Resource monitoring method, device and equipment for computer cluster
By obtaining and analyzing the hardware standardized parameters of computing equipment, the problem of resource usage monitoring in distributed computer clusters is solved, effective monitoring and scheduling optimization of computer cluster resources is achieved, and the management capabilities of cloud computing service platforms are improved.
Patent Information
- Application Number
- CN202411045424.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-31
- Publication Date
- 2024-11-29
AI Technical Summary
The prior art lacks a method to evaluate and monitor the use of hardware resources in distributed computer clusters and the use of resources by computing tasks, resulting in the inability to effectively optimize resource scheduling and charge fees.
By obtaining the calibration values and occupancy values of the hardware standardized parameters of the computing device, combining the task type and utilization rate of time periods, the hardware resource occupancy information of multiple computing devices in the computer cluster is determined, and unified monitoring of computer cluster resources is achieved.
It realizes monitoring of the hardware resources and occupation of computing equipment in computer clusters, provides data feedback for resource scheduling optimization and fee collection, and improves the resource management and scheduling capabilities of the cloud computing service platform.
Smart Images

Figure CN119046087A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of computer clusters, and in particular to a method, device and equipment for monitoring resources of a computer cluster. Background Art
[0002] As cloud-native technology gradually becomes the mainstream Internet application solution, machine learning, web services, application software, etc. often rely on cloud technology and use computer clusters geographically distributed around the world to process, transmit, and store data, and distribute computing tasks to computer clusters that provide services to calculate and return results.
[0003] Due to the diversity of computing devices in distributed computer clusters, there is currently a lack of a method to evaluate the hardware resources in a computer cluster and monitor the resource usage of the computer cluster by various computing tasks. Summary of the invention
[0004] In view of this, embodiments of the present application provide a method, apparatus, and device for monitoring resources of a computer cluster to monitor the hardware resources and occupancy of computing devices in the computer cluster.
[0005] To solve the above problems, the technical solutions provided in the embodiments of the present application are as follows:
[0006] A resource monitoring method for a computer cluster, the method comprising:
[0007] Obtaining the calibration values of hardware standardized parameters of computing devices in a computer cluster;
[0008] When the computing device executes a computing task, obtaining an occupation value and an occupation time of a hardware standardized parameter of the computing device;
[0009] The hardware resource occupancy information of the plurality of computing devices in the computer cluster is determined according to the calibration value, occupancy value and occupancy time of the hardware standardized parameter of the computing device.
[0010] In a possible implementation, obtaining the calibration value of the hardware standardized parameter of the computing device in the computer cluster includes:
[0011] A calibration program is sent to a computing device in a computer cluster, and calibration values of hardware standardization parameters of the computing device calibrated by the calibration program are obtained.
[0012] In a possible implementation, the method further includes:
[0013] According to the task type of the computing task, a predicted occupancy value of a hardware standardized parameter corresponding to the computing task is obtained.
[0014] In a possible implementation, obtaining, according to the task type of the computing task, a predicted occupancy value of a hardware standardized parameter corresponding to the computing task includes:
[0015] When the task type of the computing task is a preset type, obtaining a predicted occupancy value of a hardware standardized parameter corresponding to the computing task of a unit scale;
[0016] Obtaining the total size of the computing task;
[0017] The predicted occupancy value of the hardware standardized parameter corresponding to the computing task of the unit scale is multiplied by the total scale of the computing task to obtain the predicted occupancy value of the hardware standardized parameter corresponding to the computing task.
[0018] In a possible implementation, obtaining, according to the task type of the computing task, a predicted occupancy value of a hardware standardized parameter corresponding to the computing task includes:
[0019] When the task type of the computing task is a non-preset type, a predicted occupancy value of a hardware standardized parameter corresponding to the specified computing task is obtained.
[0020] In a possible implementation, the method further includes:
[0021] According to the occupation values of the hardware standardized parameters of different computing devices when the computing task is executed on different technical devices, the predicted occupation values of the hardware standardized parameters corresponding to the computing task are corrected.
