Calculation power aggregation method and device for common calculation power intelligent computing cloud platform

Abstract

The invention provides a computing power aggregation method and device of a common computing power intelligent computing cloud platform, and relates to the technical field of intelligent computing centers, intelligent computing centers and computing power infrastructure, the intelligent computing cloud platform comprises a logic computing power cluster and a plurality of physical computing power clusters, each physical computing power cluster comprises a plurality of computing power nodes, and each physical computing power cluster comprises a plurality of computing power nodes. The method comprises the steps that a logic computing power cluster synchronizes information of each computing power node in a physical computing power cluster in real time, the computing power nodes are mapped into virtual computing power resources, and the virtual computing power resources are put into a computing power resource pool; receiving a container creation request or a calculation task creation request of a user, and creating a container based on the request; based on the information of each virtual computing power resource in the computing power resource pool, screening out a target virtual computing power resource matched with the container from the computing power resource pool; and determining a target computing power node corresponding to the target virtual computing power resource, and scheduling the target computing power node operation container through a physical computing power cluster where the target computing power node is located.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent computing cloud platform and computing power infrastructure, and particularly relates to a computing power aggregation method and device for a facially inclusive computing power intelligent computing cloud platform. BACKGROUND

[0002] With the rapid development of artificial intelligence technology, an intelligent computing cloud platform emerges as the times require.

[0003] The intelligent computing cloud platform refers to a cloud computing platform that provides intelligent computing services based on hardware resources and software resources of an intelligent computing center.

[0004] Intelligent computing refers to computing driven by intelligent computing power (GPU, FPGA, ASIC, etc.), a general term for computing theory methods, architecture systems and technical capabilities with characteristics such as data-intensive and model-based, and aims to obtain better results by using specific computing power resources and appropriate computing methods according to the needs of tasks.

[0005] The intelligent computing center refers to a facility that provides required computing power, data and algorithms for artificial intelligence applications (such as artificial intelligence deep learning model development, model training and model inference scenarios) by using large-scale heterogeneous computing power resources, including general computing power (CPU) and intelligent computing power (GPU, FPGA, ASIC, etc.). The intelligent computing center covers facilities, hardware and software, and can provide full-stack capabilities from bottom computing power to top application enablement.

[0006] Computing power is the core of the intelligent computing cloud platform, is the ability of computer equipment or computing / data center to process information, is the ability of computer hardware and software to cooperate to execute certain computing needs, is the computing power to achieve target result output by processing information data, and is a new type of productivity integrating information computing power, network carrying capacity and data storage power, which mainly provides services to society through computing power infrastructure.

[0007] At present, a single computing power cluster cannot meet the increasing computing power demand of users in an intelligent computing center. The intelligent computing center can integrate multiple computing power clusters together through a scheduling center to provide computing power services for users. However, in order to perform computing power services and schedule computing power nodes, a related scheduling program, business service and the like need to be run on each computing power cluster. When the computing power nodes of a computing power cluster are difficult to meet the computing power demand of a certain computing power service, the scheduling center also needs to coordinate multiple computing power clusters to complete the computing power service, at this time, each computing power cluster participating in coordination needs to cooperate with the scheduling center remotely. With the increase of the scale of the computing power clusters managed by the scheduling center, the management cost and operation and maintenance cost of the computing power clusters will increase exponentially, and for users, it is also difficult to realize the cross-computing power cluster computing power use without feeling, which increases the cost of users using computing power. SUMMARY

[0008] The application provides a computing power aggregation method and device for a universal computing power intelligent computing cloud platform, which is used to solve the problem of high management cost and use cost of computing power clusters in the prior art.

[0009] In order to solve the above technical problems, the application is implemented as follows:

[0010] In a first aspect, the application provides a computing power aggregation method for a universal computing power intelligent computing cloud platform, the intelligent computing cloud platform comprising a logical computing power cluster and a plurality of physical computing power clusters connected with the logical computing power cluster, each of the physical computing power clusters comprising a plurality of computing power nodes, the method comprising:

[0011] Step S1: The logical computing power cluster synchronizes information of each computing power node in the physical computing power cluster in real time, maps the computing power node into a virtual computing power resource, and puts the virtual computing power resource into a computing power resource pool, wherein the computing power node and the virtual computing power resource in the computing power resource pool are one-to-one corresponding.

[0012] Step S2: The logical computing power cluster receives a container creation request or a computing task creation request of a user; and creates a container based on the container creation request or the computing task creation request;

[0013] Step S3: The logical computing power cluster filters out a target virtual computing power resource matched with the container from the computing power resource pool based on the information of each virtual computing power resource in the computing power resource pool maintained;

[0014] Step S4: The logical computing power cluster determines a target computing power node corresponding to the target virtual computing power resource, and schedules the target computing power node to run the container through the physical computing power cluster where the target computing power node is located.

[0015] Optionally, the step S3 comprises:

[0016] Step S3.1: in the case that the container created by the logical computing power cluster is a single container, the logical computing power cluster screens a target virtual computing power resource matching the single container from the computing power resource pool; in the case that the container created by the logical computing power cluster includes a multi-container customized resource, the logical computing power cluster converts the multi-container customized resource into a plurality of associated single containers or a plurality of independent single containers, and respectively screens a target virtual computing power resource matching each single container from the computing power resource pool.

