CN110489200A - A kind of method for scheduling task suitable for embedded container cluster - Google Patents
A kind of method for scheduling task suitable for embedded container cluster Download PDFInfo
- Publication number
- CN110489200A CN110489200A CN201810457653.0A CN201810457653A CN110489200A CN 110489200 A CN110489200 A CN 110489200A CN 201810457653 A CN201810457653 A CN 201810457653A CN 110489200 A CN110489200 A CN 110489200A
- Authority
- CN
- China
- Prior art keywords
- node
- resource
- parameter
- container cluster
- resource utilization
- 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
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
- G06F9/44—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
-
- 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
- G06F2009/45583—Memory management, e.g. access or allocation
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of method for scheduling task suitable for embedded container cluster, which comprises the resource for calculating each node of container cluster uses equilibrium degree parameter, load balancing parameter and stable degree parameter;Wherein, the resource is the resource utilization variance of each dimension of node using equilibrium degree parameter;The load balancing parameter is to carry out the resulting resource load weight of linear weighted function using each dimension resource utilization of node;The stability parameter is the node stationary value obtained according to the online rate of node and task completion rate;Three sorted lists are generated respectively using equilibrium degree parameter, load balancing parameter and stability parameter according to the resource of each node, using positional value of the node in three sorted lists as three dimensional space coordinate, choose to origin Chebyshev apart from the smallest node deployment new container.Method of the invention can reduce influence of the node dynamic change to container deployment strategy, improve container cluster overall performance.
Description
Technical field
The present invention relates to computer field more particularly to container cluster scheduling fields, in particular to a kind of to be suitable for insertion
The method for scheduling task of formula container cluster.
Background technique
With the rapid development of internet and Internet of Things, following information network will form collection " people, machine, object " Yu Yiti's
Ternary pattern of fusion information world, magnanimity heterogeneous terminals, large-scale data processing, brings greatly traditional services system processing capacity
Challenge.The Technical Architecture of the sea service proposed at present can be realized the near field resource polymerization of user side sea end node and provide.Cause
This, which is realized, has biggish meaning with regulation to effective management of extra large end embedded resource.
Docker is a kind of Lightweight Container virtualization technology, the deployment for realizing task in isomeric group that it can be convenient
With migration.Individual docker software is unable to satisfy the dispatching requirement of cluster.To solve this defect, either IT giant is still created
Industry company, enterprise customer also or commonly require an integrated docker container i.e. service management platform, allow user's energy
It is enough pellucidly to enjoy docker container bring convenience, it is finally reached the purpose for making their smooth upper cloud of application.As
The important component of the docker ecosphere, the research and development that container services are also very various, including swarm, kubernetes,
Mesos, aws, ecs etc..
Currently, there are problems for embedded container cluster management software dispatching method, it is to utilize well first
The each dimension resource of container, can not play cluster maximum potential, and customer responsiveness is not fine;Followed by resource utilization ratio compared with
Low, node dynamic is higher;The main reason for causing problem above is node resource fragment and cluster load imbalance.
Summary of the invention
The invention aims to unbalanced, the system resource utilizations that solves resource load present in embedded container cluster
The problems such as rate is lower, and node dynamic is higher proposes a kind of method for scheduling task suitable for embedded container cluster.
To achieve the goals above, the invention proposes a kind of method for scheduling task suitable for embedded container cluster,
The described method includes:
The resource for calculating each node of container cluster is joined using equilibrium degree parameter, load balancing parameter and stable degree
Amount;Wherein, the resource is the resource utilization variance of each dimension of node using equilibrium degree parameter;The load balancing parameter is
The resulting resource load weight of linear weighted function is carried out using each dimension resource utilization of node;The stability parameter is according to section
The node stationary value that the online rate of point and task completion rate obtain;
Three are generated respectively using equilibrium degree parameter, load balancing parameter and stability parameter according to the resource of each node
Sorted lists, using positional value of the node in three sorted lists as three dimensional space coordinate, choose to origin Chebyshev away from
From the smallest node deployment new container.
