CN109471725A - Resource allocation methods, device and server - Google Patents
Resource allocation methods, device and server Download PDFInfo
- Publication number
- CN109471725A CN109471725A CN201811247681.6A CN201811247681A CN109471725A CN 109471725 A CN109471725 A CN 109471725A CN 201811247681 A CN201811247681 A CN 201811247681A CN 109471725 A CN109471725 A CN 109471725A
- Authority
- CN
- China
- Prior art keywords
- application
- stock number
- resource
- ratio
- distributed
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5061—Partitioning or combining of resources
- G06F9/5077—Logical partitioning of resources; Management or configuration of virtualized resources
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
Present disclose provides a kind of resource allocation methods, device and servers, wherein this method comprises: determining the total resources in Container Management cluster;Determine the ratio for having distributed stock number and total resources of each application in Container Management cluster;Determine the smallest application of ratio;The stock number that preset quantity Pod is distributed for the smallest application of the ratio, updates the distribution stock number of the smallest application of the ratio;Judge whether the sum of stock number of distribution of all applications is greater than total resources;If judging result be it is no, continue through above-mentioned steps distribution resource;If the determination result is YES, it is the resource that each application distributes corresponding amount according to the stock number of having distributed of each application, wherein, the smallest application of ratio distributed stock number be the updated the smallest application of the ratio the stock number distributed stock number and subtracted preset quantity Pod.The disclosure can be liberally multiple allocated resources in cluster, while guarantee that assigned resource can be utilized effective and reasonablely.
Description
Technical field
This disclosure relates to technical field of virtualization, more particularly, to a kind of resource allocation methods, device and server.
Background technique
Container is a kind of lightweight, portable, self-contained Software package technology, and container can make application program almost
It can Anywhere run in an identical manner.Different from traditional virtualization technology, container operates in a certain of operating system
In user's space, other process isolations with operating system are more much smaller than virtual machine in volume.Starting container does not need to start
Whole operation system, thus container deployment and starting speed faster, expense it is smaller, be also easier to migrate.
Kubernetes is a kind of Open Source Platform for automating container operation, can be carried out to container by the platform automatic
The deployment and duplication of change, extension or shrink container scale, in groups and provide container organization to the load balancing between container, are formed and held
Device manages cluster;It can also be easily carried out simultaneously and edition upgrading is carried out to types of applications program containers in cluster, to container
Carry out the functions such as elastic telescopic.
It is carried out in deployment or renewal process to Container Management cluster (can also abbreviation cluster), Kubernetes can be mentioned
For multiple resources allocation management method, for example, predicted according to the resource requirement value that the historical data of container is container, according to
The Resources limit etc. of each container is arranged in prediction result.But these methods are substantially manually as each container configuration tool
The Resources limit of body, and the size of Resources limit is arranged according to historical data or engineer experience.If in Container Management cluster
Operation has multiple applications, this Resources limit mode being arranged according to historical data or engineer experience, be difficult be on the whole
Resource is reasonably distributed between each application, the application resource utilization easily led to is low, and some application resource deficiencies can not be just
Often operation.
Summary of the invention
In view of this, the disclosure is designed to provide a kind of resource allocation methods, device and server, to realize cluster
In be the resource allocation of multiple applications, while guaranteeing that assigned resource can be utilized effective and reasonablely.
To achieve the goals above, the disclosure the technical solution adopted is as follows:
In a first aspect, this method is applied to the main section of Container Management cluster present disclose provides a kind of resource allocation methods
Point;Method comprises determining that the total resources in Container Management cluster;Determine the money of distribution of each application in Container Management cluster
The ratio of source amount and total resources;Each application corresponds at least one Pod, and the stock number of distribution of each application is at least one
The sum of stock number of Pod;Determine the smallest application of ratio;The stock number of preset quantity Pod is distributed for the smallest application of the ratio,
Update the distribution stock number of the smallest application of the ratio;Judge whether the sum of stock number of distribution of all applications is greater than resource
Total amount;If judging result is no, the resource of distribution and total resources of each application in the determining Container Management cluster of return execution
Ratio operation;It if the determination result is YES, is that each application distributes corresponding amount according to the stock number of having distributed of each application
Resource, wherein the smallest application of ratio has distributed the distribution resource that stock number is the updated the smallest application of the ratio
Amount subtracts the stock number of preset quantity Pod.
Second aspect, present disclose provides a kind of resource allocation device, which is set to the main section of Container Management cluster
Point;Device includes: the first determining module, for determining the total resources in Container Management cluster;Second determining module, for true
Constant volume device manages the ratio for having distributed stock number and total resources of each application in cluster;Each application corresponds at least one
Pod, each application have distributed the sum of the stock number that stock number is at least one Pod;Using determining module, for determining ratio
The smallest application;Stock number distribution module is updated for distributing the stock number of preset quantity Pod for the smallest application of the ratio
The distribution stock number of the smallest application of the ratio;Judgment module, the sum of the stock number of distribution for judging all applications are
It is no to be greater than total resources;Return module, if for judging result being no, each application in the determining Container Management cluster of return execution
The ratio for having distributed resource and total resources operation;Resource distribution module, for if the determination result is YES, being answered according to each
Having distributed stock number is the resource that each application distributes corresponding amount, wherein the distribution stock number of the smallest application of ratio
For the stock number of the updated the smallest application of the ratio distributed stock number and subtracted preset quantity Pod.
The third aspect, it is machine readable present disclose provides a kind of server, including processor and machine readable storage medium
Storage medium is stored with the machine-executable instruction that can be executed by processor, and processor executes machine-executable instruction to realize
Above-mentioned resource allocation methods.
Fourth aspect, present disclose provides a kind of machine readable storage medium, machine readable storage medium is stored with machine
Executable instruction, for machine-executable instruction when being called and being executed by processor, machine-executable instruction promotes processor to realize
Above-mentioned resource allocation methods.
Above-mentioned resource allocation methods, device, server and machine readable storage medium divide for each application in cluster
During with resource, it is first determined the ratio for having distributed stock number and total resources and ratio of each application are the smallest
Using;The stock number that preset quantity Pod is distributed for the smallest application of the ratio updates the money of distribution of the smallest application of the ratio
Source amount;If the sum of stock number of distribution of all applications is less than or equal to total resources, each application has been redefined
The ratio for distributing stock number and total resources, continues to distribute;If the sum of stock number of distribution of all applications is greater than money
Source total amount is the resource that each application distributes corresponding amount according to the stock number of having distributed of each application.This mode is based on each
The ratio for having distributed stock number and total resources of application, the preferential money for selecting the smallest application distribution preset quantity Pod of ratio
Source amount can be liberally multiple allocated resources in cluster, while guarantee that assigned resource can be effective and reasonable
Ground utilizes, and avoid leads to resources idle or make because of inadequate resource using the feelings that are unable to operate normally because resource allocation is unreasonable
Condition.
