Intelligent computing power resource management method and device, equipment and storage medium

Abstract

The invention discloses an intelligent computing power resource management method and device, equipment and a storage medium. In the scheme, node information of an intelligent computing power resource node is obtained, and the node information comprises first configuration information and hardware information of the intelligent computing power resource of the node. And based on the hardware information, generating second configuration information of the node in the target tool library. And comparing the first configuration information with the second configuration information to obtain configuration difference information between the first configuration information and the second configuration information. And pushing the corresponding one or more target version intelligent computing power resource management tools in the target tool library to the node according to the configuration difference information, and installing or updating the intelligent computing power resource management tools. And performing automatic management on the intelligent computing power resources by using the intelligent computing power resource management tool at the node. According to the method, unified automatic management of heterogeneous computing power resources of different manufacturers and multiple types is realized, the management time is saved, the operation and maintenance cost is reduced, and the management safety is improved.

Description

Technical Field

[0001] The present application relates to the field of information technology, and in particular to an intelligent computing resource management method, device, equipment and storage medium. Background Art

[0002] With the continuous maturity and widespread application of technologies such as big data, cloud computing, edge computing, and deep learning, intelligent computing power has become a key factor in supporting complex computing tasks and driving business innovation. At present, intelligent computing resources cover a variety of types, such as central processing units (CPUs), graphics processing units (GPUs), field-programmable gate arrays (FPGAs), and neural network processors (NPUs), each of which has unique computing capabilities.

[0003] However, at the management level of intelligent computing resources, although each manufacturer can automate the management of its own computing resources, when faced with the management of heterogeneous computing resources across manufacturers and types, a series of management tools are needed, such as drivers, container toolkits, device plug-ins, and comprehensive monitoring tools, to achieve effective intelligent computing resource management.

[0004] This requires operation and maintenance personnel to manually install and continuously update these management tools, which is time-consuming and labor-intensive. Moreover, manual operations can easily lead to inconsistent tool versions, which in turn causes instability in the cluster environment, increasing security risks and performance risks. These problems have jointly pushed up the cost of intelligent computing resource management and increased the complexity and risk of intelligent computing resource management. Therefore, seeking a more automated and intelligent intelligent computing resource management method to save management time, reduce operation and maintenance costs, and improve management security has become an important issue that needs to be solved urgently. Summary of the invention

[0005] Based on the above problems, the present application provides an intelligent computing resource management method, device, equipment and storage medium, with the aim of achieving unified and automatic management of cross-manufacturer and multi-type intelligent computing resources, saving management time, reducing operation and maintenance costs, and improving management security.

[0006] The embodiments of the present application disclose the following technical solutions:

[0007] The first aspect of the present application provides an intelligent computing resource management method, the method comprising:

[0008] Obtaining node information of an intelligent computing resource node, the node information including first configuration information and hardware information of the intelligent computing resource of the node, the first configuration information including attribute information and version information of multiple intelligent computing resource management tools currently included in the node;

[0009] Based on the hardware information, second configuration information of the node is generated in a target tool library; the target tool library includes multiple target version intelligent computing resource management tools required for automatic management of multiple intelligent computing resources, and a correspondence between the hardware information of the multiple intelligent computing resources and the multiple target version intelligent computing resource management tools; the second configuration information includes attribute information and version information of the multiple target version intelligent computing resource management tools;

[0010] Comparing the first configuration information with the second configuration information to obtain configuration difference information between the first configuration information and the second configuration information;

[0011] Push one or more corresponding target version intelligent computing resource management tools in the target tool library to the node and install or update them according to the configuration difference information;

[0012] The intelligent computing power resource management tool at the node is used to automatically manage the intelligent computing power resources.

[0013] In an optional implementation, the target tool library is constructed in the following manner:

[0014] Obtaining hardware information of multiple intelligent computing resources, and obtaining multiple target versions of intelligent computing resource management tools required for automatic management of the multiple intelligent computing resources;

[0015] Storing the multiple target versions of intelligent computing resource management tools in the target tool library;

[0016] Establishing a correspondence between hardware information of the same intelligent computing resource and the multiple target versions of intelligent computing resource management tools;

[0017] The corresponding relationship is stored in the target tool library.

