CN108829494A - Container cloud platform intelligence method for optimizing resources based on load estimation - Google Patents
Container cloud platform intelligence method for optimizing resources based on load estimation Download PDFInfo
- Publication number
- CN108829494A CN108829494A CN201810664295.0A CN201810664295A CN108829494A CN 108829494 A CN108829494 A CN 108829494A CN 201810664295 A CN201810664295 A CN 201810664295A CN 108829494 A CN108829494 A CN 108829494A
- Authority
- CN
- China
- Prior art keywords
- load
- container
- physical node
- node
- cloud platform
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000013508 migration Methods 0.000 claims description 19
- 230000005012 migration Effects 0.000 claims description 19
- 238000005516 engineering process Methods 0.000 claims description 9
- 230000015572 biosynthetic process Effects 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 2
- 230000008676 import Effects 0.000 claims description 2
- 238000007689 inspection Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- 238000000611 regression analysis Methods 0.000 claims description 2
- 230000001373 regressive effect Effects 0.000 claims description 2
- 238000003786 synthesis reaction Methods 0.000 claims description 2
- 238000012731 temporal analysis Methods 0.000 claims description 2
- 238000000700 time series analysis Methods 0.000 claims description 2
- 230000009466 transformation Effects 0.000 claims description 2
- 230000007704 transition Effects 0.000 claims description 2
- 238000013519 translation Methods 0.000 claims description 2
- 238000005457 optimization Methods 0.000 abstract description 3
- 238000013480 data collection Methods 0.000 description 6
- 230000007423 decrease Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 239000011800 void material Substances 0.000 description 1
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
- G06F9/44—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
-
- 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]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
- G06F9/505—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals considering the load
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/60—Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network 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
- G06F2009/4557—Distribution of virtual machine instances; Migration and load balancing
Abstract
The invention discloses a kind of container cloud platform intelligence method for optimizing resources based on load estimation, belongs to container cloud platform field.This method is:According to the historic load of container instance, it is based on gray level model, predicts the loading condition of the next time window of each container instance;According to the load estimation value of containers all on each physical node, judge whether the load of section is excessively high or too low;Then corresponding dispatching algorithm is executed, the certain containers crossed on high node are moved into other node up, make the load of the node in the normal range;All container instances on load too low node are migrated to other nodes, the node is vacant out.The present invention utilizes unbalanced, scheduling of resource lag issues for current data center resources, introduce load estimation analysis, optimization is scheduled to data central loading in advance, it avoids the occurrence of the overweight bring performance loss of node load and load too low bring resource utilization is too low, improve platform resource utilization efficiency.
Description
Technical field
The present invention relates to a kind of container cloud platform intelligence method for optimizing resources based on load estimation belongs to cloud computing neck
Domain more specifically belongs to container cloud platform intelligent scheduling scope.
Background technique
Cloud computing technology rapidly develops, and the application based on cloud platform also emerges one after another.Cloud platform will by virtualization technology
Computer resource is integrated into resource pool, and elastic demand of the user to computing resource is realized in a manner of pay-for-use.Cloud computing
It is developed so far, virtualization technology is always the key technology in cloud platform, and container technique is then a kind of void emerging in recent years
Quasi-ization technology.Its appearance brings challenge to traditional virtual technology, provides new thinking to construct efficient cloud platform.
In Docker, LXC, Warden, numerous container techniques such as OpenVZ, what people had an optimistic view of the most is Docker container
Technology.Docker container technique has just had received widespread attention and has discussed since open source.Docker container technique is a kind of
Operating system layer virtualization technology, it is different from traditional virtual machine technique, it does not need operation Client OS, container with
The form of process operates in host operating system, this but also container have it is lighter than conventional virtual machine, flexibly, quickly
The advantages that deployment.Due to numerous novel characteristics and the opening of project itself, Docker obtain rapidly including Google,
The favor of the industries industry leaders such as Microsoft, VMware, and support is provided to it.Major cloud computing operator is being just now
In largely cloud platform of the building based on Docker container technique.However large-scale container cloud platform is in the prevalence of money now
The problems such as source utilization rate is low, and the utilization of resources is unbalanced, scheduling of resource lags.
The resource of different container demand different dimensions, when the resource exhaustion of an any dimension of node, if there is more
The container of dimension resource requirement is activated, then the node will not be able to satisfy the demand of creation container, cannot also run this container.
In this case, the surplus resources of other dimensions just have been idle, and cause resource utilization very low.In addition, several at present
The resource regulating method of all cloud platforms is all based on the scheduling of resource of monitoring, and cluster is by monitoring each node and appearance in real time
The loading condition of device is carried out corresponding dispatching algorithm, this dispatching method is easy to out if load value is greater than some threshold value
The problems such as existing scheduling of resource lag, until finding that overload when being scheduled, on the one hand may result under service quality
Drop, machine performance decline, seriously will lead to delay machine.It on the other hand, may be exactly some time point overload, later just just
Chang Liao.If such case is also scheduled, scheduling cost will be greatly increased.
Summary of the invention
(1) the technical issues of solving
Present invention solves the technical problem that being:How a kind of method for optimizing resources is provided, it is flat for solving present container cloud
The problems such as resource utilization existing for platform scheduling of resource is low, and the utilization of resources is unbalanced, scheduling of resource lags.
