can be one or more.

[0135] In an implementation, the electronic device further includes a communication interface 330 for communicating with external devices to transmit and exchange data. If the memory 310, the processor 320 and the communication interface 330 are independently implemented, the memory 310, the processor 320 and the communication interface 330 can be connected to each other through a bus and complete communication therebetween. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 only one thick line is used in the figure, but it does not mean that there is only one bus or only one type of bus.

[0136] Optionally, in specific implementation, if the memory 310, the processor 320 and the communication interface 330 are integrated on a chip, the memory 310, the processor 320 and the communication interface 330 can complete communication therebetween through an internal interface.

[0137] The embodiments of the present application provide a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the computing power resource allocation method provided in the above embodiments.

[0138] The embodiment of the present application further provides a chip, which comprises a processor, and is used for calling and running instructions stored in a memory, so that a communication device installed with the chip executes the method provided by the embodiment of the present application.

[0139] The embodiment of the present application further provides a chip, which comprises an input interface, an output interface, a processor and a memory, and the input interface, the output interface, the processor and the memory are connected through internal connection paths, and the processor is used for executing codes in the memory, and when the codes are executed, the processor is used for executing the method provided by the embodiment of the present application.

[0140] It should be understood that the processor mentioned above can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc. It should be noted that the processor can be a processor supporting an advanced RISC machine (ARM) architecture.

[0141] Further, the aforementioned memory can include a read-only memory, and a random access memory, and can further include a nonvolatile random access memory. The memory can be a volatile memory or a nonvolatile memory, or can include both volatile and nonvolatile memory. The nonvolatile memory can include a read-only memory (ROM), a programmable ROM (PROM), an erasable PROM (EPROM), an electrically EPROM (EEPROM), or a flash memory. The volatile memory can include a random access memory (RAM), which functions as an external cache. Many forms of RAM are available. For example, a static RAM (SRAM), a dynamic RAM (DRAM), a synchronous DRAM (SDRAM), a double data rate SDRAM (DDR SDRAM), an enhanced SDRAM (ESDRAM), a synchlink DRAM (SLDRAM), and a direct rambus RAM (DR RAM) are available.

[0142] In the above-described embodiments, all or a part can be implemented by software, hardware, firmware, or any combination thereof. When implemented as software, it can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed by a computer, all or a part of the procedures or functions according to the present disclosure are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable apparatus. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another computer-readable storage medium.

[0143] In the description of the application, the description of the terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the application. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in one or more embodiments or examples. In addition, different embodiments or examples described in the specification and the features of different embodiments or examples can be combined and combined by those skilled in the art without contradiction, if necessary.

[0144] In addition, the terms "first", "second" are only for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include at least one of the features. In the description of the application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0145] Any process or method descriptions in flow charts or described elsewhere herein can be understood as representing code modules, segments, or portions of code that include one or more executable instructions for implementing specific logic functions or other processes. And the various embodiments of the application can include additional or fewer steps or processes in comparison to those shown in the figures.

[0146] The logic and / or steps represented in flow charts or otherwise described herein, for example, can be embodied in computer-readable instructions, which can be used to cause one or more processors to perform the actions indicated in the steps. The computer-readable instructions can be stored on one or more storage media or memory devices associated with the one or more processors.

[0147] It should be understood that parts of the application can be implemented in hardware, software, firmware or a combination thereof. In the above-described embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. All or part of the steps of the above-described embodiment method can be instructed by a program to complete the relevant hardware, which can be stored in a computer-readable storage medium, and the program includes one or a combination of the steps of the method embodiment when executed.

[0148] In addition, each of the function units in each embodiment of the present application can be integrated in one processing module, or each unit can be physically present separately, or two or more units can be integrated in one module. The integrated module can be realized in the form of hardware or in the form of a software function module. When the integrated module is realized in the form of a software function module and sold or used as an independent product, it can also be stored in a computer readable storage medium. The storage medium can be a read-only memory, a magnetic disk or an optical disk, etc.

[0149] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of various changes or replacements within the technical scope disclosed in the present application, and these should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A computing resource allocation method, characterized in that, The method comprises the following steps: obtaining a tenant opening application, and determining a level type of the tenant according to a tenant property of the opening application; determining a target adaptive node quantity range according to the level type; configuring a scaling strategy of a computing power cluster node of a computing power cluster according to the target adaptive node quantity range; monitoring a function index of the computing power cluster in a process that the computing power cluster is in a use state, and dynamically adjusting a quantity of the computing power cluster node according to the function index and the scaling strategy.