Other feature and advantage of the disclosure will illustrate in the following description, alternatively, Partial Feature and advantage can be with
Deduce from specification or unambiguously determine, or by implement the disclosure above-mentioned technology it can be learnt that.
To enable the above objects, features and advantages of the present invention to be clearer and more comprehensible, better embodiment is cited below particularly, and match
Appended attached drawing is closed, is described in detail below.
Detailed description of the invention
It, below will be to specific in order to illustrate more clearly of disclosure specific embodiment or technical solution in the prior art
Embodiment or attached drawing needed to be used in the description of the prior art be briefly described, it should be apparent that, it is described below
Attached drawing is some embodiments of the disclosure, for those of ordinary skill in the art, before not making the creative labor
It puts, is also possible to obtain other drawings based on these drawings.
Fig. 1 is a kind of structural schematic diagram for Container Management cluster that disclosure embodiment provides;
Fig. 2 is a kind of flow chart for resource allocation methods that disclosure embodiment provides;
Fig. 3 is the flow chart for another resource allocation methods that disclosure embodiment provides;
Fig. 4 is a kind of structural schematic diagram for resource allocation device that disclosure embodiment provides;
Fig. 5 is the structural schematic diagram for the server that disclosure embodiment provides.
Specific embodiment
To keep the purposes, technical schemes and advantages of disclosure embodiment clearer, below in conjunction with attached drawing to this public affairs
The technical solution opened is clearly and completely described, it is clear that and described embodiment is a part of embodiment of the disclosure,
Rather than whole embodiment.Based on the embodiment in the disclosure, those of ordinary skill in the art are not making creation
Property labour under the premise of every other embodiment obtained, belong to the disclosure protection range.
In order to make it easy to understand, describing a kind of Container Management based on container operation management platform Kubernetes building first
The framework of cluster (abbreviation Kubernetes cluster), as shown in Figure 1.The Kubernetes cluster supports Docker and Rocket
(being illustrated by taking Docker as an example in Fig. 1) two kinds of container techniques, Kubernetes cluster include a host node
(Kubernetes Master) and other nodes (Node) (such as node A and node B), other nodes are connect with host node,
Node in Kubernetes cluster can be physical server, or virtual machine.
It include API Server and copy controller (Replication Controller, abbreviation RC) in host node;
Wherein, API Server is the signal centre between the api interface and cluster internal modules of cluster management, meanwhile,
API Server can also log in user and carry out authentication vs. authorization, realize and carry out security control to cluster.In order to improve performance
Or realize high availability, when the user of some application increases, the number that the application can be specified to need to replicate by RC needs
The number of duplication matches with the increased quantity of user, and creates a Pod for every part of duplication, and each user can be corresponding
Application is operated in Pod;Ensure that the Pod quantity of any time actual motion is equal with the quantity of duplication always simultaneously, if user
When reduction, the Pod of reduced user's operation can also be withdrawn by RC.Thus, it is asked by the dilatation capacity reducing that RC can solve Pod
Topic.
Other nodes are illustrated by taking two nodes as an example, respectively node A and node B for running Pod in Fig. 1;Often
A node can run multiple Pod;Pod is the basic deployment scheduling unit of Kubernetes, wherein including one group of container sum number
According to volume (being referred to as volume);It can be seen from the above, container is for running application program;And it is preserved in book and the group
The associated specific file of container or file;During container creation, book is initialized;Container during operation may be used
To use the file in book.
Pod is that the place of concrete function and service is applied and provided to actual motion;The framework of different application is different.It is right
In relatively simple application, this is can be completed using all functions in a Pod, is somebody's turn to do when user uses in application, can be every
A user distributes the Pod of a duplication, and operation has the application on the Pod of duplication, and user operates on corresponding Pod.But for
More complicated application, an application include many specific functions and service, need multiple Pod complete jointly these functions with
Service, each Pod is for completing different function and service;The various functions of be provided with same application and the Pod group of service
It is combined together completely to apply.Therefore, when user needs which function and service of the application, as user duplication pair
Operation has the function and service on the Pod answered, the Pod of duplication, and user operates on the Pod.
In addition, each Pod is provided with a key-value pair (Labels) in Fig. 1, the key-value pair is for marking Pod, with transmitting
The attribute of Pod.RC in Service or host node can be identified by key-value pair and be selected specific Pod.In addition, node
In also comprising the components such as kubelet, kube-proxy and docker;Wherein, kubelet is main node proxy;Service passes through
Link can be routed to specified Pod by kube-proxy;Docker is the container technique that Kubernetes is used, and can also be changed
For rocket.Service is that the route agent of Pod is abstract, and the Labels of Pod can find satisfactory Pod.
Kubernetes can support the multiple resources of distribution and management Pod, but in actual production environment, Mei Gerong
The resource of device management cluster is usually limited, when there are multiple in application, being difficult to determination specifically distributes to each answer in cluster
It is just most fair and reasonable with how many resources, it was not only able to satisfy the resource requirement of each application, but also can be to avoid the wasting of resources.It is based on
The above problem, the embodiment of the invention provides a kind of resource allocation methods, device and servers;The technology can be applied to container
It manages in cluster or other Container Management platforms, system;It is specifically described below by embodiment.
Firstly, disclosure embodiment provides a kind of resource allocation methods, this method is applied to Container Management cluster
Host node;In the application of Container Management cluster (being referred to as Kubernets cluster) operation, the operating space of each user is
It is generated by Pod, which is exactly the service that user gets in the application;For example, being applied for web, Yong Hu
The various operations executed on web page are all using the service provided, these services are provided by the application program run in Pod.
It can be seen from the above, Pod is usually the creation of the template according to defined in RC, for same application or same service,
Usually the stock number of the Pod of each user's distribution is fixed.And for different applications or the different corresponding Pod of service
Stock number be different, thus in Container Management clustered deploy(ment) in application, would generally in advance to application test, according to survey
Test result obtains the various stock numbers for servicing corresponding Pod in the application or application.
