CN102262567A - Virtual machine scheduling decision system, platform and method - Google Patents
Virtual machine scheduling decision system, platform and method Download PDFInfo
- Publication number
- CN102262567A CN102262567A CN2010101937884A CN201010193788A CN102262567A CN 102262567 A CN102262567 A CN 102262567A CN 2010101937884 A CN2010101937884 A CN 2010101937884A CN 201010193788 A CN201010193788 A CN 201010193788A CN 102262567 A CN102262567 A CN 102262567A
- Authority
- CN
- China
- Prior art keywords
- strategy
- virtual machine
- decision
- scheduling
- cluster
- 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
Images
Landscapes
- Stored Programmes (AREA)
Abstract
The invention discloses a virtual machine scheduling decision system, a virtual machine scheduling decision platform and a virtual machine scheduling decision method. The virtual machine scheduling decision system comprises a template management module, a strategy management module and a scheduling decision calculation module, wherein the template management module is used for selecting a strategy template customized according to user requirements; the strategy management module is used for creating and activating a strategy by setting parameters in the strategy template; and the scheduling decision calculation module is used for performing exhaustive calculation on current resource information and cluster operating data by using the activated strategy to acquire the optimal decision. In the invention, the strategy template of a script is used for performing development and deployment according to the virtual machine scheduling strategy, the development speed is high and the deployment is flexible; besides, abundant alternative strategies can be provided to meet requirements of deferent levels of users, and then the system performance of the virtual machine is improved.
Description
Technical field
The present invention relates to communication and computer realm, in particular to a kind of system, platform and method of scheduling virtual machine.
Background technology
Along with the all-IP of communication network and broadband, the development trend of mutual synthesis and ubiquit appears in telecommunication technology and IT/ Internet technology, and cloud computing more and more obtains the concern of telecommunication industry in the recent period.The telecom operation commercial city takes up to carry out multiple services operation based on cloud computing platform.The basis of cloud computing platform is a virtual platform, and one of gordian technique of virtual platform is exactly the scheduling virtual machine decision-making technic, and it is a support technology of realizing cluster virtual machine high availability, load balancing, administration of energy conservation etc.
At present, virtual platform product such as VMware ESXi etc. provide scheduling virtual machine decision making functions such as cluster virtual machine high availability, load balancing.But, the user is diversified for the demand of virtual machine scheduling policy, and the virtual machine scheduling policy that existing virtual platform provides is fixed, the not customization of support policy template and policy template, can't can not satisfy the advanced requirement of user according to the scheduling strategy of user's request customizing virtual machine for scheduling virtual machine.
The inventor finds that there is following defective in scheduling virtual machine decision-making mode in the above-mentioned correlation technique: the not customization of support policy template and policy template, can not satisfy the diversity demand of user for virtual machine scheduling policy.
Summary of the invention
Fundamental purpose of the present invention is to provide a kind of system, platform and method of scheduling virtual machine, to solve the problem of above-mentioned not support policy template and policy template customization at least.
According to an aspect of the present invention, provide a kind of scheduling virtual machine decision system, having comprised: template management module is used to select the policy template according to the user's request customization; Policy management module is used for the parameter by the Provisioning Policy template, creates and activation strategy; The scheduling decision computing module is used to utilize the strategy that is in state of activation, and current resource information and cluster service data are carried out exhaustive computing, obtains optimizing decision.
Preferably, in the technical program, policy management module also is used for strategy is made amendment, hangs up, deleted, or the priority adjustment.
Preferably, in the technical program, the scheduling virtual machine decision system also comprises: the cluster monitoring data interface module is used to obtain the cluster service data, and the cluster service data is converted into the input data of regulation engine; The resource management interface module is used to obtain current resource information, and current resource information is converted into the input data of regulation engine; The scheduling decision computing module is used to utilize the strategy that is in state of activation, adopts regulation engine that cluster service data and current resource information are carried out exhaustive computing, obtains optimizing decision.
According to a further aspect in the invention, a kind of platform that comprises above-mentioned scheduling virtual machine decision system is provided, has removed the scheduling virtual machine decision system, also comprised: cluster monitoring system, be used to monitor the service data of cluster, and the cluster service data is sent to the scheduling virtual machine decision system; Resource management system is used to manage current resource information, and current resource information is sent to the scheduling virtual machine decision system; Dispatching Control System is used for the optimizing decision of sink virtual machine scheduling decision system, carries out optimizing decision.
According to a further aspect in the invention, provide a kind of scheduling virtual machine decision-making technique, having comprised: selected policy template according to the user's request customization; By the parameter in the Provisioning Policy template, create and activation strategy; Utilization is in the strategy of state of activation, and current resource information and cluster service data are carried out exhaustive computing, obtains optimizing decision.
Preferably, in the technical program, creating also, activation strategy also comprises afterwards: strategy is made amendment, hangs up, deleted, or the priority adjustment.
Among the present invention, virtual machine scheduling policy uses the policy template of script to develop and dispose, and tempo of development is fast and dispose flexibly, can provide abundant alternate strategies to satisfy different levels user's needs.