2. The computing resource allocation method of claim 1, wherein: The step of determining the target adaptive node quantity range according to the level type comprises the following steps: matching the level type with a plurality of preset level types to determine a target level type, wherein each preset level type has a corresponding mapped level value range, and an adjustable amplitude of each level value range is different; taking a level value range mapped by the target level type as a target adaptive node quantity range.

3. The computing resource allocation method of claim 1 or 2, wherein: The step of configuring the scaling strategy of the computing power cluster node of the computing power cluster according to the target adaptive node quantity range comprises the following steps: setting a minimum value of the quantity of the computing power cluster node according to a minimum boundary value of the target adaptive node quantity range; setting a guaranteed maximum value of the quantity of the computing power cluster node according to a maximum boundary value of the target adaptive node quantity range; setting an initial value of the quantity of the computing power cluster node according to the target adaptive node quantity range, wherein the initial value is located between the minimum value and the guaranteed maximum value; configuring a monitoring threshold value of the function index according to a configuration instruction; determining the scaling strategy of the computing power cluster node of the computing power cluster according to the minimum value, the guaranteed maximum value, the initial value and the monitoring threshold value.

4. The computing resource allocation method of claim 3, wherein: The step of monitoring the function index of the computing power cluster in the process that the computing power cluster is in the use state, and dynamically adjusting the quantity of the computing power cluster node according to the function index and the scaling strategy comprises the following steps: monitoring the function index of the computing power cluster in a process that the computing power cluster node of the initial value is in the use state; performing a function index monitoring rule to compare the function index with the monitoring threshold value; when the function index is greater than or equal to the monitoring threshold value, increasing the quantity of the computing power cluster node in a range between the initial value and the guaranteed maximum value; when the function index is less than the monitoring threshold value and the function index continuously decreases, decreasing the quantity of the computing power cluster node in a range between the minimum value and the initial value.

5. The computing resource allocation method of claim 4, wherein: The step of dynamically adjusting the quantity of the computing power cluster node according to the function index and the scaling strategy further comprises the following steps: when the quantity of the computing power cluster node is increased, and the quantity of the computing power cluster node reaches the guaranteed maximum value and the function index continuously increases, performing capacity analysis on the computing power cluster to determine an available quota quantity of the computing power cluster; determining a sum value of the guaranteed maximum value and the available quota quantity, and increasing the quantity of the computing power cluster node to the sum value.

6. The computing resource allocation method of claim 4, wherein: The scaling strategy further comprises a cost threshold and a priority; and the dynamically adjusting the number of the computing cluster nodes according to the function index and the scaling strategy further comprises: calculating a total cost according to a pre-designed cost rule, the number of the computing cluster nodes, and a usage time; if the total cost is greater than the cost threshold, when the priority is that the cost rule is higher than the function index monitoring rule, reducing the number of the computing cluster nodes so that the total cost is less than or equal to the cost threshold; if the total cost is greater than the cost threshold, when the priority is that the cost rule is lower than the function index monitoring rule, performing the step of comparing the function index with the monitoring threshold.

7. The computing resource allocation method of claim 3, wherein: The dynamically adjusting the number of the computing cluster nodes according to the function index and the scaling strategy comprises: determining a preferential period and displaying it; in response to a trigger instruction, setting a trigger condition of the monitoring threshold and an adjustment value; when the actual time is the preferential period, triggering the trigger condition and reducing the value of the current monitoring threshold to the adjustment value.

8. A computing resource allocation apparatus, characterized by comprising: comprise: a first determining module configured to acquire an opening application of a tenant and determine a level type of the tenant according to a tenant property of the opening application; a second determining module configured to determine a target adaptive node number range according to the level type; a configuration module configured to configure a scaling strategy of computing cluster nodes of a computing cluster according to the target adaptive node number range; an adjusting module configured to monitor a function index of the computing cluster in a process in which the computing cluster is in a usage state, and dynamically adjust the number of the computing cluster nodes according to the function index and the scaling strategy.

9. An electronic device, comprising: comprise: a processor and a memory, the memory storing instructions, the instructions being loaded and executed by the processor to implement the method of any one of claims 1-7. 10.A computer readable storage medium, the computer readable storage medium storing a computer program, the computer program being executed to implement the method of any one of claims 1-7.

Images