[0022] In a possible implementation, the method further includes:
[0023] According to the hardware resource occupation information of the plurality of computing devices in the computer cluster, obtaining the hardware resource utilization rate of the computing devices in the target time period;
[0024] The hardware resource utilization of the computing device in a future time period is predicted according to the hardware resource utilization of the computing device in a target time period.
[0025] A resource monitoring device for a computer cluster, the device comprising:
[0026] A first acquisition unit, used to acquire the calibration value of the hardware standardization parameter of the computing device in the computer cluster;
[0027] A second acquisition unit, configured to acquire an occupation value and an occupation time of a hardware standardized parameter of the computing device when the computing device performs a computing task;
[0028] The determining unit is used to determine the hardware resource occupancy information of multiple computing devices in the computer cluster according to the calibration value, occupancy value and occupancy time of the hardware standardized parameter of the computing device.
[0029] A resource monitoring device for a computer cluster includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the resource monitoring method for the computer cluster as described above is implemented.
[0030] A computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device executes the resource monitoring method of a computer cluster as described above.
[0031] It can be seen that the embodiments of the present application have the following beneficial effects:
[0032] The embodiment of the present application adopts a set of unified hardware standardized parameters, first obtains the calibration value of the hardware standardized parameters of the computing devices in the computer cluster, and then when the computing devices perform computing tasks, the occupation value and occupation time of the hardware standardized parameters of the computing devices can be obtained. According to the calibration value, occupation value and occupation time of the hardware standardized parameters of the computing devices, the hardware resource occupation information of multiple computing devices in the computer cluster can be determined. Thus, the monitoring of the hardware resources and occupation status of the computing devices in the computer cluster is realized. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 A schematic diagram of an exemplary application scenario provided for an embodiment of the present application;
[0034] Figure 2 A flowchart of a computer cluster resource monitoring method provided in an embodiment of the present application;
[0035] Figure 3 A schematic diagram of an implementation of a resource monitoring method for a computer cluster provided in an embodiment of the present application;
[0036] Figure 4 A schematic diagram of a resource monitoring device for a computer cluster provided in an embodiment of the present application. DETAILED DESCRIPTION
[0037] In order to make the above-mentioned purposes, features and advantages of the embodiments of the present application more obvious and easy to understand, the embodiments of the present application are further described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0038] In order to facilitate understanding and explanation of the technical solution provided by the embodiments of the present application, the background technology of the embodiments of the present application will be described below.
[0039] Cloud computing service platforms can use computer clusters geographically distributed around the world to process, transmit, and store data, distribute computing tasks to computer clusters that provide services, and return results. The computing devices that make up the distributed computer clusters come from thousands of households, and there is considerable heterogeneity in their hardware configurations.
[0040] As traditional tools in the computer industry, the Windows operating system's task manager and the Linux operating system's top command can provide information about the computer's current CPU (Central Processing Unit) and memory usage. The GPU (Graphics Processing Unit) hardware produced by NVIDIA can use the nvidia-smi command to view the current graphics card's computing usage and video memory usage. On the computer side, benchmark software can perform performance testing and calibration on computing hardware, network bandwidth, hard disk storage, etc.
[0041] The existing computing equipment resource calibration and computing load monitoring technologies for computer clusters have the following problems:
[0042] The computing resource management system used in traditional computer clusters is highly specialized and can only be applied to the operating system and hardware types within the cluster, and cannot be used for other types of computing devices. For some cloud computing service platforms that need to access multiple computing devices, there is currently no universal computing resource calibration tool that can use a unified set of standardized parameters to quantify the resource contribution of different computing devices to the computing power of the cloud computing service platform.
[0043] There are different load detection tools for different hardware and task types to monitor the resource usage of computing tasks. For example, nvidia-smi can only detect the usage of the current task on the GPU computing core and video memory of the current host, but cannot provide feedback on the usage of the computing task on the overall computer cluster resources (such as memory usage, CPU usage, bandwidth usage, GPU usage of other manufacturers, etc.). Currently, there is no detection tool that can use a set of standardized parameters to quantitatively describe the resource usage of different types of computing tasks on the computing power of the cloud computing service platform.