Description of drawings
Accompanying drawing described herein is used to provide further understanding of the present invention, constitutes the application's a part, and illustrative examples of the present invention and explanation thereof are used to explain the present invention, do not constitute improper qualification of the present invention.In the accompanying drawings:
Fig. 1 is the synoptic diagram of apparatus of the present invention embodiment one scheduling virtual machine decision system;
Fig. 2 is the synoptic diagram of platform embodiment one scheduling virtual machine decision-making platform of the present invention;
Fig. 3 is the process flow diagram of the inventive method embodiment one scheduling virtual machine decision-making technique;
Fig. 4 is the process flow diagram of the inventive method embodiment two scheduling virtual machine decision-making techniques;
Fig. 5 is the process flow diagram of the inventive method embodiment three scheduling virtual machine decision-making techniques;
Fig. 6 is the synoptic diagram before the inventive method embodiment four virtual machine VM6 create;
Fig. 7 is the synoptic diagram after the inventive method embodiment four virtual machine VM6 create;
The synoptic diagram that Fig. 8 distributes for the cluster initial resource;
Fig. 9 is for adopting load balancing strategy scheduling back cluster resource distribution schematic diagram;
The synoptic diagram of Figure 10 for adopting Energy Saving Strategy scheduling back cluster resource to distribute.
Embodiment
Hereinafter will describe the present invention with reference to the accompanying drawings and in conjunction with the embodiments in detail.Need to prove that under the situation of not conflicting, embodiment and the feature among the embodiment among the application can make up mutually.
The deficiency of the scheduling virtual machine decision system that realizes in view of conventional art, the present invention proposes a kind of scheduling virtual machine decision system of customizable strategy newly, virtual machine scheduling policy uses the script develop and field, tempo of development is fast and dispose flexibly, can provide abundant alternate strategies to satisfy different levels user's needs; Support policy model customization of the present invention can be according to the quick custom strategies template of user's request, to satisfy the advanced requirement of user for scheduling virtual machine; Simultaneously, the present invention can carry out the scheduling virtual machine decision-making efficiently and accurately and calculate and analyze, and can improve the efficiency of decision-making of scheduling virtual machine, reduces the scheduling virtual machine expense of virtual platform.
Device embodiment one:
Fig. 1 is the synoptic diagram of apparatus of the present invention embodiment one scheduling virtual machine decision system.Present embodiment comprises: template management module is used to select the policy template according to the user's request customization; Policy management module is used for the parameter by the Provisioning Policy template, creates and activation strategy; The scheduling decision computing module is used to utilize the strategy that is in state of activation, and current resource information and cluster service data are carried out exhaustive computing, obtains optimizing decision.
In the present embodiment, policy template is exactly a script, can be understood as the corresponding policy template of a script.Policy template can be developed personnel or higher-level user's (this user needs to understand and to edit the policy template script) customization.Policy template is relevant with user's request, and it doesn't matter with the resource of concrete virtual platform and performance state.
In the present embodiment, virtual machine scheduling policy uses the script develop and field, and tempo of development is fast and dispose flexibly, can provide abundant alternate strategies to satisfy different levels user's needs; Support policy model customization of the present invention can be according to the quick custom strategies template of user's request, to satisfy the advanced requirement of user for scheduling virtual machine.
Device embodiment two:
Present embodiment will be on the basis of embodiment one, and the each several part of scheduling virtual machine decision system is elaborated.As shown in Figure 1, present embodiment scheduling virtual machine decision system comprises with the lower part:
Policy template storehouse: deposit the storehouse of policy template, be responsible for the conversation strategy template.Policy template is to adopt the script file exploitation, is the logical framework of strategy, and the user at first needs the template of selection strategy when construction strategy, afterwards the correlation parameter of Provisioning Policy again.
Template management module: be responsible for the management of policy template, by this module, operating personnel can increase newly or the deletion strategy template.
Policy management module: be responsible for the management of strategy, comprise construction strategy, deletion strategy, modification strategy, activation strategy, hang-up strategy etc.Strategy is the running example of policy template, and it comprises activation and hangs up two states.When strategy is in state of activation, this strategy will participate in the scheduling decision computing; When strategy is in suspended state, it will not participate in scheduling decision.Can also comprise priority attribute according to strategy, when scheduling decision, the strategy that priority is high calculates first participative decision making.
Cluster monitoring data interface module: be a data-interface, be responsible for obtaining the performance and the status data of cluster, and be the input data of regulation engine with these data conversion from cluster monitoring system.
Resource management interface module: be a data-interface, be responsible for obtaining the current resource information of virtual Warburg Pincus platform, and be translated into the input data of regulation engine from the resource management interface module.
The scheduling decision computing module: this module is the nucleus module of the scheduling virtual machine decision system of customizable strategy.It is responsible for and will carries out computing and decision-making from the data of cluster monitoring data interface module and resource management interface module input and the strategy of current activation.It comprises the two large divisions: scheduling strategy submodule and regulation engine submodule.What the scheduling strategy submodule was preserved is the scheduling strategy of cluster, the all corresponding scheduling strategy submodule of each cluster, each scheduling strategy submodule is divided into activation strategy district and expiration policy district again, the current strategy that is in state of activation is kept at participative decision making computing in the activation strategy district, and the strategy of suspended state then is kept in the expiration policy district.The regulation engine submodule is responsible for the calculating of making a strategic decision of data and strategy.
In the present embodiment, the scheduling virtual machine decision system can also comprise: the scheduling controlling interface module, it is the scheduling virtual machine decision system of customizable strategy and the interface module between the Dispatching Control System, and the responsible scheduling decision that the scheduling virtual machine decision system of customizable strategy is calculated sends to Dispatching Control System to be dispatched.