(2) technical solution
(3) beneficial effect
The present invention provides a kind of container cloud platform intelligence method for optimizing resources based on load estimation.This method is based on ash
It spends model and the loading condition of each container instance next period is predicted according to the historic load of container instance;According to some section
The resource requirement situation of all container instance future loads on point judges whether the node resource demand is excessively high or too low;Such as
Fruit Future demand is higher than the case where node load upper limit, migrates to the partial containers example on the node, prevents from taking
Quality of being engaged in decline, machine performance decline.If Future demand is lower than node load lower limit, to avoid the occurrence of node load mistake
Weight bring performance loss is executed and is dispatched the node container instance migration to other nodes, and the node is vacant out, is reduced
Resource overhead improves resource utilization.
Detailed description of the invention
Fig. 1 is the main framework of container cloud platform intelligence resource optimization;
Fig. 2 is the load estimation flow chart based on gray level model;
Fig. 3 is the load estimation schematic diagram based on time window;
Fig. 4 is the flow chart of scheduling strategy generating algorithm;
Fig. 5 is the flow chart of dispatching method;
Fig. 6 is load comparison diagram;
Fig. 7 is resource utilization comparison diagram.
Specific embodiment
In order to make the purpose of the present invention, technical solution and advantage are more clearly understood, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.
The main frame of inventive container cloud platform intelligence resource optimization is as shown in Figure 1, this approach includes the following steps:
(1) the various dimensions load data of each container instance of container cloud platform is obtained in real time, and it is corresponding to generate each dimension
Load time sequence.It specifically includes:
(1-1) runs cAdvisor tool on each container instance of container cloud platform, obtains container node in real time
Multiple dimensions such as resource utilization data, including CPU, memory, disk, network, and be sent on each physical node
Data Collection module.
(1-2) Data Collection module receives data.Take physical node number, container number and dimension name
Etc. information formed a JSON format load data.
Load data is sent to the Data Collection module that scheduling strategy generates node by (1-3).
The Data Collection module that (1-4) scheduling strategy generates node receives the load from each physical node
Data parse the load value of each dimension of JSON data, are numbered according to physical node, container number and dimension star
Load value is inserted into corresponding load time sequence back.
(2) according to the historic load of container instance, by the load estimation technology based on Time-Series analysis, prediction is each
The load of container instance future time window.As shown in Fig. 2, specifically including:
(2-1) obtains n*m newest each dimension of container instance from corresponding historic load time series
Historic load;It is divided into n time window, includes m historic load in each time window;The value of n is general
Be 30, m be time window size, according to the actual situation depending on, generally as data cycle size.
(2-2) constitutes a time series as shown in figure 3, extract i-th of historic load in each time windowIt include m historic load in each time window, so m time series of formation is
(2-3) is the validity for guaranteeing GM (1,1) model method, needs to do data necessary inspection processing.It calculatesThe grade ratio of sequence.Method is:
Wherein,Indicate i-th of historic load in k-th of time window.
If all grades can hold covering section than all falling inIt is interior, then ordered series of numbersIt can establish GM
(1,1) model and gray prediction can be carried out.
(2-4) does conversion process appropriate to data if data sequence is unsatisfactory for the requirement of GM (1,1) model method,
Such as translation transformation,By taking different c to make the grade of data sequence that can hold covering than all falling in
It is interior.
(2-5) is to examining treated sequenceUse formula
It carries out one-accumulate and generates new sequence
(2-6) establishes GM (1,1) modelWherein α is to develop gray scale, and μ is that interior generation controls gray scale.
(2-7) acquires α, the estimated value of μ with regression analysis.IfIt is solved using least square method:
Wherein
(2-8) obtains corresponding albefaction model:
(2-9) solves the differential equation, and prediction model, which can be obtained, is:
(2-10) obtains predicted value to prediction model value regressive:
(2-11) obtains the m moment of the next time window of container instance by predicting m time series
Load estimation value be
(2-12) predicts the load of multiple dimensions such as CPU, memory, disk, network by above method, is held
Load estimation value of each dimension of device example at the m moment of next time window.
(3) each object is calculated in the load estimation value of next time window according to container instance each on physical node
The integrated load for managing the next time window of node is physical node setting load mark according to integrated load.
(3-1) obtains next time window of each dimension in container instance all on physical node to be analyzed
M moment load estimation value
(3-2) carries out all container instances on physical node to be analyzed with the correspondence moment load estimation value of dimension
Summation, i.e., For i-th of moment load estimation value of q dimension,For p-th of appearance on physical node
I-th of moment load estimation value of dimension q, s indicate the number of the physical node upper container example in device example;cqIndicate the object
Manage the load value of node dimension q when not running any container instance.Obtain the next time window of physical node respective dimensions
M moment load estimation value
(3-3) basisλqIt is the weight of q dimension,It is loaded for i-th of moment of physical node q dimension
Predicted value obtains the integrated load value at m moment of the next time window of physical node to be analyzed
(3-4) basisFor the m moment setting load mark of the next time window of each physical node.Load value
Corresponding relationship with load condition is:
(4) as shown in figure 4, executing container cloud platform scheduling strategy generating algorithm, scheduling strategy is generated.Where determine migration
A little container instances move to that physical node etc..