[0017] Optionally, the information of the virtual computing power resource comprises at least one of:

[0018] an occupancy rate of the virtual computing power resource;

[0019] a usage rate of the virtual computing power resource;

[0020] whether the virtual computing power resource runs other containers;

[0021] a capacity of the virtual computing power resource;

[0022] a computing power node corresponding to the virtual computing power resource;

[0023] a model of the computing power node corresponding to the virtual computing power resource;

[0024] performance information of the computing power node corresponding to the virtual computing power resource;

[0025] wherein the information of the computing power node comprises at least one of:

[0026] a physical computing power cluster where the computing power node is located;

[0027] whether the computing power node runs a container;

[0028] an occupancy rate of the computing power node;

[0029] a usage rate of the computing power node;

[0030] a model of the computing power node;

[0031] a capacity of the computing power node;

[0032] performance information of the computing power node.

[0033] Optionally, the step S4 further comprises:

[0034] Step S5: The logical computing power cluster receives a container update request of a user, determines whether the virtual computing power resource needs to be replaced for the container based on the container update request, if the virtual computing power resource needs to be replaced for the container, screens a new target virtual computing power resource matching the container from the computing power resource pool based on the information of each virtual computing power resource in the computing power resource pool, determines a new target computing power node corresponding to the new target virtual computing power resource, and migrates the container to the new target computing power node for running through the physical computing power cluster;

[0035] And / or, the logical computing power cluster receives a container deletion request of a user, determines a target computing power node running the container based on the container deletion request, and schedules the target computing power node to delete the container through the physical computing power cluster where the target computing power node is located.

[0036] And / or, the logical computing power cluster receives a computing task update request of a user, determines whether the virtual computing power resource needs to be replaced for the container corresponding to the computing task based on the computing task update request, if the virtual computing power resource needs to be replaced for the container, screens a new target virtual computing power resource matching the container from the computing power resource pool based on the information of each virtual computing power resource in the computing power resource pool, determines a new target computing power node corresponding to the new target virtual computing power resource, and migrates the container to the new target computing power node for running through the physical computing power cluster.

[0037] And / or, the logical computing power cluster receives a computing task deletion request of a user, determines the container corresponding to the computing task based on the computing task deletion request, and determines a target computing power node running the container, and schedules the target computing power node to delete the container through the physical computing power cluster where the target computing power node is located.

[0038] Optionally, before the step S1, the method further comprises:

[0039] Step S0: The logical computing power cluster provides a plurality of logical computing power cluster external interfaces for a user, and the user can perform container creation, update and / or deletion operation through the logical computing power cluster external interfaces.

[0040] And / or, the logical computing power cluster provides other business service modules for a user, and the user can perform computing task creation, update and / or deletion operation through the other business service modules.

[0041] Optionally, the plurality of physical computing power clusters include at least two computing power clusters belonging to different intelligent computing centers.

[0042] In a second aspect, the present application provides a computing power aggregation device of a computing power intelligent computing cloud platform for universal access, comprising:

[0043] A synchronization module is configured to synchronize information of each computing power node in the physical computing power cluster in real time, map the computing power node into a virtual computing power resource, and put the virtual computing power resource into a computing power resource pool, wherein the computing power node corresponds to the virtual computing power resource in the computing power resource pool one by one.

[0044] A container creation module is configured to receive a container creation request or a computing task creation request of a user, and create a container based on the container creation request or the computing task creation request.

[0045] A computing power resource screening module is configured to screen a target virtual computing power resource matched with the container from the computing power resource pool based on information of each virtual computing power resource in the computing power resource pool.

[0046] A scheduling module is configured to determine a target computing power node corresponding to the target virtual computing power resource, and schedule the target computing power node to run the container through a physical computing power cluster where the target computing power node is located.

[0047] In a third aspect, the present application provides an electronic device, comprising a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program is executed by the processor to implement the steps of the computing power aggregation method of the computing power intelligent computing cloud platform for universal access according to the first aspect.

[0048] In a fourth aspect, the present application provides a computer readable storage medium, wherein a computer program is stored in the computer readable storage medium, and the computer program is executed by a processor to implement the steps of the computing power aggregation method of the computing power intelligent computing cloud platform for universal access according to the first aspect.

[0049] In a fifth aspect, the present application provides a computer program product comprising computer instructions, wherein the computer instructions are executed by a processor to implement the steps of the computing power aggregation method of the computing power intelligent computing cloud platform for universal access according to the first aspect.

[0050] In the present application, on one hand, a plurality of physical computing power clusters are virtually formed into a large logical computing power cluster, the computing power nodes in the computing power cluster are virtually formed into a virtual computing power resource through the logical computing power cluster, and are put into a computing power resource pool, and each virtual computing power resource in the computing power resource pool is synchronized and maintained in real time, that is, the logical computing power cluster regards each computing power node in each computing power cluster as an independent virtual computing power resource in the entire logical computing power cluster, and manages each virtual computing power resource equally.

