CN117971505B - Method and device for deploying container application - Google Patents

Method and device for deploying container application Download PDF

Info

Publication number
CN117971505B
CN117971505B CN202410376235.4A CN202410376235A CN117971505B CN 117971505 B CN117971505 B CN 117971505B CN 202410376235 A CN202410376235 A CN 202410376235A CN 117971505 B CN117971505 B CN 117971505B
Authority
CN
China
Prior art keywords
resource
resource utilization
utilization rate
index
determining
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.)
Active
Application number
CN202410376235.4A
Other languages
Chinese (zh)
Other versions
CN117971505A (en
Inventor
辛永欣
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Suzhou Metabrain Intelligent Technology Co Ltd
Original Assignee
Suzhou Metabrain Intelligent Technology Co Ltd
Filing date
Publication date
Application filed by Suzhou Metabrain Intelligent Technology Co Ltd filed Critical Suzhou Metabrain Intelligent Technology Co Ltd
Priority to CN202410376235.4A priority Critical patent/CN117971505B/en
Publication of CN117971505A publication Critical patent/CN117971505A/en
Application granted granted Critical
Publication of CN117971505B publication Critical patent/CN117971505B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Abstract

The embodiment of the application provides a method and a device for deploying container applications, wherein the method comprises the following steps: acquiring the resource utilization rate of N nodes in a target cluster on M resource indexes to obtain a target resource utilization rate set, wherein N and M are integers which are larger than or equal to 1; determining the entropy value of each of the M resource indexes through the target resource utilization rate set to obtain M entropy values, and determining M weight values according to the weights of the M resource indexes; determining the score values of each node on M resource indexes according to the M weight values to obtain N score values of N nodes, and determining a target node in the N nodes according to the N score values; the container application is deployed to the target node. The application solves the problem of larger resource allocation difference caused by the fact that the resource scheduling strategies in the prior art allocate the same weight for different indexes, thereby achieving the effect of more balanced resource allocation.

