CN104881325B - A kind of resource regulating method and resource scheduling system - Google Patents
A kind of resource regulating method and resource scheduling system Download PDFInfo
- Publication number
- CN104881325B CN104881325B CN201510225160.0A CN201510225160A CN104881325B CN 104881325 B CN104881325 B CN 104881325B CN 201510225160 A CN201510225160 A CN 201510225160A CN 104881325 B CN104881325 B CN 104881325B
- Authority
- CN
- China
- Prior art keywords
- physical node
- virtual machine
- resource
- node
- physical
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 30
- 230000001105 regulatory effect Effects 0.000 title claims abstract description 11
- 238000004891 communication Methods 0.000 claims description 30
- 230000006870 function Effects 0.000 claims description 17
- 238000012544 monitoring process Methods 0.000 claims description 17
- 230000005540 biological transmission Effects 0.000 claims description 5
- 238000007726 management method Methods 0.000 abstract description 68
- 238000013468 resource allocation Methods 0.000 abstract description 11
- 238000004364 calculation method Methods 0.000 abstract description 2
- 239000011800 void material Substances 0.000 description 8
- 238000013508 migration Methods 0.000 description 5
- 230000005012 migration Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000013480 data collection Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 238000007596 consolidation process Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Abstract
The embodiment of the present invention provides a kind of resource regulating method and resource scheduling system, is related to field of cloud calculation, can solve the problems, such as that resource allocation caused by dispatching technique is incompatible is unreasonable.The system includes management node and physical node, management node each application program corresponding resource collection when reaching the target of scheduling strategy according to the acquisition of the scheduling strategy of user, and resource collection is sent to each physical node, physical node determines the least resource amount of each virtual machine distribution on physical node according to resource collection, surplus yield is proportionately distributed to each virtual machine on physical node again, the total resources that each virtual machine obtains after distribution is sent to management node again, management node obtains the resource utilization of physical node according to the current resource usage amount of each virtual machine on the corresponding total resources of each virtual machine and physical node, and then the resource load of each physical node is dispatched according to resource utilization.The embodiment of the present invention is dispatched for virtualization resource.
Description
Technical field
The present invention relates to field of cloud calculation more particularly to a kind of resource regulating methods and resource scheduling system.
Background technology
The scheduling of infrastructure resources in data center can be divided into three levels, including:The resource tune of application layer
The scheduling of resource of degree, the scheduling of resource of virtual machine layer and physical machine layer.
Wherein, the First Input First Output for being realized in the software architecture MapReduce of large-scale dataset concurrent operation
(First Input First Output, FIFO), fair scheduling algorithm and computer capacity dispatching algorithm etc. belong to application layer
Resource scheduling algorithm, target are to optimize the task throughput of application layer or minimize task completion time;Virtual machine
What the dispatching algorithm of layer represented can be the credit scheduler algorithms in Xen systems, and this kind of algorithm is built in virtualization
The shared behavior of the different virtual machine in separate unit physical machine is adjusted in internal system, and target is right between ensureing virtual machine
The fair use of competitive resource;The dispatching algorithm of physical layer can be divided into static server integration and virtual machine live migration technology.It is quiet
State Server Consolidation is the performance of the demand combination physical server to resource according to application layer, finds fixed mapping mode
The mapping for establishing application program and physical server and virtual machine needs human intervention that will load when larger fluctuation occurs in load
On virtual machine (vm) migration to the server of light load on excessive physical server, and virtual machine live migration technology is then basis
The service condition of resources of virtual machine dynamically adjusts distribution of the virtual machine in physical machine.
The scheduling of resource that above-mentioned three kinds of algorithms are directed to some characteristics that resources of virtual machine is controlled and dispatched produces actively
Effect, still, due to each algorithm have certain limitation, various resources controls and dispatching technique are due to the mesh of itself
The characteristics such as mark and range, there may be restrictions or incompatible situation with other dispatching techniques, it is difficult to be integrated.For example, object
When the dispatching algorithm of reason machine layer needs to reach energy-efficient purpose according to the service condition of resources of virtual machine, it is dynamically adjusted virtual
Distribution of the machine in physical machine, such as the quantity of virtual machine is reduced, but if the resource scheduling algorithm needs of application layer reach
Task completion time is minimized, and needs sufficient virtual machine to complete, the restriction of two kinds of dispatching techniques is thus caused, makes
Dispatching technique it is incompatible, resource allocation is unreasonable.
Invention content
A kind of resource regulating method of offer of the embodiment of the present invention and resource scheduling system, it is incompatible can to solve dispatching technique
Caused by the unreasonable problem of resource allocation.
In a first aspect, a kind of resource scheduling system is provided, including management node and physical node, wherein:
Management node each applies journey when for reaching the target of the scheduling strategy according to the acquisition of the scheduling strategy of user
The corresponding resource collection of sequence, the resource collection include when any application program is run on virtual machine each virtual machine occupy
Stock number, and the resource collection is sent to each physical node;
The physical node, for determining that each virtual machine distribution is most on the physical node according to the resource collection
Small stock number, then the surplus yield for distributing the physical node after the least resource amount is proportionately distributed to the physics
Each virtual machine on node;
The physical node is additionally operable to the total resources for obtaining each virtual machine after distribution on physical node hair
Give the management node;
The management node, be additionally operable to according on the physical node the corresponding total resources of each virtual machine with it is described
The current resource usage amount of each virtual machine on physical node obtains the resource utilization of the physical node, and then according to institute
State the resource load that resource utilization dispatches each physical node.
With reference to first aspect, in the first mode in the cards of first aspect, the management node is specifically used for:
It is each applied when reaching the target of the scheduling strategy according to the acquisition of the utility function of the scheduling strategy of the user
The corresponding resource collection of program;
Wherein, the utility function includes:
With reference to first aspect or the first mode in the cards of first aspect, second in first aspect are possible real
In existing mode, the physical node is specifically used for:
The void is run on according to any application program described in the corresponding resource collection of each application program
The stock number that each virtual machine occupies when on quasi- machine determines the minimum that each virtual machine belonged on the physical node should distribute
Stock number, the least resource amount that any virtual machine on the physical node occupies are that any virtual machine is each answered described
With the sum of corresponding stock number in program;
The surplus yield of the physical node is obtained, the surplus yield is that the total resources of the physical node subtracts
Go the least resource amount determined for each virtual machine on the physical node;
The surplus yield is proportionately distributed to each virtual machine on the physical node, on the physical node
The surplus yield that is assigned to of any virtual machine be that the least resource amount of any virtual machine accounts for the physical node
The product of the ratio of allocated stock number and the surplus yield, the allocated stock number are on the physical node
The sum of the least resource amount of each virtual machine.
Second of mode in the cards with reference to first aspect, in the third mode in the cards of first aspect
In, the management node is specifically used for:
According to formulaThe resource utilization of the physical node is obtained,It indicates i-th
The resource utilization of physical node,Indicate the corresponding total resources of j-th of virtual machine on i-th of physical node,Indicate the corresponding total resources of i-th of physical node,Indicate j-th of void on i-th of physical node
The current resource usage amount of quasi- machine;
If the management node needs the load of balanced each physical node, each physical node is obtained
When the variance minimum of resource utilization, the resources of virtual machine distribution on each physical node;
If the management node needs to reduce the number of the physical node, first according to the money of each physical node
The sequence of source total amount from big to small arranges each physical node, then successively distributes virtual machine to physical node
On, until the total resources of the physical node currently distributed has been assigned and has been allocated again to next physical node;
It determines after being respectively assigned to all virtual machines on physical node, if there are the physics sections of unallocated virtual machine
Point, and if it exists, then close the physical node of the unallocated virtual machine.