[0044] Therefore, various cloud computing service platforms currently cannot perform resource calibration and occupancy monitoring for any connected computing hardware (computing equipment) or any running computing tasks, and thus cannot obtain a numerical basis for charging fees, nor can they optimize resource scheduling for computing tasks under such conditions.
[0045] Based on this, the embodiments of the present application provide a resource monitoring method, device and equipment for a computer cluster, and provide a technical framework for such situations, giving the cloud computing service platform more versatile computing hardware resource calibration and computing task monitoring capabilities.
[0046] In order to facilitate understanding of the resource monitoring method of the computer cluster provided in the embodiment of the present application, Figure 1 See the example scenario shown in the figure. Figure 1 As shown, this figure is a schematic diagram of an exemplary application scenario provided in an embodiment of the present application.
[0047] The method provided in the embodiment of the present application can be applied to the scheduling center node of the cloud computing service platform, and the scheduling center node can be understood as the management and control node of the cloud computing service platform. The scheduling center node can monitor the computer cluster included in the cloud computing service platform and assign the computing task scheduling to the computing devices in the computer cluster. The computer cluster can be a distributed computer cluster, which includes multiple computing devices of different types. The embodiment of the present application will provide a computing hardware resource management and computing task monitoring technology suitable for distributed computer clusters. When the cloud computing service platform is connected to a variety of different types of computing devices and runs a variety of types of computing tasks, it can quantitatively describe the hardware resources owned by the cloud computing service platform through a set of unified standardized parameters, measure the occupation of hardware resources by computing tasks, and provide data feedback for charging and scheduling optimization of the cloud computing service platform.
[0048] Those skilled in the art will understand that Figure 1 The framework diagram shown is only an example in which the embodiments of the present application can be implemented. The scope of application of the embodiments of the present application is not limited by any aspect of the framework.
[0049] To facilitate understanding of the embodiments of the present application, a computer cluster resource monitoring method provided by the embodiments of the present application is described below with reference to the accompanying drawings.
[0050] See also Figure 2 As shown in FIG. 1 , this figure is a flow chart of a resource monitoring method of a computer cluster provided in an embodiment of the present application, such as Figure 2 As shown, the method may include S201-S203:
[0051] S201: Obtain calibration values of hardware standardization parameters of computing devices in a computer cluster.
[0052] First, determine the hardware standardization parameters, through which the hardware resources of the computing device can be uniformly calibrated. Hardware standardization parameters include standardization parameters of CPU, GPU, disk, memory, and network. Among them, the standardization parameters of CPU may include: number of CPU cores, target computing task occupancy rate; the standardization parameters of GPU may include: graphics card model, number of graphics card cores, maximum frequency, real-time frequency, target computing task occupancy rate, total video memory, total occupied video memory, target computing task occupied video memory; the standardization parameters of disk may include: disk size; the standardization parameters of memory may include: memory size; the standardization parameters of network may include: network delay, network bandwidth.
[0053] The computing device does not necessarily have to be a complete physical hardware device. It can be a computer, a virtual machine, a computing cluster, etc., and the underlying hardware can also be different types of hardware from different manufacturers.
[0054] Then, the calibration values of the hardware standardized parameters of the computing devices in the computer cluster can be obtained. The calibration values of the hardware standardized parameters can be understood as the total amount of the hardware standardized parameters of the computing devices. For example, a computing device has 8 CPU cores, a disk size of 128G, a memory size of 8G, etc.
[0055] In practical applications, the calibration values of the hardware standardized parameters of various aspects of the computing device can be integrated into a computing power point state vector, where the computing power point represents a computing device that provides computing power, which can calibrate the real-time status information of the computing device.
[0056] In a possible implementation, the specific implementation of S201 acquiring the calibration values of the hardware standardized parameters of the computing devices in the computer cluster may include:
[0057] A calibration program is sent to a computing device in a computer cluster to obtain calibration values of hardware standardization parameters of the computing device calibrated by the calibration program.