[0018] In an optional implementation, after obtaining the node information of the intelligent computing power resource node, the method further includes:

[0019] According to the hardware information of the intelligent computing power resource, a label is added to the node, and the hardware information of the intelligent computing power resource at least includes the intelligent computing power resource type of the node and the intelligent computing power resource model of the node; the label is used to identify and allocate the intelligent computing power resource when automatically managing the intelligent computing power resource.

[0020] In an optional implementation, the comparing the first configuration information and the second configuration information to obtain configuration difference information between the first configuration information and the second configuration information includes:

[0021] If the second configuration information contains attribute information and version information of the intelligent computing resource management tool that does not exist in the first configuration information, the attribute information and version information of the non-existent intelligent computing resource management tool is determined as configuration difference information.

[0022] In an optional implementation, the step of pushing one or more corresponding target version intelligent computing resource management tools in the target tool library to the node and installing or updating the tools according to the configuration difference information includes:

[0023] Pushing one or more corresponding target version intelligent computing resource management tools in the target tool library to the node according to the configuration difference information;

[0024] If the node has the same intelligent computing power resource management tool as the target version intelligent computing power resource management tool, but the version is different, then the intelligent computing power resource management tool at the node is updated to the version of the target version intelligent computing power resource management tool;

[0025] If the node does not have an intelligent computing power resource management tool that is identical to the target version of the intelligent computing power resource management tool, the target version of the intelligent computing power resource management tool is installed.

[0026] In an optional implementation, the intelligent computing resource management method further includes:

[0027] If it is monitored that the target version of the intelligent computing power resource management tool in the target tool library is updated, the updated target version of the intelligent computing power resource management tool is automatically pushed to the node for installation or update.

[0028] In an optional implementation, the step of automatically managing the intelligent computing resources using the intelligent computing resource management tool at the node includes:

[0029] Using the intelligent computing power resource management tool at the node to create the intelligent computing power resource as an intelligent computing power resource instance;

[0030] The intelligent computing resource management tool at the node is used to dynamically allocate and schedule the intelligent computing resource instance.

[0031] A second aspect of the present application provides an intelligent computing resource management device, the device comprising:

[0032] A node information acquisition module, used to acquire node information of an intelligent computing resource node, wherein the node information includes first configuration information and hardware information of the intelligent computing resource of the node, wherein the first configuration information includes attribute information and version information of multiple intelligent computing resource management tools currently included in the node;

[0033] A configuration information generation module, configured to generate second configuration information of the node in a target tool library based on the hardware information; the target tool library includes multiple target version intelligent computing resource management tools required for automatic management of multiple intelligent computing resources, and a correspondence between the hardware information of the multiple intelligent computing resources and the multiple target version intelligent computing resource management tools; the second configuration information includes attribute information and version information of the multiple target version intelligent computing resource management tools;

[0034] A configuration information comparison module, used to compare the first configuration information and the second configuration information to obtain configuration difference information between the first configuration information and the second configuration information;

[0035] A tool push module, used to push one or more corresponding target version intelligent computing resource management tools in the target tool library to the node according to the configuration difference information;

[0036] A tool configuration module, used to install or update the one or more target version intelligent computing resource management tools pushed by the tool push module to the node;

[0037] A management module is used to automatically manage the intelligent computing resources using the intelligent computing resource management tool at the node.

[0038] The third aspect of the present application provides an intelligent computing resource management device, the device comprising:

[0039] a memory having a computer program stored thereon;

[0040] A processor is used to execute the computer program in the memory to implement the steps of the intelligent computing resource management method introduced in any implementation method of the first aspect.

[0041] A fourth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the intelligent computing resource management method introduced in any implementation method of the first aspect.

[0042] Compared with the prior art, this application has the following beneficial effects:

[0043] In the technical solution of the present application, firstly, the node information of the intelligent computing resource node is obtained, and the node information includes the first configuration information and the hardware information of the intelligent computing resource of the node; secondly, based on the hardware information, the second configuration information of the node is generated in the target tool library; the first configuration information and the second configuration information are compared again to obtain the configuration difference information between the first configuration information and the second configuration information; then, according to the configuration difference information, one or more corresponding target versions of the intelligent computing resource management tools in the target tool library are pushed to the node and installed or updated; finally, the intelligent computing resource management tool at the node is used to automatically manage the intelligent computing resource. It can be seen that in the technical solution of the present application, unlike the traditional solution that requires manual installation and update by operation and maintenance, the present solution uses a pre-built target tool library to generate configuration information, and automatically installs or updates the required target version of the tool after comparing the collected node information with the configuration information, thereby realizing unified and automatic management of cross-manufacturer and multi-type intelligent computing resources, saving management time and reducing operation and maintenance costs. Automatic updates and installations ensure the accuracy of the intelligent resource management tool version, avoiding cluster risks caused by version differences due to manual installation or updates, thereby improving the security of intelligent computing resource management. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0045] Figure 1 A flow chart of an intelligent computing resource management method provided in an embodiment of the present application;

[0046] Figure 2 A schematic diagram of the structure of an intelligent computing resource management device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0047] As described above, in the field of intelligent computing resource management, various manufacturers have been able to realize the automated scheduling and configuration of their internal intelligent computing resources. However, when the focus turns to the integrated management of heterogeneous intelligent computing resources across manufacturers and in multiple forms, intelligent computing resource management tools such as drivers, containerized kits, device-specific plug-ins, and comprehensive monitoring systems are indispensable, becoming a bridge to connect and optimize the management of these intelligent computing resources. However, in actual applications, the existing intelligent computing resource management methods are highly dependent on the manual operations of operation and maintenance personnel when facing the deployment, maintenance, and even continuous updating of these tools, which is not only time-consuming and labor-intensive, but also prone to increase security risks and performance hazards.

[0048] After research, the inventor proposed an intelligent computing resource management method, device, equipment and storage medium.

[0049] First, the node information of the intelligent computing power resource node is obtained, and the node information includes first configuration information and hardware information of the intelligent computing power resources of the node; secondly, based on the hardware information, the second configuration information of the node is generated in the target tool library; the first configuration information and the second configuration information are compared again to obtain the configuration difference information between the first configuration information and the second configuration information; then, according to the configuration difference information, one or more corresponding target version intelligent computing power resource management tools in the target tool library are pushed to the node and installed or updated; finally, the intelligent computing power resource management tool at the node is used to automatically manage the intelligent computing power resources.

[0050] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described 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 of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0051] See also Figure 1 , which is a flow chart of an intelligent computing resource management method provided by an embodiment of the present application. Figure 1 As shown, the method comprises the following steps:

[0052] S101. Obtain node information of an intelligent computing resource node, where the node information includes first configuration information and hardware information of the intelligent computing resource of the node.

[0053] In the embodiments of the present application, the intelligent computing resource node refers to a computing unit in the cluster that is used to provide intelligent computing power. These nodes can be one or more servers, usually equipped with high-performance CPUs, GPUs, FPGAs, etc. or other dedicated chips to support complex artificial intelligence computing tasks.

[0054] The first configuration information includes attribute information and version information of multiple intelligent computing resource management tools currently included in the node. In an exemplary implementation, the first configuration information may be in the format of {attribute information}:{version information}.

[0055] The intelligent computing resource management tool refers to software or systems used to manage, schedule and optimize intelligent computing resources, such as drivers, container toolkits, device plug-ins, and comprehensive monitoring tools.

[0056] The hardware information of the intelligent computing power resource includes at least information such as the type, model, and manufacturer of the intelligent computing power resource.

[0057] In the embodiment of the present application, the attribute information and version information of the multiple intelligent computing resource management tools currently contained in the node can be obtained through the tools or commands of the respective manufacturers of each intelligent computing resource. For example, for NVIDIA GPU, the driver version and GPU related information can be viewed through the nvidia-smi command.

[0058] In an embodiment of the present application, intelligent computing power resources can be added to the intelligent computing power resource node in a variety of ways, which are not specifically limited here. In an example implementation, the intelligent computing power resources are added to the intelligent computing power resource node in the form of an intelligent computing power card. The intelligent computing power card refers to a board that integrates a high-performance computing unit (such as a GPU, FPGA, etc.). The intelligent computing power card can be inserted into the high-speed serial computer expansion bus (Peripheral Component Interconnect Express, PCI-E) slot of the server, and cooperate with other components of the server to form a complete intelligent computing power node.

[0059] In an example implementation, when the smart computing power card is inserted into the PCI-E slot of the server, the hardware information of the smart computing power resource of the node can be obtained through the hardware number information in the / dev information.