[0051] On the other hand, after receiving the container creation request or the computing task creation request of the user, the logical computing cluster creates a container, filters a target virtual computing resource matching the container from the computing resource pool, determines a target computing node corresponding to the target virtual computing resource, and schedules the target computing node to run the container through the physical computing cluster where the target computing node is located, that is, the logical computing cluster completes the whole process of container creation, virtual computing resource filtering, computing node mapping and calling, and the physical computing cluster only needs to run the container according to the calling, and the physical computing cluster does not need to run related calling programs and business services and other services on the physical computing cluster. Even if the computing node is scheduled across the computing clusters, the physical computing cluster does not need to coordinate with the logical computing cluster across the computing clusters, and the management cost and operation and maintenance cost of the physical computing cluster will only increase linearly with the increase of the scale of the physical computing cluster. At the same time, the cross-cluster computing usage of the user can be realized at low cost, the economic cost is greatly reduced, and the wide application of the universal computing power is promoted. BRIEF DESCRIPTION OF DRAWINGS

[0052] Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The accompanying drawings are included to provide a description of the preferred embodiments and are not intended to limit the scope of the application. Moreover, the same reference numerals are used throughout the same figures. In the drawings:

[0053] Figure 1 FIG. 1 is a structural schematic diagram of the intelligent computing cloud platform of the present application.

[0054] Figure 2 FIG. 2 is a flowchart of the computing power aggregation method of the intelligent computing cloud platform for universal computing power of the present application.

[0055] Figure 3 FIG. 3 is a schematic diagram of the computing resource pool of the present application.

[0056] Figure 4 FIG. 4 is a structural schematic diagram of the computing power aggregation device of the intelligent computing cloud platform for universal computing power of the present application.

[0057] Figure 5 FIG. 5 is a structural schematic diagram of the electronic device of the present application. DETAILED DESCRIPTION

[0058] The technical solutions in the present application will be described in detail below with reference to the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0059] Firstly, the technical terms related to the present application are simply explained below.

[0060] The "computing power" described in the present application refers to: the ability of computer equipment or computing / data center to process information, the ability of computer hardware and software to jointly perform certain computing requirements, the computing power to achieve target result output by processing information data, a new type of productivity integrating information computing power, network carrying capacity and data storage power, mainly providing services to society through computing power infrastructure.

[0061] The "computing power" (Computational Power, CP) described in the present application refers to: a kind of ability of data center server to process data and realize result output, a comprehensive index for measuring the computing power of data center, including general computing power, supercomputing power and intelligent computing power. The commonly used unit of measurement is the number of floating point operations per second (FLOPS, 1EFLOPS=10^18 FLOPS), and the larger the value represents the stronger the comprehensive computing power. According to the estimation, 1 EFLOPS is about 5 Tianhe 2A or 500,000 mainstream server CPUs or 2 million mainstream notebook computing power output. The calculation formula is: CP=CP 通用 +CP 智能 +CP 超级 .

[0062] The "carrying capacity" (Network Power, NP) described in the present application refers to: the performance of data transmission capacity of computing power infrastructure, including network architecture, network bandwidth, transmission delay, intelligent management and scheduling, and other comprehensive capabilities, involving network transmission within and between data centers, and is a comprehensive index for measuring network transmission scheduling capability.

[0063] The "storage power" (Storage Power, SP) described in the present application refers to: the comprehensive ability of data center in four aspects of data storage capacity, performance, safety and reliability, and green low carbon, which is a comprehensive index for measuring the data storage capacity of data center, including external storage devices such as storage arrays and server built-in storage devices. The commonly used unit of measurement for storage capacity is exabyte (EB, 1EB=2^60bytes), the commonly used unit of measurement for performance is the number of read / write per second per unit capacity (IOPS / TB, Input / Output Operations Per Second / TB), and the disaster recovery ratio is an important performance of safety and reliability.

[0064] The "computing power infrastructure" described in the present application refers to: a new type of information infrastructure integrating information computing power, network carrying capacity and data storage power, which can realize centralized computing, storage, transmission and application of information.

[0065] The "new information infrastructure" described in the present application refers to: mainly including network infrastructure such as 5G network, fiber broadband network, backbone network, international communication network, satellite Internet, computing power infrastructure such as data center, general computing center, intelligent computing center, supercomputing center, and new technology facilities such as artificial intelligence, blockchain, quantum computing.

[0066] The "computing power" described in the present application includes: general computing power, intelligent computing power and supercomputing power.

[0067] The "general computing power" described in the present application refers to the computing power provided by servers based on CPU (Central Processing Unit) chips, which is used to support basic general computing such as cloud computing and edge computing.

[0068] The "intelligent computing power" described in the present application refers to: for various types of artificial intelligence innovation applications, based on GPU (Graphics Processing Unit), FPGA (Field Programmable Gate Array), ASIC (Application Specific Integrated Circuit) and other special chip large-scale deployment of computing platform, such as natural language processing, machine vision, etc.

[0069] The "supercomputing power" described in the present application refers to: the computing power provided by high-performance computing clusters such as supercomputers, which utilizes centralized computing resources of multiple computer systems working in parallel and through a dedicated operating system to process extremely complex or data-intensive problems, mainly used for computing in cutting-edge scientific fields, such as planet simulation, drug molecule design, gene analysis, etc.

[0070] The "intelligent computing" described in the present application refers to: computing driven by intelligent computing power (GPU, FPGA, ASIC, etc.), a general term for computing theory methods, architecture systems and technical capabilities with characteristics such as data-intensive and model-based, and its goal is to obtain better results according to the needs of the task by using specific computing power resources and appropriate computing methods. The "intelligent computing cloud platform" described in the present application refers to: an intelligent computing cloud platform based on the hardware resources and software resources of the intelligent computing center to provide intelligent computing services.