Description

Method and device for deploying container application
Technical Field
The embodiment of the application relates to the field of computers, in particular to a method and a device for deploying container applications.
Background
In recent years, with development of cloud computing technology, deployment of application services on the cloud has become a trend. Cloud computing technology can provide flexible computing resources and storage space so that application services can be deployed and extended more flexibly. Meanwhile, cloud computing can also provide high-availability and high-reliability infrastructure, and stable operation of application services is guaranteed. Container technology provides a lighter weight and simpler solution for deployment of application services.
The resource scheduling is a core module for deploying application on the cloud, and an unreasonable resource scheduling strategy can cause the problems that the overall resource utilization rate of the cloud platform is low, a user deployment request cannot respond quickly, the service quality is reduced, the software and hardware facility cost of a service provider is improved, and the like.
The existing resource scheduling strategy takes a Central Processing Unit (CPU) and a memory as indexes to filter out nodes which do not meet the requirements in a cluster, so that different application requirements cannot be met, and the two indexes adopt the same weight to schedule the resources, so that excessive inclination of resource use is easily caused.
There is currently no effective solution to the above problems.
Disclosure of Invention
The embodiment of the application provides a method and a device for at least solving the problem of larger resource allocation difference caused by the fact that the same weight is allocated to different indexes by a resource scheduling strategy in the prior art.
According to one embodiment of the present application, there is provided a method of deploying a container application, comprising: acquiring the resource utilization rate of N nodes in a target cluster on M resource indexes to obtain a target resource utilization rate set, wherein N and M are integers which are larger than or equal to 1; determining an entropy value of each of the M resource indexes through the target resource utilization rate set to obtain M entropy values, obtaining weights of the resource indexes of the M resource indexes according to the M entropy values, and determining M weight values; determining the score values of each node on the M resource indexes according to the M weight values to obtain N score values of the N nodes, and determining a target node in the N nodes according to the N score values; a container application is deployed to the target node.
In an exemplary embodiment, determining, by the target resource utilization set, an entropy value of each of the M resource indicators, to obtain M entropy values, includes: normalizing the target resource utilization rate set according to the type of each resource index in the M resource indexes to obtain a first target matrix; determining a specific gravity value of each node in each resource index in the M resource indexes to the resource index through the first target matrix to obtain a second target matrix; and determining entropy values of all the M resource indexes through the second target matrix to obtain the M entropy values.
In an exemplary embodiment, the method further comprises: determining the graphics processor, the memory, the bandwidth and the read-write rate as very large indexes under the condition that the M resource indexes comprise the graphics processor, the memory, the bandwidth and the read-write rate; and when the M resource indexes comprise a central processing unit and a magnetic disk, determining the central processing unit and the magnetic disk as very small indexes.
In an exemplary embodiment, normalizing the target resource utilization set according to the type of each resource index in the M resource indexes to obtain a first target matrix, including: under the condition that the M resource indexes comprise S maximum indexes, S first resource utilization subsets are determined in the target resource utilization rate set, wherein S is an integer which is greater than or equal to 1 and less than or equal to M, and the first resource utilization rate subsets comprise the resource utilization rates of N nodes under one maximum index; carrying out maximum index normalization processing on the first resource utilization rate subset to obtain S first vectors, wherein the first vectors comprise normalized values of the resource utilization rates of N nodes under one maximum index, and the S first vectors are included in a first target matrix; determining T second resource utilization subsets in the target resource utilization set under the condition that the M resource indexes comprise T minimum indexes, wherein T is an integer which is greater than or equal to 1 and less than or equal to M, and the second resource utilization subsets comprise the resource utilization rates of N nodes under one minimum index; and carrying out minimum index normalization processing on the second resource utilization rate subset to obtain T second vectors, wherein the second vectors comprise normalized values of the resource utilization rates of N nodes under a minimum index, and the T second vectors are included in the first target matrix.
In an exemplary embodiment, performing a large scale index normalization process on the first subset of resource utilization rates includes: determining the largest resource utilization rate in the first resource utilization rate subset as a first largest resource utilization rate, and determining the smallest resource utilization rate in the first resource utilization rate subset as a first smallest resource utilization rate; determining a difference value between an ith resource utilization rate in the first resource utilization rate subset and the first minimum resource utilization rate as a first difference value, wherein i is an integer greater than or equal to 1, and the ith resource utilization rate is any resource utilization rate in the first resource utilization rate subset; determining a difference between the first maximum resource utilization and the first minimum resource utilization as a second difference; and determining the ratio of the first difference value to the second difference value as a normalized value of the ith resource utilization rate.
In an exemplary embodiment, performing a minimum index normalization process on the second subset of resource utilization rates includes: determining the largest resource utilization rate in the second resource utilization rate subset as a second largest resource utilization rate, and determining the smallest resource utilization rate in the second resource utilization rate subset as a second smallest resource utilization rate; determining a difference between the second maximum resource utilization rate and a jth resource utilization rate in the second resource utilization rate subset as a third difference, wherein j is an integer greater than or equal to 1, and the jth resource utilization rate is any resource utilization rate in the first resource utilization rate subset; determining a difference between the second maximum resource utilization and the second minimum resource utilization as a fourth difference; and determining the ratio of the third difference value to the fourth difference value as a normalized value of the current jth resource utilization rate.
In an exemplary embodiment, determining, by the first target matrix, a specific gravity value of each node in each of the M resource indexes to the resource index, to obtain a second target matrix, including: performing the following operations on an r-th normalized value in a kth vector of the first target matrix, the kth vector including normalized values of resource utilization of the N nodes on a kth index, the r-th normalized value being normalized values of resource utilization of the r-th node on the kth index, the k being an integer greater than or equal to 1 and less than or equal to M, the r being an integer greater than or equal to 1 and less than or equal to N, the first target matrix including M vectors: determining the sum of all normalization values in the kth vector as a target sum; and determining the ratio of the r normalization value to the target sum as the specific gravity value of the r node to the k index under the k index.
In an exemplary embodiment, determining, by the second target matrix, entropy values of each of the M resource indicators, to obtain the M entropy values, includes: performing the following operation on each specific gravity value in an xth vector in the second target matrix, where x is an integer greater than or equal to 1 and less than or equal to M, the xth vector including specific gravity values of each node in the xth resource index for the xth resource index by the following formula:
Wherein, Is the entropy value of the xth resource index, N is the number of nodes in the target cluster,/>Is the specific gravity value of the y node to the x resource index under the x resource index.
In an exemplary embodiment, the obtaining the weight of each of the M resource indexes according to the M entropy values includes: performing entropy redundancy processing on each entropy value in the M entropy values to obtain M entropy redundancies; and determining the weight of each resource index in the M resource indexes through the M entropy redundancies.
In an exemplary embodiment, performing entropy redundancy processing on each entropy value in the M entropy values to obtain M entropy redundancies, including: the following operations are performed on each of the M entropy values, and the entropy value on which the following operations are performed is referred to as a current entropy value: and determining the difference value between the preset value and the current entropy value as entropy redundancy corresponding to the current entropy value.
In an exemplary embodiment, determining the score values of each node on the M resource indexes according to the M weight values, to obtain N score values of the N nodes, including: acquiring the resource quantity required by the container application on the M resource indexes to obtain M resource quantities; acquiring the total resource quantity of each node on the M resource indexes to obtain N resource total quantity sets, wherein each resource total quantity set comprises M total resource quantities of one node on the M resource indexes; acquiring the number of used resources of each node on the M resource indexes to obtain N resource sets, wherein each resource set comprises the number of used resources of one node on the M resource indexes; and obtaining N scoring values of the N nodes through the M weight values, the M resource amounts, the N resource total amount sets and the N resource sets.
