CN106878042A - Container resource regulating method and system based on SLA - Google Patents
Container resource regulating method and system based on SLA Download PDFInfo
- Publication number
- CN106878042A CN106878042A CN201510959326.1A CN201510959326A CN106878042A CN 106878042 A CN106878042 A CN 106878042A CN 201510959326 A CN201510959326 A CN 201510959326A CN 106878042 A CN106878042 A CN 106878042A
- Authority
- CN
- China
- Prior art keywords
- container
- resource
- target capabilities
- management object
- monitoring
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/50—Network service management, e.g. ensuring proper service fulfilment according to agreements
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Information Transfer Between Computers (AREA)
Abstract
The invention discloses a kind of container resource regulating method based on SLA and system, methods described includes:User service grade classification is carried out to the container in Docker based on SLA;Choose resource management object;Determine the target capabilities of resource management object;Container initialization is carried out, resource original allocation is carried out to container in host according to user service grade;Container in host is monitored;Target capabilities to the data combination resource management object of the actual performance for monitoring are analyzed, and judge whether actual performance reaches target capabilities, are scheduled according to judged result.Using this method or system, break the way of Docker fixed allocation resources, be more flexibly utilized by distributed system resource, be effectively improved resource utilization;By User Priority, it is ensured that the performance objective of different user, Consumer's Experience is enhanced.
Description
Technical field
The present invention relates to Internet technical field, more particularly to a kind of container resource based on SLA
Dispatching method and system.
Background technology
LXC (Linux Container) container is a kind of kernel Intel Virtualization Technology, can provide light weight
Level virtualization, so as to isolated process and resource, and need not provide instruction explanation facility with
And other fully virtualized complexity.Equivalent to the NameSpace (NameSpace) in C++.Hold
Effectively be divided into the resource managed by single operating in isolated group by device, with preferably
Balance has the resource use demand of conflict between isolated group.
As shown in drawings, the target of Docker projects is to realize the operating system virtualization of lightweight
Solution.On the basis of LXC, Docker has carried out further encapsulation, allows user
The management of care container need not be gone, the container of Docker is operated just as operating quick lightweight
Virtual machine it is equally simple.
The difference of Docker and traditional virtual (KVM, XEN etc.) mode is:Hold
Device is that virtualization is realized in operating system aspect, is directly multiplexed the operating system of local host,
And traditional approach is then the system for fictionalizing oneself on the basis of hardware, then on system top
Affix one's name to related APP applications.Relatively had the following advantages that with traditional VM:1) operation starts fast;
2) lightweight virtualization;3) increase income free;4) prospect and cloud are supported.
The problem that Docker cores are solved is that the function of similar VM is realized using LXC, from
And utilize the hardware resource supply more computer resources of user more saved.With Docker
Development, in a Docker host start container it is more and more.And, Docker
Number of containers and state inside host often change.How performance management is carried out to it
And experience ensures to turn into the problem for having increasing need for concern.In the prior art, performance management is only
For the minor resource corresponding to a small number of VIP users, the performance management of most of user is
Cannot be guaranteed, Consumer's Experience is poor.
The content of the invention
In view of the above problems, it is proposed that the present invention is to provide a kind of container resource based on SLA
Dispatching method and system, can carry out performance management and scheduling of resource according to the grade of service, improve
Consumer's Experience.
According to an aspect of the invention, there is provided a kind of container scheduling of resource based on SLA
Method, comprises the following steps:
User service grade classification is carried out to the container in Docker based on SLA;
Choose resource management object;
Determine the target capabilities of resource management object;
Container initialization is carried out, resource is carried out to container in host according to user service grade
Original allocation;
Container in host is monitored;
Target capabilities to the data combination resource management object of the actual performance for monitoring are divided
Analysis, judges whether actual performance reaches target capabilities, is scheduled according to judged result.
Alternatively, the resource management object is serviced at least to be provided to the application program of user
One Resource Unit.
Alternatively, the Resource Unit is divided into cpu resource unit, memory source unit, disk
Resource Unit, Internet resources unit.
Alternatively, it is described determine resource management object target capabilities include it is following at least one
Kind:
According to default performance strategy file in system, combined by some attributes for managing object
Performance strategy file determines that management object can obtain the target capabilities of performance guarantee;By custodian
Member sets the target capabilities of management object by administration interface manually.
Alternatively, it is described monitoring is carried out to container in host to include:
Periodically or routinely monitor the actual performance of each container in host.