[0071] The "intelligent computing center" described in the application refers to: by using large-scale heterogeneous computing resources, including general computing power (CPU) and intelligent computing power (GPU, FPGA, ASIC, etc.), mainly for artificial intelligence applications (such as artificial intelligence deep learning model development, model training and model inference, etc.) Provide the required computing power, data and algorithm facilities. The intelligent computing center covers facilities, hardware and software, and can provide full-stack capabilities from bottom computing to top application enablement.

[0072] The "intelligent computing center" described in the application includes but is not limited to "intelligent computing center".

[0073] The "intelligent computing center" described in the application, namely the artificial intelligence computing center, is based on artificial intelligence theory, adopts artificial intelligence computing architecture, and provides artificial intelligence application required computing power service, data service and algorithm service. A kind of computing power infrastructure.

[0074] The "computing power center" described in the application refers to: with infrastructure such as wind, fire, water and electricity and IT hardware and software equipment as the main component, with computing power, carrying capacity and storage capacity. Facilities, including general data centers, intelligent computing centers, supercomputing centers, etc.

[0075] The "supercomputing center" described in the application refers to: supercomputing data center, which is based on supercomputer or large-scale computing cluster data center, which can provide large-scale computing, storage and network services, etc. Function, widely used in aerospace, national defense, oil exploration, climate modeling and genome sequencing, etc. Application scenarios.

[0076] The "inclusive computing power" described in the application refers to: based on the requirements of equal opportunities and the principle of commercial sustainability, providing appropriate and effective computing power services to all social strata and groups with computing power service needs at an affordable cost.

[0077] The "computing power cluster" described in the application refers to: a collection of multiple computing nodes, storage nodes and network nodes, used for information computing, network carrying and data storage tasks.

[0078] The "logical computing power cluster" of the application refers to: a large cluster set virtually composed of multiple physical computing power clusters.

[0079] The "computing power resource" described in the application refers to: the technology and facilities required for the development of digital society, including but not limited to computing resources such as CPU and GPU, network resources such as switches and routers, storage resources such as storage arrays and distributed storage, security resources such as firewalls and intrusion detection systems, and supporting resources such as wind, fire, water and electricity.

[0080] The "computing power node" described in the application refers to the computing resources of a server capable of processing computing tasks.

[0081] Please refer to Figure 1 , Figure 1 The structural schematic diagram of the intelligent computing cloud platform of the application, the intelligent computing cloud platform comprises: a logical computing power cluster and a plurality of physical computing power clusters connected with the logical computing power cluster, the logical computing power cluster refers to a large cluster set virtually formed by a plurality of physical computing power clusters. Each of the physical computing power clusters comprises a plurality of computing power nodes.

[0082] In the application, the logical computing power cluster completes all processes of computing power node synchronization, mapping, container creation, virtual computing power resource screening and computing power node calling. However, the physical computing power cluster only needs to run a resource synchronizer service irrelevant to business, which is responsible for synchronizing container and computing power node information with the scheduling center of the logical computing power cluster, and is simple in logic and low in maintenance cost.

[0083] The functions of each module (scheduling center, custom resource service, other business service, logical computing power cluster external interface and computing power cluster resource synchronizer) in the logical computing power cluster in the application will be described in detail in the subsequent content.

[0084] Please refer to Figure 2 The application also provides a computing power aggregation method for the intelligent computing cloud platform of the application, the intelligent computing cloud platform comprises a logical computing power cluster and a plurality of physical computing power clusters connected with the logical computing power cluster, each of the physical computing power clusters comprises a plurality of computing power nodes, and the method comprises the following steps:

[0085] Step S1: The logical computing power cluster synchronizes information of each computing power node in the physical computing power cluster in real time, maps the computing power node into a virtual computing power resource, and puts the virtual computing power resource into a computing power resource pool, wherein the computing power node and the virtual computing power resource in the computing power resource pool correspond to each other.

[0086] In the application, optionally, the scheduling center in the logical computing power cluster synchronizes information of each computing power node in the physical computing power cluster in real time through a computing power cluster resource synchronizer, maps the computing power node into a virtual computing power resource, and puts the virtual computing power resource into a computing power resource pool.

[0087] In the application, please refer to Figure 3 , Figure 3 The schematic diagram of the computing power resource pool of the application, the logical computing power cluster maintains the computing power resource pool, the computing power resource pool comprises virtual computing power resources corresponding to all computing power nodes in all computing power clusters, each computing power node serves as an independent computing power resource in the entire logical computing power cluster, and the logical computing power cluster equally manages the computing power nodes.

[0088] Optionally, the information of the computing power node includes at least one of the following: a physical computing power cluster where the computing power node is located, whether the computing power node runs a container, an occupancy rate of the computing power node, a usage rate of the computing power node, a model of the computing power node, a capacity of the computing power node, and performance information of the computing power node.

[0089] The capacity of the computing power node includes, for example, a display memory capacity, a memory capacity, and the like.

[0090] The performance information of the computing power node may, for example, include at least one of the following: a computing power, a display memory bandwidth, an energy efficiency ratio, and a specific application scenario (such as a game, AI training, and the like).

[0091] Step S2: The logical computing power cluster receives a container creation request or a computing task creation request of a user, and creates a container based on the container creation request or the computing task creation request.

