Computing resource scheduling method and device, electronic equipment and storage medium

Abstract

Embodiments of the present application disclose a kind of computing resource scheduling method, device and electronic equipment and storage medium, based on first condition and second condition, the expected period number of each computing resource used in application scenario is determined;Wherein, the first condition is: the use rate of residual available object when the each computing resource is according to the expected period number runs is greater than target use rate;The second condition is: the total performance of the each computing resource when the each computing resource is according to the expected period number runs meets threshold value;The performance of any computing resource is positively correlated with the computing capacity of the any computing resource;According to the expected period number of the each computing resource, the each computing resource is scheduled in residual available time period.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cloud computing, and more particularly to a computing resource scheduling method and device, an electronic device and a storage medium. BACKGROUND

[0002] Cloud computing is a pay-per-use model that provides available, convenient, on-demand network access to a shared pool of configurable resources (including networks, servers, storage, applications, and services) that can be rapidly provisioned with minimal management effort or interaction with a service provider.

[0003] However, at present, the scheduling of computing resources for users can only be performed according to the computing resources ordered by the users, that is, as many computing resources as ordered by the users are scheduled for the users, and the intelligence of the scheduling of computing resources is poor. SUMMARY

[0004] The present application aims to provide a computing resource scheduling method and device, an electronic device and a storage medium, comprising the following technical solutions:

[0005] A computing resource scheduling method, the method comprising:

[0006] Based on a first condition and a second condition, the expected number of periods of each computing resource used in an application scenario is determined; wherein the first condition is that the usage rate of the remaining available objects when the each computing resource runs according to the expected number of periods is greater than the target usage rate; the second condition is that the total performance of the each computing resource when the each computing resource runs according to the expected number of periods meets a threshold value; the performance of any computing resource is positively correlated with the computing power of the any computing resource;

[0007] According to the expected number of periods of the each computing resource, the each computing resource is scheduled within the remaining available time period.

[0008] The above method, preferably, the scheduling of the each computing resource according to the expected number of periods of the each computing resource within the remaining available time period comprises:

[0009] The total performance of the each computing resource running according to the expected number of periods is obtained;

[0010] The target performance of a single period within the remaining available time period is determined according to the total performance;

[0011] The performance of the each computing resource running within the remaining available time period is close to the target performance of a single period, and the each computing resource is allocated a period corresponding to the expected number of periods within the remaining available time period.

[0012] The method, preferably, the target performance of a single period within the remaining available time period is determined according to the total performance, comprising:

[0013] The ratio of the total performance to the number of periods contained in the remaining available time period is determined as the target performance of a single period within the remaining available time period.

[0014] The method, preferably, the performance of a single period when each computing resource runs within the remaining available time period approaches the target performance as a goal, and the corresponding expected number of periods of periods are allocated to each computing resource within the remaining available time period, comprising:

[0015] The effective running time period of each computing resource within the remaining available time period is determined;

[0016] The performance of a single period when each computing resource runs within the remaining available time period approaches the target performance as a goal, and the corresponding expected number of periods of periods are allocated to each computing resource within the effective running time period of each computing resource.

[0017] The method, preferably, the performance of each computing resource under the application scenario is obtained by statistical analysis of the time consumed by each computing resource in processing the same sample data under the application scenario;

[0018] Wherein, the performance of any computing resource is negatively related to the time consumed by the any computing resource in processing the sample data.

[0019] The method, preferably, the total performance of each computing resource when each computing resource runs according to the expected number of periods is:

[0020] The weighted sum of the performance of each computing resource;

[0021] Wherein, the weight of the performance of any computing resource is the expected number of periods of the any computing resource.

[0022] The method, preferably, further comprising:

[0023] Displaying an interactive interface;

[0024] Receiving the input target available time period and the target number of available objects through the interactive interface.