Alternatively, it is described monitoring is carried out to container in host to include:
Container in scanning host, is that each container configures corresponding process;
Each Process Synchronization extracts the status information of respective container;
Status information to each container is processed, and obtains monitoring information;
The monitoring information of each container is aggregated into data queue, the monitoring information profit of each container
Made a distinction with container identification.
Alternatively, described each Process Synchronization extracts the status information of respective container including with next
Item is multinomial:
Cpu occupied informations, memory occupation information, network traffic information and disk take letter
Breath.
Alternatively, the described pair of mesh of the data combination resource management object of the actual performance for monitoring
Mark performance is analyzed, and judges whether actual performance reaches target capabilities, is entered according to judged result
Row scheduling includes:
If being unsatisfactory for target capabilities, it needs to be determined that carrying out scheduling of resource;According to user service
Grade, the performance of the container of the preferential user for ensureing that priority is higher;
If meeting target capabilities, proceed monitoring.
According to another aspect of the present invention, there is provided a kind of container resource based on SLA is adjusted
Degree system, including:
Grade of service division module, user service is carried out based on SLA to the container in Docker
Grade classification;
Resource management object select module, chooses resource management object;
Target capabilities determining module, determines the target capabilities of resource management object;
Container initialization module, carries out container initialization, according to user service grade in host
In resource original allocation is carried out to container;
Monitoring module, monitors to container in host;
Scheduler module, the target of the data combination resource management object of actual performance to monitoring
Performance is analyzed, and judges whether actual performance reaches target capabilities, is carried out according to judged result
Scheduling.
Alternatively, the resource management object is serviced at least to be provided to the application program of user
One Resource Unit.
Alternatively, the Resource Unit is divided into cpu resource unit, memory source unit, disk
Resource Unit, Internet resources unit.
Alternatively, the target capabilities determining module determines the target capabilities bag of resource management object
Include it is following at least one:
According to default performance strategy file in system, combined by some attributes for managing object
Performance strategy file determines that management object can obtain the target capabilities of performance guarantee;By custodian
Member sets the target capabilities of management object by administration interface manually.
Alternatively, the monitoring module carries out monitoring to container in host includes:
Periodically or routinely monitor the actual performance of each container in host.
Alternatively, the monitoring module includes:
Scanning submodule, scanning host in container, be each container configuration it is corresponding enter
Journey;
Submodule is extracted, each Process Synchronization extracts the status information of respective container;
Treatment submodule, the status information to each container is processed, and obtains monitoring information;
Collects submodule, the monitoring information of each container is aggregated into data queue, each container
Monitoring information made a distinction using container identification.
Alternatively, described each Process Synchronization extracts the status information of respective container including with next
Item is multinomial:
Cpu occupied informations, memory occupation information, network traffic information and disk take letter
Breath.
Alternatively, the data combination resource management of actual performance of the scheduler module to monitoring
The target capabilities of object are analyzed, and judge whether actual performance reaches target capabilities, according to sentencing
Disconnected result be scheduled including:
If being unsatisfactory for target capabilities, it needs to be determined that carrying out scheduling of resource;According to user service
Grade, the performance of the container of the preferential user for ensureing that priority is higher;
If meeting target capabilities, proceed monitoring.
The container resource regulating method based on SLA and system provided by the present invention, according to every
The User Priority of individual container sets up target capabilities;For each container creates respective process, adopt
With the mode of multi-process parallel running, extraction treatment obtains the monitoring information of each container, more has
Beneficial to the running status for reflecting each container in host;And then according to the user of each container
Priority carries out scheduling of resource.Using this method or system, Docker fixed allocations money is broken
The way in source, has more been flexibly utilized by distributed system resource, is effectively improved resource profit
With rate;By User Priority, it is ensured that the performance objective of different user, user's body is enhanced
Test.
Described above is only the general introduction of technical solution of the present invention, in order to better understand this hair
Bright technological means, and being practiced according to the content of specification, and in order to allow the present invention
Above and other objects, features and advantages can become apparent, below especially exemplified by of the invention
Specific embodiment.
Brief description of the drawings
By reading the detailed description of hereafter preferred embodiment, various other advantage and benefit
Will be clear understanding for those of ordinary skill in the art.Accompanying drawing is only used for showing to be preferable to carry out
The purpose of mode, and it is not considered as limitation of the present invention.And in whole accompanying drawing, use
Identical reference symbol represents identical part.In the accompanying drawings:
Fig. 1 diagrammatically illustrates Docker in the prior art and VM comparison diagrams;
Fig. 2 diagrammatically illustrates the container resource based on SLA according to an embodiment of the invention
The flow chart of dispatching method;
Fig. 3 diagrammatically illustrates the container resource based on SLA according to an embodiment of the invention
The flow chart monitored to container in host in scheduling;
Fig. 4 diagrammatically illustrates the container resource based on SLA according to an embodiment of the invention
The block diagram of scheduling system.