As a kind of improvement of the above method, the method is specifically included:
Step 1) obtains node di, thus the resource utilization of several dimensions of 1≤i≤K calculates each node
Resource utilization variance;Ascending order is carried out to node according to resource utilization variance to arrange to obtain sequence L1;
The resource utilization for each dimension that step 2) is obtained according to step 1) calculates the resource load power of each node
Value;Ascending order is carried out to node according to resource load weight to arrange to obtain sequence L2;
Step 3) calculates the stationary value of each node according to the historical record of each node;According to stationary value to node
Ascending order is carried out to arrange to obtain sequence L3;
Step 4) is according to node in three sequence L1、L2And L3In positional value node is mapped in three-dimensional space, calculate
Node Chebyshev's distance dis value:
Step 5) selects the smallest node of dis value as optimal node, will be on task deployment to the node.
As a kind of improvement of the above method, the step 1) is specifically included:
Step 1-1) calculate node diMemory usage M (di):
M(di)=((MenTotal-MemFree)/MemTotal) × 100%
Wherein, MemTotal is memory amount, and MemFree is memory surplus;
Step 1-2) calculate node diCpu utilization rate C (di):
C(di)=((us+sy)/(us+sy+id)) × 100%
Wherein, us is User space, and sy is kernel state, id Idle state;
Step 1-3) calculate node diNetwork load utilization rate N (di):
N(di)=((N1+N2)/M) × 100%
Wherein, N1 is to flow into flow, and N2 is outflow flow, and M is maximum network handling capacity;
Step 1-4) calculate node diAverage resource utilization rate:
Wherein, N=3;r1i=M (di), r2i=C (di), r3i=N (di);
Step 1-5) according to average resource utilization rate calculate node diResource utilization variance:
Step 1-6) according to resource utilization variance Si 2Ascending order is carried out to K node to arrange to obtain sequence L1。
As a kind of improvement of the above method, the step 2) is specifically included:
Step 2-1) obtain memory usage M (di), cpu utilization rate C (di) and network load utilization rate N (di) weight
kj, 1≤j≤N;
Step 2-2 calculate node diResource load weight W (di):
Wherein, γjFor the coefficient for adjusting weight, R1(di)=M (di), R2(di)=C (di), R3(di)=N (di);P is threshold
Value;
Step 2-3) according to resource load weight W (di) K node progress ascending order is arranged to obtain sequence L2。
As a kind of improvement of the above method, the step 1) is specifically included:
Step 3-1) calculate node diStationary value Z (di):
Wherein, M is the number of unit timeslice;njFor task quantity performed in node unit time piece, NjFor node unit
Timeslice receives total task number amount, tjFor online hours in node unit time piece, tpFor unit timeslice;
Step 3-2) according to stationary value Z (di) K node progress ascending order is arranged to obtain sequence L3。
As a kind of improvement of the above method, the step 4) specifically:
Chebyshev's distance dis that egress arrives node in three dimensions is calculated according to the following formula:
Wherein, ci1,ci2,ci3Respectively node diIn sequence L1, L2, L3In positional value.
Present invention has an advantage that
1, method of the invention can reduce influence of the node dynamic change to container deployment strategy, and it is whole to improve container cluster
Body performance;
2, the method for the present invention, which proposes Dynamic Weights, improves container cluster resource utilization, is making full use of node each
While dimension resource, equilibrium deployment container, improves the usability of cluster, has played the potential of cluster to greatest extent;
3, node division is high stable set and low stable set by method of the invention, preferentially by task deployment high steady
On node in fixed set, this method is preferably applied on the stronger embedded cluster of activity.
Detailed description of the invention
Fig. 1 is the task scheduling system architecture diagram of the embodiment of the present invention 1;
Fig. 2 is the flow chart of the container cluster method for scheduling task of the embodiment of the present invention 2.
