Resource scheduling method and device and storage medium

Abstract

The embodiment of the invention discloses a resource scheduling method and device and a storage medium. According to the technical scheme provided by the embodiment of the invention, the application resource portrait and the application resource use periodicity are determined according to the resource use statistical condition of the application, the first computing power resource demand value is determined according to the application resource portrait and the periodicity, and the recommended copy number is determined according to the application resource portrait and the first computing power resource demand value; a resource deployment strategy is determined according to the number of recommended copies, the copies are deployed in different container nodes for the application based on the resource deployment strategy, a computing power resource demand value is automatically determined, the computing power resource demand value is closer to a true value used by copy resources, the number of the copies capable of being deployed by the same container node is larger, and the deployment efficiency is improved. The packing rate of the container nodes is effectively improved, and the computing power resource utilization rate of the container nodes is effectively improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computer technology, and in particular to a resource scheduling method, device, and storage medium. Background Art

[0002] With the development of internet and cloud computing technologies, the application of cloud-based platform service systems, such as conference management systems and teaching systems, is becoming increasingly widespread. Cloud platform service systems are typically deployed in a centrally managed container cluster, which consists of multiple distributed container nodes. Cloud platform service systems typically consist of one or more microservices, each with multiple replicas. These replicas are deployed and run on container nodes, providing the corresponding services by making container resources available to the replicas. To improve container resource utilization, container resources need to be scheduled, meaning replicas are deployed across different container nodes. During container node scheduling, container nodes are filtered and scored based on the computing resource requirements set for the application replicas, ultimately selecting a container node as the deployment node for the replica.

[0003] Computing resource requirements are typically set by developers based on experience. For example, they might set the CPU and memory requirements for a single replica of an application. The container system scheduler then deploys replicas based on these requirements. However, this artificially set computing resource requirement can easily lead to a significant discrepancy between the required computing resource and the actual resource usage of the replicas, resulting in reduced computing resource utilization on container nodes. Summary of the Invention

[0004] The embodiments of the present application provide a resource scheduling method, device and storage medium to solve the technical problem in the related art that the method of manually setting the computing power resource demand value easily leads to a large difference between the computing power resource demand value and the actual value of the replica resource usage, resulting in a decrease in the computing power resource utilization rate of the container node, thereby effectively improving the computing power resource utilization rate of the container node.

[0005] In a first aspect, an embodiment of the present application provides a resource scheduling method, comprising:

[0006] Get the application's resource usage statistics;

[0007] Determine the application resource profile and the periodicity of application resource usage based on the resource usage statistics of the application;

[0008] Determine a first computing resource requirement value based on the application resource profile and the periodicity;

[0009] Determine the recommended number of replicas based on the application resource profile and the first computing resource requirement value;

[0010] A resource deployment strategy is determined according to the recommended number of replicas, and replicas are deployed for the application in different container nodes based on the resource deployment strategy.

[0011] The embodiment of the present application determines the application resource portrait and the periodicity of application resource usage based on the resource usage statistics of the application, determines the first computing power resource demand value based on the application resource portrait and the periodicity, determines the recommended number of replicas based on the application resource portrait and the first computing power resource demand value, determines the resource deployment strategy based on the recommended number of replicas, and deploys replicas for the application in different container nodes based on the resource deployment strategy, automatically determines the computing power resource demand value, the computing power resource demand value is closer to the actual value of the replica resource usage, and the same container node can deploy more replicas, effectively improving the packing rate of the container node and effectively improving the computing power resource utilization rate of the container node.

[0012] Furthermore, the determining the first computing resource requirement value according to the application resource profile and the periodicity includes:

[0013] In a case where the periodicity is that the application resource usage has periodicity, the first computing resource demand value is set to an average value of a plurality of first peak setting percentage lines within a first set time period in the application resource profile;

[0014] When the periodicity is that the application resource usage does not have periodicity, the first computing power resource demand value is set to the average value of multiple peak average values ​​within the first set time period in the application resource portrait.

[0015] In the above, by setting the first computing resource demand value to the average value of multiple first peak set percentage lines within the first set time period in the application resource portrait when the application resource usage is periodic, and setting the first computing resource demand value to the average value of multiple peak average values ​​within the first set time period in the application resource portrait when the application resource usage is not periodic, the first computing resource demand value is made closer to the actual value of the replica resource usage, effectively reducing the situation where the average utilization rate of computing resources is low due to the periodicity of business application resource utilization, and effectively improving the computing resource utilization rate of the container node.

[0016] Furthermore, determining the recommended number of replicas based on the application resource profile and the first computing resource requirement value includes:

[0017] When the first computing resource demand value is less than the set demand threshold, the computing resource utilization rate is determined based on the application resource profile and the first computing resource demand value, and the recommended number of copies is determined based on the computing resource utilization rate and the recommended utilization rate.

[0018] In the above, when the first computing resource demand value is less than the set demand threshold, the computing resource utilization rate is accurately determined according to the application resource portrait and the first computing resource demand value, thereby improving the accuracy of determining the recommended number of copies and improving the resource scheduling effect.

[0019] Furthermore, determining the computing resource usage rate according to the application resource profile and the first computing resource demand value includes:

[0020] The computing power resource utilization rate is determined based on the ratio of the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period in the application resource portrait to the total amount of application resources.

[0021] In the above, the computing power resource utilization rate is accurately determined by the ratio of the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period in the application resource portrait to the total amount of application resources, thereby improving the accuracy of determining the recommended number of copies and improving the resource scheduling effect.

[0022] Furthermore, determining the recommended number of replicas based on the computing resource utilization rate and the recommended utilization rate includes:

[0023] When the computing resource usage rate is less than the set recommended usage rate, determining the number of candidate replicas according to the percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource profile, the set scaling factor, and the second computing resource requirement value, and determining the recommended number of replicas according to the candidate number of replicas, where the second computing resource requirement value is obtained according to the first computing resource requirement value;

[0024] When the computing resource utilization rate is greater than or equal to the set recommended utilization rate, the recommended number of replicas is determined according to the computing resource utilization rate and the recommended utilization rate.