Since the assignable resource of Container Management cluster is usually limited, corresponding one is applied or serviced each
It, can be true using following resource allocation methods when the online number of application can not be predicted in the case that the stock number of Pod determines
Surely the stock number of each application is distributed to, to realize the fair allocat of the resource based on application actual demand.As shown in Fig. 2,
The resource allocation methods specifically comprise the following steps:
Step S202 determines the total resources in Container Management cluster;
All kinds of resources generally include cpu resource and memory source in said vesse management cluster, certainly can also include it
His resource, such as upstream and downstream bandwidth etc.;It can only include the total resources of a kind of resource for a Container Management cluster, it can also
With the total resources comprising multiclass resource;Wherein, cpu resource can be distributed according to the number or nucleus number of CPU;Memory source can be with
It is distributed according to the unit of memory, for example, GB, MB etc..Cpu resource is mainly used for operation control, and memory source is mainly used for storing
Data.In general, the hardware resource of Container Management cluster be it is fixed, therefore, cpu resource based on the hardware resource and memory money
Source is also fixed;In hardware resource by counting the Container Management cluster, the hardware configuration of each physical server can be with
Obtain the total resources of all kinds of resources in said vesse management cluster.
Step S204 determines the ratio for having distributed stock number and total resources of each application in Container Management cluster;Often
A application corresponds at least one Pod, and each application has distributed the sum of the stock number that stock number is at least one Pod;
The money for having distributed stock number and include specified quantity Pod of each application in the initial stage, Container Management cluster
Source amount, as this applies the stock number of a corresponding Pod;For example, if the money of the cpu resource for applying a corresponding Pod
Source amount is 1CPU, then in the initial stage, the stock number of distribution of the application includes 1CPU.If multiple using being assigned
The stock number of Pod, the then application have distributed the sum of the stock number that stock number is multiple Pod;For example, the application has been assigned with three
The stock number of a Pod, the stock number of each Pod are 1CPU, then the stock number of distribution of the application is 3CPU.
By taking cpu resource as an example, if the stock number of distribution using A is 10CPU, and the total resources of CPU is in cluster
50CPU, this is 10CPU/50CPU=0.2 using the ratio for having distributed stock number and total resources of A at this time.
Step S206 determines the smallest application of ratio;
Obtain each application in Container Management cluster distributed the ratio of stock number and total resources after, it is more each to answer
The corresponding application of the smallest ratio of numerical value is determined as the smallest application of ratio with the numerical values recited of corresponding ratio.
Step S208 distributes the stock number of preset quantity Pod for the smallest application of the ratio, updates that the ratio is the smallest to answer
Distribution stock number;
The above-mentioned the smallest application of ratio can be understood as currently occupying the smallest application of stock number, in order to keep between each application
The smallest application of current ratio is determined as when the target application in front-wheel distribution by the fairness of resource allocation, and is the application
Distribute stock number.In order to guarantee that the resource for distributing to application is used appropriately, it is not idle, can be applied according to this corresponding
The stock number of Pod is that the target application distributes resource.It is specifically as follows the stock number that the application distributes a corresponding Pod,
The stock number of multiple Pod can be distributed;In above-mentioned steps S208, for the preset quantity Pod of the ratio the smallest application distribution
Stock number is applied with this and has distributed the stock number for the specified quantity Pod for including in stock number, present count therein in the initial stage
Amount may be the same or different with specified quantity.After being assigned, the distribution resource of the smallest application of the ratio is updated
Amount.
For example, the stock number of distribution of the smallest application of the ratio is 10CPU, this applies the resource of a corresponding Pod
Amount is 1CPU, and if it is the stock number of two Pod of the application distribution, after being assigned, which has distributed stock number update
For 12CPU.
Step S210, judges whether the sum of stock number of distribution of all applications is greater than total resources;If judging result is
It is no, execute step S212;If the determination result is YES, step S214 is executed;
Step S212 returns to the ratio for having distributed resource and total resources for executing and determining each application in Container Management cluster
The operation of value;
When in Container Management cluster simultaneously operation have it is multiple in application, in order to by the total resources of the resource in cluster more
Each application is equitably distributed to, while guaranteeing that the resource of distribution is not idle again, the mode of circulation can be used to answer to be each
With distribution resource.In the distribution of every wheel, it can be one of allocated resources, after front-wheel is assigned, or work as
Before front-wheel circulation, can check whether total resources to be allocated can also be supported to continue to distribute or allocated resource
Whether amount exceeds the total resources of all kinds of resources of the cluster, to determine whether to continue as allocated resources.
Specifically, if updated all applications have distributed the sum of total resources without reaching total money in cluster
Source amount illustrates to can also continue to carry out the distribution of next round then at this time for allocated resources;If updated all applications
The sum of the total resources that distributed have arrived at total resources in cluster, illustrate can not to be further continued for for allocated resources,
The stock number that each application distributes is determined as the stock number that the application should distribute.
Step S214 is the resource that each application distributes corresponding amount according to the stock number of having distributed of each application, wherein ratio
The stock number of distribution that stock number is the updated the smallest application of the ratio of having distributed for being worth the smallest application subtracts present count
Measure the stock number of Pod.
For example, the stock number of distribution after the last round of distribution of the smallest application of the ratio is 9CPU;The ratio is the smallest
Stock number using corresponding preset quantity Pod is 3CPU;Through when front-wheel distribution after, the distribution resource of the smallest application of ratio
Amount is 12CPU, and the sum of stock number of distribution of all applications is 30CPU, and the total resources of cluster is 28CPU, at this point, explanation
Through when front-wheel distribution after, the CPU for having distributed stock number and having exceeded in the total resources of the cluster of all applications, cluster it is total
Stock number is not able to satisfy the resource allocation currently taken turns, at this point, then ending resource assigning process, the above-mentioned the smallest application of ratio is
Distribution stock number is that the stock number of distribution (i.e. 12CPU) of the updated the smallest application of the ratio subtracts preset quantity Pod's
Stock number (i.e. 3CPU) is equal to 9CPU.