(4-1) for next time window m moment there are the physical node of color=red load condition, according to
Integrated load value WloadVariation tendency judge the physical node in future time t (t generally takes 5min) load variation become
Gesture.Calculation formula is as follows:To judge whether following load imbalance can aggravate.If load imbalance will
It can aggravate, then indicating the physical node is the physical node for needing to be implemented scheduling.
(4-2) there are the physical nodes of color=green load condition for m moment of next time window, if
Green state is constantly in future time t, then indicating the physical node is the physical node for needing to be implemented scheduling.
(4-3) needs to be implemented scheduling there are color=red load condition for the m moment of next time window
Physical node, using fewest containers instance migration number as objective function, with migration after load shape of the physical node in time t
State is that color=blue is constraint condition always, using dynamic programming algorithm, finds the container instance for needing to migrate.
The container instance that (4-4) needs to migrate for some, with synthesis of the target physical node in time t after migration
Load WloadAverage value is up to objective function, with the load condition of the physical node in time t after migration always for color=
Blue is constraint, also with dynamic programming algorithm, finds the optimum target physical node of container instance migration.If do not looked for
To destination node, illustrate the physical node for not meeting transition condition in cluster, then cluster needs to increase physical node, and will increase
The target physical node that the physical node added is migrated as container instance.
(4-5) there are the physics for needing to be implemented scheduling of color=green for m moment of next time window
Node executes the target physical node algorithm for finding migration in 44 to each container instance on the node, determines migration
Optimum target physical node.
(46) container instance migrated for each needs executes scheduling strategy generating algorithm, ultimately generates complete tune
Degree strategy.
(5) as shown in figure 5, executing scheduling strategy, to the container instance that each needs to migrate, container cloud platform is called
Dispatch interface realizes container instance scheduling migration, is finally completed the scheduling of whole container cloud platform.It specifically includes:
Container instance to be migrated is packaged by (5-1) with export order, and uploads to the local container parcel of cluster.
(5-2) gets container packet generated above on target physical node from the local container parcel of cluster, uses
Import order importing is mirrored into, and disposes an application container example with this mirror image.
The data volume container for saving former container instance operation data is mounted to newly generated container instance by (5-3), and is opened
It is dynamic to run the container instance.
(5-3) deletes the container instance on original node, migrates successfully after newly generated container instance is run successfully.
(5-4) has been executed to all scheduling strategies, and load before is indicated the free physical for color=green by cluster
Node revocation falls.
(6) test effect
2 Docker clusters have been built in (6-1) this test, and each cluster includes 1 Master node and 5 node objects
Manage node.It forms as shown in table 1.
Table 1
Host name | IP address | Configuring condition |
Master | 192.168.0.101 | CPU:Tetra- core of intel core i5;Memory:8GB |
Node1 | 192.168.0.102 | CPU:Tetra- core of intel core i5;Memory:4GB |
Node2 | 192.168.0.103 | CPU:Tetra- core of intel core i5;Memory:4GB |
Node3 | 192.168.0.104 | CPU:Tetra- core of intel core i5;Memory:4GB |
Node4 | 192.168.0.105 | CPU:Tetra- core of intel core i5;Memory:4GB |
Node5 | 192.168.0.106 | CPU:Tetra- core of intel core i5;Memory:4GB |
(6-2) runs container instance as shown in Table 2 on two clusters respectively.
Table 2
(6-3) cluster 1 manages cluster using Kubernetes, and operation is based on load on the Master node of cluster 2
Decision module in the container cloud platform intelligence method for optimizing resources of prediction, Load Forcase module,
DataCollection module and container scheduler module are separately operable DataCollection mould on 5 node nodes
Block.
(6-4) is as shown in fig. 6, respectively obtained each node load situation of two clusters, each node load of cluster 1
Unbalanced, the load of node 1 has been up to 85%, has been over normal range (NR), and the load of node 5 only has 12%.Cluster 2
Each node load it is more balanced, node 1 is reduced to 71%, belongs in normal range (NR);Wherein the load of contact 5 be 0, be because
It is too low for node load, container instance thereon is moved into other node up.It can be seen that being based on load estimation
Container cloud platform intelligence method for optimizing resources can effectively solve the load imbalance problem of data center.
(6-5) as shown in fig. 7, having calculated separately the CPU of two clusters, memory, the resource utilization of hard disk can from figure
To find out, the utilization rate of each dimension of cluster 1 is generally lower, and the utilization rate of each dimension of cluster 2 is compared with cluster 1 all with obvious
Raising, it can be seen that the container cloud platform intelligence method for optimizing resources based on load estimation can effectively solve data center
The low problem of resource utilization.
Claims (9)
1. the container cloud platform intelligence method for optimizing resources based on load estimation, which is characterized in that include the following steps:
(1) the various dimensions load data of each container instance of container cloud platform is obtained in real time, and it is corresponding negative to generate each dimension
Carry time series;
(2) each container is predicted by the load estimation technology based on Time-Series analysis according to the historic load of container instance
The load of example future time window;
(3) each physics section is calculated in the load estimation value of next time window according to container instance each on physical node
The integrated load of the next time window of point is physical node setting load mark according to integrated load.
(4) container cloud platform scheduling strategy generating algorithm is executed, scheduling strategy is generated.