[0025] A computing resource scheduling device, the device comprising:

[0026] determining module, configured to determine an expected period number of each computing resource used in the application scenario based on a first condition and a second condition, wherein the first condition is that a usage rate of a remaining available object when the each computing resource runs according to the expected period number is greater than a target usage rate, and the second condition is that a total performance of the each computing resource when the each computing resource runs according to the expected period number meets a threshold; and a performance of any one computing resource is positively correlated with a computing capability of the any one computing resource;

[0027] scheduling module, configured to schedule the each computing resource in a remaining available time period according to the expected period number of the each computing resource.

[0028] An electronic device comprises:

[0029] a memory, configured to store a program;

[0030] a processor, configured to invoke and execute the program in the memory, and realize each step of the computing resource scheduling method according to the program.

[0031] A readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize each step of the computing resource scheduling method.

[0032] According to the above scheme, the computing resource scheduling method, device, electronic device and storage medium provided by the present application are provided, the expected period number of each computing resource used in the application scenario is determined according to the first condition and the second condition, and the each computing resource is scheduled in the remaining available time period according to the expected period number of the each computing resource. Since the expected period number of the computing resource is calculated by taking the usage rate of the remaining available object when the each computing resource runs according to the expected period number is greater than the target usage rate and the total performance meets the threshold as the target, the each computing resource is scheduled in the remaining available time period according to the expected period number of the each computing resource, the purpose of dynamically adjusting the computing resource based on the remaining available object and the remaining available time period is realized, the usage rate of the available object at the end of the available time period is ensured to be large, the performance of the scheduled computing resource is ensured to be optimal, and the intelligence of the computing resource scheduling is improved. BRIEF DESCRIPTION OF DRAWINGS

[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required by the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0034] Figure 1An implementation flowchart of the computing resource scheduling method provided by the embodiment of the present application;

[0035] Figure 2 An implementation flowchart of scheduling each computing resource in the remaining available time period according to the expected cycle number of each computing resource provided by the embodiment of the present application;

[0036] Figure 3 Another implementation flowchart of scheduling each computing resource in the remaining available time period according to the expected cycle number of each computing resource provided by the embodiment of the present application;

[0037] Figure 4 An effect example diagram of scheduling each computing resource in the remaining available time period according to the expected cycle number of each computing resource provided by the embodiment of the present application;

[0038] Figure 5 An implementation structure diagram of the computing resource scheduling device provided by the embodiment of the present application;

[0039] Figure 6 An implementation structure diagram of the electronic device provided by the embodiment of the present application.

[0040] The terms "first", "second", "third", "fourth" and the like (if any) in the description and claims and the above drawings are used to distinguish similar parts, and do not necessarily have to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated herein. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0042] The computing resource scheduling method provided by the embodiments of the present application can be used in a cloud container platform. The cloud container platform can provide a plurality of virtual computing resources (hereinafter referred to as computing resources), which can include at least part of the following computing resources, but are not limited to the following computing resources:

[0043] CPU (Central Processing Unit / Processor, Central Processor);

[0044] GPU (Graphics Processing Unit);

[0045] TPU (Tensor Processing Unit);

[0046] NPU (Neural Network Processing Unit);

[0047] DPU (Deep learning Processing Unit);

[0048] APU (Accelerated Processing Unit);

[0049] FPU (Floating Processing Unit);

[0050] HPU (Holographics Processing Unit);

[0051] IPU (Intelligence Processing Unit, AI Processor)

[0052] VPU (Vector Processing Unit), etc.

[0053] Among them, computing resources of the same type can be further subdivided according to their performance. For example, CPUs can be divided into low-frequency CPUs and high-frequency CPUs.

[0054] like Figure 1 The diagram shown is a flowchart of one implementation of the computing resource scheduling method provided in this application, which may include:

[0055] Step S101: Based on the first and second conditions, determine the expected number of cycles for each computing resource used in the application scenario; wherein,

[0056] The first condition is that the utilization rate of remaining available objects by each computing resource running for the expected number of cycles is greater than the target utilization rate. In other words, the first condition requires that the available objects used by each computing resource running for the expected number of cycles be as close as possible to the remaining available objects. Available objects refer to the objects consumed by the computing resource operation. As an example, the remaining available objects can be the remaining budget.