A kind of resource allocation methods that disclosure embodiment provides, in the mistake for each allocated resources in cluster
Cheng Zhong, it is first determined the smallest application of the ratio for having distributed stock number and total resources and ratio of each application;For this ratio
It is worth the stock number of the smallest application distribution preset quantity Pod, updates the distribution stock number of the smallest application of the ratio;If institute
Have the sum of stock number of distribution of application be less than or equal to total resources, then redefine each application the stock number of distribution and
The ratio of total resources continues to distribute;If the sum of stock number of distribution of all applications is greater than total resources, according to every
The stock number of having distributed of a application is the resource that each application distributes corresponding amount.This mode is provided based on the distribution of each application
The ratio of source amount and total resources, the preferential stock number for selecting the smallest application distribution preset quantity Pod of ratio, can be liberally
For multiple allocated resources in cluster, while guaranteeing that assigned resource can be utilized effective and reasonablely, avoid because
Resource allocation is unreasonable to be led to resources idle or makes because of inadequate resource using the case where being unable to operate normally.
Disclosure embodiment also provides another resource allocation methods, and this method is real on the basis of the above embodiment
It is existing;The distribution method of resource when in Container Management cluster including multiclass resource is specifically described in present embodiment;Such as Fig. 3 institute
Show, which includes the following steps:
Step S302 determines the total resources of every class resource in Container Management cluster;
It may include at least a kind of resource in the Container Management cluster, also may include multiclass resource, such as cpu resource and interior
Resource is deposited, bandwidth resources etc. can also be included certainly.If necessary to distribute a plurality of types of resources to a variety of applications simultaneously, then need
The total resources of each type in the Container Management cluster is determined respectively.
Step S304 determines each resource for having distributed stock number and respective type resource using corresponding every class resource
The ratio of total amount;Each application corresponds at least one Pod, and each application has distributed the money that stock number is at least one Pod
The sum of source amount;
If in Container Management cluster including two class resources, respectively cpu resource and memory source, in most instances,
Using usually all using cpu resource and memory source simultaneously, therefore in the stock number of every kind of corresponding Pod of application simultaneously
Cpu resource and memory source comprising preset quantity, the cpu resource only for different applications, in the stock number of a Pod
It is different with quantity, the ratio of memory source.For example, the application of game class may need more cpu resource, then the application
Resource quota in cpu resource possibly more than memory source, the resource quota of the application can be set to 3CPU, 1GB memory;It is right
In information record class application may need more memory source, then in the resource quota of the application memory source possibly more than
Cpu resource, the resource quota of the application can be set to 1CPU, 3GB memory.
For example, it is 1CPU and 3GB memory that this, which applies the stock number of a corresponding Pod, for a certain application, answered for this
With the stock number for being assigned with three Pod, then the stock number of distribution of the application is 3CPU and 9GB;Assuming that the Container Management cluster
In, the total resources of cpu resource is 30CPU, and the total resources of memory source is 45GB;At this point, the cpu resource of the application is
The ratio for distributing stock number and total resources is 3CPU/30CPU=0.1;The stock number of distribution of the memory source of the application and
The ratio of total resources is 9GB/45GB=0.2.
Step S306, for each application: having distributed every class resource of the application to stock number and the type resource
Maximum value in the ratio of total resources is determined as the ratio of the application;
For different applications, some applications may need the stock number of cpu resource larger, and some applications may need
The stock number of memory source is larger;Therefore, the total resources for having distributed stock number and the type resource of every class resource of application
Ratio in, which ratio be maximum value, then illustrate that the corresponding resource of the maximum value is the biggish resource of application demand amount;It will
The biggish resource of application demand amount is determined as the ratio of the application, and is that types of applications carries out resource allocation based on the ratio,
Can meet the needs of each application is to main resource while fair allocat resource.
For example, if the distributed stock number and the ratio of total resources of the cpu resource of a certain application are 0.1, the application
The ratio for having distributed stock number and total resources of memory source be 0.2, illustrate that the application is larger to the resource requirement of memory,
The corresponding ratio of memory is determined as to the ratio of the application.
Step S308 distributes the stock number of preset quantity Pod for the smallest application of the ratio, updates that the ratio is the smallest to answer
Distribution stock number;
In general, can support the base of a user using the stock number of CPU and memory in the stock number of a corresponding Pod
This service makes to apply when running in Pod, and CPU or memory are not idle;Therefore, above-mentioned steps S308 is for above-mentioned ratio
When the smallest allocated resources amount, it is allocated as unit of Pod;The stock number of a Pod can be specifically distributed, it can also be with
The stock number of multiple Pod is distributed, which can effectively extend the money of target application during resource fairness distributes
Source does not make the resource of distribution be wasted.
After being assigned, then the distribution stock number of the smallest application of the ratio is updated, for example, the smallest application of the ratio
Before front-wheel distribution, having distributed stock number is 8CPU and 4GB;The smallest resource using a corresponding Pod of the ratio
Amount is 2CPU and 1GB;The stock number of two Pod is distributed if it is the smallest application of the ratio, the smallest application of the ratio
It has distributed stock number and has been updated to 12CPU and 6GB.
Step S310, for every class resource: calculating the sum of stock number of distribution of all applications;If for every class resource
Distributed the sum of stock number be respectively less than or equal to respective class resource total resources, then follow the steps S304;If being directed to any sort
The total resources for having distributed the sum of stock number and being greater than such resource of resource, thens follow the steps S312.
It is appreciated that each stock number using a corresponding Pod includes the default stock number of every class resource;As above
Described, application biggish for cpu demand, it includes more CPU that this, which applies the stock number of a corresponding Pod, a such as Pod
Stock number include 3CPU and 1GB;Application biggish for memory requirements, this includes using the stock number of a corresponding Pod
The stock number of more memory, such as a Pod includes 1CPU and 3GB.
If for every class resource distributed the sum of stock number be respectively less than or equal to respective class resource total resources, explanation
All kinds of resources in Container Management cluster are not all finished by point, then continue to execute above-mentioned steps S304;And due to application
Resource allocation is all as unit of Pod, and Pod usually includes all types of resources;So if all kinds of resources in cluster
As long as the sum of the stock number of distribution of one type resource in (which specific class without limitation, any sort) is greater than such resource
Total resources, illustrate that the resource in the cluster cannot then be further continued for distributing, need to stop assigning process at this time.
Step S312 is the resource that each application distributes corresponding amount according to the stock number of having distributed of each application, wherein ratio
The stock number of distribution that stock number is the updated the smallest application of the ratio of having distributed for being worth the smallest application subtracts present count
Measure the stock number of Pod.