Present embodiment further specifies the system of scheduling virtual machine decision-making.Except that whole features with embodiment one and beneficial effect, in the present embodiment policy management module can hang up strategy, revise, operation such as deletion, thereby the dirigibility of meeting consumers' demand that strengthens more.The present invention simultaneously utilizes regulation engine to calculate, and can carry out the scheduling virtual machine decision-making efficiently and accurately and calculate and analyze, and can improve the efficiency of decision-making of scheduling virtual machine, reduces the scheduling virtual machine expense of virtual platform.
Platform embodiment one:
Fig. 2 is the synoptic diagram of platform embodiment one scheduling virtual machine decision-making platform of the present invention.As shown in Figure 2, the scheduling virtual machine decision system of customizable strategy is a subsystem of virtual platform, virtual platform comprises the plurality of sub system, the stronger system of scheduling virtual machine decision system correlativity of its customizable strategy that neutralizes comprises: resource management system, cluster monitoring system, Dispatching Control System.Its corresponding function is respectively:
Dispatching Control System: it mainly is responsible for scheduling of resource control of virtual platform, as establishment, deletion, operation, the migration of control virtual machine, the standby of main frame, wakes up etc.The scheduling operation of Dispatching Control System may cause that the virtual platform resource information changes, and Dispatching Control System need be with these change notification resource management systems.
Resource management system: it mainly is in charge of the resource information that virtual platform is managed, managed which cluster as virtual platform, each cluster comprises the resource distribution of which main frame and main frame, has moved the resource allocation conditions of which virtual machine and these virtual machines on each main frame.Resource management system also provides the resource management action inlet for the user.When the user is undertaken as operations such as establishment, deletion virtual machines by resource management system after, resource management system can send to respective operations and respective resources data the scheduling virtual machine decision system of customizable strategy and make a strategic decision, and decision information sends to Dispatching Control System and carries out corresponding operating or return user prompt the most at last.
Cluster monitoring system: the running status and the performance of the cluster that its mainly responsible monitoring virtual platform is managed, and with the scheduling virtual machine decision system that the information data that monitors the sends to customizable strategy computing of making a strategic decision, judge whether to carry out cluster resource is redistributed or dispatched, if do not need then do not do any operation, then decision information to be sent to Dispatching Control System if desired and carry out corresponding operating or return user prompt.
The scheduling virtual machine decision system of customizable strategy: it mainly is responsible for the management and the computing of making a strategic decision of strategy.Tactical management comprises: the establishment of construction strategy, deletion strategy, activation strategy, hang-up strategy and policy template and deletion etc.The decision-making computing is the Core Feature of the scheduling virtual machine decision system of customizable strategy, and it utilizes regulation engine that raw data and predetermined policy are carried out exhaustive computing, calculates optimizing decision and offers the user or directly send to Dispatching Control System and carry out.
In the present embodiment, the scheduling virtual machine decision system can be the scheduling virtual machine decision system of device embodiment one or device embodiment two, and corresponding beneficial effect with embodiment one or embodiment two, no longer repeats herein.
Method embodiment one:
Fig. 3 is the process flow diagram of the inventive method embodiment one scheduling virtual machine decision-making technique.As shown in Figure 3, present embodiment comprises:
S102: obtain current resource information and cluster service data;
S104: select policy template according to the user's request customization;
S106:, create and activation strategy by the parameter in the Provisioning Policy template;
S108: utilize the strategy that is in state of activation, current resource information and cluster service data are carried out exhaustive computing, obtain optimizing decision.
In the present embodiment, creating also, activation strategy also comprises afterwards: strategy is made amendment, hangs up, deleted, or the priority adjustment.And in follow-up flow process, utilize corresponding modification, hang-up, deletion, or the adjusted activation strategy of priority obtains optimal decision-making.
In the present embodiment, the cluster service data comprises the running status and the performance data of cluster.Also comprise after obtaining the cluster service data: the input data that the cluster service data are converted into regulation engine; Also comprise after obtaining current resource information: the input data that the current resource information of virtual platform are converted into regulation engine; Obtain optimizing decision and specifically comprise, utilize the strategy that is in state of activation, adopt regulation engine that cluster service data and current resource information are carried out exhaustive computing, obtain optimizing decision.
Present embodiment relies on the relevant apparatus of embodiment or embodiment two to realize, and has whole beneficial effects of the foregoing description, no longer repeats herein.
Method embodiment two:
Present embodiment will specify the scheduling virtual machine decision-making technique to create virtual machine instance in cluster.