(5) scheduling strategy is executed, to the container instance that each needs to migrate, calls the dispatch interface of container cloud platform to realize and holds
The scheduling migration of device example, is finally completed the scheduling of whole container cloud platform.
2. the container cloud platform intelligence method for optimizing resources according to claim 1 based on load estimation, which is characterized in that
The step (1) is specially:CAdvisor tool is run on each container instance of container cloud platform, obtains container in real time
Multiple dimensions such as the load data of example, including CPU, memory, disk, network generate the corresponding load time sequence of each dimension
Column.
3. the container cloud platform intelligence method for optimizing resources according to claim 1 based on load estimation, which is characterized in that
The step (2) specifically includes:
(21) for each dimension, newest n*m historic load is obtained from corresponding historic load time series;It divides
It include m historic load in each time window at n time window;
(22) i-th of historic load in each time window is extracted, a time series is constituted
It include m historic load in each time window, so m time series of formation is
(23) it is the validity for guaranteeing GM (1,1) model method, needs to do data necessary inspection processing.It calculatesThe grade ratio of sequence.Method is:
Wherein,Indicate i-th of historic load in k-th of time window.
If all grades can hold covering section than all falling inIt is interior, then i-th of time seriesIt can establish
GM (1,1) model and gray prediction can be carried out.Otherwise, data are converted, covers all grades than all falling in hold
In cover region.
(24) to examining treated sequenceUse formula
It carries out one-accumulate and generates new sequence
(25) establishing GM (1,1) model isWherein α is to develop gray scale, and μ is that interior generation controls gray scale.
(26) α is acquired with regression analysis, the estimated value of μ, obtaining corresponding albefaction model is:
(27) differential equation is solved, prediction model, which can be obtained, is:
(28) to prediction model value regressive, predicted value is obtained:
(29) by predicting m time series, the load for obtaining the m moment of the next time window of container instance is pre-
Measured value is
(210) load of multiple dimensions such as CPU, memory, disk, network is predicted by above method, obtains container reality
Load estimation value of each dimension of example at the m moment of next time window.
4. the container cloud platform intelligence method for optimizing resources according to claim 3 based on load estimation, which is characterized in that
The conversion process of data is translation transformation in the step 23, i.e.,By taking different c to make
The grade ratio of data all falls in and can hold in covering.
5. the container cloud platform intelligence method for optimizing resources according to claim 1 based on load estimation, which is characterized in that
According to the load estimation value of container instance, the integrated load situation of physical node is calculated, it is negative for its setting according to integrated load value
Carry mark.The step (3) specifically includes:
(31) m of next time window of each dimension in container instance all on physical node to be analyzed are obtained
The load estimation value at moment
(32) it sums to all container instances on physical node to be analyzed with the correspondence moment load estimation value of dimension,
I.e. For i-th of moment load estimation value of q dimension,It is real for p-th of container on physical node
I-th of moment load estimation value of dimension q, s indicate the number of the physical node upper container example in example;cqIndicate the physics section
The load value of point dimension q when not running any container instance.Obtain m of the next time window of physical node respective dimensions
The load estimation value at moment
(33) basisλqIt is the weight of q dimension,For i-th of moment load estimation of physical node q dimension
Value, obtains the integrated load value at m moment of the next time window of physical node to be analyzed
(34) basisFor the m moment setting load mark of the next time window of each physical node.Load value and load
The corresponding relationship of state is:
6. the container cloud platform intelligence method for optimizing resources according to claim 1 based on load estimation, which is characterized in that
Container cloud platform scheduling strategy generating algorithm is executed, scheduling strategy is generated.The step (4) specifically includes:
(41) for the m moment of next time window, there are the physical nodes of color=red load condition, according to synthesis
Load value WloadVariation tendency judge load variation tendency of the physical node in future time t.If load imbalance will
Aggravation, then indicating the physical node is the physical node for needing to be implemented scheduling.
(42) for the m moment of next time window, there are the physical nodes of color=green load condition, if not
Come in time t to be constantly in green state, then indicating the physical node is the physical node for needing to be implemented scheduling.
7. the container cloud platform intelligence method for optimizing resources according to claim 6 based on load estimation, which is characterized in that
Time t=5min.
8. the container cloud platform intelligence method for optimizing resources according to claim 1 based on load estimation, which is characterized in that
The scheduling strategy generating algorithm of the step (4) is specially:
(43) for the m moment of next time window, there are the physics for needing to be implemented scheduling of color=red load condition
Node, using fewest containers instance migration number as objective function, always with load condition of the physical node in time t after migration
It is constraint condition for color=blue, using dynamic programming algorithm, finds the container instance for needing to migrate.
(44) for the container instance that some needs to migrate, with integrated load of the target physical node in time t after migration
WloadAverage value is up to objective function, with the load condition of the physical node in time t after migration always for color=blue
The optimum target physical node of container instance migration is found also with dynamic programming algorithm for constraint.If not finding mesh
Node is marked, illustrates the physical node for not meeting transition condition in cluster, then cluster needs to increase physical node, and will be increased
The target physical node that physical node is migrated as container instance.
(45) for the m moment of next time window there are the physical node for needing to be implemented scheduling of color=green,
The target physical node algorithm for finding migration in 44 is executed to each container instance on the node, determines the best mesh of migration
Mark physical node.