In an exemplary embodiment, obtaining N score values of the N nodes by using the M weight values, the M number of resources, the N total number of resources set, and the N number of resource sets includes: the score value of the current node in the N nodes is obtained by the following formula:
Where M is the number of resource indicators, The total resource number of index C,/>Is the weight of the C-th index,Is the sum of the amount of resources required by the container application on the C-th index and the amount of resources that have been used by the current node on the C-th index.
In an exemplary embodiment, determining a target node among the N nodes according to the N score values includes: determining a node with a score value larger than or equal to a preset value in the N nodes as the target node; or determining the node with the largest score value in the N nodes as the target node.
According to another embodiment of the present application, there is provided an apparatus for deploying a container application, including: the acquisition module is used for acquiring the resource utilization rates of N nodes in the target cluster on M resource indexes to obtain a target resource utilization rate set, wherein N and M are integers which are greater than or equal to 1; the first determining module is used for determining the entropy value of each of the M resource indexes through the target resource utilization rate set to obtain M entropy values, obtaining the weight of each of the M resource indexes according to the M entropy values, and determining M weight values; the second determining module is used for determining the score values of the nodes on the M resource indexes according to the M weight values to obtain N score values of the N nodes, and determining a target node in the N nodes according to the N score values; and the deployment module is used for deploying the container application to the target node.
According to a further embodiment of the application, there is also provided a computer readable storage medium having stored therein a computer program, wherein the computer program is arranged to perform the steps of any of the method embodiments described above when run.
According to a further embodiment of the application there is also provided an electronic device comprising a memory having stored therein a computer program and a processor arranged to run the computer program to perform the steps of any of the method embodiments described above.
According to a further embodiment of the application, there is also provided a computer program product comprising a computer program which, when executed by a processor, implements the steps of any of the method embodiments described above.
According to the method and the device, the resource utilization rate of N nodes in the target cluster on M resource indexes is obtained, and a target resource utilization rate set is obtained, wherein N and M are integers which are larger than or equal to 1; determining the entropy value of each of the M resource indexes through the target resource utilization rate set to obtain M entropy values, and determining M weight values according to the weights of the M resource indexes; determining the score values of each node on M resource indexes according to M weight values to obtain N score values of N nodes, and determining a target node in the N nodes according to the N score values; the container application is deployed to the target node.
The entropy value of each resource index is obtained according to the resource utilization rate of each resource index on each node, the weight of each resource index is determined according to the entropy value, the score of each node is calculated, and the target node is determined. Therefore, the problem that the resource allocation difference is large due to the fact that the same weight is allocated to different indexes by the resource scheduling strategy in the prior art can be solved, and the effect of more balanced resource allocation is achieved.
Drawings
FIG. 1 is a flow chart of Kubernetes resource scheduling according to an embodiment of the present application;
FIG. 2 is a hardware block diagram of a server device of a method of deploying a container application according to an embodiment of the application;
FIG. 3 is a flow chart of a method of deploying a container application according to an embodiment of the application;
FIG. 4 is a flow chart of obtaining a target set of resource utilizations in accordance with an embodiment of the application;
FIG. 5 is a flow chart of resource indicator weight calculation according to an embodiment of the application;
Fig. 6 is a block diagram of an apparatus for deploying a container application in accordance with an embodiment of the application.
Detailed Description
Kubernetes is an open-source container orchestration engine, which is representative of container orchestration technology. The application of the technology is mainly based on a Docker container technology, so that the deployment and management of a containerized application program are simplified, the automation of resource management is realized, and the utilization rate of resources across a plurality of data centers is maximized. Kubernetes has complete cluster management capabilities including multi-level security and admission mechanisms, multi-tenant application support capabilities, transparent service registration and service discovery mechanisms, extensible resource scheduling mechanisms, and the like.
In the Kubernetes cluster, usually, a pod is used as a basic unit of resource scheduling, nodes which do not meet requirements are filtered according to the minimum resource requirement of the pod application submitted by a user, the resource utilization rate of the residual CPU and the memory of the nodes is used as a grading index, candidate nodes are graded, and the nodes with the highest scores are selected for deployment. The containers are automatically scheduled to different nodes according to the resource conditions in the clusters, so that load balancing and resource optimization are realized, and the requirements of running programs in products are met. Fig. 1 is a flowchart of Kubernetes resource scheduling according to an embodiment of the present application, as shown in fig. 1, the specific flow is as follows:
S01, creating a pod object to be executed with information such as mirror images of containers, resource requirements and the like;
s02, selecting a proper node for deploying pod according to the resource condition and the scheduling policy of the node;
S03, the scheduler binds the pod to the selected node, so that the running position of the pod is determined;
S04, creating and managing a container according to the description of the pod object by kubelet on the node, and ensuring that the container is always rerun;
s05, periodically monitoring the state of the node, and triggering a rescheduling process if the node fails or has insufficient resources.
Although the Kubernetes scheduling policy can automatically deploy application programs to suitable nodes and perform load balancing according to the load condition of the nodes, the scheduling policy only considers two resource indexes of a CPU and a memory in the selection of scoring indexes, so that different application requirements cannot be met, the two indexes adopt the same weight to calculate the score, the different resource requirements of pod applications cannot be met, when the number of pod application deployments is gradually increased, excessive waste of resources such as bandwidth, IO rate and the like can be caused, and therefore the problems that the overall resource utilization rate of a cloud platform is low, the deployment request of a user cannot respond quickly, the service quality is reduced, the software and hardware facility cost of a server is improved and the like are caused.
Aiming at the problems, the application provides a method and a device for deploying container application, which are based on a K8S cluster environment and ubutu 20.04.04 operating system, add evaluation indexes such as GPU, bandwidth, disk, IO rate and the like on the basis of original evaluation indexes, and simultaneously determine the weight of each index by utilizing an improved entropy weight method, thereby solving the problem of resource waste caused by unreasonable weight of resource indexes in the prior art, effectively improving the overall resource balance degree of the cluster, and avoiding the situation that single resources on nodes in the cluster are exhausted and other resources remain.
Embodiments of the present application will be described in detail below with reference to the accompanying drawings in conjunction with the embodiments.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present application and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order.
The method embodiments provided in the embodiments of the present application may be executed in a server apparatus or similar computing device. Taking the example of running on a server device, fig. 2 is a block diagram of the hardware architecture of the server device of a method for deploying a container application according to an embodiment of the present application. As shown in fig. 2, the server device may include one or more (only one is shown in fig. 2) processors 202 (the processor 202 may include, but is not limited to, a microprocessor MCU or a processing means such as a programmable logic device FPGA) and a memory 204 for storing data, wherein the server device may further include a transmission device 206 for communication functions and an input-output device 208. It will be appreciated by those of ordinary skill in the art that the structure shown in fig. 2 is merely illustrative and is not intended to limit the structure of the server apparatus described above. For example, the server device may also include more or fewer components than shown in fig. 2, or have a different configuration than shown in fig. 2.