When virtual machine is created in application in cluster, need to have in the cluster enough resources to satisfy the requirement of this virtual machine creating.If current cluster resource unreasonable distribution is even may cause having in the cluster enough resources also can't create virtual machine.Can be by the calculating of making a strategic decision of the scheduling virtual machine decision system of customizable strategy, under the prerequisite of the strategy of observing current resource management, the virtual machine of cluster is dispatched so that the resource of cluster is carried out reasonable distribution again, to satisfy the condition of creating this virtual machine.Fig. 4 is the process flow diagram of the inventive method embodiment two scheduling virtual machine decision-making techniques.As shown in Figure 4, to create the flow process of virtual machine as follows for the scheduling virtual machine decision system by customizable strategy:
Step S201, the user creates virtual machine by the resource management system application on cluster-specific, and resource management system sends to request the scheduling virtual machine decision system of customizable strategy;
Step S202, the scheduling virtual machine decision system is obtained the cluster resource data by the resource management interface module from resource management system;
Step S203 uses the strategy that is in state of activation in the corresponding colony dispatching strategy district to obtaining the resource data computing of making a strategic decision;
Step S204, according to the decision-making operation result, whether in cluster have enough resource create virtual machine, and how to create virtual machine if judging;
Step S205 if there are enough resources can create virtual machine, then calls the scheduling virtual machine interface module, creates virtual machine according to result of calculation;
Step S206 does not have enough resources can create virtual machine, then points out user virtual machine can't create and illustration.
Present embodiment is the concrete application of method embodiment one, has whole beneficial effects of this embodiment, no longer repeats herein.
Method embodiment three:
Present embodiment will be an example to carry out scheduling virtual machine in cluster, specify the scheduling virtual machine decision-making technique.During the cluster operation,, need to keep the load balancing of each main frame in the cluster in order to keep cluster running status efficient and healthful; The user wishes that toward contact cluster moves in more energy-conservation mode in addition, when the cluster load is light, can close automatically or some main frames of dormancy.These all need to finish by the virtual machine in the cluster is dispatched automatically.
Fig. 5 is the process flow diagram of the inventive method embodiment three scheduling virtual machine decision-making techniques.As shown in Figure 5, it is as follows to carry out the flow process of scheduling virtual machine decision-making by the scheduling virtual machine decision system of customizable strategy:
Step S301, at first, cluster monitoring system is monitored cluster, periodically obtains cluster operation correlation behavior, performance data;
Step S302, the cluster monitoring data interface module is obtained cluster operation related data from cluster monitoring system;
Step S303 carries out input rule engine after the pre-service with the cluster that obtains operation related data;
Step S304 obtains the cluster resource data by the resource management interface module from resource management system;
Step S305 carries out after the pre-service computing of making a strategic decision of input rule engine with the resource data that gets access to;
Step S306, whether according to predetermined strategy, calculating virtual machine needs to reschedule and how to dispatch;
Step S307 reschedules virtual machine if desired, calls the scheduling virtual machine interface module, according to the result of calculation scheduling virtual machine;
Step S308, this cycle finishes, waits for a period of time, and then the cluster monitoring of beginning next cycle and scheduling virtual machine decision-making.
Method embodiment four:
Present embodiment and following each embodiment will further be elaborated to the scheduling virtual machine decision-making technique on the basis of the foregoing description, have whole beneficial effects of the foregoing description, no longer repeat herein.
Fig. 6 is the synoptic diagram before the inventive method embodiment four virtual machine VM6 create.As shown in Figure 6, two main frames being arranged in the cluster, is respectively Host A and Host B, the memory size of Host A is 8G, is moving two virtual machines of VM1 and VM2 on it, and the EMS memory occupation of VM1 is 2G, the EMS memory occupation of VM2 is 3G, and the free memory of Host A current residual is 3G.The memory size of Host B is 8G, is moving three virtual machines of VM3, VM4 and VM5 on it, and the EMS memory occupation of VM3 is 1G, and the EMS memory occupation of VM4 is 1G, and the EMS memory occupation of VM5 is 3G, and the free memory of Host B current residual is 3G.Now desiring to create the EMS memory occupation amount in cluster is the virtual machine VM6 of 4G.
In the virtual platform of the scheduling virtual machine decision-making that conventional art is realized, because the EMS memory occupation amount of VM6 has surpassed Host A and any one the residue free memory of Host B, VM6 can't create in cluster.
Because it is 6G that two main frames of Host A and Host B remain free memory altogether,,, be VM6 enough resources of transferring to other use out so can consider by the virtual machine among scheduling Host A and the Host B greater than the needed 4G of VM6.In the scheduling virtual machine decision system of customizable strategy, can carry out following consideration:
1. check that cluster remains free memory altogether and whether can satisfy the virtual machine creating demand, if can not satisfy the demands, then refusal is created.Can improve operation efficiency like this, directly refusal is created in case this condition does not satisfy.
2. allow the virtual machine in the middle of the cluster main frame to move mutually, calculate whether there is the migration scheme, the virtual machine resource needed that the application of transferring to other use out on can any main frame in cluster is created.If no, then refusal is created.In addition, carry out virtual machine (vm) migration operation many times if desired and required resource could be transferred to other use out, may cause the result who loses more than gain.So need to consider the total degree of restriction virtual machine (vm) migration, if surpass restriction, then refusal is created virtual machine.
3. if there is multiple migration scheme can satisfy above-mentioned condition, then should select to move least number of times, and migration back cluster load on host computers the most balanced relative scheme.
Above-mentioned consideration is by script reflection becoming policy template.Then the parameter in the above-mentioned policy template (perhaps considering) specifically is provided with, just is converted into specific strategy.Specific as follows:
1. check that cluster remains free memory altogether and whether can satisfy the virtual machine creating demand, if can not satisfy the demands, then refusal is created.
2. allow the central virtual machine of cluster main frame to move mutually, it is 3 times that virtual machine (vm) migration upper limit number of times is set.The preferential scheme of selecting the migration least number of times.