Specific embodiment
To make the technical solution of the embodiment of the present invention and becoming apparent from for advantage expression, below by drawings and examples,
Technical scheme of the present invention will be described in further detail.
Technical solution of the present invention has comprehensively considered node dynamic height, load balancing and the resource fragmentation problem of cluster,
Method is optimized for embedded cluster dynamic higher characteristic simultaneously.For what is generated in cluster operational process
Resource fragmentation measures the balanced situation that resource in single node uses using resource variance.According to the variance pair of resource utilization
Node carries out ascending order and arranges to obtain sequence L1;Simultaneously aiming at the problem that cluster load balance, Dynamic Weights algorithm is taken, in selection
It deposits, the parameter of the various dimensions such as cpu, network load describes node weight.The corresponding weight of parameters is assigned, each section is calculated
The weight of point, weight is smaller to illustrate that node load is lighter.Ascending order is carried out to node according to the size of weight to arrange to obtain sequence L2。
It is main when some a certain resource dimension utilization rates are more than given threshold for more accurate reaction node load situation
The dynamic weight for improving the dimension resource.Problem higher for dynamic, the present invention propose the calculation method of node stationary value.Meter
Calculation method is based on the online rate of node and task completion rate.Ascending order is carried out to node according to the size of stationary value to arrange to obtain L3, will
Positional value of the node in above three sequence is mapped to node in three-dimensional space as three dimensional space coordinate.By container portion
Administration to origin Chebyshev on the smallest node.
Embodiment 1
As shown in Figure 1, the embodiment of the present invention 1 provides a kind of task scheduling system, which includes: swarm cluster
Module, weight module and scheduler module.
Swarm cluster module interacts, with work for providing group system basis running environment for realizing with client
Make node interaction, colony dispatching and service discovery;5 working nodes are disposed in the cluster.
Working node quantity is K=5, node di, 1≤i≤K, resource dimension N, the resource use of each node
Rate is (r1,…rN)。
Weight module makes for being responsible for the resource using information of each node of timing acquisition, and according to node Current resource
The Resource Calculation needed for situation and new container goes out the weight of each node in cluster, and by optimal node feeding back to scheduler module;
Scheduler module, for new container to be deployed in optimal node according to the feedback information of weight module.
As shown in Fig. 2, a kind of task schedule side suitable for embedded container cluster that the embodiment of the present invention 2 provides
Method, which comprises
Step 201, the container for needing to dispose is created in docker.
The information and container mirror image that container starting needs are supplied to swarm cluster management journey by docker client
Sequence.
Step 202, weight module obtains the resource information of each node.
Concretely, memory usage M (di) using obtaining surplus under Linux system/proc/meminfo catalogue
MemFree and memory amount MemTotal.
Memory usage M (di) expression formula are as follows:
M(di)=((MenTotal-MemFree)/MemTotal) × 100%
Cpu utilization rate C (di) using User space us, kernel state sy and the Idle state in top order return value under linux
id.Cpu utilization rate expression formula are as follows:
C(di)=((us+sy)/(us+sy+id)) × 100%
If inflow flow is N1, outflow flow is N2, and maximum network handling capacity is M, then network load utilization rate N (di) table
Up to formula are as follows:
N(di)=((N1+N2)/M) × 100%
Step 203, the resource utilization of each dimension obtained according to step 202, calculates all nodes in cluster
Resource utilization variance S2。
Calculate node diAverage resource utilization rate:
Wherein, N=3;r1i=M (di), r2i=C (di), r3i=N (di);
Step 1-5) according to average resource utilization rate calculate node diResource utilization variance Si 2:
Step 204, according to the resulting resource utilization variance S of step 203i 2Ascending order is carried out to node to arrange to obtain sequence
L1: L1={ d3,d4,d2,d1,d5}。
Step 205, the resource utilization of each dimension obtained according to step 202, calculates all nodes in cluster
Resource load weighted value;
For problem of load balancing, the present embodiment proposition will be that container allocation is gone out but in not used by node
The ratio of resource and the total memory of node is deposited as new resource dimension, improves the accurate journey measured for node memory resource load
Degree.It proposes simultaneously using Dynamic Weights.Due to for the weighting of each dimension resource, some dimension occupation rate will lead to
The case where rate of exchange are high, and node weight can not but embody.Therefore in order to further increase algorithm evaluation ability, propose that dynamic adjusts
The weight of relevant parameter.