Second aspect provides a kind of resource regulating method, including:
Management node obtains each application program pair when reaching the target of the scheduling strategy according to the scheduling strategy of user
The resource collection answered, the resource collection include the resource that each virtual machine occupies when any application program is run on virtual machine
Amount, and the resource collection is sent to each physical node;
The physical node determines the most petty bourgeoisie of each virtual machine distribution on the physical node according to the resource collection
Source is measured, then the surplus yield for distributing the physical node after the least resource amount is proportionately distributed to the physical node
On each virtual machine;
The total resources that each virtual machine after distribution on the physical node obtains is sent to institute by the physical node
State management node;
The management node is according to the corresponding total resources of each virtual machine and the physics section on the physical node
The current resource usage amount of each virtual machine on point obtains the resource utilization of the physical node, and then according to the resource
Utilization rate dispatches the resource load of each physical node.
In conjunction with second aspect, in the first mode in the cards of second aspect, the management node is according to user
Scheduling strategy acquisition when reaching the target of the scheduling strategy the corresponding resource collection of each application program include:
The management node obtains the mesh for reaching the scheduling strategy according to the utility function of the scheduling strategy of the user
Each application program corresponding resource collection when mark;
Wherein, the utility function includes:
In conjunction with the first of second aspect or second aspect mode in the cards, second in second aspect may be real
In existing mode, the physical node determines the minimum that each virtual machine distributes on the physical node according to the resource collection
Stock number, then the surplus yield for distributing the physical node after the least resource amount is proportionately distributed to the physics section
Each virtual machine on point includes:
The physical node is according to any application program described in the corresponding resource collection of each application program
The stock number that each virtual machine occupies when running on the virtual machine determines each virtual machine belonged on the physical node
The least resource amount that should be distributed, the least resource amount that any virtual machine on the physical node occupies are any virtual machine
The sum of corresponding stock number in each application program;
The physical node obtains the surplus yield of the physical node, and the surplus yield is the physical node
Total resources subtract on the physical node each virtual machine determine the least resource amount;
The surplus yield is proportionately distributed to each virtual machine on the physical node, institute by the physical node
The least resource amount that the surplus yield that any virtual machine on physical node is assigned to is any virtual machine is stated to account for
The product of the ratio and the surplus yield of the allocated stock number of physical node, the allocated stock number are institute
State the sum of the least resource amount of each virtual machine on physical node.
In conjunction with second of mode in the cards of second aspect, in the third mode in the cards of second aspect
In, it is virtual each of on the corresponding total resources of each virtual machine according on the physical node and the physical node
The current resource usage amount of machine obtains the resource utilization of the physical node, and then each according to resource utilization scheduling
The resource load of physical node includes:
According to formulaThe resource utilization of the physical node is obtained,It indicates i-th
The resource utilization of physical node,Indicate the corresponding total resources of j-th of virtual machine on i-th of physical node,Indicate the corresponding total resources of i-th of physical node,Indicate j-th of void on i-th of physical node
The current resource usage amount of quasi- machine;
If the management node needs the load of balanced each physical node, each physical node is obtained
When the variance minimum of resource utilization, the resources of virtual machine distribution on each physical node;
If the management node needs to reduce the number of the physical node, first according to the money of each physical node
The sequence of source total amount from big to small arranges each physical node, then successively distributes virtual machine to physical node
On, until the total resources of the physical node currently distributed has been assigned and has been allocated again to next physical node;
It determines after being respectively assigned to all virtual machines on physical node, if there are the physics sections of unallocated virtual machine
Point, and if it exists, then close the physical node of the unallocated virtual machine.
The third aspect, provides a kind of management node, including virtualization scheduler module, data memory module, communication module and
Data monitoring module, wherein:
The virtualization scheduler module, when for reaching the target of the scheduling strategy according to the acquisition of the scheduling strategy of user
The corresponding resource collection of each application program, the resource collection includes that any application program is each empty when running on virtual machine
The stock number that quasi- machine occupies;
The data memory module, for storing the corresponding resource collection of each application program;
The communication module, the resource collection for storing the data memory module are sent to each physics section
Point;
The communication module is additionally operable to receive the total resources that each virtual machine of the physical node obtains;
The data monitoring module, it is virtual each of on the physical node for being received according to the communication module
The corresponding total resources of machine obtains the physical node with the current resource usage amount of each virtual machine on the physical node
Resource utilization;
The virtualization scheduler module is additionally operable to the resource utilization scheduling obtained according to the data monitoring module
The resource load of each physical node.
Fourth aspect provides a kind of physical node, including communication module and virtualization scheduler module, wherein:
The communication module, the resource collection of each application program for receiving management node transmission, the resource set
Conjunction includes the stock number that each virtual machine occupies when any application program is run on virtual machine;
The virtualization scheduler module, for determining each virtual machine point on the physical node according to the resource collection
The least resource amount matched, then the surplus yield for distributing the physical node after the least resource amount is proportionately distributed to institute
State each virtual machine on physical node;
The communication module is additionally operable to the total resources for obtaining each virtual machine after distribution on physical node hair
The management node is given, so that the management node dispatches the resource load of each physical node according to the total resources.
A kind of resource regulating method of offer of the embodiment of the present invention and resource scheduling system, the system include management node and object
Manage node, management node each application program corresponding money when reaching the target of scheduling strategy according to the acquisition of the scheduling strategy of user
Source is gathered, and resource collection includes the stock number that each virtual machine occupies when any application program is run on virtual machine, and will money
Source set is sent to each physical node, and physical node determines that each virtual machine distribution is most on physical node according to resource collection
Each of small stock number, then the surplus yield for distributing physical node after least resource amount is proportionately distributed on physical node
The total resources that each virtual machine on physical node after distribution obtains is sent to management node by virtual machine, physical node, pipe
It is according to each virtual machine on the corresponding total resources of each virtual machine and the physical node on physical node current to manage node
Resource usage amount obtains the resource utilization of physical node, and then the resource for dispatching according to resource utilization each physical node is negative
It carries, in this way, by the resource scheduling system, the scheduling of application layer can be paid the utmost attention to, on the basis for the resource requirement for ensureing user
On, resource is scheduled from the scheduling of virtualization layer, physical layer, it is this from top and under modeling scheme, based on application layer, void
Quasi- machine and physics section and resource usage amount, unified performance prison can be carried out to the physical node in system, virtual machine node
Control and scheduling, to more reasonably distributing resource from the overall situation, can solve dispatching technique it is incompatible caused by resource allocation
Unreasonable problem.
Description of the drawings
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be in embodiment or description of the prior art
Required attached drawing is briefly described, it should be apparent that, the accompanying drawings in the following description is only some realities of the present invention
Example is applied, it for those of ordinary skill in the art, without creative efforts, can also be according to these attached drawings
Obtain other attached drawings.