The memory 204 may be used to store computer programs, such as software programs of application software and modules, such as computer programs corresponding to the methods of deploying container applications in embodiments of the present application, and the processor 202 executes the computer programs stored in the memory 204 to perform various functional applications and data processing, i.e., to implement the methods described above. Memory 204 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, memory 204 may further include memory remotely located relative to processor 202, which may be connected to the server device via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The transmission device 206 is used to receive or transmit data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider of a server device. In one example, the transmission device 206 includes a network adapter (Network Interface Controller, simply referred to as a NIC) that can connect to other network devices through a base station to communicate with the internet. In one example, the transmission device 206 may be a Radio Frequency (RF) module, which is used to communicate with the internet wirelessly.
In this embodiment, a method for deploying a container application is provided, and fig. 3 is a flowchart of a method for deploying a container application according to an embodiment of the present application, as shown in fig. 3, where the flowchart includes the following steps:
Step S302, obtaining the resource utilization rate of N nodes in a target cluster on M resource indexes to obtain a target resource utilization rate set, wherein N and M are integers which are larger than or equal to 1;
The target cluster may be a Kubernetes cluster, the nodes may be node nodes in a Kubernetes cluster environment, each node may run multiple pod, each node is provided with one or more Central Processing Units (CPUs) and/or Graphics Processing Units (GPUs), the resource index may be one or more of indexes such as CPU, memory, GPU, bandwidth, disk, IO rate, etc., and the resource utilization is a ratio of used resources to total resources. The target resource utilization rate set is obtained by obtaining the resource utilization rate of each resource index on each node in the target cluster at the current time point, and the target resource utilization rate set can be an n×m-order matrix, for example, p= [ P xy]N*M ].
As an alternative implementation manner, fig. 4 is a flowchart of obtaining a target resource utilization rate set according to an embodiment of the present application, as shown in fig. 4, a timer task device and a controller are disposed on a management node master of a cluster, a resource collector is disposed on each node of the cluster, and a resource status information script of a collection node is disposed on each node of the cluster, where the resource collector collects resource utilization rates of relevant resource indexes of each node after receiving a monitoring command, and meanwhile, a database is disposed, and the collected resource utilization rates are stored in the database by the controller.
Step S304, determining the entropy value of each resource index in the M resource indexes through the target resource utilization rate set to obtain M entropy values, obtaining the weight of each resource index in the M resource indexes according to the M entropy values, and determining M weight values;
Specifically, normalizing the target resource utilization rate set according to the type of each resource index in the M resource indexes to obtain a first target matrix; determining a specific gravity value of each node in each resource index in the M resource indexes to the resource index through the first target matrix to obtain a second target matrix; and determining entropy values of all the M resource indexes through the second target matrix to obtain the M entropy values.
The first target matrix q= [ b xy]N*M ] is a matrix obtained by normalizing each data in the target resource utilization rate set under each resource index, calculating the duty ratio of the normalized value in the index under each resource index, obtaining a second target matrix f= [ F xy]N*M ], and then calculating the entropy value e x, x=1, 2, … and M of each resource index according to the second target matrix.
The data in the target resource utilization rate set can be unified to the same scale by normalizing the data, so that the influence of the dimension and range difference of the data in different dimensions on the result is avoided, the comparability and the interpretability of the data are improved, meanwhile, the importance of each resource index on the decision result can be reflected more accurately by the entropy weight calculated after the data normalization, the deviation and the error caused by the different dimensions of the data are effectively avoided, and the accuracy and the reliability of decision are improved.
As an optional implementation manner, after obtaining the M entropy values, performing entropy redundancy processing on each entropy value in the M entropy values to obtain M entropy redundancies; and determining the weight of each resource index in the M resource indexes through the M entropy redundancies.
Specifically, the weight of each of the M resource indicators may be obtained by the following formula:
Wherein, Is the weight of the x-th resource index,/>Is the entropy redundancy of the x-th resource index of the y-th node,Is the sum of the entropy redundancy of the resource index on the y-th node.
In the first target matrix, when the normalized value of the x-th resource index of the y-th node is smaller than or equal to the average value of the normalized values of all the x-th resource indexes, the weight of the x-th resource index of the y-th node is determined to be the weight of the x-th resource index, and when the normalized value of the x-th resource index of the y-th node is larger than the average value of the normalized values of all the x-th resource indexes, the weight of the x-th resource index of the y-th node is determined to be 0, so as to obtain the weight matrix.
According to the normalized numerical value in the first target matrix, the weight of each resource index on each node is assigned, so that the problem of unreasonable assignment caused by the multiple change of the entropy weight due to the amplification of small difference when all entropy values approach 1 after the normalization processing of the entropy values can be solved, and adverse effects caused by decision or evaluation errors due to unreasonable entropy weights are avoided.
As an alternative embodiment, the following operation is performed on each of the M entropy values, and the entropy value on which the following operation is performed is referred to as a current entropy value: and determining the difference value between the preset value and the current entropy value as entropy redundancy corresponding to the current entropy value.
Wherein, the preset value may be 1, and the specific calculation formula is as follows:
In the method, in the process of the invention, Is the entropy redundancy of the xth resource indicator,Is a value of said preset value of said time period,Is the entropy value of the xth resource indicator.
The entropy value is obtained through the resource utilization rate, and the weight is determined according to the entropy value, so that the importance of different resource indexes in a data set can be better quantified, the contribution of each resource index can be better considered when a decision is made, and the rationality of node allocation and the balance degree of resources are improved. In addition, by introducing calculation of the redundancy of the entropy value, the stability, reliability and performance of the system can be improved to a certain extent.
Step S306, determining the score values of each node on the M resource indexes according to the M weight values to obtain N score values of the N nodes, and determining a target node in the N nodes according to the N score values;
Specifically, obtaining the number of resources required by the container application on the M resource indexes to obtain M resource numbers; acquiring the total resource quantity of each node on the M resource indexes to obtain N resource total quantity sets, wherein each resource total quantity set comprises M total resource quantities of one node on the M resource indexes; acquiring the number of used resources of each node on the M resource indexes to obtain N resource sets, wherein each resource set comprises the number of used resources of one node on the M resource indexes; and obtaining N scoring values of the N nodes through the M weight values, the M resource amounts, the N resource total amount sets and the N resource sets.
By comprehensively considering the actual resource condition and the use condition of the resources of the nodes, the score value of each node is calculated, the resource can be optimized and adjusted in a targeted manner, the container application with a large resource demand can be scheduled to the node with more residual resources, the problem that the whole node cannot redistribute new container applications due to over-allocation of a certain resource of a single node is avoided, the fact that each node runs a plurality of pod without exceeding the capacity of the node is ensured, and the efficiency and the performance of the whole system are improved.
As an alternative embodiment, the score value of the current node in the N nodes is obtained by the following formula:
Where M is the number of resource indicators, The total resource number of index C,/>Is the weight of the C-th index,Is the sum of the amount of resources required by the container application on the C-th index and the amount of resources that have been used by the current node on the C-th index.
Step S308, deploying the container application to the target node.
Specifically, determining a node with a score value greater than or equal to a preset value among the N nodes as the target node; or determining the node with the largest score value in the N nodes as the target node.
And determining the priority according to the scores of the nodes, and preferentially distributing the resources to the nodes with high scores, so that the deployment of container application is realized, the optimal utilization of the resources is realized, and the efficiency of the whole system is improved.
Through the steps, the resource utilization rate of N nodes in the target cluster on M resource indexes is obtained, and a target resource utilization rate set is obtained, wherein N and M are integers which are larger than or equal to 1; determining the entropy value of each of the M resource indexes through the target resource utilization rate set to obtain M entropy values, and determining M weight values according to the weights of the M resource indexes; determining the score values of each node on M resource indexes according to the M weight values to obtain N score values of N nodes, and determining a target node in the N nodes according to the N score values; the container application is deployed to the target node. The method solves the problem of large resource allocation difference caused by the fact that the same weight is allocated to different indexes by a resource scheduling strategy in the prior art, achieves the effect of more balanced resource allocation, and improves the resource balance degree.