[0025] In the above, when the computing power resource utilization rate is less than the set recommended utilization rate, the number of candidate copies is determined according to the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period in the application resource portrait, the set scaling factor and the second computing power resource demand value, and the recommended number of copies is accurately determined based on the candidate number of copies, and when the computing power resource utilization rate is greater than or equal to the set recommended utilization rate, the recommended number of copies is accurately determined based on the computing power resource utilization rate and the recommended utilization rate, thereby improving the accuracy of resource deployment.

[0026] Furthermore, determining the number of candidate replicas according to the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period in the application resource profile, setting the scaling factor and the second computing resource requirement value includes:

[0027] Rounding up the first computing resource requirement value to obtain a second computing resource requirement value;

[0028] The number of candidate copies is determined by rounding up the product of the percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource portrait and the set scaling factor, and the ratio of the second computing power resource demand value.

[0029] In the above, the second computing power resource demand value is obtained by rounding up the first computing power resource demand value, and the product of the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period in the application resource portrait and the set scaling factor is rounded up to the ratio of the second computing power resource demand value to accurately determine the number of candidate copies, thereby effectively improving the accuracy of resource deployment.

[0030] Furthermore, after rounding up the first computing resource requirement value to obtain the second computing resource requirement value, the method further includes:

[0031] In a case where the application is a set type application, the second computing resource requirement value is updated using a first set proportional coefficient.

[0032] In the above, by using the first set proportional coefficient to update the second computing power resource requirement value when the application is a set type application, taking into account the computing power resource margin reserved when the application is a set type application, the second computing power resource requirement value is determined more accurately, effectively improving the accuracy of determining the number of candidate copies.

[0033] Furthermore, determining the recommended number of replicas based on the number of candidate replicas includes:

[0034] In the case where the ratio of the maximum percentage peak value to the product of the set maximum value of replica resources and the number of candidate replicas is greater than the recommended usage rate, increasing the number of candidate replicas;

[0035] When the ratio of the maximum percentage peak value to the product of the set maximum value of replica resources and the number of candidate replicas is less than or equal to the recommended usage rate, the recommended number of replicas is determined as the recommended number of replicas.

[0036] In the above, by increasing the number of candidate copies when the ratio of the percentage peak maximum value to the product of the set maximum value of copy resources and the number of candidate copies is greater than the recommended usage rate, until the ratio of the percentage peak maximum value to the product of the set maximum value of copy resources and the number of candidate copies is less than or equal to the recommended usage rate, the recommended number of copies is determined as the recommended number of copies, thereby improving the accuracy of determining the recommended number of copies and effectively improving the accuracy of resource deployment.

[0037] Furthermore, after determining the recommended number of copies as the recommended number of copies, the method further includes:

[0038] The recommended number of copies is constrained based on a set minimum recommended copy threshold.

[0039] In the above, by using the minimum recommended replica threshold to constrain the recommended number of replicas, the recommended number of replicas is guaranteed to effectively improve the accuracy of resource deployment.

[0040] Furthermore, determining the recommended number of replicas based on the computing resource utilization rate and the recommended utilization rate includes:

[0041] When the computing power resource utilization rate is greater than or equal to the product of the set recommended utilization rate and the second set proportional coefficient, the recommended number of copies is determined by rounding up the ratio of the percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource portrait to the product of the set maximum value of the copy resource and the recommended utilization rate.

[0042] In the above, when the computing power resource utilization rate is greater than or equal to the product of the set recommended utilization rate and the second set proportional coefficient, the recommended number of copies is determined by rounding up the ratio of the percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource portrait to the product of the set maximum value of the copy resource and the recommended utilization rate, thereby effectively improving the accuracy of resource deployment.

[0043] Furthermore, after rounding up the ratio of the maximum percentage peak value of the plurality of second peak setting percentage lines within the second set time period in the application resource profile to the product of the set maximum value of the replica resource and the recommended usage rate to determine the recommended number of replicas, the method further includes:

[0044] The recommended number of copies is constrained based on a set minimum recommended copy threshold.

[0045] In the above, by using the minimum recommended replica threshold to constrain the recommended number of replicas, the recommended number of replicas is guaranteed to be deployed with sufficient replicas, thereby effectively improving the accuracy of resource deployment.

[0046] Furthermore, determining a resource deployment strategy based on the recommended number of replicas includes:

[0047] According to the recommended number of replicas, a first number of replicas is deployed to a fixed node pool, and a second number of replicas is deployed to an elastic node pool, where the sum of the first number and the second number is equal to the recommended number of replicas.

[0048] As mentioned above, by deploying some application replicas to the fixed node pool and other replicas to the elastic node pool, the peak of the fixed node pool is reduced, so that the periodic resource usage characteristic curve becomes a relatively smooth straight line, further improving the node packing rate and the actual computing power resource utilization rate of the node.

[0049] In a second aspect, an embodiment of the present application provides a resource scheduling device, including: a memory and one or more processors;

[0050] The memory is used to store one or more programs;

[0051] When the one or more programs are executed by the one or more processors, the one or more processors implement the resource scheduling method as described in the first aspect.

[0052] In a third aspect, an embodiment of the present application provides a storage medium storing computer-executable instructions, which, when executed by a computer processor, are used to execute the resource scheduling method as described in the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a flow chart of a resource scheduling method provided by an embodiment of the present application;

[0054] Figure 2 This is a schematic diagram of a first computing resource requirement value determination process provided by an embodiment of the present application;

[0055] Figure 3 This is a flowchart of another resource scheduling method provided by an embodiment of the present application;

[0056] Figure 4 This is a structural diagram of a resource scheduling device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of the present application clearer, the specific embodiments of the present application are further described in detail below in conjunction with the accompanying drawings. It is understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. It should also be noted that, for ease of description, only some, but not all, of the contents related to the present application are shown in the accompanying drawings. Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe each operation (or step) as a sequential process, many of the operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of the operations can be rearranged. The above process can be terminated when its operation is completed, but it can also have additional steps not included in the accompanying drawings. The above process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0058] In existing container node resource scheduling solutions, developers typically set the processor, memory, and other computing resource requirements for application replicas. The container scheduler then filters and scores all container nodes based on the corresponding computing resource requirements, ultimately selecting one container node as the replica's deployment node. However, because the computing resource requirements are manually set by developers, it's difficult for them to accurately set them. The resulting discrepancy between the computing resource requirements and the actual replica resource usage can be significant, leading to low computing resource utilization for container nodes. Furthermore, due to the nature of business applications, resource usage is cyclical. These periodic peaks and troughs result in low average computing resource utilization for container nodes, reducing computing resource utilization at container nodes.