Fig. 1 is a kind of configuration diagram of resource scheduling system provided in an embodiment of the present invention;
Fig. 2 is a kind of flow diagram of resource regulating method provided in an embodiment of the present invention;
Fig. 3 is a kind of structural schematic diagram of management node provided in an embodiment of the present invention;
Fig. 4 is a kind of structural schematic diagram of physical node provided in an embodiment of the present invention.
Specific implementation mode
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation describes, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art are obtained every other without creative efforts
Embodiment shall fall within the protection scope of the present invention.
In order to which the resource scheduling in virtualization system is described on the whole, the present invention proposes a kind of concentration
The resource scheduling system framework of formula, and based on this framework, it is proposed that a kind of top-down resource scheduling system, to resource
Scheduling problem is modeled, and proposes corresponding method.
In embodiments of the present invention, as shown in Figure 1, the resource scheduling system framework of above-mentioned centralization is segmented into two
Point:The ends Sensor (alternatively referred to as slave unit) and the ends Master (alternatively referred to as master control set), it is corresponding to be present in the present invention in fact
The physical node and management node in example are applied, the two constitutes the centralized architecture centered on the ends Master.The ends Master with
The ends Sensor carry out the transmission of data flow and control stream by the physical subnets of virtual cluster, common to complete to virtual cluster
Monitoring and scheduling of resource.
Wherein, the ends Sensor are located inside each physical node, run the privileged domain of domain physical node virtualization system
Dom0, responsible content may include:Monitor and collect the resource using information of other virtual machines on same physical node;According to
The information of collection adjusts the scheduling strategy inside the physical node in real time;Scheduling result and the information of collection are reported in time
Give the ends Master.The information for the scheduling result and collection that the ends Master are just reported according to the ends Sensor forms the monitoring number of cluster
According to, and then the resource allocation conditions of entire cluster are scheduled, and scheduling strategy is transferred to the ends Sensor and the ends Master
The other virtual machines of itself.
Based on above-mentioned framework, the physical node that is related to first to the resource scheduling system of the embodiment of the present invention, is answered at virtual machine
It is illustrated with layer.
For the physical node of physical layer, it is denoted as in embodiments of the present invention:
S_Node={ node1,node2,…,nodei,…,noden, it is characterized in infrastructure layer, for virtual machine
The set of the n platform physical nodes used.For each nodeiFor, there are two values:Total resources RnodeiIt is used with resource
Measure Lnodei, node is characterized respectivelyiThe ratio that upper assignable total resources and Current resource have occupied.
For the virtual machine of virtualization layer, in embodiments of the present invention, all virtual machines can be by gathering in system
S_VM is indicated, is denoted as:S_VM={ VM1,VM2,…,VMj,…,VMm}.For each VMjFor, there are two values:Resource
Total amount RVMiWith resource usage amount LVMi, VM is characterized respectivelyiThe ratio that upper assignable total resources and Current resource have occupied.
For application layer, in embodiments of the present invention, the application run on all virtual machines can be denoted as:S_APP=
{app1,app2,…,appi,…,apps, to characterize the application program of cloud computing user operation on a virtual machine.The present invention is real
Example is applied in order to describe to facilitate, runs an application program on a virtual machine, but an application program can run on it is multiple
On virtual machine.Virtual machine is assigned to this relationship of application program, and the embodiment of the present invention can be denoted as.So having:
So thatSet up
Based on above description, the embodiment of the present invention is specifically described below.
The embodiment of the present invention provides a kind of resource scheduling system, and with reference to system architecture as shown in Figure 1, which includes pipe
Node and physical node are managed, wherein:
Management node, each application program pair when reaching the target of scheduling strategy for being obtained according to the scheduling strategy of user
The resource collection answered, resource collection include the stock number that each virtual machine occupies when any application program is run on virtual machine,
And resource collection is sent to each physical node;
Physical node, the least resource amount for determining each virtual machine distribution on physical node according to resource collection, then
The surplus yield of physical node after distribution least resource amount is proportionately distributed to each virtual machine on physical node;
Physical node is additionally operable to the total resources that each virtual machine after distribution on physical node obtains being sent to management
Node;
Management node is additionally operable to according on the corresponding total resources of each virtual machine and physical node on physical node
The current resource usage amount of each virtual machine obtains the resource utilization of physical node, and then each according to resource utilization scheduling
The resource load of physical node.
Specifically, in application-level scheduling technology, management node is reaching scheduling according to the acquisition of the scheduling strategy of user
It can be specifically used for when the corresponding resource collection of each application program when the target of strategy:According to the effectiveness of the scheduling strategy of user
Function acquisition reaches the corresponding resource collection of each application program when the target of scheduling strategy.
Wherein, which may include:
f_app(appk) indicate the target of scheduling strategy, target be user buy resources of virtual machine expense it is minimum or empty
The time that quasi- machine completes task is most short, costtT-th of virtual machine VM of k-th of application program is distributed in expressiontUnit resource
Price in unit interval, RtIndicate the stock number that t-th of virtual machine for distributing to k-th of application program occupies, f_time (Rt,
missiont) indicate that t-th of virtual machine completes distributed task amount missiontWhen occupied time.
This is because, in the application of cloud computing, user needs to buy certain resources of virtual machine by paying, therefore,
In order to complete the operation submitted, user needs to weigh between resource and expense.The virtual machine of user's application is more, then resource is got over
More, the time to fulfil assignment is shorter, and certainly total expense can also improve namely how resource is chosen, and depends on the scheduling of user
Strategy.
For example, when user wishes least cost, the target of scheduling strategy can be expressed as:For each VMt, determine
Go out RtSo that f_app (appk,Rt‘)=min (f_app (appk,Rt))。
If the user desired that the time of the task of completion is most short, the target of scheduling strategy can be expressed as:For each
VMt, determine RtSo that:
min(f_time(Rt',minssiont'))=min (min (f_time (Rt,minssiont)))
Certainly, no matter the scheduling strategy of user's final choice is which kind of target be based on, it can remember that user is appkSelection
Resource collectionFor:Wherein, skIt is expressed as appkDetermining skA virtual machine,It indicates
Distribute to the s of k-th of application programkThe stock number that a virtual machine occupies.Namely resource collection includes any application program fortune
The row stock number that each virtual machine occupies when on virtual machine.
Here resource can be understood as distributing to the cpu resource of virtual machine, memory, network bandwidth and disk size etc..
After application layer is completed to the resource allocation of application program, management node needs to determine for each application program
Resource collection is sent to each physical node, in order to which physical node determines each virtual machine according to the resource collection received
Total resources, namely carry out the scheduling of resource of virtualization layer.
First, physical node can be determined first according to resource collection on physical node after receiving resource collection
The least resource amount of each virtual machine distribution, then in proportion by the surplus yield of the physical node after distribution least resource amount
Distribute to each virtual machine on the physical node.
Specifically, physical node can be run on according to any application program in the corresponding resource collection of each application program
The stock number that each virtual machine occupies when on virtual machine determines the minimum that each virtual machine belonged on the physical node should distribute
Stock number, the least resource amount that any virtual machine on the physical node occupies are that any virtual machine is right in each application program
The sum of stock number answered.That is, first to ensure the stock number of required virtual machine when each application program operation.