[0058] In actual applications, when a computing device is added to the computing power pool, a calibration program can be automatically run on the computing device to retrieve various hardware standardized parameters of the computing device, thereby incorporating the computing device into the total computing power pool.
[0059] This enables the computing power provided by a wide variety of computing devices to be described using a unified set of hardware standardized parameters, thereby building a homogeneous computing power pool platform.
[0060] S202: When the computing device executes a computing task, obtain the occupation value and occupation time of the hardware standardized parameters of the computing device.
[0061] For the computing power demand side, computing tasks can be uploaded in a variety of ways. When the computing device executes the computing task, the same hardware standardized parameters can be used to standardize the computing task, and the usage of the computing power pool by the computing task can be fed back through the occupation value and occupation time of the hardware standardized parameters.
[0062] In practical applications, after the computing task begins to execute, the occupation value and occupation time of the hardware standardized parameters of the computing task will be continuously updated, thereby providing a basis for resource allocation and billing of the cloud computing service platform. The computing task can be various tasks in cloud computing services, such as image recognition tasks, speech-to-text tasks, etc. The embodiment of the present application does not limit the type of computing tasks.
[0063] In a possible implementation, the resource monitoring method of a computer cluster provided in the embodiment of the present application may further include:
[0064] According to the task type of the computing task, a predicted occupancy value of a hardware standardized parameter corresponding to the computing task is obtained.
[0065] While metering the computing tasks, the same hardware standardized parameters can also be used to calibrate the computing tasks submitted by the demand side to feedback the computing tasks' demand for the computing power pool. The demand for calibrating computing tasks can be understood as predicting the hardware resources required for the computing tasks, that is, obtaining the predicted occupancy values of the hardware standardized parameters corresponding to the computing tasks. The predicted occupancy value may not be a set of constants, but may provide an operational range, such as a predicted occupancy value of a disk size of 1G to 2G. The scheduling center node of the cloud computing service platform can then determine a reasonable value within the range to run the computing task based on parameters such as urgency and predicted occupancy value. The embodiments of the present application do not limit scheduling algorithms and strategies.
[0066] In addition, in the embodiment of the present application, the process number of each computing task can be recorded through multi-threading and tracked in real time to provide calibration and metering for a large number of computing tasks at the same time.
[0067] Determining the predicted occupancy value of the hardware standardized parameter of the computing task can be achieved by using an automated program calibration method or a pre-specified method.
[0068] In a possible implementation, according to the task type of the computing task, the specific implementation of obtaining the predicted occupancy value of the hardware standardized parameter corresponding to the computing task may include:
[0069] A1: When the task type of the computing task is a preset type, the predicted occupancy value of the hardware standardized parameter corresponding to the computing task of the unit scale is obtained.
[0070] A2: Get the total size of the computing task.
[0071] A3: Multiply the predicted occupancy value of the hardware standardized parameter corresponding to the computing task of unit scale by the total scale of the computing task to obtain the predicted occupancy value of the hardware standardized parameter corresponding to the computing task.
[0072] In the method of using automated program calibration, for widely used computing tasks, the predicted occupancy values of the hardware standardized parameters corresponding to the computing tasks of the unit scale can be calibrated. Then, the predicted occupancy values of the hardware standardized parameters corresponding to the computing tasks can be calculated based on the total scale of the computing tasks. For example, if the computing task is an AI image generation task, the amount of hardware resources required for each image generation can be obtained through prior testing, community experience, etc., such as the predicted occupancy values of GPU utilization, CPU utilization, video memory, memory, and network bandwidth, and then the total predicted occupancy value can be obtained through its call scale.
[0073] In a possible implementation, according to the task type of the computing task, the specific implementation of obtaining the predicted occupancy value of the hardware standardized parameter corresponding to the computing task may include:
[0074] When the task type of the computing task is a non-preset type, a predicted occupancy value of a hardware standardized parameter corresponding to the specified computing task is obtained.