[0057] The second condition is that the total performance of the computing resources when each of the computing resources is running for the expected number of cycles satisfies a threshold, and the performance of any one of the computing resources is positively correlated with the computing capability of any one of the computing resources. That is, the stronger the computing capability of a computing resource, the better the performance of the computing resource. As an example, the second condition requires that, on the basis of satisfying the first condition, the total performance of the computing resources when each of the computing resources is running for the expected number of cycles is optimal.

[0058] As an example, the total performance of the computing resources when each of the computing resources is running for the expected number of cycles can be a weighted sum of the performance of each of the computing resources, and the weight of the performance of each computing resource is the expected number of cycles of the computing resource.

[0059] The application scenario refers to an application scenario to which an application program or a service currently using the cloud computing resource belongs. The application scenario can include, but is not limited to, any one of the following: a non-computation-intensive scenario (for example, a high-concurrency scenario), and a computer vision type scenario. The computing resources corresponding to different application scenarios can be different. For example, in a computer vision (CV) type scenario, the corresponding computing resources can include, but are not limited to: a TPU, a GPU, a high-frequency CPU, and a low-frequency CPU. In a high-concurrency scenario, the computing resources can only include a high-frequency CPU and a low-frequency CPU.

[0060] As an example, one day can be taken as a cycle, and of course, a cycle can also be other lengths of time, such as, two days as a cycle, and the like.

[0061] Step S102: scheduling each of the computing resources according to the expected number of cycles of each of the computing resources in the remaining available time period.

[0062] The remaining available time period is determined according to the initial available time period and the used time period. Apparently, the remaining available time period refers to the time period in the initial available time period that has not been used. For example, the initial available time period is from January 1, 2021 to December 31, 2021, and the used time period is from January 1, 2021 to June 30, 2021, and then the remaining available time period is from July 1, 2021 to December 31, 2021.

[0063] When each of the computing resources is scheduled in the remaining available time period, the number of cycles occupied by each of the computing resources is the expected number of cycles determined in the foregoing. For any one of the computing resources (denoted as computing resource A), the computing resource A can run in each cycle in the remaining available time period, or can only run in part of the cycles in the remaining available time period, as long as the sum of the running time lengths of the computing resource A in each cycle is the time length corresponding to the expected number of cycles of the computing resource.

[0064] Optionally, when scheduling each computing resource in the remaining available time period, each computing resource can be scheduled based on the total performance of the computing resources scheduled in each period being the same or substantially the same.

[0065] The computing resource scheduling method provided by the embodiments of the present application determines the expected period number of each computing resource used in the application scenario according to the first condition and the second condition, and schedules each computing resource in the remaining available time period according to the expected period number of each computing resource. Since the expected period number of each computing resource is calculated by taking the usage rate of the remaining available object by each computing resource running in the expected period number being greater than the target usage rate and the total performance meeting the threshold as the target, the scheduling of each computing resource in the remaining available time period according to the expected period number of each computing resource achieves the purpose of dynamically adjusting the computing resource based on the remaining available object and the remaining available time period, ensures that the usage rate of the available object at the end of the available time period is relatively high, and the performance of the scheduled computing resource is relatively optimal, and improves the intelligence of the computing resource scheduling.

[0066] In an optional embodiment, an implementation flowchart of the above-mentioned scheduling each computing resource in the remaining available time period according to the expected period number of each computing resource can include the following steps as shown in Figure 2

[0067] Step S201: obtaining the total performance of each computing resource running in the expected period number.

[0068] As an example, the total performance of each computing resource running in the expected period number can be the weighted sum of the performance of each computing resource, and the weight of the performance of each computing resource is the expected period number of the computing resource.

[0069] Step S202: determining the target performance of a single period in the remaining available time period according to the total performance.

[0070] Optionally, the ratio of the total performance to the number of periods contained in the remaining available time period can be determined as the target performance of a single period in the remaining available time period.