Illustratively, since the stock number on physical node is certain, it is believed that physical node is before distributing resource
Initial surplus yield R_resnodei=Rnodei,1<=j<=ni, j-th of virtual machine point on i-th of physical node
The stock number being fitted onThen since j=1, for the stock number that the distribution of each virtual machine is minimum.That is, willIt is set toWherein:
And Indicate appkIn the jth of i-th of physical node
The stock number being assigned on a virtual machine, while updating surplus yield R_resnodeiFor:
Until j=ni。
After being assigned with least resource amount for each virtual machine on physical node, physical node can obtain its residue
Stock number, the surplus yield are just that the total resources of physical node subtracts as each virtual machine determination on physical node most
Small stock number.
Then, each virtual machine that surplus yield can be proportionately distributed on physical node, on physical node
The surplus yield that any virtual machine is assigned to is that the least resource amount of any virtual machine accounts for the allocated stock number of physical node
Ratio and surplus yield product, allocated stock number be physical node on each virtual machine least resource amount it
With.
Specifically, since the least resource amount that the virtual machine on physical node is assigned to may differ, surplus resources
Amount can account for the ratio of allocated stock number to distribute to each virtual machine according to the least resource amount of virtual machine, that is, empty
Least resource amount that quasi- machine obtains it is more can continue to distribute more surplus yields, the least resource that virtual machine obtains
That measures less some can continue the few surplus yield of some of distribution, to ensure resources of virtual machine abundance.It specifically can be according to such as
Lower method distribution surplus yield R_resnodei:
For example, there are 10 cpu resources on physical node, and operation there are 2 virtual machines on the physical node, first
The least resource amount that virtual machine needs is 5 cpu resources, and the least resource amount that second virtual machine needs is 3 cpu resources,
After distributing least resource amount for each virtual machine, remaining 2 cpu resources can be first void according to the algorithm of 5/8*2
Quasi- machine continues to distribute surplus yield, continues distribution surplus yield according to the algorithm of 3/8*2 for second virtual machine, then the
The total resources that one virtual machine obtains is 5/8*2+5, and the total resources that second virtual machine obtains is 3/8*2+3.
Physical node for each virtual machine after distributing stock number, total money that physical node can obtain each virtual machine
Source amount is sent to management node, in order to which management node is scheduled physical layer according to resources of virtual machine.It specifically can basis
The corresponding total resources of each virtual machine on the physical node resource usage amount current with each virtual machine on physical node
The resource utilization of physical node is obtained, and then dispatches the resource load of each physical node according to resource utilization.
It specifically, can be according to formulaThe resource utilization of physical node is obtained,Table
Show the resource utilization of i-th of physical node,Indicate the corresponding total resources of j-th of virtual machine on i-th of physical node
Amount,Indicate the corresponding total resources of i-th of physical node,Indicate j-th of virtual machine on i-th of physical node
Current resource usage amount.
Then scheduling of resource can be carried out according to the target of physical layer scheduling, if management node needs balanced each physics section
The load of point, then when obtaining the variance minimum of the resource utilization of each physical node, the virtual machine money on each physical node
Source distribution, i.e., when the target of physical layer scheduling scheme is load balancing, utility function can be:
Map M (S_VM, S_Node) so that [Lnodei-E(Lnodei)]2It is minimum;
When the resource utilization of a certain physical node is relatively low, show the idling-resource of virtual machine on the physical node compared with
It is more, it at this time can be by the virtual machine (vm) migration on other nodes to this node;When the resource utilization of a certain physical node is higher,
Show that the idling-resource of virtual machine on the physical node is less, it at this time can be by the virtual machine (vm) migration on this node to other nodes
On, to reach the load balancing of each node.
If management node need reduce physical node number, can first according to the total resources of each physical node from
Small sequence is arrived greatly to arrange each physical node, then successively distributes virtual machine to physical node, until current
The total resources of the physical node of distribution, which has been assigned, is again allocated next physical node, then determines will be all virtual
After machine is respectively assigned on physical node, if there are the physical nodes of unallocated virtual machine, and if it exists, then closes unallocated void
The physical node of quasi- machine so can need to distribute virtual machine to physical node as few as possible, be in reducing
The physical node quantity of open state, reduces the energy consumption of physical node.
By the explanation of above-mentioned 3 levels, namely if to realize the resource dispatching strategy in virtualization system, Ke Yiyou
The scheduling scheme for first considering application layer, on the basis of ensureing the resource requirement of user, infrastructure providers are just from virtualization
Layer, physical layer angle the configuration status of resource is adjusted.Namely during entire scheduling, priority is:Using
Layer > virtualization layer > physical layers, can make scheduling system consider cloud computing user and infrastructure resources provide and
The interests of both sides, more reasonably distribute resource from the overall situation.
The embodiment of the present invention provides a kind of resource scheduling system, which includes management node and physical node, management knot
The point corresponding resource collection of each application program, resource set when reaching the target of scheduling strategy according to the acquisition of the scheduling strategy of user
Conjunction includes the stock number that each virtual machine occupies when any application program is run on virtual machine, and resource collection is sent to often
A physical node, physical node determine the least resource amount that each virtual machine distributes on physical node according to resource collection, then will
The surplus yield of physical node is proportionately distributed to each virtual machine on physical node, physics section after distribution least resource amount
The total resources that each virtual machine on physical node after distribution obtains is sent to management node by point, and management node is according to physics
The corresponding total resources of each virtual machine on node is obtained with the current resource usage amount of each virtual machine on physical node
The resource utilization of physical node, and then the resource load of each physical node is dispatched according to resource utilization, in this way, passing through this
Resource scheduling system can pay the utmost attention to the scheduling of application layer, ensure user resource requirement on the basis of, from virtualization layer,
The scheduling of physical layer is scheduled resource, it is this from top and under modeling scheme, based on application layer, virtual machine and physics section and
Resource usage amount, unified performance monitoring and scheduling can be carried out to the physical node in system, virtual machine node, to from
Resource is more reasonably distributed in the overall situation, can solve the problems, such as that resource allocation caused by dispatching technique is incompatible is unreasonable.
Further embodiment of this invention provides a kind of resource regulating method, as shown in Fig. 2, including:
201, management node obtains each application program pair when reaching the target of scheduling strategy according to the scheduling strategy of user
The resource collection answered, resource collection include the stock number that each virtual machine occupies when any application program is run on virtual machine,
And resource collection is sent to each physical node.
Specifically, management node obtains each application program when reaching the target of scheduling strategy according to the scheduling strategy of user
Corresponding resource collection may include:
Management node is each applied when reaching the target of scheduling strategy according to the acquisition of the utility function of the scheduling strategy of user
The corresponding resource collection of program;
Wherein, utility function includes:
202, physical node determines the least resource amount of each virtual machine distribution on physical node according to resource collection, then will
The surplus yield of physical node is proportionately distributed to each virtual machine on physical node after distribution least resource amount.
Specifically, physical node is run on virtually according to any application program in the corresponding resource collection of each application program
The stock number that each virtual machine occupies when on machine determines the least resource that each virtual machine belonged on physical node should distribute
It measures, the least resource amount that any virtual machine on physical node occupies is any virtual machine corresponding money in each application program
The sum of source amount, then obtains the surplus yield of physical node, surplus yield is that the total resources of physical node subtracts as object
The least resource amount that each virtual machine on node determines is managed, then surplus yield is proportionately distributed to every on physical node
A virtual machine, the surplus yield that any virtual machine on physical node is assigned to are that the least resource amount of any virtual machine accounts for object
The product of the ratio and surplus yield of the allocated stock number of node is managed, allocated stock number is each of on physical node
The sum of least resource amount of virtual machine.