[0075] In the pre-specified method, for computing tasks that are not widely used, the predicted occupancy value of the hardware standardized parameters corresponding to the specified computing task can be obtained. For example, for some computing tasks using self-developed models, the computing power scale of the computing task needs to be manually specified, and the predicted occupancy value can be modified later according to the actual situation.
[0076] S203: Determine hardware resource occupancy information of multiple computing devices in the computer cluster according to the calibration value, occupancy value and occupancy time of the hardware standardized parameters of the computing devices.
[0077] According to the calibration value, occupancy value and occupancy time of the hardware standardized parameters of the computing device, the resource occupancy of the current platform computing pool is processed and returned in real time. In practical applications, handles and status information pointing to each computing device and computing task are provided to facilitate the platform to dynamically schedule both supply and demand parties. Then, the hardware resource occupancy of multiple computing devices in the computer cluster is counted.
[0078] In a possible implementation, the resource monitoring method of a computer cluster provided in the embodiment of the present application may further include:
[0079] According to the occupied values of the hardware standardized parameters of different computing devices when the computing tasks are run on different technical devices, the predicted occupied values of the hardware standardized parameters corresponding to the computing tasks are corrected.
[0080] Based on the occupancy value of the computing task in the hardware standardized parameters of the computing device and the predicted occupancy value, a more reasonable judgment of the computing power scale can be made, so that the predicted occupancy value can be corrected.
[0081] In a possible implementation, the resource monitoring method of a computer cluster provided in the embodiment of the present application may further include:
[0082] B1: According to the hardware resource occupancy information of multiple computing devices in the computer cluster, the hardware resource utilization rate of the computing device in the target time period is obtained.
[0083] B2: Predict the hardware resource utilization of the computing device in a future time period based on the hardware resource utilization of the computing device in a target time period.
[0084] That is, under the premise of having the computing power point state vector of the computing power pool and the computing power scale required for the computing task, the computing task can be scheduled in real time by adopting the relevant scheduling algorithm. During the scheduling process, the real-time resource occupancy of the specified computing task on all computing devices and the utilization rate in different time periods are counted, so as to give a more reasonable judgment on the computing power scale and the prediction of the hardware resource utilization rate of the computing task in a certain time period in the future.
[0085] The embodiment of the present application adopts a set of unified hardware standardized parameters, first obtains the calibration value of the hardware standardized parameters of the computing devices in the computer cluster, and then when the computing devices perform computing tasks, the occupation value and occupation time of the hardware standardized parameters of the computing devices can be obtained. According to the calibration value, occupation value and occupation time of the hardware standardized parameters of the computing devices, the hardware resource occupation information of multiple computing devices in the computer cluster can be determined. Thus, the monitoring of the hardware resources and occupation status of the computing devices in the computer cluster is realized.
[0086] See also Figure 3 , which shows an implementation example diagram of a computer cluster resource monitoring method provided by an embodiment of the present application in an actual application scenario.
[0087] In practical applications, the method of implementing the embodiment of the present application can be implemented through a computing power supply-side homogenization module, a computing power demand-side task metering module, and a computing power status statistics module.
[0088] Among them, the computing power supply-side homogenization module is used to evaluate the computing power of computing devices with different hardware and different working environments on the computing power supply side, and to uniformly calibrate the computing devices with standardized hardware parameters to obtain the computing power point state vector corresponding to each computing device. The total computing power pool is composed of the computing power point state vectors. The computing power demand-side task metering module uses the same standardized hardware parameters to calibrate the computing tasks submitted by the demand side, and standardizes the metering of computing tasks by the method of "occupancy value + time occupancy of standardized hardware parameters", and feedbacks the total computing power required and usage of the computing power pool by the computing tasks. The computing power status statistics module summarizes the information of the first two modules, reflecting the overall resource supply and demand of the current computing power pool, and provides a complete adjustment interface and status feedback information for the development of computing power scheduling algorithms. At the same time, the computing power status statistics module can also automatically verify the predicted occupancy values given by the first two modules, and provide real-time feedback to help the first two modules to correct parameters.