[0071] Step S203: allocating the corresponding period of the expected period number to each computing resource in the remaining available time period, taking the performance of the computing resource in a single period running in the remaining available time period approaching the target performance as the target.

[0072] That is, when scheduling each computing resource in the remaining available time period, the total performance of the computing resource in a single period running in the remaining available time period is made to approach the target performance.

[0073] ​Optionally, for the i th (i = 1, 2, 3, …, N) period in the remaining available time period, a performance weighted sum of each computing resource in the i th period is obtained, wherein the weight x ij of the j th (j = 1, 2, 3, …, M) computing resource corresponding to the i th period is the running time length of the j th computing resource in the i th period.

[0074] Based on the sum of the running time length of the j th computing resource in each period being equal to the time length corresponding to the expected period number of the j th computing resource, and the absolute value of the difference between the performance weighted sum of each computing resource in the i th period and the target performance being less than a threshold, the running time length x ij of the j th computing resource in the i th period is determined.

[0075] N is the number of periods in the remaining available time period, and M is the number of computing resources.

[0076] In some cases, the computing resources can be set with an effective running time period, for example, some resources can only be used from 23:00 to 6:00, some can only be used on weekdays, and the like. Therefore, when scheduling each computing resource in the remaining available time period, the effective running time period of the computing resource needs to be considered. Based on this, another implementation flowchart provided by the embodiment of the application for scheduling each computing resource in the remaining available time period according to the expected period number of each computing resource is as shown in Figure 3 , which can include:

[0077] Step S301: determining the effective running time period of each computing resource in the remaining available time period.

[0078] Optionally, the effective running time period of each computing resource in the remaining available time period can be determined according to a preset correspondence between the computing resource and the effective running time period.

[0079] Step S302: allocating a period corresponding to an expected period number to each computing resource in the effective running time period of each computing resource, with the performance of each computing resource in a single period in the remaining available time period tending to the target performance as the target.

[0080] When scheduling each computing resource in the remaining available time period, each computing resource can only run in the corresponding effective running time period.

[0081] As an example, the running time length of the jth computing resource in the ith period can be determined based on that the sum of the running time lengths of the jth computing resource in the periods is equal to the time length corresponding to the expected number of periods of the jth computing resource, the absolute value of the difference between the performance weighted sum of the computing resources in the ith period and the target performance is less than a threshold, and the running time length of the jth computing resource in the ith period is less than or equal to the time length of the effective running time period of the jth computing resource in the ith period.

[0082] In an optional embodiment, the performance of each computing resource in an application scenario is obtained by statistically analyzing the time consumed by each computing resource for processing the same sample data in the application scenario. The sample data can be small sample data.

[0083] The performance of any computing resource is negatively correlated with the time consumed by the computing resource for processing the sample data, i.e., the longer the time consumed for processing the sample data, the lower the performance.

[0084] As an example, the performance of a computing resource can be the time efficiency of the computing resource. The higher the time efficiency, the higher the performance of the computing resource, and the lower the time efficiency, the lower the performance of the computing resource. Therefore, the time efficiency of any computing resource is positively correlated with the computing power of the computing resource, i.e., the stronger the computing power of the computing resource, the higher the time efficiency.

[0085] The time efficiency of a computing resource can be obtained by statistically analyzing the time consumed by each computing resource for processing the same sample data in an application scenario.

[0086] For example, in a certain CV scenario, TPU, GPU, high-frequency CPU and low-frequency CPU can be used, where the time consumed by TPU, GPU, high-frequency CPU and low-frequency CPU for processing the same sample data in the certain CV scenario and the corresponding time efficiency are shown in Table 1:

[0087] Table 1

[0088] Computing resource type Time (small sample consumption) Time efficiency TPU 2 15 GPU 5 6 High frequency CPU 10 3 Low frequency CPU 30 1

[0089] Obviously, the low-frequency CPU consumes the longest time for processing the small sample data and has the lowest time efficiency, and the TPU consumes the shortest time for processing the same small sample data and has the highest time efficiency.