The main execution body of the above steps may be a server, a terminal, or the like, but is not limited thereto.
By increasing the resource indexes such as GPU, bandwidth, disk, IO rate and the like and improving the weight of each resource index obtained by the entropy weight method, proper nodes are selected for container application to be deployed, the overall resource balance of the cluster is effectively improved, and the situation that single resources on the nodes in the cluster are exhausted and other resources remain is avoided.
As an optional implementation manner, in the case that the M resource indexes include a graphics processor, a memory, a bandwidth and a read-write rate, determining the graphics processor, the memory, the bandwidth and the read-write rate as very large indexes; and when the M resource indexes comprise a central processing unit and a magnetic disk, determining the central processing unit and the magnetic disk as very small indexes.
As an optional implementation manner, in a case that the M resource indexes include S maximum indexes, determining S first resource utilization subsets in the target resource utilization set, where S is an integer greater than or equal to 1 and less than or equal to M, where the first resource utilization subsets include resource utilization of N nodes under one maximum index; and carrying out maximum index normalization processing on the first resource utilization rate subset to obtain S first vectors, wherein the first vectors comprise normalized values of the resource utilization rates of N nodes under one maximum index, and the S first vectors are included in a first target matrix.
Optionally, performing a large scale index normalization process on the first resource utilization subset, including: determining the largest resource utilization rate in the first resource utilization rate subset as a first largest resource utilization rate, and determining the smallest resource utilization rate in the first resource utilization rate subset as a first smallest resource utilization rate; determining a difference value between an ith resource utilization rate in the first resource utilization rate subset and the first minimum resource utilization rate as a first difference value, wherein i is an integer greater than or equal to 1, and the ith resource utilization rate is any resource utilization rate in the first resource utilization rate subset; determining a difference between the first maximum resource utilization and the first minimum resource utilization as a second difference; and determining the ratio of the first difference value to the second difference value as a normalized value of the ith resource utilization rate.
The specific calculation formula is as follows:
Wherein, Is a normalized value of the ith resource utilization,/>Is the i-th resource utilization in the first subset of resource utilization,/>Is the first minimum resource utilization,/>Is the first maximum resource utilization.
As an optional implementation manner, in the case that the M resource indexes include T minimum indexes, determining T second resource utilization subsets in the target resource utilization set, where T is an integer greater than or equal to 1 and less than or equal to M, where the second resource utilization subsets include resource utilization of N nodes under one minimum index; and carrying out minimum index normalization processing on the second resource utilization rate subset to obtain T second vectors, wherein the second vectors comprise normalized values of the resource utilization rates of N nodes under a minimum index, and the T second vectors are included in the first target matrix.
Optionally, performing a minimum index normalization process on the second subset of resource utilization rates, including: determining the largest resource utilization rate in the second resource utilization rate subset as a second largest resource utilization rate, and determining the smallest resource utilization rate in the second resource utilization rate subset as a second smallest resource utilization rate; determining a difference between the second maximum resource utilization rate and a jth resource utilization rate in the second resource utilization rate subset as a third difference, wherein j is an integer greater than or equal to 1, and the jth resource utilization rate is any resource utilization rate in the first resource utilization rate subset; determining a difference between the second maximum resource utilization and the second minimum resource utilization as a fourth difference; and determining the ratio of the third difference value to the fourth difference value as a normalized value of the current jth resource utilization rate.
The specific calculation formula is as follows:
Wherein, Is a normalized value of the j-th resource utilization,/>Is the j-th resource utilization in the second subset of resource utilization,/>Is the second maximum resource utilization,/>Is the second minimum resource utilization.
As an optional implementation manner, determining, by the first target matrix, a specific gravity value of each node in each of the M resource indexes to the resource index, to obtain a second target matrix, where the second target matrix includes: performing the following operations on an r-th normalized value in a kth vector of the first target matrix, the kth vector including normalized values of resource utilization of the N nodes on a kth index, the r-th normalized value being normalized values of resource utilization of the r-th node on the kth index, the k being an integer greater than or equal to 1 and less than or equal to M, the r being an integer greater than or equal to 1 and less than or equal to N, the first target matrix including M vectors: determining the sum of all normalization values in the kth vector as a target sum; and determining the ratio of the r normalization value to the target sum as the specific gravity value of the r node to the k index under the k index.
The specific calculation formula is as follows:
In the method, in the process of the invention, Is the specific gravity value of the r node to the k index under the k index,/>Is the r < th > normalized value,/>Is the target sum.
As an optional implementation manner, determining, by the second target matrix, an entropy value of each of the M resource indexes, to obtain the M entropy values, including: performing the following operation on each specific gravity value in an xth vector in the second target matrix, where x is an integer greater than or equal to 1 and less than or equal to M, the xth vector including specific gravity values of each node in the xth resource index for the xth resource index by the following formula:
Wherein, Is the entropy value of the xth resource index, N is the number of nodes in the target cluster,/>Is the specific gravity value of the y node to the x resource index under the x resource index.
As an alternative implementation, fig. 5 is a flowchart of calculation of weights of resource indicators according to an embodiment of the present application, as shown in fig. 5,
Step 1, normalizing node resource information data;
And constructing a matrix X= (X ij )n×m,（i=1,2,…,n;j=1,2,…,m）,xij represents the resource utilization rate of the jth resource index on the ith node) by monitoring the real-time data of the resource utilization rate acquired by the acquisition unit.
Meanwhile, considering that the measurement units of each resource index are different, before calculating the comprehensive resource index, carrying out standardization processing, namely converting the absolute value of the resource index into a relative value, and normalizing the dimensionless value of the dimensionless value to be { b ij } according to the resource utilization rate x ij of the jth resource index on the ith node. As for large model training tasks, indexes such as GPU resources, bandwidth, memory and IO rate have larger influence on performance compared with a CPU and a disk, the resource indexes are classified and calculated.
1) The method is characterized in that the larger the value, the better the effect is, and the method is calculated as:
2) The normalization of the minimum index is used for processing the CPU and the disk index parameters, and the decision parameters are characterized in that the smaller the value is, the better the effect is, and the calculation is as follows:
Wherein, And/>The maximum value and the minimum value of the j-th resource index are respectively represented, and the closer the index x value is to 1, the better the performance of the evaluation object is represented.
Step 2, calculating index value proportion f ij of the ith node under the jth index;
Step3, calculating the entropy value of the j index;
step 4, calculating the weight of each index.
Note that, for the index value lower than or equal to the average level,Assigning entropy value of corresponding j-th index/>For index values above average,/>The value is 0. /(I)
From the description of the above embodiments, it will be clear to a person skilled in the art that the method according to the above embodiments may be implemented by means of software plus the necessary general hardware platform, but of course also by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. ROM/RAM, magnetic disk, optical disk) comprising instructions for causing a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to perform the method according to the embodiments of the present application.
The embodiment also provides a device for deploying container applications, which is used for implementing the foregoing embodiments and preferred embodiments, and is not described in detail. As used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function. While the means described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and contemplated.
Fig. 6 is a block diagram of an apparatus for deploying a container application according to an embodiment of the application, as shown in fig. 6, the apparatus comprising: an obtaining module 602, configured to obtain resource utilization rates of N nodes in a target cluster on M resource indexes, to obtain a target resource utilization rate set, where N and M are integers greater than or equal to 1; a first determining module 604, configured to determine, according to the target resource utilization set, an entropy value of each of the M resource indexes, obtain M entropy values, obtain weights of each of the M resource indexes according to the M entropy values, and determine M weight values; a second determining module 606, configured to determine, according to the M weight values, a score value of each node on the M resource indexes, obtain N score values of the N nodes, and determine a target node among the N nodes according to the N score values; a deployment module 608 for deploying the container application to the target node.