One threshold value is set for each dimension first, weighs the resource when dimension resource utilization arrival threshold value
It is increased to original k times again, k can be specified by user oneself.Preferred setting γjIt is 2.Although improving to a certain extent
The resource load of the node, but more reasonably reacted actual conditions.The resource load of egress is calculated according to the following formula
Weight W (di):
Wherein, kjIndicate the weighted value of each dimension resource, γjFor readjusting the coefficient of weight according to upper loading limit,
R1(di)=M (di), R2(di)=C (di), R3(di)=N (di);P is threshold value;
Step 206, ascending order is carried out to node according to the resulting resource load weight of step 205 to arrange to obtain sequence L2, example
Such as L2={ d5,d3,d1,d2,d4}。
Step 207, the node historical record obtained according to step 202, calculates the stationary value of egress.
Stability grade classification rule are as follows:
Wherein, M is the number of unit timeslice;njFor task quantity performed in node unit time piece, NjFor node unit
Timeslice receives total task number amount, tjFor online hours in node unit time piece, tpFor unit timeslice;Z(di) it is that node is steady
Definite value;WhenWhen be high stable grade node, be otherwise low stationary level node,For stability threshold value.
Step 208, ascending order is carried out to node according to the resulting stationary value of step 207 to arrange to obtain sequence L3, such as L3=
{d3,d5,d4,d2,d1}。
Step 209, node is mapped in three-dimensional space according to positional value of the node in three arrangements, calculate node
Dis value:
Chebyshev's distance dis that egress arrives node in three dimensions is calculated according to the following formula:
Wherein, ci1,ci2,ci3Respectively diIn L1, L2, L3In positional value.
Step 210, it is minimum side by side to judge whether there is multinode dis value, if it does not exist, then step 211 is jumped to, if
In the presence of then jumping to step 212.
Step 211, the smallest node deployment container of dis value is selected.
Step 212, dis value the smallest node deployment container side by side is randomly choosed.
Method proposed by the present invention has comprehensively considered resource fragmentation and two kinds of typical problems of load balancing in cluster.It proposes
Dynamic Weights improve container cluster resource utilization, while making full use of node each dimension resource, to greatest extent
Balanced deployment container improves the usability of cluster, has played the potential of cluster.And by node division be high stable set with it is low
Stable set applies this method preferably in activity preferentially by task deployment on the node in high stable set
On the stronger embedded cluster of property.
It should be noted last that the above examples are only used to illustrate the technical scheme of the present invention and are not limiting.Although ginseng
It is described the invention in detail according to embodiment, those skilled in the art should understand that, to technical side of the invention
Case is modified or replaced equivalently, and without departure from the spirit and scope of technical solution of the present invention, should all be covered in the present invention
Scope of the claims in.
Claims (6)
1. a kind of method for scheduling task suitable for embedded container cluster, which comprises
The resource for calculating each node of container cluster uses equilibrium degree parameter, load balancing parameter and stable degree parameter;Its
In, the resource is the resource utilization variance of each dimension of node using equilibrium degree parameter;The load balancing parameter is to use
Each dimension resource utilization of node carries out the resulting resource load weight of linear weighted function;The stability parameter is to be existed according to node
The node stationary value that line rate and task completion rate obtain;
Three sequences are generated respectively using equilibrium degree parameter, load balancing parameter and stability parameter according to the resource of each node
List is chosen to origin Chebyshev distance most using positional value of the node in three sorted lists as three dimensional space coordinate
Small node deployment new container.