[0059] Based on this, a resource scheduling method is provided in an embodiment of the present application to solve the technical problem that the existing method of manually setting the computing power resource demand value easily leads to a large difference between the computing power resource demand value and the actual value of the replica resource usage, resulting in a reduction in the computing power resource utilization rate of the container node.

[0060] Figure 1 A flowchart of a resource scheduling method provided in an embodiment of the present application is given. The resource scheduling method provided in an embodiment of the present application can be implemented by hardware and / or software and integrated into a resource scheduling device.

[0061] The following description is based on an example of a resource scheduling device executing a resource scheduling method. Figure 1 , the resource scheduling method includes:

[0062] S110: Obtain resource usage statistics of the application, and determine the application resource profile and the periodicity of application resource usage based on the resource usage statistics of the application.

[0063] Exemplarily, the resource usage statistics of the set application within the set time period are obtained in real time. The resource usage provided by this solution can be the usage of computing resources (number of processors and / or amount of memory, etc.) by the application. Exemplarily, the usage of computing resources by the application in multiple time periods within a first set time period (for example, 3 days, 7 days, 15 days, etc.) is collected, and the resource usage statistics of the application are determined based on the usage of computing resources by the application in multiple time periods within the first set time period. Furthermore, the application resource profile and the periodicity of application resource usage are determined based on the resource usage statistics of the application.

[0064] Among them, containers are a lightweight virtualization technology that can be used to package and deploy applications and their dependencies. Optionally, the containers provided by this solution can be Kubernetes containers, Docker containers, Linux containers, etc., and the nodes corresponding to Kubernetes containers are k8s nodes. Multiple container nodes can be distributed on different physical machines, virtual machines, or container orchestration platforms. The replicas (pods) provided by this solution are the smallest deployment units of services on container nodes. In cloud platform service systems, replicas can be multiple instances of a service, and each replica is usually deployed in a container node.

[0065] In one embodiment, the application resource profile provided by this solution may include the maximum value of computing power resources used by a single replica of the application within a first set time period (the maximum value of a single Pod, max), the average value (avg) of computing power resources used by each replica of the application, the peak average value (peakTimeAvg), a first peak setting percentage line (e.g., peak 95 line peakTime95), a plurality of second peak setting percentage lines for second time periods, the number of replicas of the application (number of pods), etc. Optionally, the second peak setting percentage lines for the plurality of second time periods may be the second peak setting percentage lines for different time intervals of each day within the first time period, and the plurality of second time periods may be the morning peak (7:30-11:30), noon (11:30-13:30), afternoon (13:30-17:30), evening (17:30-22:00), midnight (22:00-7:30 the next day), etc.

[0066] Whether application resource usage exhibits periodicity can be determined using a Fourier transform and / or autocorrelation function. For example, a Fourier transform can be performed on the data corresponding to the application resource profile. The Fourier transform can convert the time series of the application resource profile into a corresponding frequency domain representation, decomposing a time series into different frequency components. By analyzing these frequency components, it is possible to determine whether periodicity exists in the time series corresponding to the application resource profile. After Fourier transforming the application resource profile to obtain the coefficients of each frequency component in the application resource profile, if the coefficient of a frequency component is large, it indicates that the frequency component plays a significant role in the time series. These coefficients are then examined to see if they exhibit a periodic pattern. If so, it indicates that the original time series exhibits periodicity. The autocorrelation function is a method for measuring the similarity between a time series and itself at different time delays. For a periodic time series, the autocorrelation function will exhibit peaks at specific time delays. These peaks correspond to the recurrence of periodic signals. By analyzing the values ​​of the autocorrelation function, it is possible to determine whether there are peaks associated with the periodicity of the time series. If significant peaks are present, it indicates that the time series exhibits periodicity. Optionally, the analysis results of Fourier transform and autocorrelation function can be combined to more accurately determine whether the data is periodic. For example, when the Fourier transform results show the presence of significant frequency components and the autocorrelation function exhibits a peak at a specific time delay, it can be determined that the application resource usage is periodic and the corresponding application is a periodic application.

[0067] S120: Determine a first computing resource demand value based on the application resource profile and periodicity.

[0068] Exemplarily, the first computing resource demand value is determined periodically based on the application resource profile determined above, wherein the first computing resource demand value is the application's demand value (request value) for computing resources (eg, processor CPU and / or memory).

[0069] In one possible embodiment, Figure 2 As shown in a flowchart of a first computing resource requirement value determination process, the resource scheduling method provided in this solution includes the following steps when determining the first computing resource requirement value based on application resource profiles and periodically:

[0070] S121: When the periodicity is that the application resource usage has periodicity, the first computing power resource demand value is set to the average value of multiple first peak setting percentage lines within a first set time period in the application resource portrait.

[0071] S122: When the periodicity is that the application resource usage does not have periodicity, the first computing power resource demand value is set to the average value of multiple peak value average values ​​within a first set time period in the application resource portrait.

[0072] Exemplarily, when it is determined that the application resource usage is periodic, the first computing resource demand value can be set to the average value of multiple first peak setting percentage lines within a first set time period (e.g., 3 days, 7 days, 15 days, etc.) in the application resource portrait. For example, the first computing resource demand value is set to the average value of multiple first peak setting percentage lines (e.g., peak setting 95 line peakTime95) within 7 days in the application resource portrait. When the application resource usage is not periodic, the first computing resource demand value can be set to the average value of multiple peak average values ​​within a first set time period (e.g., 3 days, 7 days, 15 days, etc.) in the application resource portrait. For example, the first computing resource demand value is set to the average value of multiple peak average values ​​(peakTimeAvg) within 7 days in the application resource portrait. Optionally, the copies corresponding to the non-periodic applications can be deployed to the set node pool, and the non-periodic applications can be deployed to an additional node pool, separately from the periodic applications, to improve the management efficiency of the applications.