As an optional implementation manner, normalizing the target resource utilization rate set according to the type of each resource index in the M resource indexes to obtain a first target matrix; determining a specific gravity value of each node in each resource index in the M resource indexes to the resource index through the first target matrix to obtain a second target matrix; and determining entropy values of all the M resource indexes through the second target matrix to obtain the M entropy values.
As an optional implementation manner, in the case that the M resource indexes include a graphics processor, a memory, a bandwidth and a read-write rate, determining the graphics processor, the memory, the bandwidth and the read-write rate as very large indexes; and when the M resource indexes comprise a central processing unit and a magnetic disk, determining the central processing unit and the magnetic disk as very small indexes.
As an optional implementation manner, in a case that the M resource indexes include S maximum indexes, determining S first resource utilization subsets in the target resource utilization set, where S is an integer greater than or equal to 1 and less than or equal to M, where the first resource utilization subsets include resource utilization of N nodes under one maximum index; carrying out maximum index normalization processing on the first resource utilization rate subset to obtain S first vectors, wherein the first vectors comprise normalized values of the resource utilization rates of N nodes under one maximum index, and the S first vectors are included in a first target matrix; determining T second resource utilization subsets in the target resource utilization set under the condition that the M resource indexes comprise T minimum indexes, wherein T is an integer which is greater than or equal to 1 and less than or equal to M, and the second resource utilization subsets comprise the resource utilization rates of N nodes under one minimum index; and carrying out minimum index normalization processing on the second resource utilization rate subset to obtain T second vectors, wherein the second vectors comprise normalized values of the resource utilization rates of N nodes under a minimum index, and the T second vectors are included in the first target matrix.
As an optional implementation manner, determining the largest resource utilization rate in the first resource utilization rate subset as the first largest resource utilization rate, and determining the smallest resource utilization rate in the first resource utilization rate subset as the first smallest resource utilization rate; determining a difference value between an ith resource utilization rate in the first resource utilization rate subset and the first minimum resource utilization rate as a first difference value, wherein i is an integer greater than or equal to 1, and the ith resource utilization rate is any resource utilization rate in the first resource utilization rate subset; determining a difference between the first maximum resource utilization and the first minimum resource utilization as a second difference; and determining the ratio of the first difference value to the second difference value as a normalized value of the ith resource utilization rate.
As an optional implementation manner, determining the largest resource utilization rate in the second resource utilization rate subset as the second largest resource utilization rate, and determining the smallest resource utilization rate in the second resource utilization rate subset as the second smallest resource utilization rate; determining a difference between the second maximum resource utilization rate and a jth resource utilization rate in the second resource utilization rate subset as a third difference, wherein j is an integer greater than or equal to 1, and the jth resource utilization rate is any resource utilization rate in the first resource utilization rate subset; determining a difference between the second maximum resource utilization and the second minimum resource utilization as a fourth difference; and determining the ratio of the third difference value to the fourth difference value as a normalized value of the current jth resource utilization rate.
As an alternative embodiment, the following operations are performed on an nth normalized value in a kth vector of the first target matrix, where the kth vector includes normalized values of resource utilization of the N nodes on a kth index, the nth normalized value is normalized values of resource utilization of the kth node on the kth index, k is an integer greater than or equal to 1 and less than or equal to M, and r is an integer greater than or equal to 1 and less than or equal to N, and the first target matrix includes M vectors: determining the sum of all normalization values in the kth vector as a target sum; and determining the ratio of the r normalization value to the target sum as the specific gravity value of the r node to the k index under the k index.
As an alternative embodiment, the following operations are performed on each specific gravity value in an xth vector in the second target matrix by the following formula, where x is an integer greater than or equal to 1 and less than or equal to M, and the xth vector includes the specific gravity value of each node in the xth resource index:
Wherein, Is the entropy value of the xth resource index, N is the number of nodes in the target cluster,/>Is the specific gravity value of the y node to the x resource index under the x resource index.
As an optional implementation manner, performing entropy redundancy processing on each entropy value in the M entropy values to obtain M entropy redundancies; and determining the weight of each resource index in the M resource indexes through the M entropy redundancies.
As an alternative embodiment, the following operation is performed on each of the M entropy values, and the entropy value on which the following operation is performed is referred to as a current entropy value: and determining the difference value between the preset value and the current entropy value as entropy redundancy corresponding to the current entropy value.
As an optional implementation manner, obtaining the number of resources required by the container application on the M resource indexes, to obtain M number of resources; acquiring the total resource quantity of each node on the M resource indexes to obtain N resource total quantity sets, wherein each resource total quantity set comprises M total resource quantities of one node on the M resource indexes; acquiring the number of used resources of each node on the M resource indexes to obtain N resource sets, wherein each resource set comprises the number of used resources of one node on the M resource indexes; and obtaining N scoring values of the N nodes through the M weight values, the M resource amounts, the N resource total amount sets and the N resource sets.
As an alternative embodiment, the score value of the current node in the N nodes is obtained by the following formula:
Where M is the number of resource indicators, The total resource number of index C,/>Is the weight of index C,/>Is the sum of the amount of resources required by the container application on the C-th index and the amount of resources that have been used by the current node on the C-th index.
As an optional implementation manner, determining a node with a score value greater than or equal to a preset value in the N nodes as the target node; or determining the node with the largest score value in the N nodes as the target node.
It should be noted that each of the above modules may be implemented by software or hardware, and for the latter, it may be implemented by, but not limited to: the modules are all located in the same processor; or the above modules may be located in different processors in any combination.
Embodiments of the present application also provide a computer readable storage medium having a computer program stored therein, wherein the computer program is arranged to perform the steps of any of the method embodiments described above when run.
In one exemplary embodiment, the computer readable storage medium may include, but is not limited to: a usb disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory RAM), a removable hard disk, a magnetic disk, or an optical disk, or other various media capable of storing a computer program.
An embodiment of the application also provides an electronic device comprising a memory having stored therein a computer program and a processor arranged to run the computer program to perform the steps of any of the method embodiments described above.
In an exemplary embodiment, the electronic device may further include a transmission device connected to the processor, and an input/output device connected to the processor.
Embodiments of the application also provide a computer program product comprising a computer program which, when executed by a processor, implements the steps of any of the method embodiments described above.
Specific examples in this embodiment may refer to the examples described in the foregoing embodiments and the exemplary implementation, and this embodiment is not described herein.
It will be appreciated by those skilled in the art that the modules or steps of the application described above may be implemented in a general purpose computing device, they may be concentrated on a single computing device, or distributed across a network of computing devices, they may be implemented in program code executable by computing devices, so that they may be stored in a storage device for execution by computing devices, and in some cases, the steps shown or described may be performed in a different order than that shown or described herein, or they may be separately fabricated into individual integrated circuit modules, or multiple modules or steps of them may be fabricated into a single integrated circuit module. Thus, the present application is not limited to any specific combination of hardware and software.
The above description is only of the preferred embodiments of the present application and is not intended to limit the present application, but various modifications and variations can be made to the present application by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the principle of the present application should be included in the protection scope of the present application.