[0060] S102: Generate second configuration information of the node in the target tool library based on the hardware information.

[0061] In the technical solution of the present application, a target tool library is creatively constructed. The target tool library includes multiple target versions of intelligent computing resource management tools required for automatic management of multiple intelligent computing resources, and the corresponding relationship between the hardware information of the multiple intelligent computing resources and the multiple target versions of intelligent computing management tools. The construction method of the target tool library is described below. In order to construct the target tool library, the hardware information of multiple intelligent computing resources is first obtained, and the multiple intelligent computing resources include but are not limited to Huawei NPU, Intel graphics card, AMD graphics card, Nvidia graphics card and other multi-vendor, multi-type intelligent computing resources, and the multiple target versions of intelligent computing resource management tools required for automatic management of the multiple intelligent computing resources are obtained. The multiple target versions of intelligent computing resource management tools are stored in the target tool library. Then, a corresponding relationship between the hardware information of the same intelligent computing resource and the multiple target versions of intelligent computing resource management tools is established. The corresponding relationship is stored in the target tool library.

[0062] In an embodiment of the present application, the target version of the intelligent computing power resource management tool is a stable version of the intelligent computing power resource management tool required for automatic management of the corresponding intelligent computing power resources.

[0063] The second configuration information includes attribute information and version information of the multiple target versions of the intelligent computing resource management tools.

[0064] Based on the hardware information obtained in the previous step as a query condition, a matching target version intelligent computing resource management tool is searched in the pre-built target tool library. Optionally, this process can be performed by traversing the corresponding relationship between the hardware information in the tool library and multiple target version intelligent computing resource management tools.

[0065] The second configuration information of the node is generated according to the attribute information and version information of the selected multiple target versions of the intelligent computing resource management tool. In an exemplary implementation, the second configuration information may be in the format of {attribute information}:{version information}.

[0066] In an embodiment of the present application, the second configuration information includes the attribute information and version information of the stable version of the intelligent computing resource management tool required for automatic management of all intelligent computing resources corresponding to the above-mentioned hardware information, so as to facilitate subsequent automated deployment or management.

[0067] S103: Compare the first configuration information and the second configuration information to obtain configuration difference information between the first configuration information and the second configuration information.

[0068] In the embodiment of the present application, the first configuration information and the second configuration information are compared to identify the difference in configuration information between the two. The difference in configuration information is caused by the lack of intelligent computing resource management tools or the existence of outdated intelligent computing resource management tools.

[0069] In an embodiment of the present application, if the second configuration information contains attribute information and version information of an intelligent computing resource management tool that does not exist in the first configuration information, the attribute information and version information of the non-existent intelligent computing resource management tool is determined as configuration difference information.

[0070] For example, the first configuration information is:

[0071] {Docker Engine-Community}:{20.10.11};

[0072] {Prometheus}:{2.30.3};

[0073] {Grafana}:{8.3.3}.

[0074] The second configuration information is:

[0075] {Intel GPU driver}:{31.0.101.3943};

[0076] {Docker Engine-Community}:{20.10.12};

[0077] {Intel GPU device plugin}:{1.5.0};

[0078] {Prometheus}:{2.30.3};

[0079] {Grafana}:{8.3.3}.

[0080] By comparing the first configuration information with the second configuration information, the configuration difference information between the first configuration information and the second configuration information is obtained as follows:

[0081] {Intel GPU driver}:{31.0.101.3943};

[0082] {Docker Engine-Community}:{20.10.12};

[0083] {Intel GPU device plugin}:{1.5.0}.

[0084] S104. Push one or more corresponding target version intelligent computing resource management tools in the target tool library to the node and install or update them according to the configuration difference information.

[0085] In an embodiment of the present application, the target version of the intelligent computing resource management tool that needs to be pushed and installed or updated is determined according to the configuration difference information obtained in the previous step. For example, the target version of the intelligent computing resource management tool that needs to be pushed and installed or updated according to the above configuration difference information is: Intel GPU driver version 31.0.101.3943, Docker Engine-Community version 20.10.12, and Intel GPU device plug-in version 1.5.0.