[0090] For example, in a non-computation-intensive scenario, such as a high-concurrency scenario, only high-frequency CPU and low-frequency CPU can be used, where the time consumed by the high-frequency CPU and the low-frequency CPU for processing the same sample data and the corresponding time efficiency are shown in Table 2:

[0091] Table 2

[0092] Computing resource type Time (small sample consumption) Time efficiency High frequency CPU 1 3 Low frequency CPU 3 1

[0093] Obviously, the low-frequency CPU consumes more time in processing small sample data than the high-frequency CPU, so the time efficiency of the low-frequency CPU is lower than that of the high-frequency CPU.

[0094] In an optional embodiment, the information processing method provided by the embodiment of the application can further include:

[0095] displaying an interactive interface;

[0096] receiving an input target available time period and a target number of available objects through the interactive interface.

[0097] After obtaining the target available time period and the target number of available objects, the scheduling of the computing resources can be performed based on the computing resource allocation method provided by the application. As can be seen, based on the application, the user only needs to input the use time period of the computing resources and the target number of available objects according to the needs, so as to realize the computing resource scheduling method with optimal performance adapted to the input target available time period and target number of available objects, thereby reducing the difficulty of selecting computing resources for the user.

[0098] For example, the target available time period represents the use duration of the computing resources, and the target number of available objects can refer to the cost (or budget) of purchasing the computing resources.

[0099] Next, taking the performance of the computing resources as the time efficiency of the computing resources as an example, an implementation of determining the expected cycle number of each computing resource used in the application scenario based on the first condition and the second condition will be described below in combination with Table 1.

[0100] Please refer to Table 2. The expected cycle numbers of the computing resources shown in Table 1 are as follows: the expected cycle number of the TPU is denoted as X1, the expected cycle number of the GPU is denoted as X2, the expected cycle number of the high-frequency CPU is denoted as X3, and the expected cycle number of the low-frequency CPU is denoted as X4. The use rates of the TPU, GPU, high-frequency CPU, and low-frequency CPU on the available objects (if the available object is the cost, the use rate of the available object is the price of a single cycle) are 20, 10, 5, and 2, respectively.

[0101] Table 3

[0102]

[0103]

[0104] The single-cycle price in Table 3 refers to the single-cycle price of 1 core computing resource. For example, the single-cycle price of 1 core TPU is 20, and the single-cycle price of 1 core GPU is 10.

[0105] Optionally, the used fee can be determined according to the usage parameter of the computing resource (including which computing resource is used, the used time length, etc.), the total fee is subtracted by the used fee, and the current remaining fee is obtained, denoted as Y.

[0106] It is assumed that there are N periods in the current remaining available time period.

[0107] The fee y used by each computing resource shown in Table 3 according to the corresponding expected period number is represented by the formula:

[0108] y = 20 * X1 + 10 * X2 + 5 * X3 + 2 * X4

[0109] The total time efficiency e of each computing resource shown in Table 3 according to the corresponding expected period number is represented by the formula:

[0110] e = 15 * X1 + 6 * X2 + 3 * X3 + 1 * X4

[0111] Based on y equal to Y and e taking the maximum value, the values of X1, X2, X3 and X4 are obtained. Alternatively, based on the difference between Y and y being less than a threshold value and e taking the maximum value, the values of X1, X2, X3 and X4 are obtained.

[0112] The maximum value of e is divided by N to obtain the target time efficiency (denoted as eo) of each period in the remaining available time period.

[0113] For the i-th (i = 1, 2, 3, …, N) period in the remaining available time period, the time efficiency weighted sum (denoted as ei) of each computing resource in the i-th period is obtained:

[0114] ei = 15 * xi1 + 6 * xi2 + 3 * xi3 + 1 * xi4

[0115] Wherein, xij (j = 1, 2, 3, 4) represents the running time length of the j-th computing resource in the i-th period.