3. preferential migration and the most balanced scheme of virtual machine creating back loading selected.
Calculate decision-making according to above strategy.Fig. 7 is the synoptic diagram after the inventive method embodiment four virtual machine VM6 create.As shown in Figure 7, specifically comprise: 1) VM1 virtual machine (vm) migration among the Host A is moved to Host B; 2) on Host A, create the VM6 virtual machine.
Method embodiment five:
Present embodiment is the embodiment of cluster load balancing.The synoptic diagram that Fig. 8 distributes for the cluster initial resource.As shown in Figure 8, cluster source material distribution situation is as follows:
Four host computer is arranged in the cluster, is respectively Host A, Host B, Host C and Host D, and the memory size of Host A is 8G, moving two virtual machines of VM1 and VM2 on it, the EMS memory occupation of VM1 is 4G, and the EMS memory occupation of VM2 is 4G, the current no remaining free memory of Host A.The memory size of Host B is 8G, is moving three virtual machines of VM3, VM4 and VM5 on it, and the EMS memory occupation of VM3 is 3G, and the EMS memory occupation of VM4 is 1G, and the EMS memory occupation of VM5 is 4G, the current no remaining free memory of Host B.The memory size of Host C is 8G, is moving two virtual machines of VM6 and VM7 on it, and the EMS memory occupation of VM6 is 2G, and the EMS memory occupation of VM7 is 1G, and the free memory of Host C current residual is 5G.The memory size of Host D is 8G, is moving two virtual machines of VM8 and VM9 on it, and the EMS memory occupation of VM8 is 2G, and the EMS memory occupation of VM9 is 1G, and the free memory of Host D current residual is 5G.
As can be seen from Figure 8, the resource distribution of this cluster is very unbalanced, and Host A and Host B do not have free memory, approaching operating at full capacity, and the load of Host C and Host D is very light.In order to guarantee cluster running status efficient and healthful, the load on Host A and the Host B need be shared on Host C and the Host D.
In the scheduling virtual machine decision system of customizable strategy, can carry out following consideration: when 1) need considering to satisfy which kind of condition, cluster needs the working load balance policy that virtual machine is rescheduled; 2) virtual machine in the middle of the cluster main frame moves mutually, calculates the virtual machine (vm) migration scheme, can make the All hosts load in the cluster reach balanced as much as possible after using this scheme; 3) carry out virtual machine (vm) migration operation many times if desired and just can reach effect of load balance, also may cause migration overhead excessive, produce the result who loses more than gain.So need to consider the total degree of restriction virtual machine (vm) migration.Take place to reach effect of load balance as far as possible under the limited migration number of times prerequisite at virtual machine.
Above-mentioned consideration is by script reflection becoming policy template.Then the parameter in the above-mentioned policy template (perhaps considering) specifically is provided with, just is converted into specific strategy.Specific as follows:
1. the threshold condition of host resource distribution degree of unbalancedness in the setting cluster needs the working load balance policy that scheduling virtual machine is made a strategic decision when surpassing threshold value.
2. setting and allowing the maximum times of the virtual machine of migration in the load balancing strategy is 3 times.
3. select the most balanced scheme of load.
Calculate decision-making according to above strategy, Fig. 9 is for adopting load balancing strategy scheduling back cluster resource distribution schematic diagram.As shown in Figure 9: 1) VM3 virtual machine (vm) migration among the Host B is moved to Host C; 2) VM1 virtual machine (vm) migration among the Host A is moved to Host D; 3) VM7 virtual machine (vm) migration among the Host C is moved to Host A.
Method embodiment six:
Present embodiment is the embodiment of cluster administration of energy conservation.Cluster source material distribution situation as shown in Figure 8, the as can be seen from the figure maldistribution of the resources of cluster, the load of Host C and Host D is very light, can consider that the virtual machine that moves on the main frame that load is very light concentrates, virtual machine on some main frame is all transferred on other main frames, and, can reach energy-conservation effect with these Host Shutdowns.
In the scheduling virtual machine decision system of customizable strategy, can carry out following consideration:
1. when need considering to satisfy which kind of condition, cluster need use Energy Saving Strategy that virtual machine is rescheduled.
2. allow the virtual machine in the middle of the cluster main frame to move mutually, calculate the virtual machine (vm) migration scheme, use the load centralization that can make after this scheme in the cluster to the least possible main frame, remaining idle Host Shutdown.
3. need to consider the total degree of restriction virtual machine (vm) migration.Take place to reach the effect of load centralization as far as possible under the limited migration number of times prerequisite at virtual machine.
Above-mentioned consideration is by script reflection becoming policy template.Then the parameter in the above-mentioned policy template (perhaps considering) specifically is provided with, just is converted into specific strategy, content is specific as follows:
1. the threshold condition of host resource load in the setting cluster need use Energy Saving Strategy that scheduling virtual machine is made a strategic decision when surpassing threshold value.
2. allowing the maximum times of the virtual machine of migration in the setting Energy Saving Strategy is 3 times.
3. need to select the minimum scheme of main frame of operation.
4. the preferential the most balanced scheme of migration back loading of selecting.
Calculate decision-making according to above strategy, the synoptic diagram of Figure 10 for adopting Energy Saving Strategy scheduling back cluster resource to distribute.As shown in figure 10, specifically comprise: 1) VM8 virtual machine (vm) migration among the Host D is moved to Host C; 2) VM4 virtual machine (vm) migration among the Host B is moved to Host C.