[0089] In this way, the embodiment of the present application provides a unified calibration method for distributed computer clusters with various hardware types, abstracts complex computing equipment conditions into unified hardware standardized parameters for description, and realizes the additivity and simplicity of computing resource calibration. For various computing tasks, it is possible to predict and detect the demand for hardware resources of the computing task, and describe it with the same set of hardware standardized parameters plus time data. It provides sufficient data feedback for status monitoring, billing basis and scheduling optimization of cloud computing service platforms running in complex environments.
[0090] Based on the resource monitoring method of a computer cluster provided by the above method embodiment, the embodiment of the present application also provides a resource monitoring device of a computer cluster, which will be described below in conjunction with the accompanying drawings.
[0091] See also Figure 4 As shown in FIG. 1 , this figure is a schematic diagram of the structure of a resource monitoring device for a computer cluster provided in an embodiment of the present application. Figure 4 As shown, the resource monitoring device of the computer cluster includes:
[0092] A first acquisition unit 401 is used to acquire the calibration value of the hardware standardization parameter of the computing device in the computer cluster;
[0093] A second acquisition unit 402 is used to acquire an occupation value and an occupation time of a hardware standardized parameter of the computing device when the computing device performs a computing task;
[0094] The determining unit 403 is used to determine the hardware resource occupancy information of the plurality of computing devices in the computer cluster according to the calibration value, occupancy value and occupancy time of the hardware standardized parameter of the computing device.
[0095] In a possible implementation manner, the first acquiring unit is specifically configured to:
[0096] A calibration program is sent to a computing device in a computer cluster, and calibration values of hardware standardization parameters of the computing device calibrated by the calibration program are obtained.
[0097] In a possible implementation manner, the device further includes:
[0098] The third acquisition unit is used to acquire the predicted occupancy value of the hardware standardized parameter corresponding to the computing task according to the task type of the computing task.
[0099] In a possible implementation manner, the third obtaining unit is specifically configured to:
[0100] When the task type of the computing task is a preset type, obtaining a predicted occupancy value of a hardware standardized parameter corresponding to the computing task of a unit scale;
[0101] Obtaining the total size of the computing task;
[0102] The predicted occupancy value of the hardware standardized parameter corresponding to the computing task of the unit scale is multiplied by the total scale of the computing task to obtain the predicted occupancy value of the hardware standardized parameter corresponding to the computing task.
[0103] In a possible implementation manner, the third obtaining unit is specifically configured to:
[0104] When the task type of the computing task is a non-preset type, a predicted occupancy value of a hardware standardized parameter corresponding to the specified computing task is obtained.
[0105] In a possible implementation manner, the device further includes:
[0106] The correction unit is used to correct the predicted occupancy value of the hardware standardized parameter corresponding to the computing task according to the occupancy value of the hardware standardized parameter of different computing devices when the computing task is running on different technical devices.
[0107] In a possible implementation manner, the device further includes:
[0108] A fourth acquisition unit, configured to acquire the hardware resource utilization rate of the computing device in a target time period according to the hardware resource occupation information of the plurality of computing devices in the computer cluster;
[0109] The prediction unit is used to predict the hardware resource utilization of the computing device in a future time period according to the hardware resource utilization of the computing device in a target time period.
[0110] In addition, an embodiment of the present application also provides a resource monitoring device for a computer cluster, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the resource monitoring method for a computer cluster as described in any one of the above items.
[0111] The embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device executes the resource monitoring method for a computer cluster as described in any one of the above items.
[0112] The embodiment of the present application further provides a computer program product, including computer program instructions. When the computer program instructions are executed on a computer, the computer executes any one of the resource monitoring methods for a computer cluster as described above.