[0073] This solution sets the first computing resource demand value to the average value of multiple first peak set percentage lines within the first set time period in the application resource portrait when the application resource usage is periodic, and sets the first computing resource demand value to the average value of multiple peak average values ​​within the first set time period in the application resource portrait when the application resource usage is not periodic. This makes the first computing resource demand value closer to the actual value of replica resource usage, effectively reduces the situation where the average computing resource utilization rate is low due to the periodicity of business application resource utilization, and effectively improves the computing resource utilization rate of container nodes.

[0074] S130: Determine the recommended number of replicas based on the application resource profile and the first computing resource demand value.

[0075] Exemplarily, after determining the first computing resource requirement value, a recommended number of replicas is determined based on the determined application resource profile and the first computing resource requirement value, wherein the recommended number of replicas is the number of replicas recommended for application deployment.

[0076] In one possible embodiment, the corresponding recommended number of copies can be determined in advance based on the combination of different application resource portraits and the first computing resource demand value, and when the current application resource portrait and the first computing resource demand value are determined, the corresponding recommended number of copies can be determined based on the current application resource portrait and the first computing resource demand value.

[0077] S140: Determine a resource deployment strategy according to the recommended number of replicas, and deploy replicas for the application in different container nodes based on the resource deployment strategy.

[0078] Exemplarily, after determining the recommended number of replicas for an application, a resource deployment strategy is determined based on the recommended number of replicas. The resource deployment strategy can be used to indicate the number of replicas to deploy for the application and the locations (e.g., node pools) where the replicas are deployed. The locations where different replicas are deployed can be the same or different, meaning that different replicas can be deployed in the same or different node pools. Alternatively, the recommended number of replicas can be determined as the number of replicas to deploy for the application as indicated by the resource deployment strategy.

[0079] After determining the resource deployment strategy, replicas can be deployed for the application in different container nodes based on the resource deployment strategy. For example, according to the number of replicas and deployment locations indicated by the resource deployment strategy, replicas are deployed for the application through the scheduler of the container system (such as the kube-scheduler scheduler of container k8s). The computing power resource demand value determined based on the application resource profile and the periodicity is closer to the actual value of replica resource usage, so that more replicas can be deployed on the same container node, effectively improving the packing rate of the container node. Among them, the packing rate of the container node can be used to measure the efficiency of container resource utilization. It represents the ratio of the amount of resources actually used on the container node to the total amount of resources of the node within a set time, and the higher the packing rate, the higher the resource utilization on the container node.

[0080] In one embodiment, this solution can configure multiple applications on a server, and can determine the resource deployment strategy based on the resource scheduling method provided by this solution for the resource usage statistics of each application, and deploy copies for each application in different container nodes based on the resource deployment strategy.

[0081] In the above, the application resource profile and the periodicity of application resource usage are determined based on the resource usage statistics of the application, and the first computing resource demand value is determined based on the application resource profile and the periodicity. The recommended number of replicas is determined based on the application resource profile and the first computing resource demand value, and the resource deployment strategy is determined based on the recommended number of replicas. Based on the resource deployment strategy, replicas are deployed for the application in different container nodes, and the computing resource demand value is automatically determined. The computing resource demand value is closer to the actual value of the replica resource usage, and more replicas can be deployed on the same container node, which effectively improves the packing rate of the container node and effectively improves the computing resource utilization rate of the container node.

[0082] Based on the above embodiments, Figure 3 A flowchart of another resource scheduling method provided in an embodiment of the present application is given, which is a specific implementation of the above resource scheduling method. Figure 3 , the resource scheduling method includes:

[0083] S210: Determine an application resource profile and a periodicity of application resource usage based on the application resource usage statistics.

[0084] S220: Determine the first computing power resource demand value based on the application resource profile and periodicity.

[0085] S230: Determine the recommended number of replicas based on the application resource profile and the first computing resource demand value.

[0086] In one possible embodiment, the resource scheduling method provided by this solution, when determining the recommended number of copies based on the application resource profile and the first computing resource demand value, can be: when the first computing resource demand value is less than the set demand threshold, determine the computing resource utilization rate based on the application resource profile and the first computing resource demand value, and determine the recommended number of copies based on the computing resource utilization rate and the recommended utilization rate.

[0087] Exemplarily, after determining the first computing power resource demand value, the first computing power resource demand value is compared with a set demand threshold (for example, the set demand threshold is set to 2). When the first computing power resource demand value is less than the set demand threshold, the computing power resource utilization rate is determined based on the application resource profile and the first computing power resource demand value, and the recommended number of copies is determined based on the computing power resource utilization rate and the recommended utilization rate. When the first computing power resource demand value is greater than or equal to the set demand threshold, it can be considered that the current resource deployment strategy for the application is reasonable, the current resource deployment strategy for the application can be maintained, and the resource scheduling for the current application can be terminated, and the resource scheduling process for the next application can be carried out. This solution improves the accuracy of determining the recommended number of copies and improves the resource scheduling effect by accurately determining the computing power resource utilization rate based on the application resource profile and the first computing power resource demand value when the first computing power resource demand value is less than the set demand threshold.

[0088] In one possible embodiment, the resource scheduling method provided by this solution, when determining the computing power resource utilization rate based on the application resource portrait and the first computing power resource demand value, can be: determining the computing power resource utilization rate based on the ratio of the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period in the application resource portrait and the total amount of application resources.

[0089] This solution provides multiple second set time periods. For example, multiple second set time periods can be different time intervals of a day, such as the morning rush hour (7:30-11:30), noon (11:30-13:30), afternoon (13:30-17:30), evening (17:30-22:00), and midnight (22:00-7:30 the next day). When performing resource scheduling, the corresponding recommended number of replicas for the application can be determined for each second set time period, the resource deployment strategy can be determined based on the recommended number of replicas, and replicas can be deployed for the application in different container nodes based on the resource deployment strategy.