[0086] In an example implementation, the process of pushing the above-mentioned multiple target versions of intelligent computing power management tools to nodes and installing or updating them can use a file transfer tool to transfer the installation package of the target version of the intelligent computing power management tool to the target node for installation or updating.

[0087] S105. Use the intelligent computing power resource management tool at the node to automatically manage the intelligent computing power resources.

[0088] The technical solution of this application has automatically completed the installation and update of the intelligent computing resource management tools at the node in the above steps. Therefore, these tools can be directly used to efficiently and automatically manage the intelligent computing resources.

[0089] For example, in the management of intelligent computing resources, the intelligent computing resource management tools at the node include: drivers, container toolkits, device plug-ins, and comprehensive monitoring tools. These tools together realize an efficient and flexible automated management system.

[0090] Among them, the driver acts as a bridge between the hardware and the operating system, automatically detecting and optimizing the performance of intelligent computing resources to ensure that they run in the best condition.

[0091] Container toolkits, such as Docker and Kubernetes, encapsulate intelligent computing resources into independent container instances, implement isolation and dynamic scheduling of intelligent computing resources, and flexibly adjust the allocation of intelligent computing resources according to business needs to adapt to different computing loads.

[0092] As an extension mechanism, device plug-ins incorporate intelligent computing resources such as GPUs and FPGAs on nodes into a unified resource management platform, allowing these intelligent computing resources to be recognized and utilized by various applications in the cluster, thereby achieving efficient integration and sharing of intelligent computing resources.

[0093] Comprehensive monitoring tools monitor the operation status of intelligent computing resources in real time, collect and analyze various performance indicators and resource utilization data, and provide accurate data support for intelligent computing resource scheduling and optimization. When resource utilization reaches the preset threshold or a failure occurs, the monitoring tool can quickly issue an alarm notification to remind management personnel to intervene in time.

[0094] This automated management system not only improves the utilization rate and response speed of intelligent computing resources, but also greatly reduces operation and maintenance costs and manpower burden, providing powerful computing power support for various computing-intensive applications.

[0095] In the embodiment of the present application, by using the pre-built target tool library to generate configuration information, and automatically installing or updating the required target version of the tool after comparing the collected node information with the configuration information, unified automatic management of cross-manufacturer and multi-type intelligent computing resources is achieved, saving management time and reducing operation and maintenance costs. The accuracy of the intelligent resource management tool version is guaranteed through automatic update and installation, avoiding cluster risks caused by version differences caused by manual installation or update, thereby improving the security of intelligent computing resource management.

[0096] In an optional implementation of the technical solution of the present application, in order to more conveniently manage and allocate intelligent computing resources, after step S101, the method further includes:

[0097] According to the hardware information of the intelligent computing power resource, a label is added to the node, and the hardware information of the intelligent computing power resource at least includes the intelligent computing power resource type of the node and the intelligent computing power resource model of the node; the label is used to identify and allocate the intelligent computing power resource when automatically managing the intelligent computing power resource.

[0098] In an embodiment of the present application, corresponding labels are added to the nodes according to hardware information such as the type and model of the intelligent computing power resources. These labels can identify the type and model of the intelligent computing power resources, and can be used to quickly identify and allocate appropriate intelligent computing power resources in the subsequent automatic management of intelligent computing power resources. For example, for nodes whose intelligent computing power resources at the node are Intel arc, add the label: intel.com / arc; for nodes whose intelligent computing power resources at the node are Huawei Ascend npu, add the label: huawei.com / npu; for nodes whose intelligent computing power resources at the node are NVIDIA graphics cards, add the label: nvidia.com / gpu; for nodes whose intelligent computing power resources at the node are AMD graphics cards, add the label: amd.com / gpu.

[0099] In an embodiment of the present application, by identifying complex and diverse intelligent computing resources in the form of concise and clear labels, rapid classification and identification of intelligent computing resources is achieved. This method not only simplifies the complexity of intelligent computing resource scheduling, but also enables the system to automatically match and allocate the most suitable intelligent computing resources based on task requirements, thereby improving overall operating efficiency. At the same time, labeling management also enhances the transparency and flexibility of intelligent computing resource management. Managers can easily grasp the distribution and usage of intelligent computing resources, which facilitates refined management and optimization. In addition, with the continuous increase and update of intelligent computing resource types, the labeling method can also flexibly adapt to changes to ensure the continued effectiveness and scalability of the resource management system.