2. the method for scheduling task according to claim 1 suitable for embedded container cluster, which is characterized in that the side
Method specifically includes:
Step 1) obtains node di, thus the resource utilization of several dimensions of 1≤i≤K calculates the resource of each node
Utilization rate variance;Ascending order is carried out to node according to resource utilization variance to arrange to obtain sequence L1;
The resource utilization for each dimension that step 2) is obtained according to step 1) calculates the resource load weight of each node;
Ascending order is carried out to node according to resource load weight to arrange to obtain sequence L2;
Step 3) calculates the stationary value of each node according to the historical record of each node;Node is carried out according to stationary value
Ascending order arranges to obtain sequence L3;
Step 4) is according to node in three sequence L1、L2And L3In positional value node is mapped in three-dimensional space, calculate node
Chebyshev's distance dis value:
Step 5) selects the smallest node of dis value as optimal node, will be on task deployment to the node.
3. the method for scheduling task according to claim 2 suitable for embedded container cluster, which is characterized in that the step
It is rapid 1) to specifically include:
Step 1-1) calculate node diMemory usage M (di):
M(di)=((MenTotal-MemFree)/MemTotal) × 100%
Wherein, MemTotal is memory amount, and MemFree is memory surplus;
Step 1-2) calculate node diCpu utilization rate C (di):
C(di)=((us+sy)/(us+sy+id)) × 100%
Wherein, us is User space, and sy is kernel state, id Idle state;
Step 1-3) calculate node diNetwork load utilization rate N (di):
N(di)=((N1+N2)/M) × 100%
Wherein, N1 is to flow into flow, and N2 is outflow flow, and M is maximum network handling capacity;
Step 1-4) calculate node diAverage resource utilization rate:
Wherein, N=3;r1i=M (di), r2i=C (di), r3i=N (di);
Step 1-5) according to average resource utilization rate calculate node diResource utilization variance:
Step 1-6) according to resource utilization variance Si 2Ascending order is carried out to K node to arrange to obtain sequence L1。
4. the method for scheduling task according to claim 3 suitable for embedded container cluster, which is characterized in that the step
It is rapid 2) to specifically include:
Step 2-1) obtain memory usage M (di), cpu utilization rate C (di) and network load utilization rate N (di) weight kj, 1≤
j≤N;
Step 2-2 calculate node diResource load weight W (di):
Wherein, γjFor the coefficient for adjusting weight, R1(di)=M (di), R2(di)=C (di), R3(di)=N (di);P is threshold value;
Step 2-3) according to resource load weight W (di) K node progress ascending order is arranged to obtain sequence L2。
5. the method for scheduling task according to claim 4 suitable for embedded container cluster, which is characterized in that the step
It is rapid 1) to specifically include:
Step 3-1) calculate node diStationary value Z (di):
Wherein, M is the number of unit timeslice;njFor task quantity performed in node unit time piece, NjFor the node unit time
Piece receives total task number amount, tjFor online hours in node unit time piece, tpFor unit timeslice;
Step 3-2) according to stationary value Z (di) K node progress ascending order is arranged to obtain sequence L3。
6. the method for scheduling task according to claim 5 suitable for embedded container cluster, which is characterized in that the step
It is rapid 4) specifically:
Chebyshev's distance dis that egress arrives node in three dimensions is calculated according to the following formula:
Wherein, ci1,ci2,ci3Respectively node diIn sequence L1, L2, L3In positional value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810457653.0A CN110489200B (en) | 2018-05-14 | 2018-05-14 | Task scheduling method suitable for embedded container cluster |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810457653.0A CN110489200B (en) | 2018-05-14 | 2018-05-14 | Task scheduling method suitable for embedded container cluster |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110489200A true CN110489200A (en) | 2019-11-22 |