203, the total resources that each virtual machine on physical node after distribution obtains is sent to management knot by physical node
Point.
204, management node according on physical node the corresponding total resources of each virtual machine with it is every on physical node
The current resource usage amount of a virtual machine obtains the resource utilization of physical node, and then dispatches each object according to resource utilization
Manage the resource load of node.
It specifically, can be according to formulaThe resource utilization of physical node is obtained,Table
Show the resource utilization of i-th of physical node,Indicate the corresponding total resources of j-th of virtual machine on i-th of physical node
Amount,Indicate the corresponding total resources of i-th of physical node,Indicate j-th of virtual machine on i-th of physical node
Current resource usage amount;
If management node needs the load of balanced each physical node, the resource utilization of each physical node is obtained
When variance minimum, the resources of virtual machine distribution on each physical node;
If management node need reduce physical node number, first according to each physical node total resources from greatly to
Small sequence arranges each physical node, then successively distributes virtual machine to physical node, until current distribution
The total resources of physical node be assigned next physical node be allocated again, and then determine by all virtual machines
After being respectively assigned on physical node, if there are the physical nodes of unallocated virtual machine, and if it exists, then close unallocated virtual
The physical node of machine.
It is real can to refer to above-mentioned resource scheduling system for management node and the specific implementation of physical node in the present embodiment
The elaboration in example is applied, is not being repeated here.
The embodiment of the present invention provides a kind of resource regulating method, is applied to resource scheduling system, which includes that management is tied
Point and physical node, management node obtain each application program pair when reaching the target of scheduling strategy according to the scheduling strategy of user
The resource collection answered, resource collection include the stock number that each virtual machine occupies when any application program is run on virtual machine,
And resource collection is sent to each physical node, physical node determines each virtual machine point on physical node according to resource collection
The least resource amount matched, then the surplus yield of physical node after distribution least resource amount is proportionately distributed on physical node
Each virtual machine, physical node by the total resources that each virtual machine on physical node after distribution obtains be sent to management knot
Point, management node are worked as according to the corresponding total resources of each virtual machine on physical node with each virtual machine on physical node
Preceding resource usage amount obtains the resource utilization of physical node, and then the money of each physical node is dispatched according to resource utilization
Source loads, in this way, by the resource scheduling system, can pay the utmost attention to the scheduling of application layer, is ensureing the resource requirement of user
On the basis of, resource is scheduled from the scheduling of virtualization layer, physical layer, it is this from top and under modeling scheme, based on application
Layer, virtual machine and physics section and resource usage amount, unified property can be carried out to the physical node in system, virtual machine node
Can monitor and dispatch, to more reasonably distributing resource from the overall situation, can solve dispatching technique it is incompatible caused by resource
The problem of unreasonable distribution.
Further embodiment of this invention provides a kind of management node 3, as shown in figure 3, the management node includes master control
(Master) device 30, master control set 30 include virtual machine monitor (Virtual Machine Manager, VMM) interface
301, interface module 302, data collection module 303, data memory module 304, data monitoring module 305, virtualization scheduling mould
Block 306, communication module 307 and alarm module 308.
Wherein, interface module 302 and communication module 307 are received for other physical nodes in collection management node and cluster
The data of collection to form the monitoring data of cluster, and are centrally stored in data memory module 304.Virtualize scheduler module 306
For according to history and real-time monitoring data, being adjusted to the resource allocation conditions of entire cluster, and adjustable strategies are led to
It crosses communication module 307 and is transferred to physical node and other virtual machines of management node itself.
In embodiments of the present invention, scheduler module 306 is virtualized, for reaching scheduling according to the acquisition of the scheduling strategy of user
Each application program corresponding resource collection when the target of strategy, resource collection includes that any application program is run on virtual machine
When the stock number that occupies of each virtual machine;
Data memory module 304, for storing the corresponding resource collection of each application program;
Communication module 307, the resource collection for storing data memory module are sent to each physical node;
Communication module 307 is additionally operable to receive the total resources that each virtual machine of physical node obtains;
Data monitoring module 305, each virtual machine on the physical node for being received according to communication module are corresponding
Total resources obtains the resource utilization of physical node with the current resource usage amount of each virtual machine on physical node;
Scheduler module 306 is virtualized, is additionally operable to dispatch each physics according to the resource utilization that data monitoring module obtains
The resource load of node.
Optionally, virtualization scheduler module 306 can be specifically used for:
Each application program corresponds to when reaching the target of scheduling strategy according to the acquisition of the utility function of the scheduling strategy of user
Resource collection;
Wherein, utility function includes:
Optionally, virtualization scheduler module 306 can be specifically used for:
According to formulaThe resource utilization of physical node is obtained,Indicate i-th of physics
The resource utilization of node,Indicate the corresponding total resources of j-th of virtual machine on i-th of physical node,It indicates
The corresponding total resources of i-th of physical node,Indicate that the current resource of j-th of virtual machine on i-th of physical node makes
Dosage;
If management node needs the load of balanced each physical node, the resource utilization of each physical node is obtained
When variance minimum, the resources of virtual machine distribution on each physical node;
If management node need reduce physical node number, first according to each physical node total resources from greatly to
Small sequence arranges each physical node, then successively distributes virtual machine to physical node, until current distribution
The total resources of physical node be assigned next physical node be allocated again;
It determines after being respectively assigned to all virtual machines on physical node, if there are the physics sections of unallocated virtual machine
Point, and if it exists, then close the physical node of unallocated virtual machine.
In the embodiment of the present invention specific implementation of each module can refer to above-mentioned resource scheduling system embodiment and
Specific descriptions in embodiment of the method, which is not described herein again.
The embodiment of the present invention provides a kind of management node, the virtualization scheduler module in the management node, for according to
The scheduling strategy acquisition at family reaches the corresponding resource collection of each application program when the target of scheduling strategy, and resource collection includes appointing
The one application program stock number that each virtual machine occupies when running on virtual machine, data memory module, for storing each answer
With the corresponding resource collection of program, communication module, the resource collection for storing data memory module is sent to each physics
Node, communication module are additionally operable to receive the total resources that each virtual machine of physical node obtains;Data monitoring module is used for
The corresponding total resources of each virtual machine on the physical node received according to communication module and each void on physical node
The current resource usage amount of quasi- machine obtains the resource utilization of physical node;Scheduler module is virtualized, is additionally operable to be supervised according to data
The resource utilization that control module obtains dispatches the resource load of each physical node, in this way, the scheduling of application layer can be paid the utmost attention to,
On the basis of ensureing the resource requirement of user, resource is scheduled from the scheduling of virtualization layer, physical layer, it is this from top and
Under modeling scheme, based on application layer, virtual machine and physics section and resource usage amount, can be to the object in resource scheduling system
Reason node, virtual machine node carry out unified performance monitoring and scheduling, can to more reasonably distribute resource from the overall situation
Solve the problems, such as that resource allocation caused by dispatching technique is incompatible is unreasonable.