[0100] In an optional implementation of the technical solution of the present application, in order to improve the flexibility and response speed of management, the method further includes:

[0101] If it is monitored that the target version of the intelligent computing power resource management tool in the target tool library is updated, the updated target version of the intelligent computing power resource management tool is automatically pushed to the node for installation or update.

[0102] In an embodiment of the present application, by real-time monitoring of whether the target version of the intelligent computing resource management tool in the target tool library is updated, once an update is detected, the updated target version of the intelligent computing resource management tool is automatically pushed to the corresponding node for installation or update, ensuring that the node always uses the latest version of the intelligent computing resource management tool.

[0103] In an optional implementation of the technical solution of the present application, step S104 includes:

[0104] Pushing one or more corresponding target version intelligent computing resource management tools in the target tool library to the node according to the configuration difference information;

[0105] If the node has an intelligent computing power resource management tool that is the same as the target version of the intelligent computing power resource management tool, but the version is different, the intelligent computing power resource management tool at the node is updated to the version of the target version of the intelligent computing power resource management tool.

[0106] If the node does not have an intelligent computing power resource management tool that is identical to the target version of the intelligent computing power resource management tool, the target version of the intelligent computing power resource management tool is installed.

[0107] In the implementation of this application, through a flexible push, update and installation mechanism, it is ensured that the node will have all the intelligent computing resource management tools required to manage its intelligent computing resources, thereby achieving effective scheduling and efficient utilization of intelligent computing resources.

[0108] In an optional implementation of the technical solution of the present application, step S105 includes:

[0109] Step 1: Use the intelligent computing power resource management tool at the node to create the intelligent computing power resource as an intelligent computing power resource instance.

[0110] This step transforms abstract intelligent computing resources into operational and manageable intelligent computing resource instances.

[0111] In an example implementation, the specific implementation process of this step can be to encapsulate these intelligent computing resources into independent intelligent computing resource instances through the intelligent computing management tool. Each intelligent computing resource instance contains the necessary configuration information, status monitoring interface, and operation interface for subsequent management and scheduling. Finally, the created intelligent computing resource instance is registered in the intelligent computing resource pool so that other services or applications can discover and request these intelligent computing resources.

[0112] Step 2: Utilize the intelligent computing resource management tool at the node to dynamically allocate and schedule the intelligent computing resource instance.

[0113] In an embodiment of the present application, after the intelligent computing power resources are instantiated, the intelligent computing power resource management tool can receive resource requests from different applications or services and analyze the specific requirements of these requests, such as resource type, quantity, performance requirements, etc.

[0114] Based on the results of the demand analysis, the intelligent computing resource management tool will search for qualified intelligent computing resource instances in the intelligent computing resource pool. If multiple qualified intelligent computing resource instances are found, the intelligent computing resource management tool may select the most suitable instance based on certain policies (such as load balancing, priority, etc.).

[0115] Once the intelligent computing resource instance to be allocated is determined, the corresponding intelligent computing resource management tool will allocate it to the requester and update the status information of the intelligent computing resource pool.

[0116] After the intelligent computing resources are allocated, the intelligent computing resource management tool may also dynamically schedule the allocated intelligent computing resource instances based on the execution of the tasks and the usage of the intelligent computing resources. For example, if a task is completed ahead of schedule and the intelligent computing resources are released, the intelligent computing resource management tool may reallocate these intelligent computing resources to other waiting tasks.

[0117] Based on the intelligent computing power resource management method provided in the aforementioned embodiment, the present application also provides an intelligent computing power resource management device accordingly. Figure 2 This is a schematic diagram of the structure of an intelligent computing resource management device provided in an embodiment of the present application. Figure 2 As shown, the intelligent computing resource management device includes:

[0118] The node information acquisition module 201 is used to acquire the node information of the intelligent computing resource node, wherein the node information includes the first configuration information and the hardware information of the intelligent computing resource of the node, wherein the first configuration information includes the attribute information and version information of multiple intelligent computing resource management tools currently included in the node;