[0092] The container in the present application is a lightweight virtualization technology, which allows an application and its dependencies to be packaged into a portable container that can run on any server with a Docker engine installed, including Linux and Windows machines.

[0093] Optionally, the following steps are performed by a scheduling center in the logical computing power cluster: receiving a container creation request or a computing task creation request of a user, and creating a container based on the container creation request or the computing task creation request.

[0094] In the present application, if the user sends a container creation request, the logical computing power cluster directly creates a container according to the information of the container carried in the container creation request. If the user sends a computing task creation request, the information of the container to be created is determined according to the information in the computing task creation request, so as to create the container.

[0095] Step S3: The logical computing power cluster filters out a target virtual computing power resource matched with the container from the computing power resource pool based on the information of each virtual computing power resource in the maintained computing power resource pool.

[0096] Optionally, the above step S3 is performed by a scheduling center in the logical computing power cluster.

[0097] Step S4: The logical computing power cluster determines a target computing power node corresponding to the target virtual computing power resource, and schedules the target computing power node to run the container through a physical computing power cluster where the target computing power node is located.

[0098] Optionally, the above step S4 is performed by a scheduling center in the logical computing power cluster.

[0099] In the present application, on the one hand, a plurality of physical computing power clusters are virtually formed into a large logical computing power cluster, the computing power nodes in the computing power cluster are virtually formed into a virtual computing power resource by the logical computing power cluster, and are put into a computing power resource pool, and each virtual computing power resource in the computing power resource pool is synchronized and maintained in real time, that is, the logical computing power cluster regards each computing power node in each computing power cluster as an independent virtual computing power resource in the entire logical computing power cluster, and manages each virtual computing power resource equally.

[0100] On the other hand, after the logical computing power cluster receives a container creation request or a computing task creation request of a user, the container is created, the target virtual computing power resource matched with the container is screened out from the computing power resource pool, the target computing power node corresponding to the target virtual computing power resource is determined, and the target computing power node is dispatched to run the container by the physical computing power cluster where the target computing power node is located, that is, the logical computing power cluster completes the whole process of container creation, virtual computing power resource screening, computing power node mapping and calling, and the physical computing power cluster only needs to run the container according to the calling, and the related calling program and business service and other services do not need to be run on the physical computing power cluster, even if the computing power nodes are dispatched across the computing power clusters, the physical computing power cluster also does not need to coordinate with the logical computing power cluster across the computing power clusters, and the management cost and operation and maintenance cost of the physical computing power cluster will only increase linearly with the increase of the size of the physical computing power cluster, at the same time, the cross-cluster computing power use of the user can be realized at low cost, the economic cost is greatly reduced, and the wide application of the universal computing power is promoted.

[0101] In the present application, when the logical computing power cluster creates a container, the container can be created based on the preset resources such as Pod, Deployment, Service, Volume provided by the logical computing power cluster, however, in some special scenarios, the existing preset resources cannot meet the needs of the business, and the resources required by the container can be created through a custom resource (Custom Resource Definition, CRD).

[0102] The resource created through the custom resource can be a multi-container custom resource.

[0103] In the present application, optionally, the step S3 comprises:

[0104] Step S3.1: in the case that the container created by the logical computing power cluster is a single container, the logical computing power cluster screens a target virtual computing power resource matching the single container from the computing power resource pool; in the case that the container created by the logical computing power cluster includes a multi-container customized resource, the logical computing power cluster converts the multi-container customized resource into a plurality of associated single containers or a plurality of independent single containers, and respectively screens a target virtual computing power resource matching each single container from the computing power resource pool.

[0105] Optionally, in the present application, the plurality of associated single containers can be scheduled in one computing power cluster or across computing power clusters. However, the plurality of independent single containers can be scheduled in one computing power cluster.

[0106] Optionally, in the present application, the step of converting the multi-container customized resource into a plurality of associated single containers or a plurality of independent single containers in the step S3.1 is performed by a customized resource service of the logical computing power cluster. Other steps in the step S3.1 are performed by a scheduling center of the logical computing power cluster.

[0107] In the present application, for a single container, the logical computing power cluster can directly perform efficient single container scheduling. For a multi-container customized resource, the logical computing power cluster converts the multi-container customized resource into a plurality of associated single containers or a plurality of independent single containers. For a computing power cluster, the processing is all single containers, and the management is simple.

[0108] Optionally, in the present application, the information of the virtual computing power resource includes at least one of the following:

[0109] The occupancy rate of the virtual computing power resource;

[0110] The usage rate of the virtual computing power resource;

[0111] Whether the virtual computing power resource runs other containers;

[0112] The capacity of the virtual computing power resource;

[0113] The computing power node corresponding to the virtual computing power resource;

[0114] The model of the computing power node corresponding to the virtual computing power resource;

[0115] The performance information of the computing power node corresponding to the virtual computing power resource.

[0116] Optionally, in the present application, the strategy of screening the target virtual computing power resource can include at least one of the following: selecting a virtual computing power resource not running other containers, selecting a virtual computing power resource with an occupancy rate less than a preset threshold, and selecting a virtual computing power resource of a certain model.