Fig. 5 diagrammatically illustrates the container resource based on SLA according to an embodiment of the invention
The monitoring module block diagram of scheduling system.
Specific embodiment
The invention will be further described with specific embodiment below in conjunction with the accompanying drawings.
As shown in Fig. 2 the container resource regulating method based on SLA of one embodiment of the invention
Comprise the following steps:
101, the grade of service is divided
By SLA (Service Level Agrement, clothes before resource management object is chosen
Business level protocol) division of the grade of service is carried out come the application program to the user in Docker,
The object of the division of the grade of service for the application program of user provide service at least one
Resource Unit, that is, manage object.Different QoS (the Quality of of different grade of service correspondences
Service, service quality).
102, choose management object
Choose management object in Docker, the management object is for user in Docker
Application program provides the Resource Unit of service.Specifically, the Resource Unit can be divided into CPU
Resource Unit, memory source unit, disk resource unit, Internet resources unit etc., for for
The application program of user provides the services such as calculating, storage, transmission.
103, determine target capabilities
After it is determined that managing object, it is determined that the target capabilities of management object.
Specifically, can be according to default performance strategy file in system, by managing object
Some attribute binding ability strategy files determine that management object can obtain the Objective of performance guarantee
Can, it is also possible to which the target capabilities of management object are set manually by administrative staff by administration interface.
In addition, it is also possible to QoS has been had predetermined that in the grade of service, for example, above-mentioned
If it have been determined that the target capabilities of the grade of service during segmentation service grade in step 101, then
The target capabilities of the grade of service can be defined as the management object of selection in the grade of service
Target capabilities.
The type of target capabilities is a lot, can include but is not limited to CPU usage, response delay,
Read-write number of times per second, message transmission rate etc.;Can be single parameter, or many seed ginsengs
Several combinations.
104, container initialization is carried out, container is carried out in host according to user service grade
Resource original allocation;
The application program of user is existed in the form of container in Docker technologies, container
Between be isolated from each other, mutually it is invisible.
105, container in host is monitored
Periodically or routinely monitor the actual performance of each container in host.The reality of monitoring
The type of border performance can be identical with target type, it is also possible to different.Specifically, such as Objective
1ms can be less than for propagation delay time, the type of the actual performance information of monitoring may also be time delay, such as
Monitor that the actual time delay for managing object is 0.5ms.Additionally, monitoring actual performance information with
The different situation of target type also likely to be present.
As shown in figure 3, in the present embodiment, container in host is carried out monitoring include it is following
Sub-step:
1051, the container in scanning host is that each container configures corresponding process;
Specifically, place in local monitor program, timing scan host is started inside host
It is that it creates corresponding process in the main frame of existing state, and by Container Name record in container
In list.By above-mentioned steps, multiple processes are created, executed in parallel is obtaining each container
Monitoring information.
Step 1052, each Process Synchronization extracts the status information of respective container;
The status information that each process is extracted is including but not limited to following several:Cpu status informations,
Internal storage state information, network state information and Disk State information, this several information are reflections
The key index information of container running status.
Having in Docker hosts under assigned catalogue cgroup, cgroup a catalogue has some texts
Part is real-time update, and cpu status informations and internal storage state information can be by reading these files
The data of record are obtained.And network state information and Disk State information can be by calling Docker
Exec () interface obtain.
The process extracts described under the assigned catalogue cgroup of Docker application container engines
The interrupting value of all cpu in host inside;The process obtains the cpu shared by corresponding container
Total interrupting value;And according to the cpu sharing policies of the host, to shared by the container
The total interrupting values of cpu carry out processing the cpu occupied informations for obtaining the container.
The process extracts internal memory shape under the assigned catalogue cgroup of docker application container engines
State information, to internal memory status information process obtaining memory occupation information;
The process calls exec () interface of docker application container engines, and extraction obtains network
Status information and/or Disk State information, to the network state information and/or Disk State information
Process and obtain network traffic information and/or disk occupied information.
Step 1053, the status information to each container is processed, and obtains monitoring information;
The monitoring information that the treatment of each process is obtained is including but not limited to following several:Cpu takes
Information, memory occupation information, network traffic information and disk occupied information.