(46) container instance migrated for each needs executes scheduling strategy generating algorithm, ultimately generates complete scheduling plan
Slightly.
9. step (5) described in the container cloud platform intelligence method for optimizing resources according to claim 1 based on load estimation,
It is characterized in that, perform claim requires 8 scheduling strategies generated, to the container instance that each needs to migrate, container cloud is called
The dispatch interface of platform realizes container instance scheduling migration, is finally completed the scheduling of whole container cloud platform.It specifically includes:
(51) container instance to be migrated is packaged with export order, and uploads to the local container parcel of cluster.
(52) on target physical node, container packet generated above is got from the local container parcel of cluster, is used
Import order importing is mirrored into, and disposes an application container example with this mirror image.
(53) the data volume container for saving former container instance operation data is mounted to newly generated container instance, and starts operation
The container instance.
(53) after newly generated container instance is run successfully, the container instance on original node is deleted, is migrated successfully.
(54) it has been executed to all scheduling strategies, cluster indicates load before to be picked for the free physical node of color=green
It removes.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810664295.0A CN108829494B (en) | 2018-06-25 | 2018-06-25 | Container cloud platform intelligent resource optimization method based on load prediction |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810664295.0A CN108829494B (en) | 2018-06-25 | 2018-06-25 | Container cloud platform intelligent resource optimization method based on load prediction |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108829494A true CN108829494A (en) | 2018-11-16 |
CN108829494B CN108829494B (en) | 2020-09-29 |
Family
ID=64138242
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810664295.0A Active CN108829494B (en) | 2018-06-25 | 2018-06-25 | Container cloud platform intelligent resource optimization method based on load prediction |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108829494B (en) |
Cited By (33)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109918198A (en) * | 2019-02-18 | 2019-06-21 | 中国空间技术研究院 | A kind of emulation cloud platform load dispatch system and method based on user characteristics prediction |
CN110058923A (en) * | 2019-03-29 | 2019-07-26 | 华中科技大学 | It is a kind of based on dynamically prefetch strategy Docker container COW mechanism optimization method and system |
CN110286996A (en) * | 2019-05-17 | 2019-09-27 | 平安科技(深圳)有限公司 | Container instance IP switching method, device, computer equipment and storage medium |
CN110377395A (en) * | 2019-07-03 | 2019-10-25 | 无锡华云数据技术服务有限公司 | A kind of Pod moving method in Kubernetes cluster |
CN110413380A (en) * | 2019-08-02 | 2019-11-05 | 北京智芯微电子科技有限公司 | The dispatching method of container cluster |
CN110457131A (en) * | 2019-07-31 | 2019-11-15 | 华中科技大学 | Electric system supercomputer platform method for scheduling task based on Docker container |
CN110753120A (en) * | 2019-10-25 | 2020-02-04 | 北京浪潮数据技术有限公司 | Stateful load creating method, device, equipment and medium |
CN110888741A (en) * | 2019-11-27 | 2020-03-17 | 新华三大数据技术有限公司 | Resource scheduling method and device for application container, server and storage medium |
CN110990159A (en) * | 2019-12-25 | 2020-04-10 | 浙江大学 | Historical data analysis-based container cloud platform resource quota prediction method |
CN111124619A (en) * | 2019-12-25 | 2020-05-08 | 浙江大学 | Container scheduling method for secondary scheduling |
CN111143070A (en) * | 2019-12-27 | 2020-05-12 | 北京奇艺世纪科技有限公司 | Resource scheduling method and device |
CN111209083A (en) * | 2020-01-08 | 2020-05-29 | 中国联合网络通信集团有限公司 | Container scheduling method, device and storage medium |
CN111367632A (en) * | 2020-02-14 | 2020-07-03 | 重庆邮电大学 | Container cloud scheduling method based on periodic characteristics |
CN111638958A (en) * | 2020-06-02 | 2020-09-08 | 中国联合网络通信集团有限公司 | Cloud host load processing method and device, control equipment and storage medium |
WO2020181896A1 (en) * | 2019-03-14 | 2020-09-17 | 深圳先进技术研究院 | Multi-agent reinforcement learning scheduling method and system and electronic device |
CN112019628A (en) * | 2020-09-01 | 2020-12-01 | 江西凌峰售电有限公司 | Low-delay low-energy-consumption intelligent platform system based on Internet of things |
CN112087504A (en) * | 2020-08-31 | 2020-12-15 | 浪潮通用软件有限公司 | Dynamic load balancing method and device based on working load characteristics |
CN112187894A (en) * | 2020-09-17 | 2021-01-05 | 杭州谐云科技有限公司 | Container dynamic scheduling method based on load correlation prediction |
CN112416596A (en) * | 2020-12-01 | 2021-02-26 | 新华三人工智能科技有限公司 | Node scheduling method, device and equipment |