CN110489200B CN110489200B (en) | 2022-03-08 |
Family
ID=68544858
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810457653.0A Active CN110489200B (en) | 2018-05-14 | 2018-05-14 | Task scheduling method suitable for embedded container cluster |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110489200B (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111046102A (en) * | 2019-11-27 | 2020-04-21 | 复旦大学 | High-performance block chain service system based on ether house |
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 |
CN112114950A (en) * | 2020-09-21 | 2020-12-22 | 中国建设银行股份有限公司 | Task scheduling method and device and cluster management system |
WO2021120550A1 (en) * | 2019-12-19 | 2021-06-24 | Huawei Technologies Co., Ltd. | Methods and apparatus for resource scheduling of resource nodes of a computing cluster or a cloud computing platform |
CN113377495A (en) * | 2021-05-17 | 2021-09-10 | 杭州中港科技有限公司 | Method for optimizing docker cluster deployment based on heuristic ant colony algorithm |
US20220398134A1 (en) * | 2021-06-11 | 2022-12-15 | International Business Machines Corporation | Allocation of services to containers |
CN116132447A (en) * | 2022-12-21 | 2023-05-16 | 天翼云科技有限公司 | Load balancing method and device based on Kubernetes |
CN117076142A (en) * | 2023-10-17 | 2023-11-17 | 阿里云计算有限公司 | Multi-tenant resource pool configuration method and multi-tenant service system |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060062148A1 (en) * | 2004-09-22 | 2006-03-23 | Nec Corporation | System utilization rate managing apparatus and system utilization rate managing method to be employed for it, and its program |
CN103207804A (en) * | 2013-04-07 | 2013-07-17 | 杭州电子科技大学 | MapReduce load simulation method based on cluster job logging |
US20140181839A1 (en) * | 2012-02-09 | 2014-06-26 | Tencent Technology (Shenzhen) Company Limited | Capacity-based multi-task scheduling method, apparatus and system |
CN105607947A (en) * | 2015-12-11 | 2016-05-25 | 西北工业大学 | Novel cloud environment virtual machine scheduling method |
CN106776005A (en) * | 2016-11-23 | 2017-05-31 | 华中科技大学 | A kind of resource management system and method towards containerization application |
-
2018
- 2018-05-14 CN CN201810457653.0A patent/CN110489200B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20060062148A1 (en) * | 2004-09-22 | 2006-03-23 | Nec Corporation | System utilization rate managing apparatus and system utilization rate managing method to be employed for it, and its program |
US20140181839A1 (en) * | 2012-02-09 | 2014-06-26 | Tencent Technology (Shenzhen) Company Limited | Capacity-based multi-task scheduling method, apparatus and system |
CN103207804A (en) * | 2013-04-07 | 2013-07-17 | 杭州电子科技大学 | MapReduce load simulation method based on cluster job logging |
CN105607947A (en) * | 2015-12-11 | 2016-05-25 | 西北工业大学 | Novel cloud environment virtual machine scheduling method |
CN106776005A (en) * | 2016-11-23 | 2017-05-31 | 华中科技大学 | A kind of resource management system and method towards containerization application |
Non-Patent Citations (1)
Title |
---|
黄凯: "基于Docker Swarm集群的动态加权调度策略", 《计算机应用》 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111046102A (en) * | 2019-11-27 | 2020-04-21 | 复旦大学 | High-performance block chain service system based on ether house |
CN111046102B (en) * | 2019-11-27 | 2023-10-31 | 复旦大学 | High performance blockchain service system |
WO2021120550A1 (en) * | 2019-12-19 | 2021-06-24 | Huawei Technologies Co., Ltd. | Methods and apparatus for resource scheduling of resource nodes of a computing cluster or a cloud computing platform |
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 |
CN112114950A (en) * | 2020-09-21 | 2020-12-22 | 中国建设银行股份有限公司 | Task scheduling method and device and cluster management system |