Step 1054, the monitoring information of all containers is aggregated into data queue, each container
Monitoring information made a distinction using container identification;
Step 106, the data combining target performance of the actual performance to monitoring is analyzed,
Judge whether actual performance reaches target capabilities.
If being unsatisfactory for target capabilities, it needs to be determined that carrying out scheduling of resource;According to user service
Grade, the performance of the container of the preferential user for ensureing that priority is higher.
Such as, when actual performance IO time delays are exceeded, then scheduling of resource is carried out, increases this service and step on
The resource proportioning of the container of note, such as CPU, caching can also be by limiting lower priority
The service traffics of container of user of the grade of service meet the grade of service of higher priority
Demand.Have benefited from cgroup, can easily dynamic adjustresources limitation.
If meeting target capabilities, return to step 105, proceed monitoring, in order to
Its performance is unsatisfactory for carrying out performance management during target capabilities.
The present embodiment can also set the frequency of scheduling of resource, such as 10s/ times.The frequency of scheduling of resource
Rate is relevant with the turnover rate of monitoring data, monitors that the turnover rate of data is at least above scheduling of resource
Frequency.Further, the turnover rate of monitoring data is the pre- timing performed by the present embodiment method
Between be spaced determine, optionally, the scheduled time according to systematic function and set.In system fortune
Go in the case of normal, aboundresources, administrative staff are without often to container resource in host
It is scheduled, such scheduled time setting can be more long;System operation is abnormal, resource is nervous
When, administrative staff need the container resource in Real-Time Scheduling host, so pre- timing
Between setting can be shorter, lifting monitoring data turnover rate, and then improve scheduling of resource frequency.
As shown in figure 4, the container resource scheduling system based on SLA of one embodiment of the invention,
Including:
Grade of service division module,
By SLA (Service Level Agrement, clothes before resource management object is chosen
Business level protocol) division of the grade of service is carried out come the application program to the user in Docker,
The object of the division of the grade of service for the application program of user provide service at least one
Resource Unit, that is, manage object.Different QoS (the Quality of of different grade of service correspondences
Service, service quality).
Management object select module,
Choose management object in Docker, the management object is for user in Docker
Application program provides the Resource Unit of service.Specifically, the Resource Unit can be divided into CPU
Resource Unit, memory source unit, disk resource unit, Internet resources unit etc., for for
The application program of user provides the services such as calculating, storage, transmission.
Target capabilities determining module, after it is determined that managing object, it is determined that the Objective of management object
Energy.
Specifically, can be according to default performance strategy file in system, by managing object
Some attribute binding ability strategy files determine that management object can obtain the Objective of performance guarantee
Can, it is also possible to which the target capabilities of management object are set manually by administrative staff by administration interface.
In addition, it is also possible to QoS has been had predetermined that in the grade of service, if for example, clothes
Business grade classification module is in segmentation service grade if it have been determined that the Objective of the grade of service
Can, then the target capabilities of the grade of service can be defined as the pipe of selection in the grade of service
Manage the target capabilities of object.
The type of target capabilities is a lot, can include but is not limited to CPU usage, response delay,
Read-write number of times per second, message transmission rate etc.;Can be single parameter, or many seed ginsengs
Several combinations.
Container initialization module, resource is carried out according to user service grade in host to container
Original allocation;
The application program of user is existed in the form of container in Docker technologies, container
Between be isolated from each other, mutually it is invisible.
Monitoring module, monitors to container in host
Periodically or routinely monitor the actual performance of each container in host.The reality of monitoring
The type of border performance can be identical with target type, it is also possible to different.Specifically, such as Objective
1ms can be less than for propagation delay time, the type of the actual performance information of monitoring may also be time delay, such as
Monitor that the actual time delay for managing object is 0.5ms.Additionally, monitoring actual performance information with
The different situation of target type also likely to be present.
As shown in figure 5, in the present embodiment, monitoring module includes:
Scanning submodule, the container in scanning host is that each container configures corresponding process;
Specifically, place in local monitor program, timing scan host is started inside host
It is that it creates corresponding process in the main frame of existing state, and by Container Name record in container
In list.By above-mentioned steps, multiple processes are created, executed in parallel is obtaining each container
Monitoring information.
Submodule is extracted, each Process Synchronization extracts the status information of respective container;
The status information that each process is extracted is including but not limited to following several:Cpu status informations,
Internal storage state information, network state information and Disk State information, this several information are reflections
The key index information of container running status.