Further embodiment of this invention provides a kind of physical node 4, as shown in figure 4, include subordinate (Sensor) device 40, it should
Slave unit 40 includes VMM interfaces 401, interface module 402, data collection and memory module 403, virtualization scheduler module 404
With communication module 405.
Wherein, data collection and memory module 403 can be by VMM interfaces 401 to all virtual machine and physical node
Resource be monitored, and by the data transfer being collected into scheduler module 404 to virtualization, virtualize 404 basis of scheduler module
Scheduling strategy is made whether the decision changed to current Resource Allocation Formula.Communication module 405 is responsible for VMM or pipe
Node transmission control command or data information are managed, so that management node makes comprehensive scheduling strategy, while also by physics section
Resource dispatching strategy inside point is placed in inside physical node, can simplify the interaction between the virtual machine of different physical nodes.
In embodiments of the present invention, communication module 405 can be used for receiving each application program that management node is sent
Resource collection, resource collection include the stock number that each virtual machine occupies when any application program is run on virtual machine;
Scheduler module 404 is virtualized, the minimum for determining each virtual machine distribution on physical node according to resource collection
Stock number, then the surplus yield of physical node after distribution least resource amount is proportionately distributed to each void on physical node
Quasi- machine;
Communication module 405 is additionally operable to the total resources that each virtual machine after distribution on physical node obtains being sent to
Management node, so that management node dispatches the resource load of each physical node according to total resources.
Optionally, virtualization scheduler module 404 can be specifically used for:
According to each virtual when any application program is run on virtual machine in the corresponding resource collection of each application program
The stock number that machine occupies, determines the least resource amount that each virtual machine for belonging on physical node should distribute, on physical node
The least resource amount that any virtual machine occupies is any virtual machine the sum of corresponding stock number in each application program;
The surplus yield of physical node is obtained, surplus yield is that the total resources of physical node subtracts as physical node
On each virtual machine determine least resource amount
Surplus yield is proportionately distributed to each virtual machine on physical node, any virtual machine on physical node
The surplus yield being assigned to is that the least resource amount of any virtual machine accounts for the ratio of the allocated stock number of physical node and remains
The product of remaining stock number, allocated stock number are the sum of the least resource amount of each virtual machine on physical node.
The specific implementation of each module in the embodiment of the present invention may refer to above-mentioned resource scheduling system and method is real
Being specifically described in example is applied, which is not described herein again.
The embodiment of the present invention provides a kind of physical node, communication module therein, for receiving the every of management node transmission
The resource collection of a application program, resource collection include when any application program is run on virtual machine each virtual machine occupy
Stock number;Scheduler module is virtualized, the least resource for determining each virtual machine distribution on physical node according to resource collection
Amount, then the surplus yield for distributing physical node after least resource amount each of is proportionately distributed on physical node virtual
Machine;Communication module is additionally operable to the total resources that each virtual machine after distribution on physical node obtains being sent to management node,
So that management node dispatches the resource load of each physical node according to total resources, in this way, the tune of application layer can be paid the utmost attention to
Degree, ensure user resource requirement on the basis of, resource is scheduled from the scheduling of virtualization layer, physical layer, it is this from
Top and under modeling scheme, based on application layer, virtual machine and physics section and resource usage amount, can be in resource scheduling system
Physical node, virtual machine node carry out unified performance monitoring and scheduling, to more reasonably distributing resource from the overall situation,
It can solve the problems, such as that resource allocation caused by dispatching technique is incompatible is unreasonable.
In several embodiments provided herein, it should be understood that disclosed node and method can pass through it
Its mode is realized.For example, the apparatus embodiments described above are merely exemplary, for example, the division of the unit, only
Only a kind of division of logic function, formula that in actual implementation, there may be another division manner, such as multiple units or component can be tied
Another system is closed or is desirably integrated into, or some features can be ignored or not executed.Another point, it is shown or discussed
Mutual coupling, direct-coupling or communication connection can be the INDIRECT COUPLING or logical by some interfaces, device or unit
Letter connection can be electrical, machinery or other forms.
The unit illustrated as separating component may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, you can be located at a place, or may be distributed over multiple
In network element.Some or all of unit therein can be selected according to the actual needs to realize the mesh of this embodiment scheme
's.
In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, it can also
It is that the independent physics of each unit includes, it can also be during two or more units be integrated in one unit.Above-mentioned integrated list
The form that hardware had both may be used in member is realized, can also be realized in the form of hardware adds SFU software functional unit.
The above-mentioned integrated unit being realized in the form of SFU software functional unit can be stored in one and computer-readable deposit
In storage media.Above-mentioned SFU software functional unit is stored in a storage medium, including some instructions are used so that a computer
Equipment (can be personal computer, server or the network equipment etc.) executes the portion of each embodiment the method for the present invention
Step by step.And storage medium above-mentioned includes:USB flash disk, mobile hard disk, read-only memory (Read-Only Memory, abbreviation
ROM), random access memory (Random Access Memory, abbreviation RAM), magnetic disc or CD etc. are various to store
The medium of program code.
Finally it should be noted that:The above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations;Although
Present invention has been described in detail with reference to the aforementioned embodiments, it will be understood by those of ordinary skill in the art that:It still may be used
With technical scheme described in the above embodiments is modified or equivalent replacement of some of the technical features;
And these modifications or replacements, various embodiments of the present invention technical solution that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (8)
1. a kind of resource scheduling system, which is characterized in that including management node and physical node, wherein:
Management node, each application program pair when reaching the target of the scheduling strategy for being obtained according to the scheduling strategy of user
The resource collection answered, the resource collection include the resource that each virtual machine occupies when any application program is run on virtual machine
Amount, and the resource collection is sent to each physical node;
The physical node, the most petty bourgeoisie for determining each virtual machine distribution on the physical node according to the resource collection
Source is measured, then the surplus yield for distributing the physical node after the least resource amount is proportionately distributed to the physical node
On each virtual machine;
The physical node is additionally operable to the total resources that each virtual machine after distribution on the physical node obtains being sent to
The management node;
The management node is additionally operable to according to the corresponding total resources of each virtual machine and the physics on the physical node
The current resource usage amount of each virtual machine on node obtains the resource utilization of the physical node, and then according to the money
Source utilization rate dispatches the resource load of each physical node,
The management node is specifically used for:
Each application program when reaching the target of the scheduling strategy is obtained according to the utility function of the scheduling strategy of the user
Corresponding resource collection;
Wherein, the utility function includes:
f_app(appk) indicate the scheduling strategy
Target, the target are that the user buys the expense minimum of the resources of virtual machine or the time of virtual machine completion task
It is most short, costtT-th of virtual machine VM of k-th of application program is distributed in expressiontThe unit resource unit interval in price, Rt
Indicate the stock number that t-th of virtual machine for distributing to k-th of application program occupies, f_time (Rt,missiont) indicate institute
It states t-th of virtual machine and completes distributed task amount missiontWhen occupied time.
2. system according to claim 1, which is characterized in that the physical node is specifically used for:
The virtual machine is run on according to any application program described in the corresponding resource collection of each application program
The stock number that each virtual machine occupies when upper determines the least resource that each virtual machine belonged on the physical node should distribute
It measures, the least resource amount that any virtual machine on the physical node occupies is any virtual machine described each using journey
The sum of corresponding stock number in sequence;
Obtain the surplus yield of the physical node, the surplus yield be the total resources of the physical node subtract for
The least resource amount that each virtual machine on the physical node determines;
The surplus yield is proportionately distributed to each virtual machine on the physical node, appointing on the physical node
The surplus yield that one virtual machine is assigned to accounts for the physical node for the least resource amount of any virtual machine and has divided
The product of the ratio for the stock number matched and the surplus yield, the allocated stock number are every on the physical node
The sum of least resource amount of a virtual machine.