Obviously, those skilled in the art should be understood that, above-mentioned each module of the present invention or each step can realize with the general calculation device, they can concentrate on the single calculation element, perhaps be distributed on the network that a plurality of calculation element forms, alternatively, they can be realized with the executable program code of calculation element, thereby, they can be stored in the memory storage and carry out by calculation element, and in some cases, can carry out step shown or that describe with the order that is different from herein, perhaps they are made into each integrated circuit modules respectively, perhaps a plurality of modules in them or step are made into the single integrated circuit module and realize.Like this, the present invention is not restricted to any specific hardware and software combination.
Be the preferred embodiments of the present invention only below, be not limited to the present invention, for a person skilled in the art, the present invention can have various changes and variation.Within the spirit and principles in the present invention all, any modification of being done, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (9)
1. a scheduling virtual machine decision system is characterized in that, comprising:
Template management module is used to select the policy template according to the user's request customization;
Policy management module is used for creating and activation strategy by the parameter of described policy template is set;
The scheduling decision computing module is used to utilize the strategy that is in state of activation, and current resource information and cluster service data are carried out exhaustive computing, obtains optimizing decision.
2. system according to claim 1 is characterized in that:
Described policy management module also is used for described strategy is made amendment, hangs up, deleted, or the priority adjustment.
3. system according to claim 1 is characterized in that, also comprises:
The cluster monitoring data interface module is used to obtain described cluster service data, and described cluster service data is converted into the input data of regulation engine;
The resource management interface module is used to obtain described current resource information, and described current resource information is converted into the input data of regulation engine;
Described scheduling decision computing module is used to utilize the strategy that is in state of activation, adopts regulation engine that described cluster service data and current resource information are carried out exhaustive computing, obtains optimizing decision.
4. system according to claim 3 is characterized in that: described scheduling decision computing module further comprises:
The scheduling strategy submodule is used to preserve the scheduling strategy of cluster, comprises activation strategy district and expiration policy district, and described activation strategy district is used to preserve activation strategy, and described expiration policy district is used to preserve expiration policy;
The regulation engine submodule is used for cluster service data, current resource information are utilized regulation engine and described activation strategy, obtains optimizing decision.
5. according to each described system among the claim 1-4, it is characterized in that:
Described policy template is the procedure script according to the user's request customization, and has the parameter that one or more users of need are provided with.
6. a virtual platform that comprises each scheduling virtual machine decision system among the claim 1-5 is characterized in that, also comprises:
Cluster monitoring system is used to monitor the service data of cluster, and described cluster service data is sent to described scheduling virtual machine decision system;
Resource management system is used to manage current resource information, and described current resource information is sent to described scheduling virtual machine decision system;
Dispatching Control System is used to receive the optimizing decision of described scheduling virtual machine decision system, and carries out described optimizing decision.
7. a scheduling virtual machine decision-making technique is characterized in that, comprising:
Selection is according to the policy template of user's request customization;
By the parameter in the described policy template is set, create and activation strategy;
Utilization is in the strategy of state of activation, and current resource information and cluster service data are carried out exhaustive computing, obtains optimizing decision.
8. method according to claim 7 is characterized in that, also comprises:
Described strategy is made amendment, hangs up, deleted, or the priority adjustment.
9. method according to claim 7 is characterized in that:
Described carrying out also comprises before the exhaustive computing: obtain described cluster service data, and described cluster service data is converted into the input data of regulation engine;
Described carrying out also comprises before the exhaustive computing: obtain described current resource information, and described current resource information is converted into the input data of regulation engine;