Having in Docker hosts under assigned catalogue cgroup, cgroup a catalogue has some texts
Part is real-time update, and cpu status informations and internal storage state information can be by reading these files
The data of record are obtained.And network state information and Disk State information can be by calling Docker
Exec () interface obtain.
The process extracts described under the assigned catalogue cgroup of Docker application container engines
The interrupting value of all cpu in host inside;The process obtains the cpu shared by corresponding container
Total interrupting value;And according to the cpu sharing policies of the host, to shared by the container
The total interrupting values of cpu carry out processing the cpu occupied informations for obtaining the container.
The process extracts internal memory shape under the assigned catalogue cgroup of docker application container engines
State information, to internal memory status information process obtaining memory occupation information;
The process calls exec () interface of docker application container engines, and extraction obtains network
Status information and/or Disk State information, to the network state information and/or Disk State information
Process and obtain network traffic information and/or disk occupied information.
Treatment submodule, the status information to each container is processed, and obtains monitoring information;
The monitoring information that the treatment of each process is obtained is including but not limited to following several:Cpu takes
Information, memory occupation information, network traffic information and disk occupied information.
Collects submodule, the monitoring information of all containers is aggregated into data queue, each appearance
The monitoring information of device is made a distinction using container identification;
Scheduler module, the data combining target performance of the actual performance to monitoring is analyzed,
Judge whether actual performance reaches target capabilities.
If being unsatisfactory for target capabilities, it needs to be determined that carrying out scheduling of resource;According to user service
Grade, the performance of the container of the preferential user for ensureing that priority is higher.
Such as, when actual performance IO time delays are exceeded, then scheduling of resource is carried out, increases this service and step on
The resource proportioning of the container of note, such as CPU, caching can also be by limiting lower priority
The service traffics of container of user of the grade of service meet the grade of service of higher priority
Demand.Have benefited from cgroup, can easily dynamic adjustresources limitation.
If meeting target capabilities, proceed monitoring, in order to be unsatisfactory for mesh in its performance
Performance management is carried out during mark performance.
The present embodiment can also set the frequency of scheduling of resource, such as 10s/ times.The frequency of scheduling of resource
Rate is relevant with the turnover rate of monitoring data, monitors that the turnover rate of data is at least above scheduling of resource
Frequency.Further, the turnover rate of monitoring data is the pre- timing performed by the present embodiment method
Between be spaced determine, optionally, the scheduled time according to systematic function and set.In system fortune
Go in the case of normal, aboundresources, administrative staff are without often to container resource in host
It is scheduled, such scheduled time setting can be more long;System operation is abnormal, resource is nervous
When, administrative staff need the container resource in Real-Time Scheduling host, so pre- timing
Between setting can be shorter, lifting monitoring data turnover rate, and then improve scheduling of resource frequency.
The container resource regulating method based on SLA and system provided by the present embodiment, according to
The User Priority of each container sets up target capabilities;For each container creates respective process,
By the way of multi-process parallel running, extraction treatment obtains the monitoring information of each container, more
Be conducive to reflecting the running status of each container host Nei;And then according to the use of each container
Family priority carries out scheduling of resource.Using this method or system, Docker fixed allocations are broken
The way of resource, has more been flexibly utilized by distributed system resource, is effectively improved resource
Utilization rate;By User Priority, it is ensured that the performance objective of different user, user is enhanced
Experience.
All parts embodiment of the invention can realize with hardware, or with one or many
The software module run on individual processor is realized, or is realized with combinations thereof.This area
It will be appreciated by the skilled person that microprocessor or digital signal processor can be used in practice
(DSP) come in realizing the container resource scheduling system based on SLA according to embodiments of the present invention
Some or all parts some or all functions.The present invention be also implemented as
Perform method as described herein some or all equipment or program of device (for example,
Computer program and computer program product).It is such to realize that program of the invention be stored
On a computer-readable medium, or can have one or more signal form.So
Signal can be downloaded from internet website and obtain, or provided on carrier signal, or
There is provided in any other form.
" one embodiment ", " embodiment " or " one or more implementation referred to herein
Example " is it is meant that the special characteristic, structure or the characteristic that describe are included in the present invention in conjunction with the embodiments
At least one embodiment in.Further, it is noted that word " in one embodiment " here
Example is not necessarily all referring to same embodiment.