Due to the total resources for having distributed the sum of stock number and being greater than such resource of any sort resource, illustrate through when front-wheel point
After matching, at least a kind of resource has exceeded the total resources of such resource;Then illustrate that the wheel is the distribution of the smallest application of ratio
The stock number of preset quantity Pod cannot achieve, need to subtract the stock number as the preset quantity Pod of front-wheel distribution, after subtracting
Final distribution stock number of the result as the smallest application of the ratio.
In above-mentioned steps S312, after each application has been distributed stock number, in order to by the distribution resource of each application
Amount veritably distributes to each application, can be realized by the way that different application is deployed in different NameSpaces.In Container Management
Deployment avoids influencing each other between different application in application, for the ease of unified management in cluster, can will usually apply deployment
In individual NameSpace, the stock number entirely applied is managed by NameSpace.Different order spaces are understood that
For different name formats, i.e., application is named using different name formats.NameSpace can specifically be passed through
ResourceQuota object limits the resource that the application in the NameSpace can obtain.It is set in the ResourceQuota object
It is equipped with limit parameter, cluster manages the resource of each application according to the limit parameter.It is distributed through the above steps for each application
After resource, resource quota object in NameSpace where the stock number of distribution that application should distribute being set to corresponding application
Limit parameter in ResourceQuota, so that the total resources for each application distribution can applied through the above way
The effect of resource management is really played in actual moving process.
In aforesaid way, after every wheel is assigned, it is equal to or greatly if existed in the total resources of distribution of all kinds of resources
In the total resources of corresponding class resource, then stop resource allocation, otherwise continues next round distribution;And then it again will be final each
It is set as using the total resources that should be distributed using the stock number that should be distributed;Which can be more public by resources all kinds of in cluster
Level land is distributed in multiple applications, meanwhile, guarantee that assigned resource can be utilized effective and reasonablely.
For a further understanding of the resource allocation methods in above embodiment, embodiment further provides another for the disclosure
Kind resource allocation methods;This embodiment offers a kind of specific calculations of above-mentioned resource allocation methods;This method is based on
Following assumed conditions are realized:
Condition a: in Container Management cluster comprising n apply, respectively A1, A2 ..., An.
Condition b: the total resources using Ai (1≤i≤n, i are natural numbers) current demand is Ui=< Ui-cpu, Ui-mem
>, wherein Ui-cpu is the cpu resource total amount using Ai, the memory source total amount of Ui-mem application Ai;The initial value of Ui is 0,
It is 0 using the CPU of Ai and the resource quota of memory i.e. when there is no user using Ai.
Condition c: using the stock number of a Pod of Ai are as follows: ci CPU and miGB memory can be expressed as < ciCPU,
miGB>。
Condition d: the assignable total nucleus number of CPU is C in cluster.
Condition e: assignable memory sum is MGB.
Condition f: allocated CPU core number is Rcpu in cluster, and allocated memory sum is Rmem.
Condition g: the maximum of corresponding resource total amount percentage in cluster is accounted for respectively using the current acquired all kinds of resources of Ai
Value Si, Si=max { Ui-cpu/C, Ui-mem/M }, the initial value of Si are 0.
Based on above-mentioned assumed condition, above-mentioned resource allocation methods include the following steps:
Step 11, in first run assigning process, the Si=max { ci/C, mi/M } of each application in computing cluster, by Si value
The smallest application is determined as target application;
Step 12, the stock number of the corresponding Pod of the target application is distributed for the target application;
Step 13, in computing cluster each application Si=max { Ui-cpu/C, Ui-mem/M }, answer Si value is the smallest
With being determined as target application;
Step 14, if judgement increases the stock number for applying a corresponding Pod newly to target application Ai, if be more than collection
The total resources of group;Judge Rcpu+ci≤C and whether Rmem+mi≤M is true;
Step 15, if set up, the Ui value of target application Ai, i.e. Ui-cpu=Ui-cpu+ci, Ui-mem=Ui- are updated
mem+mi;Update Rcpu=Rcpu+ci, Rmem=Rmem+mi;
Step 16, it continues cycling through and executes step 3 to step 5, until in step 4, Rcpu+ci≤C and Rmem+mi≤M is not
It sets up, jumps out circulation, terminate assigning process;
Step 17, resource quota in NameSpace where setting corresponding application for the total resources of distribution of each application
Limit parameter in object ResourceQuota.
Based on above embodiment, embodiment further provides a kind of calculating use-cases of resource allocation methods for the disclosure, false
If total resources is 36CPU in current cluster, 64GB memory disposes three applications, respectively using A, using B in the cluster
With apply C;Wherein, using the stock number of a Pod of A be<1CPU, 1GB>, using B a Pod stock number be<
2CPU, 4GB>, using C a Pod stock number be<3CPU, 1GB>.
Following table 1 illustrates the process of resource allocation, and this time resource allocation is assigned through 22 wheels altogether;Wherein, in table 1
Primary resource can be understood as the biggish resource of application demand amount, primary resource accounting can be understood as having distributed to the application
The stock number of the biggish resource of demand occupies the ratio of the total resources of the resource;In first run assigning process, due to applying A
A Pod stock number in primary resource (CPU or memory) it is minimum, divide so resource will be carried out as target object using A
Match, after distribution, the primary resource accounting using A is 1/36;In subsequent allocation wheel, using the smallest application of primary resource accounting as
Target application, it is minimum using the primary resource accounting of B after the 5th wheel distribution, it is 4/64, then in the 6th wheel, will makees using B
For target application carry out resource allocation, for application B distribution using B a Pod stock number, i.e.,<2CPU, 4GB>.
After above-mentioned more wheel distribution, the Resource Allocation Formula of the final cluster are as follows: distributing to using the resource of A is 12CPU
With 12GB memory;Distributing to using the resource of B is 12CPU and 24GB memory;Distributing to using the resource of B is in 12CPU and 4GB
It deposits.
In above-mentioned resource allocation methods, calculated using DRF (Dominant Resource Fairness, superior resources are fair)
Resources all kinds of in cluster can be more distributed fairly in multiple applications by the resource respectively applied in method computing cluster, meanwhile,
Guarantee that assigned resource can be utilized effective and reasonablely.
Table 1
Resource allocation methods described in above embodiment can be applied to Container Management cluster build-time, for the collection
The application that group will run carries out resource allocation, also can be applied to newly-increased to the cluster in application, for newly-increased application and the collection
Original application in group distributes resource unitedly again;In above-mentioned resource allocation methods, can be fair give Kubernetes cluster
The allocated resources of middle deployment optimize original allocation scheme, improve the utilization rate of resource, avoid distributing to a certain application excessive
The less resource of demand distributes very few demand more resource.