3. system according to claim 2, which is characterized in that the management node is specifically used for:
According to formulaThe resource utilization of the physical node is obtained,Indicate i-th of physics
The resource utilization of node,Indicate the corresponding total resources of j-th of virtual machine on i-th of physical node,
Indicate the corresponding total resources of i-th of physical node,Indicate j-th of virtual machine on i-th of physical node
Current resource usage amount;
If the management node needs the load of balanced each physical node, the resource of each physical node is obtained
When the variance minimum of utilization rate, the resources of virtual machine distribution on each physical node;
If the management node needs to reduce the number of the physical node, first total according to the resource of each physical node
The sequence of amount from big to small arranges each physical node, then successively distributes virtual machine to physical node,
Until the total resources of the physical node currently distributed has been assigned and has been allocated again to next physical node;
It determines after dividing equally all virtual machines and being assigned on physical node, if there are the physical node of unallocated virtual machine,
If in the presence of the physical node of the unallocated virtual machine is closed.
4. a kind of resource regulating method, which is characterized in that including:
Each application program is corresponding when management node reaches the target of the scheduling strategy according to the acquisition of the scheduling strategy of user
Resource collection, the resource collection include the stock number that each virtual machine occupies when any application program is run on virtual machine,
And the resource collection is sent to each physical node;
The physical node determines the least resource amount of each virtual machine distribution on the physical node according to the resource collection,
The surplus yield for distributing the physical node after the least resource amount is proportionately distributed on the physical node again
Each virtual machine;
The total resources that each virtual machine after distribution on the physical node obtains is sent to the pipe by the physical node
Manage node;
The management node is according on the corresponding total resources of each virtual machine and the physical node on the physical node
The current resource usage amount of each virtual machine obtain the resource utilization of the physical node, and then used according to the resource
Rate dispatches the resource load of each physical node,
The management node obtains each application program pair when reaching the target of the scheduling strategy according to the scheduling strategy of user
The resource collection answered includes:
When the management node reaches the target of the scheduling strategy according to the acquisition of the utility function of the scheduling strategy of the user
The corresponding resource collection of each application program;
Wherein, the utility function includes:
f_app(appk) indicate the scheduling strategy
Target, the target are that the user buys the expense minimum of the resources of virtual machine or the time of virtual machine completion task
It is most short, costtT-th of virtual machine VM of k-th of application program is distributed in expressiontThe unit resource unit interval in price, Rt
Indicate the stock number that t-th of virtual machine for distributing to k-th of application program occupies, f_time (Rt,missiont) indicate institute
It states t-th of virtual machine and completes distributed task amount missiontWhen occupied time.
5. according to the method described in claim 4, it is characterized in that, the physical node is according to described in resource collection determination
The least resource amount of each virtual machine distribution on physical node, then the surplus of the physical node after the least resource amount will be distributed
Remaining stock number is proportionately distributed to each virtual machine on the physical node:
The physical node is run according to any application program described in the corresponding resource collection of each application program
The stock number that each virtual machine occupies when on the virtual machine determines that each virtual machine belonged on the physical node should divide
The least resource amount matched, the least resource amount that any virtual machine on the physical node occupies are any virtual machine in institute
State the sum of corresponding stock number in each application program;
The physical node obtains the surplus yield of the physical node, and the surplus yield is the total of the physical node
Stock number subtracts the least resource amount determined for each virtual machine on the physical node;
The surplus yield is proportionately distributed to each virtual machine on the physical node, the object by the physical node
The surplus yield that any virtual machine on reason node is assigned to is described in the least resource amount of any virtual machine accounts for
The product of the ratio and the surplus yield of the allocated stock number of physical node, the allocated stock number are the object
Manage the sum of the least resource amount of each virtual machine on node.
6. according to the method described in claim 5, it is characterized in that, each virtual machine pair according on the physical node
The total resources the answered resource usage amount current with each virtual machine on the physical node obtains the money of the physical node
Source utilization rate, and then the resource load for dispatching according to the resource utilization each physical node includes:
According to formulaThe resource utilization of the physical node is obtained,Indicate i-th of physics
The resource utilization of node,Indicate the corresponding total resources of j-th of virtual machine on i-th of physical node,
Indicate the corresponding total resources of i-th of physical node,Indicate that j-th of virtual machine on i-th of physical node is worked as
Preceding resource usage amount;
If the management node needs the load of balanced each physical node, the resource of each physical node is obtained
When the variance minimum of utilization rate, the resources of virtual machine distribution on each physical node;
If the management node needs to reduce the number of the physical node, first total according to the resource of each physical node
The sequence of amount from big to small arranges each physical node, then successively distributes virtual machine to physical node,
Until the total resources of the physical node currently distributed has been assigned and has been allocated again to next physical node;
It determines after dividing equally all virtual machines and being assigned on physical node, if there are the physical node of unallocated virtual machine,
If in the presence of the physical node of the unallocated virtual machine is closed.
7. a kind of management node, which is characterized in that including virtualization scheduler module, data memory module, communication module and data
Monitoring module, wherein:
The virtualization scheduler module, the utility function acquisition for the scheduling strategy according to user reach the scheduling strategy
Each application program corresponding resource collection when target, when the resource collection includes that any application program is run on virtual machine
The stock number that each virtual machine occupies;
The data memory module, for storing the corresponding resource collection of each application program;
Communication module, the resource collection for storing the data memory module are sent to each physical node;
The communication module is additionally operable to receive the total resources that each virtual machine of the physical node obtains;
The data monitoring module, each virtual machine pair on the physical node for being received according to the communication module
The total resources the answered resource usage amount current with each virtual machine on the physical node obtains the money of the physical node
Source utilization rate;
It is each to be additionally operable to the resource utilization scheduling obtained according to the data monitoring module for the virtualization scheduler module
The resource load of physical node;
Wherein, the utility function includes:
f_app(appk) indicate the scheduling strategy
Target, the target are that the user buys the expense minimum of the resources of virtual machine or the time of virtual machine completion task
It is most short, costtT-th of virtual machine VM of k-th of application program is distributed in expressiontThe unit resource unit interval in price, Rt
Indicate the stock number that t-th of virtual machine for distributing to k-th of application program occupies, f_time (Rt,missiont) indicate institute
It states t-th of virtual machine and completes distributed task amount missiontWhen occupied time.