[0090] Exemplarily, for a second set time period corresponding to a time interval, the computing power resource utilization rate is determined based on the ratio of the maximum percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource portrait to the total amount of application resources, wherein the percentage peak value of the multiple second peak setting percentage lines within the second set time period can be the percentage peak value of the second peak setting percentage lines corresponding to multiple second set time periods within the first set time period (for example, 7 days).

[0091] For example, taking the computing power resource as the number of processors (CPUs) as an example, the computing power resource utilization rate can be expressed as: cpuRate = peakTimeSumMax / appCpuLimits, where peakTimeSumMax is the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period, and appCpuLimits is the total amount of application resources. Optionally, the total amount of application resources can be determined based on the product of the maximum computing power resource value of each copy of the application configuration and the current number of copies of the application. For example, the total amount of application resources can be expressed as: appCpulimits = limitsCpu * current number of copies, where limitsCpu is the maximum computing power resource value of each copy of the application configuration. This solution accurately determines the computing power resource utilization rate based on the ratio of the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period in the application resource portrait to the total amount of application resources, thereby improving the accuracy of determining the recommended number of copies and improving the resource scheduling effect.

[0092] In one possible embodiment, the resource scheduling method provided by this solution, when determining the recommended number of replicas based on computing resource utilization and recommended utilization, includes:

[0093] S231: When the computing power resource utilization rate is less than the set recommended utilization rate, the number of candidate copies is determined based on the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period in the application resource portrait, the set scaling factor and the second computing power resource demand value, and the recommended number of copies is determined based on the number of candidate copies. The second computing power resource demand value is obtained based on the first computing power resource demand value.

[0094] S232: When the computing resource utilization rate is greater than or equal to the set recommended utilization rate, determine the recommended number of replicas based on the computing resource utilization rate and the recommended utilization rate.

[0095] Exemplarily, after determining the computing power resource utilization rate, the computing power resource utilization rate is compared with the set recommended utilization rate (for example, 60%). If the computing power resource utilization rate is less than the recommended utilization rate, the number of candidate replicas is determined based on the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period in the application resource profile, the set scaling factor (the set scaling factor can be between 0.9 and 1.1, for example, the set scaling factor can be set to 1), and the second computing power resource demand value, and the recommended number of replicas is determined based on the number of candidate replicas. For example, the product of the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period and the set scaling factor is used as the ratio of the second computing power resource demand value as the candidate number of replicas.

[0096] Among them, the second computing power resource requirement value provided by this solution can be obtained based on the first computing power resource requirement value. For example, the first computing power resource requirement value can be directly used as the second computing power resource requirement value, or the product of the first computing power resource requirement value and the set proportional coefficient can be used as the second computing power resource requirement value.

[0097] In one embodiment, when the computing power resource utilization rate is greater than or equal to the set recommended utilization rate, the recommended number of replicas is determined based on the computing power resource utilization rate and the recommended utilization rate. This solution improves resource deployment accuracy by determining the candidate number of replicas based on the percentage peak maximum value of multiple second peak set percentage lines within a second set time period in the application resource profile, the set scaling factor, and the second computing power resource demand value when the computing power resource utilization rate is less than the set recommended utilization rate, and accurately determining the recommended number of replicas based on the candidate number of replicas. When the computing power resource utilization rate is greater than or equal to the set recommended utilization rate, the recommended number of replicas is accurately determined based on the computing power resource utilization rate and the recommended utilization rate.

[0098] In one possible embodiment, the resource scheduling method provided by this solution, when determining the number of candidate replicas based on the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period in the application resource profile, the set scaling factor, and the second computing power resource requirement value, includes:

[0099] S2311: Round up the first computing resource requirement value to obtain a second computing resource requirement value.

[0100] S2312: The number of candidate copies is determined by rounding up the product of the percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource portrait and the set scaling factor, and the ratio of the second computing power resource demand value.

[0101] Exemplarily, after determining the computing resource utilization rate, the first computing resource demand value is rounded up to obtain the second computing resource demand value. For example, the second computing resource demand value can be expressed as: desireRequestCpu = ceil(requestCpu), where ceil() is a round-up operation and requestCpu is the first computing resource demand value.

[0102] In one possible embodiment, the resource scheduling method provided by this solution further includes, after rounding up the first computing power resource demand value to obtain the second computing power resource demand value: when the application is a set type application, using the first set proportional coefficient to update the second computing power resource demand value.

[0103] The set type application provided by this solution can be a NodeJs application, where a single process of the NodeJs application can run on at most one core. The first set proportional coefficient can be set according to the computing power resource margin reserved by the NodeJs application. For example, the first set proportional coefficient can be set to 0.7.

[0104] Exemplarily, after determining the second computing power resource requirement value, if the current application is a non-set type application, the second computing power resource requirement value can be directly used to determine the number of candidate copies. If the application is a set type application, the first set proportional coefficient is used to update the second computing power resource requirement value. For example, the product of the second computing power resource requirement value and the first set proportional coefficient is used as the latest second computing power resource requirement value, and the latest second computing power resource requirement value is used to determine the number of candidate copies. For example, for a set type application, updating the second computing power resource requirement value using the first set proportional coefficient can be expressed as: desireRequestCpu*=a, where a is the first set proportional coefficient. This solution updates the second computing power resource requirement value using the first set proportional coefficient when the application is a set type application, taking into account the computing power resource margin reserved when the application is a set type application, and more accurately determines the second computing power resource requirement value, thereby effectively improving the accuracy of determining the number of candidate copies.