In specification mentioned herein, numerous specific details are set forth.However, can manage
Solution, embodiments of the invention can be put into practice in the case of without these details.One
In a little examples, known method, structure and technology is not been shown in detail, so as not to fuzzy to this
The understanding of specification.
It should be noted that above-described embodiment the present invention will be described is carried out rather than to the present invention
Limit, and those skilled in the art without departing from the scope of the appended claims may be used
Design alternative embodiment.In the claims, any reference that will should not be located between bracket
Symbol construction is into limitations on claims.Word "comprising" do not exclude the presence of be not listed in right will
Element or step in asking.Word "a" or "an" before element does not exclude the presence of many
Individual such element.The present invention by means of the hardware for including some different elements and can be borrowed
Help properly programmed computer to realize.If in the unit claim for listing equipment for drying,
Several in these devices can be embodied by same hardware branch.Word first,
Second and third use do not indicate that any order.These words can be construed to title.
Furthermore, it should also be noted that the language used in this specification primarily to it is readable and
The purpose of teaching and select, selected rather than in order to explain or limit subject of the present invention
's.Therefore, in the case of without departing from the scope of the appended claims and spirit, for this
Many modifications and changes will be apparent from for the those of ordinary skill of technical field.For
The scope of the present invention, is illustrative and not restrictive, sheet to the disclosure that the present invention is done
The scope of invention is defined by the appended claims.
Claims (10)
1. a kind of container resource regulating method based on SLA, it is characterised in that including following
Step:
User service grade classification is carried out to the container in Docker based on SLA;
Choose resource management object;
Determine the target capabilities of resource management object;
Container initialization is carried out, resource is carried out to container in host according to user service grade
Original allocation;
Container in host is monitored;
Target capabilities to the data combination resource management object of the actual performance for monitoring are divided
Analysis, judges whether actual performance reaches target capabilities, is scheduled according to judged result.
2. the container resource regulating method of SLA, its feature are based on as claimed in claim 1
It is,
The resource management object for the application program of user provide service at least one resource
Unit;
The Resource Unit is divided into cpu resource unit, memory source unit, disk resource list
Unit, Internet resources unit.
3. the container resource regulating method of SLA, its feature are based on as claimed in claim 1
Be, it is described determine resource management object target capabilities include it is following at least one:
According to default performance strategy file in system, combined by some attributes for managing object
Performance strategy file determines that management object can obtain the target capabilities of performance guarantee;By custodian
Member sets the target capabilities of management object by administration interface manually.
4. the container resource regulating method of SLA, its feature are based on as claimed in claim 1
It is, it is described monitoring is carried out to container in host to include:
Periodically or routinely monitor the actual performance of each container in host;
Container in scanning host, is that each container configures corresponding process;
Each Process Synchronization extracts the status information of respective container;
Status information to each container is processed, and obtains monitoring information;
The monitoring information of each container is aggregated into data queue, the monitoring information profit of each container
Made a distinction with container identification;
The status information includes that cpu occupied informations, EMS memory occupation are believed following one or more
Breath, network traffic information and disk occupied information.
5. the container resource regulating method of SLA, its feature are based on as claimed in claim 1
It is, the described pair of target capabilities of the data combination resource management object of the actual performance for monitoring
It is analyzed, judges whether actual performance reaches target capabilities, is scheduled according to judged result
Including:
If being unsatisfactory for target capabilities, it needs to be determined that carrying out scheduling of resource;According to user service
Grade, the performance of the container of the preferential user for ensureing that priority is higher;
If meeting target capabilities, proceed monitoring.
6. a kind of container resource scheduling system based on SLA, it is characterised in that including:
Grade of service division module, user service is carried out based on SLA to the container in Docker
Grade classification;
Resource management object select module, chooses resource management object;
Target capabilities determining module, determines the target capabilities of resource management object;
Container initialization module, carries out container initialization, according to user service grade in host
In resource original allocation is carried out to container;
Monitoring module, monitors to container in host;
Scheduler module, the target of the data combination resource management object of actual performance to monitoring
Performance is analyzed, and judges whether actual performance reaches target capabilities, is carried out according to judged result
Scheduling.
7. the container resource scheduling system of SLA, its feature are based on as claimed in claim 6
It is,
The resource management object for the application program of user provide service at least one resource
Unit;
The Resource Unit is divided into cpu resource unit, memory source unit, disk resource list
Unit, Internet resources unit.