It should be noted that above-mentioned each method embodiment is described in a progressive manner, each embodiment emphasis
What is illustrated is all the difference with other embodiments, and same and similar part refers to each other i.e. between each embodiment
It can.
Corresponding to above method embodiment, resource allocation device shown in Fig. 4 is participated in, which is set to Container Management
The host node of cluster;The device includes:
First determining module 40, for determining the total resources in Container Management cluster;
Second determining module 41, for determining the stock number of distribution and total resources of each application in Container Management cluster
Ratio;Each application corresponds at least one Pod, each application distributed stock number that stock number is at least one Pod it
With;
Using determining module 42, for determining the smallest application of ratio;
Stock number distribution module 43, for distributing the stock number of preset quantity Pod for the smallest application of the ratio, updating should
The distribution stock number of the smallest application of ratio;
Whether judgment module 44, the sum of stock number of distribution for judging all applications are greater than total resources;
Return module 45, if for judging result be it is no, return execute determination Container Management cluster in each application
Distribute the operation of the ratio of resource and total resources;
Resource distribution module 46, for being each to answer according to the stock number of distribution of each application if the determination result is YES
With the resource of distribution corresponding amount, wherein the stock number of distribution of the smallest application of ratio, which is that the updated ratio is the smallest, answers
The stock number that stock number subtracts preset quantity Pod is distributed.
A kind of resource allocation device that disclosure embodiment provides is each in cluster by way of taking turns distribution more
Allocated resources;In every wheel assigning process, it is first determined total resources, each application in Container Management cluster have divided
The smallest application of ratio and ratio with stock number and total resources;Preset quantity Pod is distributed for the smallest application of the ratio
Stock number, update the distribution stock number of the smallest application of the ratio;If the sum of the stock number of distribution of all applications is small
In or equal to total resources, then continue next round distribution, if the sum of stock number of distribution of all applications is greater than resource
Total amount is the resource that each application distributes corresponding amount according to the stock number of having distributed of each application.Which can liberally be
Multiple allocated resources in cluster, while guaranteeing that assigned resource can be utilized effective and reasonablely, it avoids because of money
Source unreasonable distribution leads to resources idle or makes because of inadequate resource using the case where being unable to operate normally.
Further, the initial stage, the stock number of distribution of each application includes specified quantity in Container Management cluster
The stock number of Pod.
It further, include multiclass resource in said vesse management cluster;First determining module, is also used to: determining container
Manage the total resources of every class resource in cluster;Second determining module, is also used to: determining each using corresponding every class resource
The ratio of the total resources of stock number and respective type resource is distributed.
Further, above-mentioned to apply determining module, it is also used to: for each application: by every class resource of the application
The maximum value in the ratio of the total resources of stock number and the type resource is distributed, the ratio of the application is determined as;It is answered from each
Minimum value is selected in ratio, and the corresponding application of minimum value is determined into the smallest application of ratio.
Further, above-mentioned each stock number using a corresponding Pod includes: the default stock number of every class resource;
Judgment module is also used to: being directed to every class resource: being calculated the sum of stock number of distribution of all applications;If being directed to any sort resource
Distributed the sum of stock number be greater than such resource total resources, it is determined that judging result is yes;If for every class resource
Distributed the sum of stock number be respectively less than or equal to respective class resource total resources, then judging result is no.
Present embodiments provide for a kind of servers corresponding with above method embodiment.Fig. 5 is the server
Structural schematic diagram, as shown in figure 5, the equipment includes processor 501 and memory 502;Wherein, memory 502 is for storing one
Item or a plurality of computer instruction, one or more computer instruction are executed by processor, to realize above-mentioned resource allocation methods.
Server shown in fig. 5 further includes bus 503 and forwarding chip 504, processor 501, forwarding chip 504 and storage
Device 502 is connected by bus 503.The server can be network edge device.
Wherein, memory 502 may include high-speed random access memory (RAM, Random Access Memory),
It may further include non-labile memory (non-volatile memory), for example, at least a magnetic disk storage.Bus
503 can be isa bus, pci bus or eisa bus etc..The bus can be divided into address bus, data/address bus, control always
Line etc..Only to be indicated with a four-headed arrow in Fig. 5, it is not intended that an only bus or a type of convenient for indicating
Bus.
Forwarding chip 504 is used to connect by network interface at least one user terminal and other network units, will seal
The IPv4 message or IPv6 message installed is sent to the user terminal by network interface.
Processor 501 may be a kind of IC chip, the processing capacity with signal.It is above-mentioned during realization
Each step of method can be completed by the integrated logic circuit of the hardware in processor 501 or the instruction of software form.On
The processor 501 stated can be general processor, including central processing unit (Central Processing Unit, abbreviation
CPU), network processing unit (Network Processor, abbreviation NP) etc.;It can also be digital signal processor (Digital
Signal Processing, abbreviation DSP), specific integrated circuit (Application Specific Integrated
Circuit, abbreviation ASIC), ready-made programmable gate array (Field-Programmable Gate Array, abbreviation FPGA) or
Person other programmable logic device, discrete gate or transistor logic, discrete hardware components.It may be implemented or execute sheet
Disclosed each method, step and logic diagram in invention embodiment.General processor can be microprocessor or this at
Reason device is also possible to any conventional processor etc..The step of method in conjunction with disclosed in embodiment of the present invention, can direct body
Now executes completion for hardware decoding processor, or in decoding processor hardware and software module combine and execute completion.It is soft
Part module can be located at random access memory, and flash memory, read-only memory, programmable read only memory or electrically erasable programmable are deposited
In the storage medium of this fields such as reservoir, register maturation.The storage medium is located at memory 502, and processor 501 reads storage
Information in device 502, in conjunction with its hardware complete aforementioned embodiments method the step of.
Embodiment of the present invention additionally provides a kind of machine readable storage medium, and machine readable storage medium storage is organic
Device executable instruction, for the machine-executable instruction when being called and being executed by processor, machine-executable instruction promotes processor
Realize above-mentioned resource allocation methods, specific implementation can be found in method implementation, and details are not described herein.
The technical effect and preceding method of server provided by embodiment of the present invention, realization principle and generation are implemented
Mode is identical, and to briefly describe, device embodiments part does not refer to place, can refer in corresponding in preceding method embodiment
Hold.