Claims (13)

1. A method of deploying a container application, comprising:
Acquiring the resource utilization rate of N nodes in a target cluster on M resource indexes to obtain a target resource utilization rate set, wherein N and M are integers which are larger than or equal to 1;
Normalizing the target resource utilization rate set according to the type of each resource index in the M resource indexes to obtain a first target matrix; determining a specific gravity value of each node in each resource index in the M resource indexes to the resource index through the first target matrix to obtain a second target matrix; determining entropy values of all the M resource indexes through the second target matrix to obtain M entropy values, and performing entropy redundancy processing on all the entropy values in the M entropy values to obtain M entropy redundancies; determining weights of the resource indexes in the M resource indexes through the M entropy redundancies to obtain M weight values;
Acquiring the resource quantity required by the container application on the M resource indexes to obtain M resource quantities; acquiring the total resource quantity of each node on the M resource indexes to obtain N resource total quantity sets, wherein each resource total quantity set comprises M total resource quantities of one node on the M resource indexes; acquiring the number of used resources of each node on the M resource indexes to obtain N resource sets, wherein each resource set comprises the number of used resources of one node on the M resource indexes; obtaining N score values of the N nodes through the M weight values, the M resource amounts, the N resource total amount sets and the N resource sets, and determining a target node in the N nodes according to the N score values;
deploying a container application to the target node;
Wherein the following operations are performed on respective specific gravity values in an xth vector in the second target matrix, where x is an integer greater than or equal to 1 and less than or equal to M, the xth vector including specific gravity values of the xth resource index for each of the nodes under the xth resource index:
，
Wherein, Is the entropy value of the xth resource index, N is the number of nodes in the target cluster,/>Is the specific gravity value of the y node to the x resource index under the x resource index.
2. The method according to claim 1, wherein the method further comprises:
Determining the graphics processor, the memory, the bandwidth and the read-write rate as very large indexes under the condition that the M resource indexes comprise the graphics processor, the memory, the bandwidth and the read-write rate;
and when the M resource indexes comprise a central processing unit and a magnetic disk, determining the central processing unit and the magnetic disk as very small indexes.
3. The method of claim 2, wherein normalizing the set of target resource utilization rates according to the type of each of the M resource metrics to obtain a first target matrix comprises:
Under the condition that the M resource indexes comprise S maximum indexes, S first resource utilization subsets are determined in the target resource utilization rate set, wherein S is an integer which is greater than or equal to 1 and less than or equal to M, and the first resource utilization rate subsets comprise the resource utilization rates of N nodes under one maximum index; carrying out maximum index normalization processing on the first resource utilization rate subset to obtain S first vectors, wherein the first vectors comprise normalized values of the resource utilization rates of N nodes under one maximum index, and the S first vectors are included in a first target matrix;
Determining T second resource utilization subsets in the target resource utilization set under the condition that the M resource indexes comprise T minimum indexes, wherein T is an integer which is greater than or equal to 1 and less than or equal to M, and the second resource utilization subsets comprise the resource utilization rates of N nodes under one minimum index; and carrying out minimum index normalization processing on the second resource utilization rate subset to obtain T second vectors, wherein the second vectors comprise normalized values of the resource utilization rates of N nodes under a minimum index, and the T second vectors are included in the first target matrix.
4. A method according to claim 3, wherein performing a maximum index normalization process on the first subset of resource utilizations comprises:
determining the largest resource utilization rate in the first resource utilization rate subset as a first largest resource utilization rate, and determining the smallest resource utilization rate in the first resource utilization rate subset as a first smallest resource utilization rate;
determining a difference value between an ith resource utilization rate in the first resource utilization rate subset and the first minimum resource utilization rate as a first difference value, wherein i is an integer greater than or equal to 1, and the ith resource utilization rate is any resource utilization rate in the first resource utilization rate subset;
determining a difference between the first maximum resource utilization and the first minimum resource utilization as a second difference;
And determining the ratio of the first difference value to the second difference value as a normalized value of the ith resource utilization rate.
5. A method according to claim 3, wherein performing minimal index normalization on the second subset of resource utilizations comprises:
determining the largest resource utilization rate in the second resource utilization rate subset as a second largest resource utilization rate, and determining the smallest resource utilization rate in the second resource utilization rate subset as a second smallest resource utilization rate;
determining a difference between the second maximum resource utilization rate and a jth resource utilization rate in the second resource utilization rate subset as a third difference, wherein j is an integer greater than or equal to 1, and the jth resource utilization rate is any resource utilization rate in the first resource utilization rate subset;
Determining a difference between the second maximum resource utilization and the second minimum resource utilization as a fourth difference;
and determining the ratio of the third difference value to the fourth difference value as a normalized value of the current j-th resource utilization rate.
6. The method of claim 1, wherein determining, by the first target matrix, a specific gravity value of each node in each of the M resource indexes to the resource index, to obtain a second target matrix, includes:
Performing the following operations on an r-th normalized value in a kth vector of the first target matrix, the kth vector including normalized values of resource utilization of the N nodes on a kth index, the r-th normalized value being normalized values of resource utilization of the r-th node on the kth index, the k being an integer greater than or equal to 1 and less than or equal to M, the r being an integer greater than or equal to 1 and less than or equal to N, the first target matrix including M vectors:
Determining the sum of all normalized values in the kth vector as a target sum;
And determining the ratio of the r normalization value to the target sum as the specific gravity value of the r node to the k index under the k index.
7. The method of claim 1, wherein performing entropy redundancy processing on each of the M entropy values to obtain M entropy redundancies, comprises:
The following operations are performed on each of the M entropy values, and the entropy value on which the following operations are performed is referred to as a current entropy value:
and determining the difference value between the preset value and the current entropy value as entropy redundancy corresponding to the current entropy value.
8. The method of claim 1, wherein obtaining N score values for the N nodes from the M weight values, the M number of resources, the N total number of resources set, the N number of resources set, comprises:
The score value of the current node in the N nodes is obtained by the following formula:
，
Where M is the number of resource indicators, The total resource number of index C,/>Is the weight of index C,/>Is the sum of the amount of resources required by the container application on the C-th index and the amount of resources that have been used by the current node on the C-th index.
9. The method of claim 1, wherein determining a target node among the N nodes based on the N score values comprises:
determining a node with a score value larger than or equal to a preset value in the N nodes as the target node; or alternatively
And determining the node with the largest score value in the N nodes as the target node.
10. An apparatus for deploying a container application, comprising:
The acquisition module is used for acquiring the resource utilization rates of N nodes in the target cluster on M resource indexes to obtain a target resource utilization rate set, wherein N and M are integers which are greater than or equal to 1;
The first determining module is used for carrying out normalization processing on the target resource utilization rate set according to the types of each resource index in the M resource indexes to obtain a first target matrix; determining a specific gravity value of each node in each resource index in the M resource indexes to the resource index through the first target matrix to obtain a second target matrix; determining entropy values of all the M resource indexes through the second target matrix to obtain M entropy values, and performing entropy redundancy processing on all the entropy values in the M entropy values to obtain M entropy redundancies; determining weights of the resource indexes in the M resource indexes through the M entropy redundancies to obtain M weight values;
The second determining module is used for obtaining the resource quantity required by the container application on the M resource indexes to obtain M resource quantities; acquiring the total resource quantity of each node on the M resource indexes to obtain N resource total quantity sets, wherein each resource total quantity set comprises M total resource quantities of one node on the M resource indexes; acquiring the number of used resources of each node on the M resource indexes to obtain N resource sets, wherein each resource set comprises the number of used resources of one node on the M resource indexes; obtaining N score values of the N nodes through the M weight values, the M resource amounts, the N resource total amount sets and the N resource sets, and determining a target node in the N nodes according to the N score values;
a deployment module for deploying a container application to the target node;
The apparatus is further configured to perform the following operation on each specific gravity value in an xth vector in the second target matrix, where x is an integer greater than or equal to 1 and less than or equal to M, and the xth vector includes specific gravity values of each of the nodes under the xth resource index for the xth resource index:
，
Wherein, Is the entropy value of the xth resource index, N is the number of nodes in the target cluster,/>Is the specific gravity value of the y node to the x resource index under the x resource index.
11. A computer-readable storage medium comprising,
The computer readable storage medium has stored therein a computer program, wherein the computer program when executed by a processor realizes the steps of the method as claimed in any of claims 1 to 9.
12. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that,
The processor, when executing the computer program, implements the steps of the method as claimed in any one of claims 1 to 9.
13. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method as claimed in any one of claims 1 to 9.
CN202410376235.4A 2024-03-29 Method and device for deploying container application Active CN117971505B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410376235.4A CN117971505B (en) 2024-03-29 Method and device for deploying container application