[0116] The sum of the running time length of the j-th computing resource in each period (denoted as Tj) is:

[0117]

[0118] Without considering the effective running time period of the computing resource, the value of xij can be determined based on Tj = Xj and ei = eo, or based on Tj = Xj and the absolute value of the difference between ei and eo being less than a threshold value.

[0119] In consideration of the effective running time period of the computing resource, assuming that the length of the effective running time period of the jth computing resource in the ith cycle is tij, the value of xij can be determined based on Tj=Xj, xij being less than or equal to tij, and ei=eo, or based on Tj=Xj, xij being less than or equal to tij, and the absolute value of the difference between ei and eo being less than a threshold.

[0120] As shown in Table 3, the single-cycle price of each computing resource is calculated according to 1 core. In actual applications, a computing resource can need multiple cores to operate simultaneously. In this case, for the jth computing resource, the use time of the jth resource in the ith cycle can be divided into K sub-time periods according to the number of cores K actually used by the jth computing resource in the ith cycle, and the running time of each core of the jth computing resource in the ith cycle is one of the sub-time periods.

[0121] As shown in Table 3, the single-cycle price of each computing resource is calculated according to 1 core. In actual applications, a computing resource can need multiple cores to operate simultaneously. In this case, for the jth computing resource, the use time of the jth resource in the ith cycle can be divided into K sub-time periods according to the number of cores K actually used by the jth computing resource in the ith cycle, and the running time of each core of the jth computing resource in the ith cycle is one of the sub-time periods. Figure 4 As shown in Table 3, the single-cycle price of each computing resource is calculated according to 1 core. In actual applications, a computing resource can need multiple cores to operate simultaneously. In this case, for the jth computing resource, the use time of the jth resource in the ith cycle can be divided into K sub-time periods according to the number of cores K actually used by the jth computing resource in the ith cycle, and the running time of each core of the jth computing resource in the ith cycle is one of the sub-time periods.

[0122] The computing resource scheduling method of the present application can be used to build a private cloud service.

[0123] Corresponding to the method embodiment, the present application also provides a computing resource scheduling device. A structure diagram of the computing resource scheduling device provided by the present application is shown in Figure 5 As shown in Table 3, the single-cycle price of each computing resource is calculated according to 1 core. In actual applications, a computing resource can need multiple cores to operate simultaneously. In this case, for the jth computing resource, the use time of the jth resource in the ith cycle can be divided into K sub-time periods according to the number of cores K actually used by the jth computing resource in the ith cycle, and the running time of each core of the jth computing resource in the ith cycle is one of the sub-time periods.

[0124] The determining module 501 and the scheduling module 502; wherein,

[0125] The determining module 501 is configured to determine the expected period number of each computing resource used in the application scenario based on a first condition and a second condition, wherein the first condition is that the usage of the remaining available object by the each computing resource running at the expected period number is greater than a target usage, and the second condition is that the total performance of the each computing resource running at the expected period number meets a threshold; the performance of any computing resource is positively correlated with the computing capability of the any computing resource.

[0126] The scheduling module 502 is configured to schedule the each computing resource in the remaining available time period according to the expected period number of the each computing resource.

[0127] The computing resource scheduling apparatus provided by the embodiment of the present application schedules the each computing resource in the remaining available time period according to the expected period number of the each computing resource. Since the expected period number of the computing resource is calculated by taking the usage of the remaining available object by the each computing resource running at the expected period number being greater than the target usage and the total performance meeting the threshold as the target, the each computing resource is scheduled in the remaining available time period according to the expected period number of the each computing resource, the purpose of dynamically adjusting the computing resource based on the remaining available object and the remaining available time period is achieved, the usage of the available object at the end of the available time period is ensured to be greater, and the performance of the scheduled computing resource is ensured to be better, thereby improving the intelligence of the computing resource scheduling.

[0128] In an optional embodiment, the scheduling module 502 can be configured to:

[0129] obtain the total performance of the each computing resource running at the expected period number;

[0130] determine the target performance of a single period in the remaining available time period according to the total performance;

[0131] allocate a period corresponding to the expected period number for the each computing resource in the remaining available time period, so that the performance of the each computing resource in the single period in the remaining available time period approaches the target performance.