[0105] In one embodiment, after determining the second computing resource requirement value, the number of candidate replicas is determined by rounding up the product of the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period in the application resource profile and the set scaling factor to the ratio of the second computing resource requirement value. For example, the number of candidate replicas can be expressed as:

[0106] newReplicas=ceil(peakTimeSumMax*ratio / desireRequestCpu)

[0107] Among them, peakTimeSumMax is the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period, ratio is the set scaling factor, for example, ratio = 1, and desireRequestCpu is the second computing power resource demand value. This solution obtains the second computing power resource demand value by rounding up the first computing power resource demand value, and rounds up the product of the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period in the application resource profile and the set scaling factor to the ratio of the second computing power resource demand value to accurately determine the number of candidate replicas, effectively improving the accuracy of resource deployment.

[0108] In one possible embodiment, the resource scheduling method provided by this solution, when determining the recommended number of replicas based on the number of candidate replicas, includes:

[0109] S2313: When the ratio of the maximum percentage peak value to the product of the set maximum value of replica resources and the number of candidate replicas is greater than the recommended usage rate, increase the number of candidate replicas.

[0110] S2314: When the ratio of the maximum percentage peak value to the product of the set maximum value of replica resources and the number of candidate replicas is less than or equal to the recommended usage rate, the recommended number of replicas is determined as the recommended number of replicas.

[0111] For example, after determining the number of candidate replicas, the ratio of the percentage peak maximum value to the product of the set maximum value of replica resources and the number of candidate replicas is determined, and this ratio is compared with the set recommended utilization rate. For example, a determination is made as to whether (peakTimeSumMax / (limitsCpu*newReplicas))>recommendCpuRate holds. If so, the ratio of the percentage peak maximum value to the product of the set maximum value of replica resources and the number of candidate replicas is determined to be greater than the recommended utilization rate. Where recommendedCpuRate is the recommended utilization rate, for example, recommendedCpuRate may be 60%.

[0112] In one embodiment, when it is determined that the ratio of the percentage peak maximum value to the product of the set maximum value of replica resources and the number of candidate replicas is greater than the recommended usage rate, the current number of candidate replicas is increased (for example, the current number of candidate replicas is increased by one, that is, newReplicas+=1), and the ratio of the percentage peak maximum value to the product of the set maximum value of replica resources and the latest number of candidate replicas is continued to be determined based on the increased number of candidate replicas. Whether it is greater than the recommended usage rate, until the ratio of the percentage peak maximum value to the product of the set maximum value of replica resources and the number of candidate replicas is less than or equal to the recommended usage rate.

[0113] In one embodiment, when it is determined that the ratio of the maximum percentage peak value to the product of the set maximum number of replica resources and the number of candidate replicas is less than or equal to the recommended usage rate, the current recommended number of replicas is determined as the recommended number of replicas. The current recommended number of replicas may be the initial recommended number of replicas (i.e., the ratio of the maximum percentage peak value to the product of the set maximum number of replica resources and the initially determined number of candidate replicas is less than or equal to the recommended usage rate) or an updated recommended number of replicas.

[0114] This solution increases the number of candidate replicas when the ratio of the maximum percentage peak value to the product of the set maximum value of replica resources and the number of candidate replicas is greater than the recommended utilization rate, until the ratio of the maximum percentage peak value to the product of the set maximum value of replica resources and the number of candidate replicas is less than or equal to the recommended utilization rate, and determines the recommended number of replicas as the recommended number of replicas, thereby improving the accuracy of determining the recommended number of replicas and effectively improving the accuracy of resource deployment.

[0115] In one possible embodiment, the resource scheduling method provided by this solution may further constrain the recommended number of replicas based on a set minimum recommended replica threshold after determining the recommended number of replicas as the recommended number of replicas. For example, after determining the recommended number of replicas, the set minimum recommended replica threshold is used to constrain the minimum value of the recommended number of replicas. That is, when the determined recommended number of replicas is less than the minimum recommended replica threshold, the recommended number of replicas is set to the minimum recommended replica threshold. By constraining the recommended number of replicas using the minimum recommended replica threshold, this solution provides a safety net for the recommended number of replicas, thereby effectively improving resource deployment accuracy.

[0116] In one possible embodiment, the resource scheduling method provided by this solution determines the recommended number of copies based on the computing power resource utilization rate and the recommended utilization rate, including: when the computing power resource utilization rate is greater than or equal to the product of the set recommended utilization rate and the second set proportional coefficient, the recommended number of copies is determined by rounding up the ratio of the percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource portrait to the product of the set maximum value of the copy resource and the recommended utilization rate.

[0117] For example, when the computing power resource utilization rate is greater than or equal to the set recommended utilization rate, it is further determined that the computing power resource utilization rate is greater than or equal to the product of the set recommended utilization rate and a second set proportional coefficient (the second set proportional coefficient is greater than or equal to 1, for example, the second set proportional coefficient can be 1.1).

[0118] In one embodiment, if the computing power resource utilization rate is less than the product of the set recommended utilization rate and the second set proportional coefficient (that is, the computing power resource utilization rate is between one times and a multiple of the second set proportional coefficient of the recommended utilization rate), the current resource deployment strategy for the application can be considered reasonable, the current resource deployment strategy for the application can be maintained, and the resource scheduling of the current application can be ended, and the resource scheduling processing of the next application can be carried out.

[0119] In one embodiment, when the computing power resource utilization rate is greater than or equal to the product of the set recommended utilization rate and the second set proportional coefficient, the ratio of the percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource portrait to the product of the set maximum value of the replica resource and the recommended utilization rate is rounded up to determine the recommended number of replicas.

[0120] For example, the recommended number of replicas can be expressed as: recommendReplicas = ceil(peakTimeSumMax / (limitsCpu*recommendCpuRate)), where peakTimeSumMax is the percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource profile, limitsCpu is the maximum value of the replica resource, and recommendCpuRate is the recommended utilization rate. This solution determines the recommended number of replicas by rounding up the ratio of the percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource profile to the product of the set maximum value of the replica resource and the recommended utilization rate when the computing power resource utilization rate is greater than or equal to the product of the set recommended utilization rate and the second set proportional coefficient, effectively improving resource deployment accuracy.