In several embodiments provided herein, it should be understood that disclosed device and method can also lead to
Other modes are crossed to realize.Device embodiments described above are only schematical, for example, the flow chart in attached drawing and
Block diagram shows the system in the cards of the device of multiple embodiments according to the present invention, method and computer program product
Framework, function and operation.In this regard, each box in flowchart or block diagram can represent a module, program segment or generation
A part of code, a part of the module, section or code include one or more for realizing defined logic function
Executable instruction.It should also be noted that function marked in the box can also be in some implementations as replacement
Occur different from the sequence marked in attached drawing.For example, two continuous boxes can actually be basically executed in parallel, they
Sometimes it can also execute in the opposite order, this depends on the function involved.It is also noted that block diagram and or flow chart
In each box and the box in block diagram and or flow chart combination, can function or movement as defined in executing it is special
Hardware based system is realized, or can be realized using a combination of dedicated hardware and computer instructions.
Finally, it should be noted that embodiment described above, the only specific embodiment of the disclosure, to illustrate this public affairs
The technical solution opened, rather than its limitations, the protection scope of the disclosure are not limited thereto, although referring to aforementioned embodiments pair
The disclosure is described in detail, those skilled in the art should understand that: any technology for being familiar with the art
Personnel can still modify to technical solution documented by aforementioned embodiments in the technical scope that the disclosure discloses
Or variation or equivalent replacement of some of the technical features can be readily occurred in;And these modifications, variation or replacement,
The spirit and scope for disclosure embodiment technical solution that it does not separate the essence of the corresponding technical solution, should all cover in this public affairs
Within the protection scope opened.Therefore, the protection scope of the disclosure shall be subject to the protection scope of the claim.
Claims (12)
1. a kind of resource allocation methods, which is characterized in that the method is applied to the host node of Container Management cluster;The method
Include:
Determine the total resources in Container Management cluster;
Determine the ratio for having distributed stock number and the total resources of each application in Container Management cluster;Each application
At least one corresponding Pod, each application have distributed the sum of the stock number that stock number is at least one Pod;
Determine the smallest application of the ratio;
The stock number that preset quantity Pod is distributed for the smallest application of the ratio updates the money of distribution of the smallest application of the ratio
Source amount;
Judge whether the sum of stock number of distribution of all applications is greater than the total resources;
If judging result is no, the resource of distribution and the money of each application in the return execution determining Container Management cluster
The operation of the ratio of source total amount;
If the determination result is YES, then corresponding amount is distributed for each application according to the stock number of having distributed of each application
Resource, wherein the smallest application of ratio distributed stock number be the updated the smallest application of the ratio
Distribution stock number subtracts the stock number of the preset quantity Pod.
2. the method according to claim 1, wherein the initial stage, each application in the Container Management cluster
The stock number for having distributed stock number and include specified quantity Pod.
3. method according to claim 1 or 2, which is characterized in that include at least a kind of money in the Container Management cluster
Source;
Total resources in the determining Container Management cluster, comprising: determine that the resource of every class resource in Container Management cluster is total
Amount;
The ratio for having distributed stock number and the total resources of each application in the determining Container Management cluster, comprising: really
The ratio of fixed each total resources for having distributed stock number and respective type resource using corresponding every class resource.
4. according to the method described in claim 3, it is characterized in that, the smallest application of the determination ratio, comprising:
For each application: by the ratio of the total resources for having distributed stock number and the type resource of every class resource of the application
In maximum value, be determined as the ratio of the application;
Minimum value is selected from the ratio of each application, the corresponding application of the minimum value is determined that the ratio is the smallest
Using.
5. according to the method described in claim 3, it is characterized in that, each stock number packet using a corresponding Pod
It includes: the default stock number of every class resource;
Whether the sum of described stock number of distribution for judging all applications is greater than the total resources, comprising:
For every class resource: calculating the sum of the stock number of distribution of all applications;
If for the total resources for having distributed the sum of stock number and being greater than such resource of any sort resource, it is determined that described to sentence
Disconnected result is yes;
If for every class resource it is described distributed the sum of stock number be respectively less than or equal to respective class resource total resources, institute
It is no for stating judging result.
6. a kind of resource allocation device, which is characterized in that described device is set to the host node of Container Management cluster;Described device
Include:
First determining module, for determining the total resources in Container Management cluster;
Second determining module has distributed stock number and the total resources for determine each application in Container Management cluster
Ratio;Each application corresponds at least one Pod, and each application has distributed the money that stock number is at least one Pod
The sum of source amount;
Using determining module, for determining the smallest application of the ratio;
Stock number distribution module updates the ratio most for distributing the stock number of preset quantity Pod for the smallest application of the ratio
The distribution stock number of small application;
Whether judgment module, the sum of stock number of distribution for judging all applications are greater than the total resources;
Return module, if for judging result be it is no, return to having divided for each application in the execution determining Container Management cluster
The operation of ratio with resource and the total resources;
Resource distribution module, for being described each according to the stock number of distribution of each application if the determination result is YES
Using the resource of distribution corresponding amount, wherein the stock number of distribution of the smallest application of ratio is described updated ratio
It is worth the stock number of the smallest application distributed stock number and subtracted the preset quantity Pod.
7. device according to claim 6, which is characterized in that initial stage, each application in the Container Management cluster
The stock number for having distributed stock number and include specified quantity Pod.
8. device according to claim 6 or 7, which is characterized in that include at least a kind of money in the Container Management cluster
Source;
First determining module, is also used to: determining the total resources of every class resource in Container Management cluster;
Second determining module, is also used to: determining each stock number of distribution and respective class using corresponding every class resource
The ratio of the total resources of type resource.
9. device according to claim 8, which is characterized in that it is described to apply determining module, it is also used to:
For each application: by the ratio of the total resources for having distributed stock number and the type resource of every class resource of the application
In maximum value, be determined as the ratio of the application;
Minimum value is selected from the ratio of each application, the corresponding application of the minimum value is determined that the ratio is the smallest
Using.
10. device according to claim 8, which is characterized in that each stock number packet using a corresponding Pod
It includes: the default stock number of every class resource;
The judgment module, is also used to: being directed to every class resource: calculating the sum of stock number of distribution of all applications;
If for the total resources for having distributed the sum of stock number and being greater than such resource of any sort resource, it is determined that described to sentence
Disconnected result is yes;
If for every class resource it is described distributed the sum of stock number be respectively less than or equal to respective class resource total resources, institute
It is no for stating judging result.