CN113377495A (en) * | 2021-05-17 | 2021-09-10 | 杭州中港科技有限公司 | Method for optimizing docker cluster deployment based on heuristic ant colony algorithm |
CN113377495B (en) * | 2021-05-17 | 2024-02-27 | 杭州中港科技有限公司 | Dock cluster deployment optimization method based on heuristic ant colony algorithm |
US20220398134A1 (en) * | 2021-06-11 | 2022-12-15 | International Business Machines Corporation | Allocation of services to containers |
CN116132447A (en) * | 2022-12-21 | 2023-05-16 | 天翼云科技有限公司 | Load balancing method and device based on Kubernetes |
CN117076142A (en) * | 2023-10-17 | 2023-11-17 | 阿里云计算有限公司 | Multi-tenant resource pool configuration method and multi-tenant service system |
CN117076142B (en) * | 2023-10-17 | 2024-01-30 | 阿里云计算有限公司 | Multi-tenant resource pool configuration method and multi-tenant service system |
Also Published As
Publication number | Publication date |
---|---|
CN110489200B (en) | 2022-03-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110489200A (en) | A kind of method for scheduling task suitable for embedded container cluster | |
CN102780759B (en) | Based on the cloud computing resource scheduling method in regulation goal space | |
CN107045455B (en) | A kind of Docker Swarm cluster resource method for optimizing scheduling based on load estimation | |
CN106233276B (en) | The coordination admission control of network-accessible block storage device | |
WO2016082370A1 (en) | Distributed node intra-group task scheduling method and system | |
CN106385468B (en) | The predictable dynamic load balancing method of Web cluster | |
Liu et al. | Minimum-cost cloud storage service across multiple cloud providers | |
CN103793272B (en) | Periodical task scheduling method and periodical task scheduling system | |
CN108667878A (en) | Server load balancing method and device, storage medium, electronic equipment | |
CN108551489A (en) | A kind of application server load balancing method, system, device and storage medium | |
CN103491024B (en) | Job scheduling method and device for streaming data | |
CN107426332B (en) | A kind of load-balancing method and system of web server cluster | |
CN106325976B (en) | A kind of rendering task scheduling processing method and server | |
CN105892996A (en) | Assembly line work method and apparatus for batch data processing | |
CN110287245A (en) | Method and system for scheduling and executing distributed ETL (extract transform load) tasks | |
CN104899100B (en) | A kind of resource regulating method for cloud system | |
CN102591873B (en) | A kind of information recommendation method and equipment | |
CN103902384B (en) | The method and device of physical machine is distributed for virtual machine | |
CN105744006A (en) | Particle swarm optimization user request dispatching method facing multi-type service | |
CN106293533A (en) | A kind of based on the load-balancing method between storage multipath and system | |
CN106412124A (en) | Task allocation system and task allocation method for parallel ordering cloud service platform | |
CN110362379A (en) | Based on the dispatching method of virtual machine for improving ant group algorithm | |
CN106998340B (en) | Load balancing method and device for board resources | |
CN108427602A (en) | A kind of coordinated dispatching method and device of distributed computing task | |
CN100388183C (en) | Server load equalization method for implementing weighted minimum linked allocation |
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 | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20210803 Address after: Room 1601, 16th floor, East Tower, Ximei building, No. 6, Changchun Road, high tech Industrial Development Zone, Zhengzhou, Henan 450001 Applicant after: Zhengzhou xinrand Network Technology Co.,Ltd. Address before: 100190, No. 21 West Fourth Ring Road, Beijing, Haidian District Applicant before: INSTITUTE OF ACOUSTICS, CHINESE ACADEMY OF SCIENCES |
|
TA01 | Transfer of patent application right | ||
GR01 | Patent grant | ||
GR01 | Patent grant |