[0132] In an optional embodiment, when the scheduling module 502 determines the target performance of a single period in the remaining available time period according to the total performance, the scheduling module 502 is configured to:

[0133] determine the ratio of the total performance to the number of periods contained in the remaining available time period as the target performance of a single period in the remaining available time period.

[0134] In an optional embodiment, the scheduling module 502 targets the performance of a single cycle of each of the computing resources in the remaining available time period to approach the target performance, and allocates, for each of the computing resources, a corresponding expected number of cycles in the remaining available time period for running the corresponding computing resource.

[0135] determines an effective running time period of each of the computing resources in the remaining available time period;

[0136] allocates, for each of the computing resources, a corresponding expected number of cycles in the effective running time period of the corresponding computing resource for running the corresponding computing resource.

[0137] In an optional embodiment, the performance of each of the computing resources in the application scenario is obtained by statistically analyzing the time consumed by each of the computing resources for processing the same sample data in the application scenario.

[0138] wherein the performance of any computing resource is negatively related to the time consumed by the any computing resource for processing the sample data.

[0139] In an optional embodiment, the total performance of each of the computing resources when running according to the expected number of cycles is:

[0140] a weighted sum of the performance of each of the computing resources;

[0141] wherein the weight of the performance of any computing resource is the expected number of cycles of the any computing resource.

[0142] In an optional embodiment, the computing resource scheduling apparatus provided in the present application further comprises:

[0143] an interaction module configured to display an interaction interface and receive, through the interaction interface, an input target available time period and a target number of available objects.

[0144] Corresponding to the method embodiments, the present application further provides an electronic device, which can be a cloud server. A structural schematic diagram of the electronic device is shown in Figure 6 The electronic device can include at least one processor 1, at least one communication interface 2, at least one memory 3, and at least one communication bus 4.

[0145] In the embodiments of the present application, the number of the processor 1, the communication interface 2, the memory 3, and the communication bus 4 is at least one, and the processor 1, the communication interface 2, and the memory 3 complete communication with each other through the communication bus 4.

[0146] The processor 1 can be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement one or more embodiments of the application.

[0147] The memory 3 can include a high-speed RAM memory, and can also include a non-volatile memory, such as at least one disk memory.

[0148] The memory 3 stores a program, and the processor 1 can invoke the program stored in the memory 3, and the program is used for:

[0149] Based on the first condition and the second condition, determining an expected period number of each computing resource used in the application scenario; wherein the first condition is that the usage of the remaining available objects when the each computing resource runs according to the expected period number is greater than a target usage; the second condition is that the total performance of the each computing resource when the each computing resource runs according to the expected period number meets a threshold; the performance of any computing resource is positively correlated with the computing capability of the any computing resource;

[0150] According to the expected period number of the each computing resource, scheduling the each computing resource within the remaining available time period.

[0151] Optionally, the refinement function and the extension function of the program can refer to the description above.

[0152] The embodiments of the application also provide a storage medium which can store a program suitable for a processor to execute, and the program is used for:

[0153] Based on the first condition and the second condition, determining an expected period number of each computing resource used in the application scenario; wherein the first condition is that the usage of the remaining available objects when the each computing resource runs according to the expected period number is greater than a target usage; the second condition is that the total performance of the each computing resource when the each computing resource runs according to the expected period number meets a threshold; the performance of any computing resource is positively correlated with the computing capability of the any computing resource;

[0154] According to the expected period number of the each computing resource, scheduling the each computing resource within the remaining available time period.

[0155] Optionally, the refinement function and the extension function of the program can refer to the description above.

[0156] Those skilled in the art can understand that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0157] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0158] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e. they can be located in one place or distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0159] In addition, each functional unit in each embodiment of the present application can be integrated into a processing unit, or each unit can exist physically, or two or more units can be integrated into one unit.