[0121] In one possible embodiment, the resource scheduling method provided by this solution can also constrain the recommended number of copies based on a set minimum recommended copy threshold after rounding up the ratio of the maximum percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource portrait to the product of the set maximum value of the copy resource and the recommended usage rate to determine the recommended number of copies. Exemplarily, after determining the recommended number of copies, the set minimum recommended copy threshold is used to constrain the minimum value of the recommended number of copies, that is, when the determined recommended number of copies is less than the minimum recommended copy threshold, the recommended number of copies is set to the minimum recommended copy threshold. This solution uses the minimum recommended copy threshold to constrain the recommended number of copies, provides a safety net for the recommended number of copies, ensures that sufficient copies are deployed, and effectively improves the accuracy of resource deployment.

[0122] S240: Deploy a first number of replicas to a fixed node pool and a second number of replicas to an elastic node pool according to the recommended number of replicas. The sum of the first number and the second number is equal to the recommended number of replicas, and obtain a resource deployment strategy.

[0123] Exemplarily, after determining the recommended number of replicas, a first number and a second number are determined based on the recommended number of replicas, where the sum of the first number and the second number equals the recommended number of replicas. Furthermore, the first number of replicas is deployed to a fixed node pool and the second number of replicas is deployed to an elastic node pool to obtain a resource deployment strategy. Accordingly, when deploying replicas for an application on different container nodes based on the resource deployment strategy, the first number of replicas is configured for the application in the fixed node pool, and the second number of replicas is configured for the elastic node pool.

[0124] Optionally, the first and second quantities are determined based on the recommended number of replicas. This can be done based on a pre-set correspondence between different combinations of the recommended number of replicas and the first and second quantities. Alternatively, a first proportional coefficient corresponding to the first quantity can be pre-set, the product of the recommended number of replicas and the first proportional coefficient can be rounded up to obtain the first quantity, and the second quantity can be determined based on the difference between the recommended number of replicas and the first quantity. For example, when the recommended number of replicas is 16, the first and second quantities can be determined as 10 and 6, respectively, indicating that 10 replicas are configured in the fixed node pool and 6 replicas are configured in the elastic node pool.

[0125] This solution deploys some application replicas to a fixed node pool and other replicas to an elastic node pool, thereby shaving the peak load on the fixed node pool. This transforms the periodic resource usage characteristic curve into a smoother straight line, further improving the node packing rate and the actual computing resource utilization rate of the node.

[0126] S250: Deploy replicas for the application in different container nodes based on the resource deployment strategy.

[0127] In the above, by determining an application resource profile and the periodicity of application resource usage based on application resource usage statistics, determining a first computing resource requirement based on the application resource profile and the periodicity, determining a recommended number of replicas based on the application resource profile and the first computing resource requirement, determining a resource deployment strategy based on the recommended number of replicas, and deploying application replicas on different container nodes based on the resource deployment strategy, the computing resource requirement is automatically determined. This computing resource requirement is closer to the actual value of replica resource usage, and a larger number of replicas can be deployed on the same container node, effectively improving the container node packing rate and computing resource utilization rate. Simultaneously, some application replicas are deployed to a fixed node pool, while others are deployed to an elastic node pool, achieving peak load shaving for the fixed node pool, transforming the periodic resource usage characteristic curve into a smoother straight line, further improving the node packing rate and the actual computing resource utilization rate of the node. Furthermore, the system enables the scheduling of application replicas from being limited to a single node pool to being able to be scheduled to multiple node pools, making resource scheduling more flexible and efficient. An application can be served by multiple replicas, improving the application's high availability and scalability.

[0128] The embodiment of the present application also provides a resource scheduling device, Figure 4 This is a schematic diagram of the structure of a resource scheduling device provided by an embodiment of the present application. Figure 4 The resource scheduling device includes: an input device 43, an output device 44, a memory 42, and one or more processors 41; the memory 42 is used to store one or more programs; when the one or more programs are executed by the one or more processors 41, the one or more processors 41 implement the resource scheduling method provided in the above embodiment. The input device 43, the output device 44, the memory 42, and the processor 41 can be connected by a bus or other means. Figure 4 The bus connection is taken as an example.

[0129] The memory 42 is a computing device readable storage medium that can be used to store software programs, computer executable programs and modules, such as the program instructions / modules corresponding to the resource scheduling method provided in any embodiment of the present application. The memory 42 may mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system and application programs required for at least one function; the data storage area can store data created according to the use of the device, etc. In addition, the memory 42 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 42 may further include a memory remotely located relative to the processor 41, and these remote memories can be connected to the device via a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network and a combination thereof.

[0130] The input device 43 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the device. The output device 44 may include a display device such as a display screen.

[0131] The processor 41 executes various functional applications and data processing of the device by running software programs, instructions and modules stored in the memory 42, that is, implements the above-mentioned resource scheduling method.

[0132] The resource scheduling device and computer provided above can be used to execute the resource scheduling method provided in any of the above embodiments, and have corresponding functions and beneficial effects.

[0133] An embodiment of the present application also provides a storage medium for storing computer-executable instructions, which, when executed by a computer processor, are used to execute a resource scheduling method as provided in the above embodiment, the resource scheduling method comprising: obtaining resource usage statistics of an application; determining an application resource profile and the periodicity of application resource usage based on the application resource usage statistics; determining a first computing power resource demand value based on the application resource profile and the periodicity; determining a recommended number of replicas based on the application resource profile and the first computing power resource demand value; determining a resource deployment strategy based on the recommended number of replicas, and deploying replicas for the application in different container nodes based on the resource deployment strategy.

[0134] Storage medium - any of various types of memory devices or storage devices. The term "storage medium" is intended to include: installation media, such as CD-ROMs, floppy disks, or tape drives; computer system memory or random access memory, such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media (such as hard disks or optical storage); registers or other similar types of memory elements, etc. Storage media may also include other types of memory or combinations thereof. In addition, the storage medium may be located in the first computer system in which the program is executed, or it may be located in a different second computer system that is connected to the first computer system via a network (such as the Internet). The second computer system can provide program instructions to the first computer for execution. The term "storage medium" may include two or more storage media that can reside in different locations (e.g., in different computer systems connected via a network). The storage medium can store program instructions (e.g., embodied as a computer program) that can be executed by one or more processors.