CN112905343A (en) * | 2021-02-09 | 2021-06-04 | 重庆大学 | Resource scheduling system based on load characteristics in industrial cloud environment |
CN113055199A (en) * | 2019-12-26 | 2021-06-29 | 中国移动通信集团重庆有限公司 | Gateway access method and device and gateway equipment |
CN113110914A (en) * | 2021-03-02 | 2021-07-13 | 西安电子科技大学 | Internet of things platform construction method based on micro-service architecture |
CN113626282A (en) * | 2021-07-16 | 2021-11-09 | 济南浪潮数据技术有限公司 | Cloud computing physical node load monitoring method and device, terminal and storage medium |
CN113791863A (en) * | 2021-08-10 | 2021-12-14 | 北京中电飞华通信有限公司 | Virtual container-based power internet of things agent resource scheduling method and related equipment |
CN113946438A (en) * | 2021-09-03 | 2022-01-18 | 西安电子科技大学 | Online migration scheduling method, device, equipment and system based on container integration |
CN113965574A (en) * | 2021-09-27 | 2022-01-21 | 西安交通大学 | Cloud service data center backup virtual resource scheduling method and device |
CN114401194A (en) * | 2021-12-29 | 2022-04-26 | 山东省计算中心(国家超级计算济南中心) | Dynamic expansion method and platform supporting network function virtualization and computer |
CN114595029A (en) * | 2022-02-15 | 2022-06-07 | 黄淮学院 | Cloud container resource scheduling method and system based on automatic learning |
CN114615177A (en) * | 2022-03-03 | 2022-06-10 | 腾讯科技(深圳)有限公司 | Load detection method and device of cloud platform, electronic equipment and storage medium |
CN114661448A (en) * | 2022-05-17 | 2022-06-24 | 中电云数智科技有限公司 | Kubernetes resource scheduling method and scheduling component |
CN115665157A (en) * | 2022-11-14 | 2023-01-31 | 杭州谐云科技有限公司 | Balanced scheduling method and system based on application resource types |
WO2023103349A1 (en) * | 2021-12-08 | 2023-06-15 | 深圳前海微众银行股份有限公司 | Load adjustment method, management node, and storage medium |
CN113946438B (en) * | 2021-09-03 | 2024-04-30 | 西安电子科技大学 | Online migration scheduling method, device, equipment and system based on container integration |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP3086229A1 (en) * | 2015-04-20 | 2016-10-26 | Repsol, S.A. | Managing hydrocarbon energy production while proactively maintaining a balanced workload |
CN106933650A (en) * | 2017-03-03 | 2017-07-07 | 北方工业大学 | load management method and system of cloud application system |
CN106970831A (en) * | 2017-05-15 | 2017-07-21 | 金航数码科技有限责任公司 | The resources of virtual machine dynamic scheduling system and method for a kind of facing cloud platform |
CN107070965A (en) * | 2016-12-22 | 2017-08-18 | 广东石油化工学院 | A kind of Multi-workflow resource provision method virtualized under container resource |
CN107707431A (en) * | 2017-10-31 | 2018-02-16 | 河南科技大学 | The data safety monitoring method and system of a kind of facing cloud platform |
US20180074860A1 (en) * | 2016-09-14 | 2018-03-15 | Microsoft Technology Licensing, Llc | Outage detection and compute resource protection in a computer service environment |
-
2018
- 2018-06-25 CN CN201810664295.0A patent/CN108829494B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP3086229A1 (en) * | 2015-04-20 | 2016-10-26 | Repsol, S.A. | Managing hydrocarbon energy production while proactively maintaining a balanced workload |
US20180074860A1 (en) * | 2016-09-14 | 2018-03-15 | Microsoft Technology Licensing, Llc | Outage detection and compute resource protection in a computer service environment |
CN107070965A (en) * | 2016-12-22 | 2017-08-18 | 广东石油化工学院 | A kind of Multi-workflow resource provision method virtualized under container resource |
CN106933650A (en) * | 2017-03-03 | 2017-07-07 | 北方工业大学 | load management method and system of cloud application system |
CN106970831A (en) * | 2017-05-15 | 2017-07-21 | 金航数码科技有限责任公司 | The resources of virtual machine dynamic scheduling system and method for a kind of facing cloud platform |
CN107707431A (en) * | 2017-10-31 | 2018-02-16 | 河南科技大学 | The data safety monitoring method and system of a kind of facing cloud platform |
Non-Patent Citations (1)
Title |
---|
杨鹏飞: "基于Kubernetes的资源动态调度的研究与实现", 《中国优秀硕士学位论文全文库》 * |
Cited By (53)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109918198B (en) * | 2019-02-18 | 2020-02-11 | 中国空间技术研究院 | Simulation cloud platform load scheduling system and method based on user characteristic prediction |
CN109918198A (en) * | 2019-02-18 | 2019-06-21 | 中国空间技术研究院 | A kind of emulation cloud platform load dispatch system and method based on user characteristics prediction |
WO2020181896A1 (en) * | 2019-03-14 | 2020-09-17 | 深圳先进技术研究院 | Multi-agent reinforcement learning scheduling method and system and electronic device |
CN110058923A (en) * | 2019-03-29 | 2019-07-26 | 华中科技大学 | It is a kind of based on dynamically prefetch strategy Docker container COW mechanism optimization method and system |
CN110058923B (en) * | 2019-03-29 | 2021-03-26 | 华中科技大学 | Docker container COW mechanism optimization method and system based on dynamic prefetching strategy |