[0160] It should be understood that the features in the embodiments of the present application, the various embodiments, the features can be combined with each other, and can achieve the solution to the foregoing technical problems.

[0161] If the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the essential part or part of the technical solutions that make contributions to the prior art can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0162] The foregoing description of the disclosed embodiments enables a person skilled in the art to make or use the application. Modifications of these embodiments will occur to persons of skill in the art, and that the generic principles defined herein can be applied to other embodiments without departing from the spirit or scope of the application. Therefore, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for computing resource scheduling, the method comprising: determining expected cycle numbers of each computing resource used in an application scenario based on a first condition and a second condition, wherein the first condition is that a usage rate of a remaining available object by the each computing resource running at the expected cycle numbers is greater than a target usage rate, and the second condition is that a total performance of the each computing resource running at the expected cycle numbers meets a threshold, wherein a performance of any computing resource is positively correlated with a computing capability of the any computing resource, and the available object is an object consumed by a computing resource running; and scheduling the each computing resource in a remaining available time period according to the expected cycle numbers of the each computing resource. 2.The method of claim 1, wherein the scheduling the each computing resource in the remaining available time period according to the expected cycle numbers of the each computing resource comprises: obtaining a total performance of the each computing resource running at the expected cycle numbers; determining a target performance of a single cycle in the remaining available time period according to the total performance; and allocating a cycle corresponding to the expected cycle numbers to the each computing resource in the remaining available time period, with a performance of a single cycle of the each computing resource running in the remaining available time period tending to the target performance as a target. 3.The method of claim 2, wherein the determining the target performance of the single cycle in the remaining available time period according to the total performance comprises: determining a ratio of the total performance to a number of cycles contained in the remaining available time period as the target performance of the single cycle in the remaining available time period. 4.The method of claim 2, wherein the allocating the cycle corresponding to the expected cycle numbers to the each computing resource in the remaining available time period, with the performance of the single cycle of the each computing resource running in the remaining available time period tending to the target performance as the target, comprises: determining an effective running time period of the each computing resource in the remaining available time period; and allocating a cycle corresponding to the expected cycle numbers to the each computing resource in the effective running time period of the each computing resource, with the performance of the single cycle of the each computing resource running in the effective running time period tending to the target performance as the target. 5.The method of claim 1, wherein the performance of the each computing resource in the application scenario is obtained by statistically analyzing time consumed by the each computing resource processing same sample data in the application scenario, and wherein the performance of any computing resource is negatively correlated with time consumed by the any computing resource processing the sample data. 6.The method of claim 1, wherein the total performance of the each computing resource running at the expected cycle numbers is a weighted sum of the performance of the each computing resource, and wherein a weight of the performance of any computing resource is the expected cycle number of the any computing resource. 7.The method of any one of claims 1-5, further comprising: displaying an interactive interface. ​ ​ ​ ​ ​ ​ wherein ​ ​ ​ wherein ​ ​ ​ The target available time period and the target number of available objects are received by the interactive interface.

8. A computing resource scheduling apparatus, comprising: a determining module configured to determine an expected cycle number of each computing resource used in an application scenario based on a first condition and a second condition, wherein the first condition is that a usage rate of a remaining available object when the each computing resource runs according to the expected cycle number is greater than a target usage rate, and the second condition is that a total performance of the each computing resource when the each computing resource runs according to the expected cycle number meets a threshold value, a performance of any computing resource is positively correlated with a computing capability of the any computing resource, and the available object refers to an object consumed by a computing resource running; a scheduling module configured to schedule the each computing resource in a remaining available time period according to the expected cycle number of the each computing resource.

9. An electronic device, comprising: a memory configured to store a program; a processor configured to invoke and execute the program in the memory, and implement each step of a computing resource scheduling method according to any one of claims 1-7 by executing the program.

10. A readable storage medium having a computer program stored thereon, and the computer program, when executed by a processor, implements each step of a computing resource scheduling method according to any one of claims 1-7.

Images