11. a kind of server, which is characterized in that including processor and machine readable storage medium, the machine readable storage is situated between
Matter is stored with the machine-executable instruction that can be executed by the processor, and the processor executes the machine-executable instruction
To realize method described in any one of claim 1 to 5.
12. a kind of machine readable storage medium, which is characterized in that the machine readable storage medium is stored with the executable finger of machine
It enables, for the machine-executable instruction when being called and being executed by processor, the machine-executable instruction promotes the processor
Realize method described in any one of claim 1 to 5.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811247681.6A CN109471725A (en) | 2018-10-24 | 2018-10-24 | Resource allocation methods, device and server |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811247681.6A CN109471725A (en) | 2018-10-24 | 2018-10-24 | Resource allocation methods, device and server |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109471725A true CN109471725A (en) | 2019-03-15 |
Family
ID=65665910
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811247681.6A Pending CN109471725A (en) | 2018-10-24 | 2018-10-24 | Resource allocation methods, device and server |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109471725A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021012956A1 (en) * | 2019-07-22 | 2021-01-28 | 腾讯科技(深圳)有限公司 | Resource processing method for cloud platform, and related device and storage medium |
CN112882821A (en) * | 2019-11-29 | 2021-06-01 | 北京国双科技有限公司 | Method, device and equipment for allocating processor resources |
WO2021103646A1 (en) * | 2019-11-26 | 2021-06-03 | 华为技术有限公司 | Pod deployment method and device |
CN113127203A (en) * | 2021-04-25 | 2021-07-16 | 华南理工大学 | Deep learning distributed compiler for cloud edge computing and construction method |
CN115617517A (en) * | 2022-10-12 | 2023-01-17 | 中航信移动科技有限公司 | Data processing system for application pod control |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106576114A (en) * | 2014-08-08 | 2017-04-19 | 甲骨文国际公司 | Policy based resource management and allocation system |
CN107688322A (en) * | 2017-08-31 | 2018-02-13 | 天津中新智冠信息技术有限公司 | A kind of containerization management system |
CN108519911A (en) * | 2018-03-23 | 2018-09-11 | 上饶市中科院云计算中心大数据研究院 | The dispatching method and device of resource in a kind of cluster management system based on container |
-
2018
- 2018-10-24 CN CN201811247681.6A patent/CN109471725A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106576114A (en) * | 2014-08-08 | 2017-04-19 | 甲骨文国际公司 | Policy based resource management and allocation system |
CN107688322A (en) * | 2017-08-31 | 2018-02-13 | 天津中新智冠信息技术有限公司 | A kind of containerization management system |
CN108519911A (en) * | 2018-03-23 | 2018-09-11 | 上饶市中科院云计算中心大数据研究院 | The dispatching method and device of resource in a kind of cluster management system based on container |
Non-Patent Citations (3)
Title |
---|
ALI GHODSI: "Dominant Resource Fairness:Fair Allocation of Multiple Resource Types", 《NSDI》 * |
唐瑞: "基于Kubernetes的容器云平台资源调度策略研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
张子龙: "面向大数据的多租户关键技术研究", 《中国优秀硕士学位论文全文数据库 经济与管理科学辑》 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2021012956A1 (en) * | 2019-07-22 | 2021-01-28 | 腾讯科技(深圳)有限公司 | Resource processing method for cloud platform, and related device and storage medium |
US11966792B2 (en) | 2019-07-22 | 2024-04-23 | Tencent Technology (Shenzhen) Company Limited | Resource processing method of cloud platform, related device, and storage medium |
WO2021103646A1 (en) * | 2019-11-26 | 2021-06-03 | 华为技术有限公司 | Pod deployment method and device |
CN112882821A (en) * | 2019-11-29 | 2021-06-01 | 北京国双科技有限公司 | Method, device and equipment for allocating processor resources |
CN113127203A (en) * | 2021-04-25 | 2021-07-16 | 华南理工大学 | Deep learning distributed compiler for cloud edge computing and construction method |
CN113127203B (en) * | 2021-04-25 | 2022-06-14 | 华南理工大学 | Deep learning distributed compiler for cloud edge computing and construction method |
CN115617517A (en) * | 2022-10-12 | 2023-01-17 | 中航信移动科技有限公司 | Data processing system for application pod control |
CN115617517B (en) * | 2022-10-12 | 2023-11-10 | 中航信移动科技有限公司 | Data processing system for applying pod control |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109471725A (en) | Resource allocation methods, device and server | |
US11704144B2 (en) | Creating virtual machine groups based on request | |
CN104468803B (en) | A kind of virtual data center method for mapping resource and equipment | |
US10505796B2 (en) | Network function virtualization | |
CN106385329B (en) | Processing method, device and the equipment of resource pool | |
JP5510556B2 (en) | Method and system for managing virtual machine storage space and physical hosts | |
CN108737325A (en) | A kind of multi-tenant data partition method, apparatus and system | |
CN106301829A (en) | A kind of method and apparatus of Network dilatation | |
CN104679594B (en) | A kind of middleware distributed computing method | |
CN104995604A (en) | Resource allocation method of virtual machine and device thereof | |
CN103455363B (en) | Command processing method, device and physical host of virtual machine | |
CN108370328A (en) | A kind of management method and device of NFV MANO policy depictions symbol | |
CN104506669B (en) | The IP address distribution system and method for a kind of Based on Distributed network simulation platform | |
US11093288B2 (en) | Systems and methods for cluster resource balancing in a hyper-converged infrastructure | |
CN104601680A (en) | Resource management method and device | |
CN104331332A (en) | Virtual resource preallocation algorithm based on SLA (Service Level Agreement) | |
CN109522090A (en) | Resource regulating method and device | |
CN110275760A (en) | Process based on fictitious host computer processor hangs up method and its relevant device | |
CN107977773A (en) | A kind of method for the entry resource amount for managing multiple cloud platforms | |
CN108376103A (en) | A kind of the equilibrium of stock control method and server of cloud platform | |
CN109788061B (en) | Computing task deployment method and device | |
CN109976870A (en) | Creation method, device, equipment and the medium of virtual machine | |
CN110286961A (en) | Process based on physical host processor hangs up method and relevant device | |
CN109995571A (en) | Server configuration applies matched method and device with VNF | |
CN110417777A (en) | A kind of method and device communicated between micro services of optimization |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190315 |