8. the container resource scheduling system of SLA, its feature are based on as claimed in claim 6
It is that it is following that the target capabilities determining module determines that the target capabilities of resource management object include
In at least one:
According to default performance strategy file in system, combined by some attributes for managing object
Performance strategy file determines that management object can obtain the target capabilities of performance guarantee;By custodian
Member sets the target capabilities of management object by administration interface manually.
9. the container resource scheduling system of SLA, its feature are based on as claimed in claim 6
It is that the monitoring module carries out monitoring to container in host to be included:
Periodically or routinely monitor the actual performance of each container in host.
The monitoring module includes:
Scanning submodule, scanning host in container, be each container configuration it is corresponding enter
Journey;
Submodule is extracted, each Process Synchronization extracts the status information of respective container;
Treatment submodule, the status information to each container is processed, and obtains monitoring information;
Collects submodule, the monitoring information of each container is aggregated into data queue, each container
Monitoring information made a distinction using container identification.
The status information includes that cpu occupied informations, EMS memory occupation are believed following one or more
Breath, network traffic information and disk occupied information.
10. the container resource scheduling system of SLA, its feature are based on as claimed in claim 6
It is that the scheduler module is to the data combination resource management object of actual performance that monitors
Target capabilities are analyzed, and judge whether actual performance reaches target capabilities, according to judged result
Be scheduled including:
If being unsatisfactory for target capabilities, it needs to be determined that carrying out scheduling of resource;According to user service
Grade, the performance of the container of the preferential user for ensureing that priority is higher;
If meeting target capabilities, proceed monitoring.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510959326.1A CN106878042A (en) | 2015-12-18 | 2015-12-18 | Container resource regulating method and system based on SLA |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510959326.1A CN106878042A (en) | 2015-12-18 | 2015-12-18 | Container resource regulating method and system based on SLA |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106878042A true CN106878042A (en) | 2017-06-20 |
Family
ID=59238939
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510959326.1A Pending CN106878042A (en) | 2015-12-18 | 2015-12-18 | Container resource regulating method and system based on SLA |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106878042A (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107291551A (en) * | 2017-06-22 | 2017-10-24 | 北京奇艺世纪科技有限公司 | The load equilibration scheduling method and device of a kind of container |
CN107436801A (en) * | 2017-06-23 | 2017-12-05 | 努比亚技术有限公司 | I/O resource regulating methods, terminal and computer-readable recording medium |
CN109783533A (en) * | 2018-12-13 | 2019-05-21 | 平安科技(深圳)有限公司 | Collecting method, device, computer equipment and storage 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 |
CN110990119A (en) * | 2019-12-23 | 2020-04-10 | 中通服咨询设计研究院有限公司 | Method for improving service capability of Iaas cloud platform based on container technology |
CN113220423A (en) * | 2021-06-04 | 2021-08-06 | 恒为科技(上海)股份有限公司 | Multi-chip management method and device based on container |
US11249809B1 (en) | 2021-02-05 | 2022-02-15 | International Business Machines Corporation | Limiting container CPU usage based on network traffic |
CN115292024A (en) * | 2022-10-09 | 2022-11-04 | 苏州市中地行信息技术有限公司 | Multi-target oriented data screening method and system |
CN116185642A (en) * | 2023-04-24 | 2023-05-30 | 安徽海马云科技股份有限公司 | Container memory optimization method and device, storage medium and electronic device |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101443761A (en) * | 2005-04-06 | 2009-05-27 | 国际商业机器公司 | QOS-enabled lifecycle management for file systems |
CN102195886A (en) * | 2011-05-30 | 2011-09-21 | 兰雨晴 | Service scheduling method on cloud platform |
CN103649920A (en) * | 2011-06-27 | 2014-03-19 | 微软公司 | Resource management for cloud computing platforms |
CN103763130A (en) * | 2013-12-31 | 2014-04-30 | 华为数字技术(苏州)有限公司 | Method, device and system for managing large-scale cluster |
CN104899126A (en) * | 2015-06-12 | 2015-09-09 | 北京奇虎科技有限公司 | Method, device and system for performing local real-time monitoring on containers in host |
CN104991830A (en) * | 2015-07-10 | 2015-10-21 | 山东大学 | YARN resource allocation and energy-saving scheduling method and system based on service level agreement |