[0117] Optionally, the step S4 further comprises:

[0118] Step S5: The logical computing power cluster receives a container update request of a user, determines whether the virtual computing power resource needs to be replaced for the container based on the container update request, if the virtual computing power resource needs to be replaced for the container, screens out a new target virtual computing power resource matched with the container from the computing power resource pool based on the information of each virtual computing power resource in the computing power resource pool, determines a new target computing power node corresponding to the new target virtual computing power resource, and migrates the container to the new target computing power node for running through the physical computing power cluster;

[0119] And / or, the logical computing power cluster receives a container deletion request of a user, determines a target computing power node running the container based on the container deletion request, and schedules the target computing power node to delete the container through the physical computing power cluster where the target computing power node is located;

[0120] And / or, the logical computing power cluster receives a computing task update request of a user, determines whether the virtual computing power resource needs to be replaced for the container corresponding to the computing task based on the computing task update request, if the virtual computing power resource needs to be replaced for the container, screens out a new target virtual computing power resource matched with the container from the computing power resource pool based on the information of each virtual computing power resource in the computing power resource pool, determines a new target computing power node corresponding to the new target virtual computing power resource, and migrates the container to the new target computing power node for running through the physical computing power cluster;

[0121] And / or, the logical computing power cluster receives a computing task deletion request of a user, determines the container corresponding to the computing task based on the computing task deletion request, and determines a target computing power node running the container, and schedules the target computing power node to delete the container through the physical computing power cluster where the target computing power node is located.

[0122] Through the above steps, the update and deletion of the container can be realized.

[0123] Optionally, the step S1 further comprises:

[0124] Step S0: The logical computing power cluster provides a plurality of logical computing power cluster external interfaces for a user, and the user can perform the creation, update and / or deletion operation of the container through the logical computing power cluster external interface;

[0125] And / or, the logic computing cluster provides a user with other business service modules, and the user can create, update and / or delete a computing task through the other business service modules.

[0126] In the application, for a user who wants to use the logic computing cluster external interface, it is not necessary to face multiple physical computing clusters, but only needs to connect with one logic computing cluster, thereby improving the development efficiency of the user and saving the development cost of the user.

[0127] In the application, optionally, the multiple physical computing clusters include at least two computing clusters belonging to different intelligent computing centers, that is, the computing clusters in the logic computing cluster of the application can be located in the same intelligent computing center or in different intelligent technology centers.

[0128] In the application, when the user uses the computing node of the intelligent computing cloud platform, it is not necessary to pay attention to which intelligent technology center the computing node is located in, or which physical computing cluster the computing node is located in, or which server the computing node is located in, or which GPU is used, and the user only needs to focus on the business service, thereby greatly reducing the threshold for the user to use the computing node. When the business service of the user runs, it can cross the intelligent technology center or cross the computing cluster, but it is not necessary for the user to sense.

[0129] Please refer to Figure 4 The application further provides an algorithm aggregation device 10 of the algorithm intelligent computing cloud platform, which comprises:

[0130] A synchronization module 11 is configured to synchronize information of each computing node in the physical computing cluster in real time, map the computing node into a virtual algorithm resource, and put the virtual algorithm resource into an algorithm resource pool, wherein the computing node and the virtual algorithm resource in the algorithm resource pool are in one-to-one correspondence.

[0131] A container creation module 12 is configured to receive a container creation request or a computing task creation request of a user, and create a container based on the container creation request or the computing task creation request.

[0132] An algorithm resource screening module 13 is configured to screen a target virtual algorithm resource matched with the container from the algorithm resource pool based on information of each virtual algorithm resource in the algorithm resource pool.

[0133] A scheduling module 14 is configured to schedule a target computing node corresponding to the target virtual algorithm resource, and schedule the target computing node to run the container through a physical computing cluster where the target computing node is located.

[0134] In the present application, on the one hand, a plurality of physical computing power clusters are virtually formed into a large logical computing power cluster, the computing power nodes in the computing power cluster are virtually formed into a virtual computing power resource by the logical computing power cluster, and are put into a computing power resource pool, and each virtual computing power resource in the computing power resource pool is synchronized and maintained in real time, that is, the logical computing power cluster regards each computing power node in each computing power cluster as an independent virtual computing power resource in the entire logical computing power cluster, and equally manages each virtual computing power resource.

[0135] On the other hand, after the logical computing power cluster receives a container creation request or a computing task creation request of a user, a container is created, a target virtual computing power resource matching the container is screened out from the computing power resource pool, a target computing power node corresponding to the target virtual computing power resource is determined, and the target computing power node is dispatched to run the container by the physical computing power cluster where the target computing power node is located, that is, the logical computing power cluster completes the entire process of container creation, virtual computing power resource screening, computing power node mapping and calling, the physical computing power cluster only needs to run the container according to the calling, and the physical computing power cluster does not need to run related calling programs and business services and the like on the physical computing power cluster, even if the computing power nodes are dispatched across the computing power clusters, the physical computing power cluster does not need to coordinate and cooperate with the logical computing power cluster across the computing power clusters, and the management cost and operation and maintenance cost of the physical computing power cluster will only linearly increase with the increase of the size of the physical computing power cluster, while the cross-cluster computing power use of the user can be realized at a low cost, the economic cost is greatly reduced, and the wide application of the universal computing power is promoted.

[0136] In the present application, optionally, the computing power resource screening module 13 is configured to, in a case where the container created by the logical computing power cluster is a single container, screen out a target virtual computing power resource matching the single container from the computing power resource pool; in a case where the container created by the logical computing power cluster includes a plurality of containers of a custom resource, convert the plurality of containers of the custom resource into a plurality of associated single containers or a plurality of independent single containers, and screen out a target virtual computing power resource matching each single container from the computing power resource pool.