[0113] The embodiment of the present application adopts a set of unified hardware standardized parameters, first obtains the calibration value of the hardware standardized parameters of the computing devices in the computer cluster, and then when the computing devices perform computing tasks, the occupation value and occupation time of the hardware standardized parameters of the computing devices can be obtained. According to the calibration value, occupation value and occupation time of the hardware standardized parameters of the computing devices, the hardware resource occupation information of multiple computing devices in the computer cluster can be determined. Thus, the monitoring of the hardware resources and occupation status of the computing devices in the computer cluster is realized.
[0114] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the system or device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.
[0115] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.
[0116] It should also be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0117] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
[0118] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for monitoring resources of a computer cluster, characterized in that: The method comprises: Obtaining the calibration values of hardware standardized parameters of computing devices in a computer cluster; When the computing device executes a computing task, obtaining an occupation value and an occupation time of a hardware standardized parameter of the computing device; The hardware resource occupancy information of the plurality of computing devices in the computer cluster is determined according to the calibration value, occupancy value and occupancy time of the hardware standardized parameter of the computing device.
2. The method according to claim 1, characterized in that: The step of obtaining the calibration values of the hardware standardized parameters of the computing devices in the computer cluster includes: A calibration program is sent to a computing device in a computer cluster, and calibration values of hardware standardization parameters of the computing device calibrated by the calibration program are obtained.
3. The method according to claim 1, characterized in that The method further comprises: According to the task type of the computing task, a predicted occupancy value of a hardware standardized parameter corresponding to the computing task is obtained.
4. The method according to claim 3, characterized in that The obtaining, according to the task type of the computing task, a predicted occupancy value of a hardware standardized parameter corresponding to the computing task includes: When the task type of the computing task is a preset type, obtaining a predicted occupancy value of a hardware standardized parameter corresponding to the computing task of a unit scale; Obtaining the total size of the computing task; The predicted occupancy value of the hardware standardized parameter corresponding to the computing task of the unit scale is multiplied by the total scale of the computing task to obtain the predicted occupancy value of the hardware standardized parameter corresponding to the computing task.
5. The method according to claim 3, characterized in that: The obtaining, according to the task type of the computing task, a predicted occupancy value of a hardware standardized parameter corresponding to the computing task includes: When the task type of the computing task is a non-preset type, a predicted occupancy value of a hardware standardized parameter corresponding to the specified computing task is obtained.
6. The method according to any one of claims 3 to 5, characterized in that: The method further comprises: According to the occupation values of the hardware standardized parameters of different computing devices when the computing task is executed on different technical devices, the predicted occupation values of the hardware standardized parameters corresponding to the computing task are corrected.
7. The method according to claim 1, characterized in that The method further comprises: According to the hardware resource occupation information of the plurality of computing devices in the computer cluster, obtaining the hardware resource utilization rate of the computing devices in the target time period; The hardware resource utilization of the computing device in a future time period is predicted according to the hardware resource utilization of the computing device in a target time period.
8. A resource monitoring device for a computer cluster, characterized in that: The device comprises: A first acquisition unit, used to acquire the calibration value of the hardware standardization parameter of the computing device in the computer cluster; A second acquisition unit, configured to acquire an occupation value and an occupation time of a hardware standardized parameter of the computing device when the computing device performs a computing task; The determining unit is used to determine the hardware resource occupancy information of multiple computing devices in the computer cluster according to the calibration value, occupancy value and occupancy time of the hardware standardized parameter of the computing device.
9. A resource monitoring device for a computer cluster, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the resource monitoring method for a computer cluster as described in any one of claims 1 to 7 is implemented.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device executes the resource monitoring method for a computer cluster according to any one of claims 1 to 7.
Citation Information
Patent Citations
Computing power scheduling method and related device
CN114003370A
Optimization method and optimization device for computing power resource allocation, electronic equipment and medium
CN116541176A
Management and control architecture and method based on centralized large-scale computing power network
CN116633845A
Computing power resource settlement method, device and equipment and computer readable storage medium
CN117573335A
Calculation power scheduling method and system based on task demand prediction
CN117931446A
Cited By
Computer information monitoring method and system based on Internet of Things
CN120256099A
A computer information monitoring method and system based on the Internet of Things
CN120256099B