[0119] The configuration information generation module 202 is used to generate the second configuration information of the node in the target tool library based on the hardware information obtained by the node information acquisition module 201; the target tool library includes multiple target version intelligent computing resource management tools required for automatic management of multiple intelligent computing resources, and the corresponding relationship between the hardware information of the multiple intelligent computing resources and the multiple target version intelligent computing resource management tools; the second configuration information includes the attribute information and version information of the multiple target version intelligent computing resource management tools;

[0120] A configuration information comparison module 203 is used to compare the first configuration information with the second configuration information to obtain configuration difference information between the first configuration information and the second configuration information;

[0121] A tool push module 204 is used to push one or more corresponding target version intelligent computing resource management tools in the target tool library to the node according to the configuration difference information;

[0122] A tool configuration module 205 is used to install or update the one or more target version intelligent computing resource management tools pushed by the tool push module to the node;

[0123] The management module 206 is used to automatically manage the intelligent computing resources using the intelligent computing resource management tool at the node.

[0124] Optionally, the configuration information comparison module 203 is specifically used to: if the second configuration information contains attribute information and version information of the intelligent computing resource management tool that does not exist in the first configuration information, determine the attribute information and version information of the non-existent intelligent computing resource management tool as configuration difference information;

[0125] Optionally, the tool configuration module 205 is used to: push one or more target version intelligent computing resource management tools corresponding to the target tool library to the node according to the configuration difference information; if the node has the same intelligent computing resource management tool as the target version intelligent computing resource management tool, but the version is different, then update the intelligent computing resource management tool at the node to the version of the target version intelligent computing resource management tool; if the node does not have the same intelligent computing resource management tool as the target version intelligent computing resource management tool, then install the target version intelligent computing resource management tool;

[0126] Optionally, the management module 206 is specifically used to: use the intelligent computing power resource management tool at the node to create the intelligent computing power resource as an intelligent computing power resource instance; use the intelligent computing power resource management tool at the node to dynamically allocate and schedule the intelligent computing power resource instance.

[0127] Optionally, the device further comprises:

[0128] A construction module is used to construct a target tool library; specifically used to obtain hardware information of multiple intelligent computing resources, and obtain multiple target versions of intelligent computing resource management tools required for automatic management of the multiple intelligent computing resources; store the multiple target versions of intelligent computing resource management tools in the target tool library; establish a correspondence between the hardware information of the same type of intelligent computing resources and the multiple target versions of intelligent computing resource management tools; and store the correspondence in the target tool library.

[0129] A label module, used to add a label to the node according to the hardware information of the intelligent computing resource, wherein the hardware information of the intelligent computing resource at least includes the intelligent computing resource type of the node and the intelligent computing resource model of the node; the label is used to identify and allocate the intelligent computing resource when automatically managing the intelligent computing resource;

[0130] A monitoring module, used to monitor whether a target version of the intelligent computing power management tool in the target tool library is updated;

[0131] The automatic update configuration module is used to automatically push the updated target version of the intelligent computing resource management tool to the node for installation or update if it is detected that the target version of the intelligent computing resource management tool in the target tool library is updated.

[0132] In addition, an embodiment of the present application also provides an intelligent computing resource management device, which includes a processor and a memory.

[0133] The memory is used to store computer program code and transmit the program code to the processor;

[0134] The processor is used to execute the steps of the intelligent computing power resource management method introduced in any of the above method embodiments according to the instructions in the program code.

[0135] In addition, an embodiment of the present application also provides a computer-readable storage medium, in which a computer program is stored. When the program is executed by a processor, the intelligent computing power resource management method described in any of the method embodiments is implemented.

[0136] It should be noted that each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The device embodiment described above is merely schematic, in which the unit described as a separate component may or may not be physically separated, and the component prompted as a unit may or may not be a physical unit, that is, it may be located in one place, or it may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative work.

[0137] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An intelligent computing resource management method, characterized in that: The method comprises: Obtaining node information of an intelligent computing resource node, the node information including first configuration information and hardware information of the intelligent computing resource of the node, the first configuration information including attribute information and version information of multiple intelligent computing resource management tools currently included in the node; Based on the hardware information, second configuration information of the node is generated in a target tool library; the target tool library includes multiple target version intelligent computing resource management tools required for automatic management of multiple intelligent computing resources, and a correspondence between the hardware information of the multiple intelligent computing resources and the multiple target version intelligent computing resource management tools; the second configuration information includes attribute information and version information of the multiple target version intelligent computing resource management tools; Comparing the first configuration information with the second configuration information to obtain configuration difference information between the first configuration information and the second configuration information; Push one or more corresponding target version intelligent computing resource management tools in the target tool library to the node and install or update them according to the configuration difference information; The intelligent computing power resource management tool at the node is used to automatically manage the intelligent computing power resources.