CN110286996A (en) * | 2019-05-17 | 2019-09-27 | 平安科技(深圳)有限公司 | Container instance IP switching method, device, computer equipment and storage medium |
WO2020233002A1 (en) * | 2019-05-17 | 2020-11-26 | 平安科技(深圳)有限公司 | Container instance ip switching method and apparatus, computer device, and storage medium |
CN110286996B (en) * | 2019-05-17 | 2023-08-18 | 平安科技(深圳)有限公司 | Container instance IP switching method, device, computer equipment and storage medium |
CN110377395A (en) * | 2019-07-03 | 2019-10-25 | 无锡华云数据技术服务有限公司 | A kind of Pod moving method in Kubernetes cluster |
CN110377395B (en) * | 2019-07-03 | 2021-11-02 | 华云数据控股集团有限公司 | Pod migration method in Kubernetes cluster |
CN110457131B (en) * | 2019-07-31 | 2021-08-20 | 华中科技大学 | Task scheduling method for supercomputing platform of power system based on Docker container |
CN110457131A (en) * | 2019-07-31 | 2019-11-15 | 华中科技大学 | Electric system supercomputer platform method for scheduling task based on Docker container |
CN110413380A (en) * | 2019-08-02 | 2019-11-05 | 北京智芯微电子科技有限公司 | The dispatching method of container cluster |
CN110753120B (en) * | 2019-10-25 | 2022-06-10 | 北京浪潮数据技术有限公司 | Stateful load creating method, device, equipment and medium |
CN110753120A (en) * | 2019-10-25 | 2020-02-04 | 北京浪潮数据技术有限公司 | Stateful load creating method, device, equipment and medium |
CN110888741A (en) * | 2019-11-27 | 2020-03-17 | 新华三大数据技术有限公司 | Resource scheduling method and device for application container, server and storage medium |
CN111124619B (en) * | 2019-12-25 | 2023-07-21 | 浙江大学 | Container scheduling method for secondary scheduling |
CN110990159B (en) * | 2019-12-25 | 2023-05-23 | 浙江大学 | Historical data analysis-based container cloud platform resource quota prediction method |
CN111124619A (en) * | 2019-12-25 | 2020-05-08 | 浙江大学 | Container scheduling method for secondary scheduling |
CN110990159A (en) * | 2019-12-25 | 2020-04-10 | 浙江大学 | Historical data analysis-based container cloud platform resource quota prediction method |
CN113055199B (en) * | 2019-12-26 | 2023-08-18 | 中国移动通信集团重庆有限公司 | Gateway access method and device and gateway equipment |
CN113055199A (en) * | 2019-12-26 | 2021-06-29 | 中国移动通信集团重庆有限公司 | Gateway access method and device and gateway equipment |
CN111143070A (en) * | 2019-12-27 | 2020-05-12 | 北京奇艺世纪科技有限公司 | Resource scheduling method and device |
CN111209083A (en) * | 2020-01-08 | 2020-05-29 | 中国联合网络通信集团有限公司 | Container scheduling method, device and storage medium |
CN111209083B (en) * | 2020-01-08 | 2023-05-09 | 中国联合网络通信集团有限公司 | Container scheduling method, device and storage medium |
CN111367632A (en) * | 2020-02-14 | 2020-07-03 | 重庆邮电大学 | Container cloud scheduling method based on periodic characteristics |
CN111367632B (en) * | 2020-02-14 | 2023-04-18 | 重庆邮电大学 | Container cloud scheduling method based on periodic characteristics |
CN111638958B (en) * | 2020-06-02 | 2024-04-05 | 中国联合网络通信集团有限公司 | Cloud host load processing method and device, control equipment and storage medium |
CN111638958A (en) * | 2020-06-02 | 2020-09-08 | 中国联合网络通信集团有限公司 | Cloud host load processing method and device, control equipment and storage medium |
CN112087504A (en) * | 2020-08-31 | 2020-12-15 | 浪潮通用软件有限公司 | Dynamic load balancing method and device based on working load characteristics |
CN112019628A (en) * | 2020-09-01 | 2020-12-01 | 江西凌峰售电有限公司 | Low-delay low-energy-consumption intelligent platform system based on Internet of things |
CN112187894A (en) * | 2020-09-17 | 2021-01-05 | 杭州谐云科技有限公司 | Container dynamic scheduling method based on load correlation prediction |
CN112187894B (en) * | 2020-09-17 | 2022-06-10 | 杭州谐云科技有限公司 | Container dynamic scheduling method based on load correlation prediction |
CN112416596A (en) * | 2020-12-01 | 2021-02-26 | 新华三人工智能科技有限公司 | Node scheduling method, device and equipment |
CN112905343B (en) * | 2021-02-09 | 2023-09-26 | 重庆大学 | Resource scheduling system based on load characteristics in industrial cloud environment |
CN112905343A (en) * | 2021-02-09 | 2021-06-04 | 重庆大学 | Resource scheduling system based on load characteristics in industrial cloud environment |
CN113110914A (en) * | 2021-03-02 | 2021-07-13 | 西安电子科技大学 | Internet of things platform construction method based on micro-service architecture |
CN113626282B (en) * | 2021-07-16 | 2023-12-22 | 济南浪潮数据技术有限公司 | Cloud computing physical node load monitoring method, device, terminal and storage medium |
CN113626282A (en) * | 2021-07-16 | 2021-11-09 | 济南浪潮数据技术有限公司 | Cloud computing physical node load monitoring method and device, terminal and storage medium |
CN113791863B (en) * | 2021-08-10 | 2024-01-23 | 北京中电飞华通信有限公司 | Virtual container-based power Internet of things proxy resource scheduling method and related equipment |
CN113791863A (en) * | 2021-08-10 | 2021-12-14 | 北京中电飞华通信有限公司 | Virtual container-based power internet of things agent resource scheduling method and related equipment |