-
2015
- 2015-12-18 CN CN201510959326.1A patent/CN106878042A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101443761A (en) * | 2005-04-06 | 2009-05-27 | 国际商业机器公司 | QOS-enabled lifecycle management for file systems |
CN102195886A (en) * | 2011-05-30 | 2011-09-21 | 兰雨晴 | Service scheduling method on cloud platform |
CN103649920A (en) * | 2011-06-27 | 2014-03-19 | 微软公司 | Resource management for cloud computing platforms |
CN103763130A (en) * | 2013-12-31 | 2014-04-30 | 华为数字技术(苏州)有限公司 | Method, device and system for managing large-scale cluster |
CN104899126A (en) * | 2015-06-12 | 2015-09-09 | 北京奇虎科技有限公司 | Method, device and system for performing local real-time monitoring on containers in host |
CN104991830A (en) * | 2015-07-10 | 2015-10-21 | 山东大学 | YARN resource allocation and energy-saving scheduling method and system based on service level agreement |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107291551A (en) * | 2017-06-22 | 2017-10-24 | 北京奇艺世纪科技有限公司 | The load equilibration scheduling method and device of a kind of container |
CN107291551B (en) * | 2017-06-22 | 2020-06-26 | 北京奇艺世纪科技有限公司 | Load balancing scheduling method and device for container |
CN107436801A (en) * | 2017-06-23 | 2017-12-05 | 努比亚技术有限公司 | I/O resource regulating methods, terminal and computer-readable recording medium |
CN109783533A (en) * | 2018-12-13 | 2019-05-21 | 平安科技(深圳)有限公司 | Collecting method, device, computer equipment and storage 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 |
CN110990119A (en) * | 2019-12-23 | 2020-04-10 | 中通服咨询设计研究院有限公司 | Method for improving service capability of Iaas cloud platform based on container technology |
US11249809B1 (en) | 2021-02-05 | 2022-02-15 | International Business Machines Corporation | Limiting container CPU usage based on network traffic |
CN113220423A (en) * | 2021-06-04 | 2021-08-06 | 恒为科技(上海)股份有限公司 | Multi-chip management method and device based on container |
CN115292024A (en) * | 2022-10-09 | 2022-11-04 | 苏州市中地行信息技术有限公司 | Multi-target oriented data screening method and system |
CN115292024B (en) * | 2022-10-09 | 2023-03-24 | 苏州市中地行信息技术有限公司 | Multi-target oriented data screening method and system |
CN116185642A (en) * | 2023-04-24 | 2023-05-30 | 安徽海马云科技股份有限公司 | Container memory optimization method and device, storage medium and electronic device |
CN116185642B (en) * | 2023-04-24 | 2023-07-18 | 安徽海马云科技股份有限公司 | Container memory optimization method and device, storage medium and electronic device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106878042A (en) | Container resource regulating method and system based on SLA | |
US11842208B2 (en) | Virtual provisioning with implementation resource boundary awareness | |
US10003500B2 (en) | Systems and methods for resource sharing between two resource allocation systems | |
AU2011314183B2 (en) | Virtual resource cost tracking with dedicated implementation resources | |
US9436508B1 (en) | Provisioning virtual resource on a server based on label associated with virtual resource and servers | |
CN105103506B (en) | For the method and system for the non-homogeneous bandwidth request allocation bandwidth in system for cloud computing | |
CN108829469B (en) | Application page display method and device | |
CN108304250A (en) | Method and apparatus for the node for determining operation machine learning task | |
US20120084113A1 (en) | Virtual resource cost tracking with dedicated implementation resources | |
US20200174844A1 (en) | System and method for resource partitioning in distributed computing | |
CN106933664B (en) | Hadoop cluster resource scheduling method and device | |
CN111506404A (en) | Kubernetes-based shared GPU (graphics processing Unit) scheduling method | |
WO2017010922A1 (en) | Allocation of cloud computing resources | |
US20200348977A1 (en) | Resource scheduling methods, device and system, and central server | |
CN107040578A (en) | Method of data synchronization, device and system | |
CN105677342B (en) | A kind of combined table top virtual method solving heterogeneous operating system | |
CN110166507A (en) | More resource regulating methods and device | |
US11005718B2 (en) | Determining capabilities of cognitive entities in a distributed network based on application of cognitive protocols | |
Islam et al. | SLA-based scheduling of spark jobs in hybrid cloud computing environments | |
Senjab et al. | A survey of Kubernetes scheduling algorithms | |
CN106874102A (en) | Resource regulating method and device based on container work property | |
CN111475251A (en) | Cluster container scheduling method, system, terminal and storage medium | |
US11016816B1 (en) | Compute capacity allocation based on multi-factor analysis | |
CN116450290A (en) | Computer resource management method and device, cloud server and storage medium | |
WO2019119951A1 (en) | Device resource management |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170620 |