[0135] Of course, the computer-executable instructions of the storage medium provided in the embodiment of the present application are not limited to the resource scheduling method provided above, and can also execute related operations in the resource scheduling method provided in any embodiment of the present application.

[0136] The resource scheduling device and storage medium provided in the above embodiments can execute the resource scheduling method provided in any embodiment of the present application. For technical details not described in detail in the above embodiments, please refer to the resource scheduling method provided in any embodiment of the present application.

[0137] The above are only preferred embodiments of the present application and the technical principles employed. The present application is not limited to the specific embodiments provided herein, and various obvious changes, readjustments, and substitutions that can be made by those skilled in the art will not depart from the scope of protection of the present application. Therefore, although the present application has been described in detail through the above embodiments, the present application is not limited to the above embodiments. Without departing from the concept of the present application, it may also include more other equivalent embodiments, and the scope of the present application is determined by the scope of the claims.

Claims

1. A resource scheduling method, characterized in that: include: Get the application's resource usage statistics; Determine the application resource profile and the periodicity of application resource usage based on the resource usage statistics of the application; Determine a first computing resource requirement value based on the application resource profile and the periodicity; Determine the recommended number of replicas based on the application resource profile and the first computing resource requirement value; A resource deployment strategy is determined according to the recommended number of replicas, and replicas are deployed for the application in different container nodes based on the resource deployment strategy.

2. The resource scheduling method according to claim 1, characterized in that: The determining the first computing resource requirement value according to the application resource profile and the periodicity includes: In a case where the periodicity is that the application resource usage has periodicity, the first computing resource demand value is set to an average value of a plurality of first peak setting percentage lines within a first set time period in the application resource profile; When the periodicity is that the application resource usage does not have periodicity, the first computing power resource demand value is set to the average value of multiple peak average values ​​within the first set time period in the application resource portrait.

3. The resource scheduling method according to claim 1, characterized in that: The determining the recommended number of replicas according to the application resource profile and the first computing resource requirement value includes: When the first computing resource demand value is less than the set demand threshold, the computing resource utilization rate is determined based on the application resource profile and the first computing resource demand value, and the recommended number of copies is determined based on the computing resource utilization rate and the recommended utilization rate.

4. The resource scheduling method according to claim 3, characterized in that: The determining the computing resource usage rate according to the application resource profile and the first computing resource demand value includes: The computing power resource utilization rate is determined based on the ratio of the percentage peak maximum value of multiple second peak setting percentage lines within the second set time period in the application resource portrait to the total amount of application resources.

5. The resource scheduling method according to claim 3, characterized in that: The determining the recommended number of replicas according to the computing resource utilization rate and the recommended utilization rate includes: When the computing resource usage rate is less than the set recommended usage rate, determining the number of candidate replicas according to the percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource profile, the set scaling factor, and the second computing resource requirement value, and determining the recommended number of replicas according to the candidate number of replicas, where the second computing resource requirement value is obtained according to the first computing resource requirement value; When the computing resource utilization rate is greater than or equal to the set recommended utilization rate, the recommended number of replicas is determined according to the computing resource utilization rate and the recommended utilization rate.

6. The resource scheduling method according to claim 5, characterized in that: The determining of the number of candidate replicas according to the percentage peak maximum value of the plurality of second peaks within the second set time period in the application resource profile, setting the scaling factor and the second computing resource requirement value includes: Rounding up the first computing resource requirement value to obtain a second computing resource requirement value; The number of candidate copies is determined by rounding up the product of the percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource portrait and the set scaling factor, and the ratio of the second computing power resource demand value.

7. The resource scheduling method according to claim 6, characterized in that: After rounding up the first computing resource requirement value to obtain the second computing resource requirement value, the method further includes: In a case where the application is a set type application, the second computing resource requirement value is updated using a first set proportional coefficient.

8. The resource scheduling method according to claim 5, characterized in that: Determining the recommended number of replicas according to the number of candidate replicas includes: In the case where the ratio of the maximum percentage peak value to the product of the set maximum value of replica resources and the number of candidate replicas is greater than the recommended usage rate, increasing the number of candidate replicas; When the ratio of the maximum percentage peak value to the product of the set maximum value of replica resources and the number of candidate replicas is less than or equal to the recommended usage rate, the recommended number of replicas is determined as the recommended number of replicas.

9. The resource scheduling method according to claim 8, characterized in that: After determining the recommended number of copies as the recommended number of copies, the method further includes: The recommended number of copies is constrained based on a set minimum recommended copy threshold.

10. The resource scheduling method according to claim 5, characterized in that: The determining the recommended number of replicas according to the computing resource utilization rate and the recommended utilization rate includes: When the computing power resource utilization rate is greater than or equal to the product of the set recommended utilization rate and the second set proportional coefficient, the recommended number of copies is determined by rounding up the ratio of the percentage peak value of multiple second peak setting percentage lines within the second set time period in the application resource portrait to the product of the set maximum value of the copy resource and the recommended utilization rate.

11. The resource scheduling method according to claim 10, characterized in that: After determining the recommended number of replicas by rounding up the ratio of the maximum percentage peak value of the plurality of second peak setting percentage lines within the second set time period in the application resource profile to the product of the set maximum replica resource value and the recommended usage rate, the method further includes: The recommended number of copies is constrained based on a set minimum recommended copy threshold.

12. The resource scheduling method according to claim 1, characterized in that: Determining a resource deployment strategy according to the recommended number of replicas includes: According to the recommended number of replicas, a first number of replicas is deployed to a fixed node pool, and a second number of replicas is deployed to an elastic node pool, where the sum of the first number and the second number is equal to the recommended number of replicas.

13. A resource scheduling device, characterized in that: include: memory and one or more processors; The memory is used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the resource scheduling method according to any one of claims 1 to 12.

14. A storage medium storing computer executable instructions, characterized in that: When the computer executable instructions are executed by a computer processor, they are used to execute the resource scheduling method according to any one of claims 1 to 12.

Images