2. The method according to claim 1, characterized in that The target tool library is constructed in the following manner: Obtaining hardware information of multiple intelligent computing resources, and obtaining multiple target versions of intelligent computing resource management tools required for automatic management of the multiple intelligent computing resources; Storing the multiple target versions of intelligent computing resource management tools in the target tool library; Establishing a correspondence between hardware information of the same intelligent computing resource and the multiple target versions of intelligent computing resource management tools; The corresponding relationship is stored in the target tool library.

3. The method according to claim 1, characterized in that After obtaining the node information of the intelligent computing power resource node, the method further includes: According to the hardware information of the intelligent computing power resource, a label is added to the node, and the hardware information of the intelligent computing power resource at least includes the intelligent computing power resource type of the node and the intelligent computing power resource model of the node; the label is used to identify and allocate the intelligent computing power resource when automatically managing the intelligent computing power resource.

4. The method according to claim 1, characterized in that The comparing the first configuration information and the second configuration information to obtain configuration difference information between the first configuration information and the second configuration information includes: If the second configuration information contains attribute information and version information of the intelligent computing resource management tool that does not exist in the first configuration information, the attribute information and version information of the non-existent intelligent computing resource management tool is determined as configuration difference information.

5. The method according to claim 1, characterized in that The step of pushing one or more corresponding target version intelligent computing resource management tools in the target tool library to the node and installing or updating the tools according to the configuration difference information includes: Pushing one or more corresponding target version intelligent computing resource management tools in the target tool library to the node according to the configuration difference information; If the node has the same intelligent computing power resource management tool as the target version intelligent computing power resource management tool, but the version is different, then the intelligent computing power resource management tool at the node is updated to the version of the target version intelligent computing power resource management tool; If the node does not have an intelligent computing power resource management tool that is identical to the target version of the intelligent computing power resource management tool, the target version of the intelligent computing power resource management tool is installed.

6. The method according to any one of claims 1 to 5, characterized in that: Also includes: If it is monitored that the target version of the intelligent computing power resource management tool in the target tool library is updated, the updated target version of the intelligent computing power resource management tool is automatically pushed to the node for installation or update.

7. The method according to claim 1, characterized in that The using the intelligent computing resource management tool at the node to automatically manage the intelligent computing resources includes: Using the intelligent computing power resource management tool at the node to create the intelligent computing power resource as an intelligent computing power resource instance; The intelligent computing resource management tool at the node is used to dynamically allocate and schedule the intelligent computing resource instance.

8. An intelligent computing power management device, characterized in that: The device comprises: A node information acquisition module, used to acquire node information of an intelligent computing resource node, wherein the node information includes first configuration information and hardware information of the intelligent computing resource of the node, wherein the first configuration information includes attribute information and version information of multiple intelligent computing resource management tools currently included in the node; A configuration information generation module, configured to generate second configuration information of the node in a target tool library based on the hardware information; the target tool library includes multiple target version intelligent computing resource management tools required for automatic management of multiple intelligent computing resources, and a correspondence between the hardware information of the multiple intelligent computing resources and the multiple target version intelligent computing resource management tools; the second configuration information includes attribute information and version information of the multiple target version intelligent computing resource management tools; A configuration information comparison module, used to compare the first configuration information and the second configuration information to obtain configuration difference information between the first configuration information and the second configuration information; A tool push module, used to push one or more corresponding target version intelligent computing resource management tools in the target tool library to the node according to the configuration difference information; A tool configuration module, used to install or update the one or more target version intelligent computing resource management tools pushed by the tool push module to the node; A management module is used to automatically manage the intelligent computing resources using the intelligent computing resource management tool at the node.

9. An intelligent computing resource management device, characterized in that: The device comprises: a memory having a computer program stored thereon; A processor is used to execute the computer program in the memory to implement the steps of the intelligent computing resource management method described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the intelligent computing resource management method described in any one of claims 1-7 are implemented.

Images