The described optimizing decision of obtaining specifically comprises, utilizes the strategy that is in state of activation, adopts regulation engine that described cluster service data and current resource information are carried out exhaustive computing, obtains optimizing decision.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101937884A CN102262567A (en) | 2010-05-24 | 2010-05-24 | Virtual machine scheduling decision system, platform and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101937884A CN102262567A (en) | 2010-05-24 | 2010-05-24 | Virtual machine scheduling decision system, platform and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102262567A true CN102262567A (en) | 2011-11-30 |
Family
ID=45009204
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010101937884A Pending CN102262567A (en) | 2010-05-24 | 2010-05-24 | Virtual machine scheduling decision system, platform and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102262567A (en) |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102609309A (en) * | 2012-01-19 | 2012-07-25 | 中兴通讯股份有限公司 | Strategy scheduling system for cloud computing and strategy scheduling method for cloud computing |
| CN102662754A (en) * | 2012-04-20 | 2012-09-12 | 浙江大学 | Multi-field supportable virtual machine dispatching device and multi-field supportable virtual machine dispatching method |
| CN102722413A (en) * | 2012-05-16 | 2012-10-10 | 上海兆民云计算科技有限公司 | Distributed resource scheduling method for desktop cloud cluster |
| CN102915248A (en) * | 2012-09-28 | 2013-02-06 | 用友软件股份有限公司 | Automatic deploying system and automatic deploying method of application system |
| CN103064743A (en) * | 2012-12-27 | 2013-04-24 | 深圳先进技术研究院 | Resource scheduling method for multiple robots and resource scheduling system thereof |
| CN103186410A (en) * | 2011-12-28 | 2013-07-03 | 英业达股份有限公司 | Virtual machine configuration method and server system using same |
| CN103186411A (en) * | 2011-12-28 | 2013-07-03 | 英业达股份有限公司 | Virtual machine configuration method and server system using same |
| WO2013190562A1 (en) * | 2012-06-22 | 2013-12-27 | Hewlett-Packard Development Company, L.P. | Optimal assignment of virtual machines and virtual disks using multiary tree |
| CN103607305A (en) * | 2013-11-26 | 2014-02-26 | 北京华胜天成科技股份有限公司 | Distributed network strategy implementation method and device |
| CN103780688A (en) * | 2014-01-16 | 2014-05-07 | 博元森禾信息科技（北京）有限公司 | Migration method and device |
| CN104239159A (en) * | 2013-06-11 | 2014-12-24 | 鸿富锦精密工业（深圳）有限公司 | Virtual machine maintenance system and method |
| CN105589734A (en) * | 2015-12-15 | 2016-05-18 | 国云科技股份有限公司 | Method for creating applications through self-defined templates |
| CN106255957A (en) * | 2014-04-30 | 2016-12-21 | 瑞典爱立信有限公司 | The distribution of cloud computing resources |
| CN106789186A (en) * | 2016-12-02 | 2017-05-31 | 山东中创软件商用中间件股份有限公司 | A kind of regulation management method and device, UMP monitoring systems and monitoring method |
| CN107562502A (en) * | 2017-09-06 | 2018-01-09 | 郑州云海信息技术有限公司 | A kind of creation method of elastic telescopic device and elastic telescopic strategy |
| CN108833528A (en) * | 2018-06-11 | 2018-11-16 | 郑州云海信息技术有限公司 | A kind of cloud platform colony dispatching method and apparatus |
| CN109002348A (en) * | 2018-07-26 | 2018-12-14 | 郑州云海信息技术有限公司 | Load-balancing method and device in a kind of virtualization system |
| CN104899026B (en) * | 2015-05-14 | 2019-08-30 | 青岛环智信息科技有限公司 | A kind of general cloud application system O& middleware of commercial value driving |
| CN111614746A (en) * | 2020-05-15 | 2020-09-01 | 北京金山云网络技术有限公司 | Load balancing method and device of cloud host cluster and server |
| CN112346815A (en) * | 2019-08-08 | 2021-02-09 | 中移(苏州)软件技术有限公司 | Virtual machine processing method and device and computer readable storage medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101458631A (en) * | 2007-12-14 | 2009-06-17 | 联想(北京)有限公司 | Method for scheduling self-adapting virtual machine and computer |
-
2010
- 2010-05-24 CN CN2010101937884A patent/CN102262567A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101458631A (en) * | 2007-12-14 | 2009-06-17 | 联想(北京)有限公司 | Method for scheduling self-adapting virtual machine and computer |
Non-Patent Citations (3)
| Title |
|---|
| 《中国优秀硕士学位论文全文数据库》 20100515 袁金艳 多虚拟机快速部署机制的研究 第14、15、26、28页 5，6 , 第5期 * |
| 欧阳星明等: "虚拟机的可定制生成及其动态优化", 《计算机工程与科学》 * |
| 袁金艳: "多虚拟机快速部署机制的研究", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (31)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103186410B (en) * | 2011-12-28 | 2016-08-17 | 英业达股份有限公司 | The collocation method of virtual machine and the server system using the method |
| CN103186410A (en) * | 2011-12-28 | 2013-07-03 | 英业达股份有限公司 | Virtual machine configuration method and server system using same |
| CN103186411A (en) * | 2011-12-28 | 2013-07-03 | 英业达股份有限公司 | Virtual machine configuration method and server system using same |
| WO2013107141A1 (en) * | 2012-01-19 | 2013-07-25 | 中兴通讯股份有限公司 | Policy scheduling system and method for cloud computing |
| CN102609309A (en) * | 2012-01-19 | 2012-07-25 | 中兴通讯股份有限公司 | Strategy scheduling system for cloud computing and strategy scheduling method for cloud computing |
| CN102662754A (en) * | 2012-04-20 | 2012-09-12 | 浙江大学 | Multi-field supportable virtual machine dispatching device and multi-field supportable virtual machine dispatching method |
| CN102722413A (en) * | 2012-05-16 | 2012-10-10 | 上海兆民云计算科技有限公司 | Distributed resource scheduling method for desktop cloud cluster |
| CN102722413B (en) * | 2012-05-16 | 2017-12-29 | 上海兆民云计算科技有限公司 | The distributed resource scheduling method that a kind of desktop cloud cluster uses |
| CN104604187A (en) * | 2012-06-22 | 2015-05-06 | 惠普发展公司，有限责任合伙企业 | Optimal assignment of virtual machines and virtual disks using multiary tree |
| WO2013190562A1 (en) * | 2012-06-22 | 2013-12-27 | Hewlett-Packard Development Company, L.P. | Optimal assignment of virtual machines and virtual disks using multiary tree |