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410376235.4A CN117971505B (en) 2024-03-29 Method and device for deploying container application

Publications (2)

Publication Number Publication Date
CN117971505A CN117971505A (en) 2024-05-03
CN117971505B true CN117971505B (en) 2024-06-07

Family

ID=

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017045424A1 (en) * 2015-09-18 2017-03-23 乐视控股（北京）有限公司 Application program deployment system and deployment method
CN112925611A (en) * 2021-03-11 2021-06-08 南京邮电大学 Distributed container scheduling method and system based on shared GPU
CN113010270A (en) * 2021-04-08 2021-06-22 桂林电子科技大学 Kubernetes platform-based dynamic resource load balancing scheduling method and system
CN113342477A (en) * 2021-07-08 2021-09-03 河南星环众志信息科技有限公司 Container group deployment method, device, equipment and storage medium
CN115729671A (en) * 2022-11-16 2023-03-03 浙江大华技术股份有限公司 Resource scheduling method and related device
CN116932198A (en) * 2022-11-23 2023-10-24 中国移动通信有限公司研究院 Resource scheduling method, device, electronic equipment and readable storage medium
CN116991558A (en) * 2023-09-22 2023-11-03 苏州元脑智能科技有限公司 Computing power resource scheduling method, multi-architecture cluster, device and storage medium
CN117349023A (en) * 2023-11-15 2024-01-05 中国联合网络通信集团有限公司 Application deployment method, device and storage medium
CN117435337A (en) * 2023-08-16 2024-01-23 中国工商银行股份有限公司 Resource allocation method, device, electronic equipment and storage medium
WO2024016596A1 (en) * 2022-07-21 2024-01-25 天翼云科技有限公司 Container cluster scheduling method and apparatus, device, and storage medium

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2017045424A1 (en) * 2015-09-18 2017-03-23 乐视控股（北京）有限公司 Application program deployment system and deployment method
CN112925611A (en) * 2021-03-11 2021-06-08 南京邮电大学 Distributed container scheduling method and system based on shared GPU
CN113010270A (en) * 2021-04-08 2021-06-22 桂林电子科技大学 Kubernetes platform-based dynamic resource load balancing scheduling method and system
CN113342477A (en) * 2021-07-08 2021-09-03 河南星环众志信息科技有限公司 Container group deployment method, device, equipment and storage medium
WO2024016596A1 (en) * 2022-07-21 2024-01-25 天翼云科技有限公司 Container cluster scheduling method and apparatus, device, and storage medium
CN115729671A (en) * 2022-11-16 2023-03-03 浙江大华技术股份有限公司 Resource scheduling method and related device
CN116932198A (en) * 2022-11-23 2023-10-24 中国移动通信有限公司研究院 Resource scheduling method, device, electronic equipment and readable storage medium
CN117435337A (en) * 2023-08-16 2024-01-23 中国工商银行股份有限公司 Resource allocation method, device, electronic equipment and storage medium
CN116991558A (en) * 2023-09-22 2023-11-03 苏州元脑智能科技有限公司 Computing power resource scheduling method, multi-architecture cluster, device and storage medium
CN117349023A (en) * 2023-11-15 2024-01-05 中国联合网络通信集团有限公司 Application deployment method, device and storage medium

Similar Documents

Publication Publication Date Title
US20200287961A1 (en) Balancing resources in distributed computing environments
CN109918198B (en) Simulation cloud platform load scheduling system and method based on user characteristic prediction
CN109218355B (en) Load balancing engine, client, distributed computing system and load balancing method
CN112783649A (en) Cloud computing-oriented interactive perception containerized micro-service resource scheduling method
US11496413B2 (en) Allocating cloud computing resources in a cloud computing environment based on user predictability
CN110287245A (en) Method and system for scheduling and executing distributed ETL (extract transform load) tasks
CN114356587B (en) Calculation power task cross-region scheduling method, system and equipment
CN110221920B (en) Deployment method, device, storage medium and system
CN116541134B (en) Method and device for deploying containers in multi-architecture cluster
CN115543577B (en) Covariate-based Kubernetes resource scheduling optimization method, storage medium and device
CN115134371A (en) Scheduling method, system, equipment and medium containing edge network computing resources
CN111930493A (en) NodeManager state management method and device in cluster and computing equipment
CN116467082A (en) Big data-based resource allocation method and system
CN111767145A (en) Container scheduling system, method, device and equipment
CN106506229B (en) A kind of SBS cloud application adaptive resource optimizes and revises system and method
CN117971505B (en) Method and device for deploying container application
CN115665157B (en) Balanced scheduling method and system based on application resource types
Ravandi et al. A black-box self-learning scheduler for cloud block storage systems
CN117971505A (en) Method and device for deploying container application
CN113301087B (en) Resource scheduling method, device, computing equipment and medium
CN114090201A (en) Resource scheduling method, device, equipment and storage medium
CN112612579A (en) Virtual machine deployment method, storage medium, and computer device
CN113127289A (en) Resource management method based on YARN cluster, computer equipment and storage medium
CN111459651A (en) Load balancing method, device, storage medium and scheduling system
CN114443258B (en) Resource scheduling method, device, equipment and storage medium for virtual machine

Legal Events

Date Code Title Description
PB01 Publication
SE01 Entry into force of request for substantive examination
GR01 Patent grant