[0137] In the present application, optionally, the information of the virtual computing power resource includes at least one of the following:

[0138] The occupancy rate of the virtual computing power resource;

[0139] The usage rate of the virtual computing power resource;

[0140] Whether the virtual computing power resource runs other containers;

[0141] The capacity of the virtual computing power resource;

[0142] The computing power node corresponding to the virtual computing power resource;

[0143] The model of the computing node corresponding to the virtual computing resources;

[0144] The performance information of the computing nodes corresponding to the virtual computing resources.

[0145] In this invention, optionally, the information of the computing node includes at least one of the following:

[0146] The physical computing cluster where the computing power node is located;

[0147] Does the computing node run a container?

[0148] The utilization rate of the computing nodes;

[0149] The utilization rate of the computing nodes;

[0150] The model of the computing node;

[0151] The capacity of the computing nodes;

[0152] The performance information of the computing nodes.

[0153] Optionally, in this invention, the computing power aggregation device 10 of the universal computing power intelligent computing cloud platform further includes:

[0154] The processing module is used to receive a user's container update request, determine whether the virtual computing resources for the container need to be replaced based on the container update request, and if the virtual computing resources for the container need to be replaced, filter out a new target virtual computing resource that matches the container from the computing resource pool based on the information of each virtual computing resource in the maintained computing resource pool; determine a new target computing node corresponding to the new target virtual computing resource, and migrate the container to the new target computing node for operation through the physical computing cluster.

[0155] And / or, the logical computing power cluster receives a user's container deletion request, determines the target computing power node running the container based on the container deletion request, and schedules the target computing power node to delete the container through the physical computing power cluster where the target computing power node is located;

[0156] And / or, the logic computing cluster receives a user's computing task update request, determines whether the virtual computing resource needs to be replaced for the container corresponding to the computing task based on the computing task update request, if the virtual computing resource needs to be replaced for the container, filters out a new target virtual computing resource matched with the container from the computing resource pool based on the information of each virtual computing resource in the computing resource pool; determines a new target computing node corresponding to the new target virtual computing resource, and migrates the container to the new target computing node for running through the physical computing cluster;

[0157] And / or, the logic computing cluster receives a user's computing task update request, determines whether the virtual computing resource needs to be replaced for the container corresponding to the computing task based on the computing task update request, if the virtual computing resource needs to be replaced for the container, filters out a new target virtual computing resource matched with the container from the computing resource pool based on the information of each virtual computing resource in the computing resource pool; determines a new target computing node corresponding to the new target virtual computing resource, and migrates the container to the new target computing node for running through the physical computing cluster.

[0158] In the application, optionally, the computing power aggregation device 10 of the face-to-face universal computing power intelligent computing cloud platform further comprises:

[0159] The interface providing module is configured to provide a plurality of external interfaces of the logic computing cluster for a user, and the user can perform a container creation, update and / or deletion operation through the external interface of the logic computing cluster.

[0160] And / or, the business service body providing module is configured to provide a user with other business service modules, and the user can perform a computing task creation, update and / or deletion operation through the other business service modules.

[0161] In the application, optionally, the plurality of physical computing clusters include at least two computing clusters belonging to different intelligent computing centers.

[0162] Please refer to Figure 5 The application further provides an electronic device 20, which comprises a processor 21, a memory 22, and a computer program stored in the memory 22 and capable of running on the processor 21, wherein the computer program is executed by the processor 21 to implement each process of the computing power aggregation method of the face-to-face universal computing power intelligent computing cloud platform and achieve the same technical effects, and thus repeated description is omitted here.

[0163] The application further provides a computer readable storage medium, which stores a computer program, and each process of the above-mentioned face-to-face universal computing power intelligent cloud platform computing power aggregation method embodiment is realized when the computer program is executed by a processor, and the same technical effects can be achieved.

[0164] The application further provides a computer program product, which comprises computer instructions, and each process of the above-mentioned face-to-face universal computing power intelligent cloud platform computing power aggregation method embodiment is realized when the computer instructions are executed by a processor, and the same technical effects can be achieved. Figure 1 The application further provides a computer program product, which comprises computer instructions, and each process of the above-mentioned face-to-face universal computing power intelligent cloud platform computing power aggregation method embodiment is realized when the computer instructions are executed by a processor, and the same technical effects can be achieved.

[0165] It should be noted that, in this paper, the term "comprise", "include" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitations, the element defined by the statement "comprises a" does not exclude the presence of another identical element in the process, method, article or device including the element.

[0166] From the above description of the embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment method can be realized by means of software and necessary general hardware platform, of course, it can also be realized by hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes a plurality of instructions for making a terminal (which can be a mobile phone, computer, server, air conditioner or network device, etc.) execute the method described in each embodiment of the present application.

[0167] The embodiments of the application are described above in combination with the drawings, but the application is not limited to the above-mentioned specific embodiments, and the above-mentioned specific embodiments are only illustrative, not limiting, and those skilled in the art can make many forms under the inspiration of the present application without departing from the purpose of the present application and the scope protected by the claims, which all belong to the protection of the present application.