CN113946438A (en) * | 2021-09-03 | 2022-01-18 | 西安电子科技大学 | Online migration scheduling method, device, equipment and system based on container integration |
CN113946438B (en) * | 2021-09-03 | 2024-04-30 | 西安电子科技大学 | Online migration scheduling method, device, equipment and system based on container integration |
CN113965574A (en) * | 2021-09-27 | 2022-01-21 | 西安交通大学 | Cloud service data center backup virtual resource scheduling method and device |
CN113965574B (en) * | 2021-09-27 | 2022-07-12 | 西安交通大学 | Cloud service data center backup virtual resource scheduling method and device |
WO2023103349A1 (en) * | 2021-12-08 | 2023-06-15 | 深圳前海微众银行股份有限公司 | Load adjustment method, management node, and storage medium |
CN114401194A (en) * | 2021-12-29 | 2022-04-26 | 山东省计算中心(国家超级计算济南中心) | Dynamic expansion method and platform supporting network function virtualization and computer |
CN114401194B (en) * | 2021-12-29 | 2023-08-01 | 山东省计算中心(国家超级计算济南中心) | Dynamic expansion method, platform and computer supporting network function virtualization |
CN114595029A (en) * | 2022-02-15 | 2022-06-07 | 黄淮学院 | Cloud container resource scheduling method and system based on automatic learning |
CN114615177B (en) * | 2022-03-03 | 2023-10-13 | 腾讯科技(深圳)有限公司 | Load detection method and device of cloud platform, electronic equipment and storage medium |
CN114615177A (en) * | 2022-03-03 | 2022-06-10 | 腾讯科技(深圳)有限公司 | Load detection method and device of cloud platform, electronic equipment and storage medium |
CN114661448A (en) * | 2022-05-17 | 2022-06-24 | 中电云数智科技有限公司 | Kubernetes resource scheduling method and scheduling component |
CN115665157A (en) * | 2022-11-14 | 2023-01-31 | 杭州谐云科技有限公司 | Balanced scheduling method and system based on application resource types |
Also Published As
Publication number | Publication date |
---|---|
CN108829494B (en) | 2020-09-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108829494A (en) | Container cloud platform intelligence method for optimizing resources based on load estimation | |
EP3847549B1 (en) | Minimizing impact of migrating virtual services | |
CN106502792B (en) | A kind of multi-tenant priority scheduling of resource method towards different type load | |
CN106776005B (en) | Resource management system and method for containerized application | |
CN113490254B (en) | VNF migration method based on bidirectional GRU resource demand prediction in federal learning | |
CN106534318B (en) | A kind of OpenStack cloud platform resource dynamic scheduling system and method based on flow compatibility | |
CN104050042B (en) | The resource allocation methods and device of ETL operations | |
CN103916438B (en) | Cloud testing environment scheduling method and system based on load forecast | |
CN105468435A (en) | NFV dynamic resource distribution method | |
CN108182105A (en) | Local dynamic station moving method and control system based on Docker container techniques | |
CN106020934A (en) | Optimized deploying method based on virtual cluster online migration | |
CN109492774A (en) | A kind of cloud resource dispatching method based on deep learning | |
CN105373432B (en) | A kind of cloud computing resource scheduling method based on virtual resource status predication | |
CN103500123B (en) | Parallel computation dispatching method in isomerous environment | |
CN104104621B (en) | A kind of virtual network resource dynamic self-adapting adjusting method based on Nonlinear Dimension Reduction | |
CN103475538A (en) | Multi-interface-based self-adaptive cloud service test method | |
WO2023109068A1 (en) | Automatic virtual machine migration decision-making method based on user experience in multi-cloud environment | |
CN105740059B (en) | A kind of population dispatching method towards Divisible task | |
WO2020134133A1 (en) | Resource allocation method, substation, and computer-readable storage medium | |
CN106656555A (en) | Dynamic adjustment method of service resources of cloud computing system | |
CN111752678A (en) | Low-power-consumption container placement method for distributed collaborative learning in edge computing | |
CN105607943A (en) | Dynamic deployment mechanism of virtual machine under cloud environment | |
Xu et al. | Enhancing Kubernetes Automated Scheduling with Deep Learning and Reinforcement Techniques for Large-Scale Cloud Computing Optimization | |
CN108874508A (en) | A kind of cloud computing virtual server system load equilibration scheduling method | |
CN104346220B (en) | A kind of method for scheduling task and system |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: Intelligent Resource Optimization Method for Container Cloud Platform Based on Load Prediction Granted publication date: 20200929 Pledgee: Hangzhou High-tech Financing Guarantee Co.,Ltd. Pledgor: HANGZHOU HARMONYCLOUD TECHNOLOGY Co.,Ltd. Registration number: Y2024980003766 |
|
PE01 | Entry into force of the registration of the contract for pledge of patent right |