| CN102915248B (en) * | 2012-09-28 | 2016-07-06 | 用友网络科技股份有限公司 | The automatic deployment system of application system and automatic deployment method |
| CN102915248A (en) * | 2012-09-28 | 2013-02-06 | 用友软件股份有限公司 | Automatic deploying system and automatic deploying method of application system |
| CN103064743A (en) * | 2012-12-27 | 2013-04-24 | 深圳先进技术研究院 | Resource scheduling method for multiple robots and resource scheduling system thereof |
| CN103064743B (en) * | 2012-12-27 | 2017-09-26 | 深圳先进技术研究院 | A kind of resource regulating method and its resource scheduling system for multirobot |
| CN104239159A (en) * | 2013-06-11 | 2014-12-24 | 鸿富锦精密工业（深圳）有限公司 | Virtual machine maintenance system and method |
| CN103607305A (en) * | 2013-11-26 | 2014-02-26 | 北京华胜天成科技股份有限公司 | Distributed network strategy implementation method and device |
| CN103607305B (en) * | 2013-11-26 | 2017-03-15 | 北京华胜天成科技股份有限公司 | A kind of distributed network strategy implementation method and device |
| CN103780688A (en) * | 2014-01-16 | 2014-05-07 | 博元森禾信息科技（北京）有限公司 | Migration method and device |
| CN106255957A (en) * | 2014-04-30 | 2016-12-21 | 瑞典爱立信有限公司 | The distribution of cloud computing resources |
| CN104899026B (en) * | 2015-05-14 | 2019-08-30 | 青岛环智信息科技有限公司 | A kind of general cloud application system O& middleware of commercial value driving |
| CN105589734B (en) * | 2015-12-15 | 2019-03-22 | 国云科技股份有限公司 | A kind of method of self-defined template creation application |
| CN105589734A (en) * | 2015-12-15 | 2016-05-18 | 国云科技股份有限公司 | Method for creating applications through self-defined templates |
| CN106789186A (en) * | 2016-12-02 | 2017-05-31 | 山东中创软件商用中间件股份有限公司 | A kind of regulation management method and device, UMP monitoring systems and monitoring method |
| CN107562502A (en) * | 2017-09-06 | 2018-01-09 | 郑州云海信息技术有限公司 | A kind of creation method of elastic telescopic device and elastic telescopic strategy |
| CN108833528A (en) * | 2018-06-11 | 2018-11-16 | 郑州云海信息技术有限公司 | A kind of cloud platform colony dispatching method and apparatus |
| CN109002348A (en) * | 2018-07-26 | 2018-12-14 | 郑州云海信息技术有限公司 | Load-balancing method and device in a kind of virtualization system |
| CN109002348B (en) * | 2018-07-26 | 2021-03-19 | 郑州云海信息技术有限公司 | Load balancing method and device in virtualization system |
| CN112346815A (en) * | 2019-08-08 | 2021-02-09 | 中移(苏州)软件技术有限公司 | Virtual machine processing method and device and computer readable storage medium |
| CN112346815B (en) * | 2019-08-08 | 2023-04-14 | 中移(苏州)软件技术有限公司 | Virtual machine processing method and device and computer readable storage medium |
| CN111614746A (en) * | 2020-05-15 | 2020-09-01 | 北京金山云网络技术有限公司 | Load balancing method and device of cloud host cluster and server |
| CN111614746B (en) * | 2020-05-15 | 2022-03-22 | 北京金山云网络技术有限公司 | Load balancing method and device of cloud host cluster and server |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102262567A (en) | Virtual machine scheduling decision system, platform and method | |
| CN102546700B (en) | Resource scheduling and resource migration methods and equipment | |
| CN103248659B (en) | A kind of cloud computing resource scheduling method and system | |
| CN104657221B (en) | The more queue flood peak staggered regulation models and method of task based access control classification in a kind of cloud computing | |
| Zhao et al. | A new energy-aware task scheduling method for data-intensive applications in the cloud | |
| US8694679B2 (en) | Control device, method and program for deploying virtual machine | |
| CN103559072B (en) | Virtual machine two-way automatic telescopic service implementing method and system thereof | |
| CN101938416B (en) | Cloud computing resource scheduling method based on dynamic reconfiguration virtual resources | |
| CN101706743B (en) | Dispatching method of virtual machine under multi-core environment | |
| CN103885831B (en) | The system of selection of virtual machine host machine and device | |
| CN103179048B (en) | Main frame qos policy transform method and the system of cloud data center | |
| CN103856512A (en) | Cloud computing management server, working host, idle host and resource scheduling method | |
| CN104468407A (en) | Method and device for performing service platform resource elastic allocation | |
| CN102117226A (en) | Resource dispatching system and resource dispatching method | |
| CN105893113A (en) | Management system and management method of virtual machine | |
| Salimian et al. | Survey of energy efficient data centers in cloud computing | |
| CN108023958A (en) | A kind of resource scheduling system based on cloud platform resource monitoring | |
| CN107203255A (en) | Power-economizing method and device are migrated in a kind of network function virtualized environment | |
| CN103491151A (en) | Method and device for dispatching cloud computing resources and cloud computing platform | |
| CN103235744A (en) | Application resource management system for smart TV (television) | |
| CN106909462A (en) | A kind of cloud resource regulating method and device | |
| CN112559122A (en) | Virtualization instance management and control method and system based on electric power special security and protection equipment | |
| Lin et al. | Novel resource allocation model and algorithms for cloud computing | |
| Najm et al. | Towards cost-aware VM migration to maximize the profit in federated clouds | |
| CN107203256A (en) | Energy-conservation distribution method and device under a kind of network function virtualization scene |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20111130 |