Claims

1. A computing power aggregation method for a computing power intelligent computing cloud platform, characterized in that, The intelligent computing cloud platform comprises a logical computing cluster and a plurality of physical computing clusters connected with the logical computing cluster, each of the physical computing clusters comprising a plurality of computing nodes, and the method comprises: Step S1: The logical computing cluster synchronizes information of each computing node in the physical computing cluster in real time, maps the computing nodes into virtual computing resources, and puts the virtual computing resources into a computing resource pool, wherein the computing nodes correspond to the virtual computing resources in the computing resource pool one by one; Step S2: The logical computing cluster receives a container creation request or a computing task creation request of a user, and creates a container based on the container creation request or the computing task creation request; Step S3: The logical computing cluster screens a target virtual computing resource matching the container from the computing resource pool based on information of each virtual computing resource in the computing resource pool maintained by the logical computing cluster; Step S4: The logical computing cluster determines a target computing node corresponding to the target virtual computing resource, and schedules the target computing node to run the container through a physical computing cluster where the target computing node is located.

2. The method of claim 1, wherein, The step S3 comprises: Step S3.1: In a case where the container created by the logical computing cluster is a single container, the logical computing cluster screens a target virtual computing resource matching the single container from the computing resource pool; in a case where the container created by the logical computing cluster comprises a plurality of containers of a custom resource, the logical computing cluster converts the plurality of containers of the custom resource into a plurality of associated single containers or a plurality of independent single containers, and screens a target virtual computing resource matching each single container from the computing resource pool respectively.

3. The method of claim 1, wherein: The information of the virtual computing resource comprises at least one of: an occupancy rate of the virtual computing resource; a usage rate of the virtual computing resource; whether the virtual computing resource runs other containers; a capacity of the virtual computing resource; a computing node corresponding to the virtual computing resource; a model of the computing node corresponding to the virtual computing resource; performance information of the computing node corresponding to the virtual computing resource; The information of the computing node comprises at least one of: a physical computing cluster where the computing node is located; whether the computing node runs a container; an occupancy rate of the computing node; a usage rate of the computing node; a model of the computing node; a capacity of the computing node; performance information of the computing node.

4. The method of claim 1, wherein, The step S4 further comprises: Step S5: The logical computing power cluster receives a container update request of a user, determines whether the virtual computing power resource needs to be replaced for the container based on the container update request, if the virtual computing power resource needs to be replaced for the container, screens a new target virtual computing power resource matching the container from the computing power resource pool based on the information of each virtual computing power resource in the computing power resource pool, determines a new target computing power node corresponding to the new target virtual computing power resource, and migrates the container to the new target computing power node for running through the physical computing power cluster; And / or, the logical computing power cluster receives a container deletion request of a user, determines a target computing power node running the container based on the container deletion request, and schedules the target computing power node to delete the container through the physical computing power cluster where the target computing power node is located; And / or, the logical computing power cluster receives a computing task update request of a user, determines whether the virtual computing power resource needs to be replaced for the container corresponding to the computing task based on the computing task update request, if the virtual computing power resource needs to be replaced for the container, screens a new target virtual computing power resource matching the container from the computing power resource pool based on the information of each virtual computing power resource in the computing power resource pool, determines a new target computing power node corresponding to the new target virtual computing power resource, and migrates the container to the new target computing power node for running through the physical computing power cluster; And / or, the logical computing power cluster receives a computing task deletion request of a user, determines the container corresponding to the computing task based on the computing task deletion request, and determines a target computing power node running the container, and schedules the target computing power node to delete the container through the physical computing power cluster where the target computing power node is located.

5. The method of claim 1, wherein, Before the step S1, further comprising: Step S0: The logical computing power cluster provides a plurality of logical computing power cluster external interfaces for a user, and the user can perform container creation, update and / or deletion operation through the logical computing power cluster external interface; And / or, the logical computing power cluster provides other business service modules for a user, and the user can perform computing task creation, update and / or deletion operation through the other business service modules.

6. The method of claim 1, wherein, The plurality of physical computing power clusters include at least two computing power clusters belonging to different intelligent computing centers.

7. A computing power aggregation device of a computing power intelligent computing cloud platform, characterized in that, Comprise: A synchronization module for synchronizing information of each computing power node in the physical computing power cluster in real time, mapping the computing power node into a virtual computing power resource, and putting the virtual computing power resource into a computing power resource pool, wherein the computing power node and the virtual computing power resource in the computing power resource pool correspond one by one; A container creation module for receiving a container creation request or a computing task creation request of a user; creating a container based on the container creation request or the computing task creation request; A computing power resource screening module for screening a target virtual computing power resource matching the container from the computing power resource pool based on the information of each virtual computing power resource in the computing power resource pool; A container creation module for receiving a container creation request or a computing task creation request of a user; creating a container based on the container creation request or the computing task creation request; The scheduling module is configured to determine a target computing power node corresponding to the target virtual computing power resource, and schedule the target computing power node to run the container through a physical computing power cluster where the target computing power node is located.

8. An electronic device, comprising: The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the steps of the computing power aggregation method of the universal computing power intelligent computing cloud platform according to any one of claims 1 to 6. The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the steps of the computing power aggregation method of the universal computing power intelligent computing cloud platform according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the steps of the computing power aggregation method of the universal computing power intelligent computing cloud platform according to any one of claims 1 to 6.

10. A computer program product, characterised in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the steps of the computing power aggregation method of the universal computing power intelligent computing cloud platform according to any one of claims 1 to 6.

Images