8. a kind of physical node, which is characterized in that including communication module and virtualization scheduler module, wherein:
The communication module, the resource collection of each application program for receiving management node transmission, the resource collection packet
Include the stock number that each virtual machine occupies when any application program is run on virtual machine;
The virtualization scheduler module, for determining each virtual machine distribution on the physical node according to the resource collection
Least resource amount, then the surplus yield for distributing the physical node after the least resource amount is proportionately distributed to the object
Manage each virtual machine on node;
The communication module is additionally operable to the total resources that each virtual machine after distribution on the physical node obtains being sent to
The management node, so that the management node dispatches the resource load of each physical node according to the total resources.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510225160.0A CN104881325B (en) | 2015-05-05 | 2015-05-05 | A kind of resource regulating method and resource scheduling system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510225160.0A CN104881325B (en) | 2015-05-05 | 2015-05-05 | A kind of resource regulating method and resource scheduling system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104881325A CN104881325A (en) | 2015-09-02 |
CN104881325B true CN104881325B (en) | 2018-09-21 |
Family
ID=53948831
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510225160.0A Active CN104881325B (en) | 2015-05-05 | 2015-05-05 | A kind of resource regulating method and resource scheduling system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104881325B (en) |
Families Citing this family (28)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105392068B (en) * | 2015-11-04 | 2019-01-01 | 合一网络技术(北京)有限公司 | The parallel distribution method of the more transmission channel network direct broadcasting videos of distribution and system |
EP3402084B1 (en) * | 2016-02-03 | 2020-04-22 | Huawei Technologies Co., Ltd. | Resource scheduling method in dsl system, apparatus and system |
CN107203255A (en) * | 2016-03-20 | 2017-09-26 | 田文洪 | Power-economizing method and device are migrated in a kind of network function virtualized environment |
CN105912403B (en) * | 2016-04-14 | 2019-08-06 | 聚好看科技股份有限公司 | The method for managing resource and device of Docker container |
CN107301093B (en) * | 2016-04-15 | 2021-02-09 | 华为技术有限公司 | Method and device for managing resources |
CN107346264A (en) * | 2016-05-05 | 2017-11-14 | 北京金山云网络技术有限公司 | A kind of method, apparatus and server apparatus of virtual machine load balance scheduling |
CN107493485B (en) * | 2016-06-13 | 2021-11-05 | 中兴通讯股份有限公司 | Resource control method and device and IPTV server |
CN106126323B (en) * | 2016-06-17 | 2019-11-22 | 广州商品清算中心股份有限公司 | Real-time task scheduling method based on cloud platform |
CN107634978B (en) * | 2016-07-19 | 2020-11-06 | 华为技术有限公司 | Resource scheduling method and device |
CN106412124B (en) * | 2016-12-01 | 2019-10-29 | 广州高能计算机科技有限公司 | A kind of and sequence cloud service platform task distribution system and method for allocating tasks |
CN108268317B (en) * | 2016-12-30 | 2020-07-28 | 华为技术有限公司 | Resource allocation method and device |
CN108667859A (en) * | 2017-03-27 | 2018-10-16 | 中兴通讯股份有限公司 | A kind of method and device for realizing scheduling of resource |
CN107528914B (en) * | 2017-09-12 | 2020-05-12 | 电子科技大学 | Resource requisition scheduling method for data fragmentation |
CN108287748B (en) * | 2017-10-25 | 2021-10-22 | 国云科技股份有限公司 | Virtual machine and container parallel scheduling method |
CN108037993B (en) * | 2017-11-07 | 2020-08-04 | 大国创新智能科技(东莞)有限公司 | Cloud computing scheduling method and system based on big data and deep learning neural network |
CN108647090B (en) * | 2018-04-03 | 2021-09-14 | 联想(北京)有限公司 | Resource allocation method and device and server cluster |
CN108768877B (en) * | 2018-07-20 | 2021-06-22 | 网宿科技股份有限公司 | Distribution method and device of burst traffic and proxy server |
CN109189548B (en) * | 2018-08-01 | 2023-05-26 | 平安科技(深圳)有限公司 | Resource scheduling method, device, computer equipment and computer readable storage medium |
CN109308193B (en) * | 2018-09-06 | 2019-08-09 | 广州市品高软件股份有限公司 | A kind of multi-tenant function calculates the concurrency control method of service |
CN109445282A (en) * | 2018-11-07 | 2019-03-08 | 北京航空航天大学 | A kind of Optimization Scheduling towards basic device processing technology |
CN109298949B (en) * | 2018-12-04 | 2021-08-20 | 国网辽宁省电力有限公司大连供电公司 | Resource scheduling system of distributed file system |
CN109600642B (en) * | 2018-12-17 | 2020-10-30 | 广州华多网络科技有限公司 | CDN resource scheduling method and device |
CN110109758A (en) * | 2019-04-30 | 2019-08-09 | 温州职业技术学院 | A kind of cloud computing resources distribution method |
CN110620818B (en) * | 2019-09-18 | 2022-04-05 | 东软集团股份有限公司 | Method, device and related equipment for realizing node distribution |
CN111143060B (en) * | 2019-12-18 | 2021-01-26 | 重庆紫光华山智安科技有限公司 | GPU resource scheduling method and device and GPU |
CN112130997B (en) * | 2020-09-22 | 2022-09-20 | 新华三大数据技术有限公司 | Resource allocation method and device |
CN113641461A (en) * | 2021-10-14 | 2021-11-12 | 云宏信息科技股份有限公司 | Resource allocation method, medium and system for KVM virtualization system |
CN115392641B (en) * | 2022-07-29 | 2023-06-23 | 武汉达梦数据技术有限公司 | Resource scheduling method and device |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103455378A (en) * | 2013-09-05 | 2013-12-18 | 华为技术有限公司 | Resource distribution method of virtual clusters, device and system |
CN103823709A (en) * | 2014-02-28 | 2014-05-28 | 华为技术有限公司 | Virtualized cluster system, resource distributing method and management node |
CN104244422A (en) * | 2014-09-23 | 2014-12-24 | 中国联合网络通信集团有限公司 | Resource allocation method and device |
CN104572301A (en) * | 2015-01-04 | 2015-04-29 | 中国联合网络通信集团有限公司 | Resource distribution method and system |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104011685B (en) * | 2012-12-26 | 2016-05-25 | 华为技术有限公司 | A kind of method for managing resource, dummy machine system and device of dummy machine system |
-
2015
- 2015-05-05 CN CN201510225160.0A patent/CN104881325B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103455378A (en) * | 2013-09-05 | 2013-12-18 | 华为技术有限公司 | Resource distribution method of virtual clusters, device and system |
CN103823709A (en) * | 2014-02-28 | 2014-05-28 | 华为技术有限公司 | Virtualized cluster system, resource distributing method and management node |
CN104244422A (en) * | 2014-09-23 | 2014-12-24 | 中国联合网络通信集团有限公司 | Resource allocation method and device |
CN104572301A (en) * | 2015-01-04 | 2015-04-29 | 中国联合网络通信集团有限公司 | Resource distribution method and system |
Also Published As
Publication number | Publication date |
---|---|
CN104881325A (en) | 2015-09-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104881325B (en) | A kind of resource regulating method and resource scheduling system | |
CN103207814B (en) | Managing and task scheduling system and dispatching method across cluster resource of a kind of decentration | |
CN107066319A (en) | A kind of multidimensional towards heterogeneous resource dispatches system | |
CN109491790A (en) | Industrial Internet of Things edge calculations resource allocation methods and system based on container | |
CN107404523A (en) | Cloud platform adaptive resource dispatches system and method | |
CN104750557B (en) | A kind of EMS memory management process and memory management device | |
CN103827825A (en) | Virtual resource object component | |
CN107548549A (en) | Equilibrium of stock in DCE | |
CN110221920B (en) | Deployment method, device, storage medium and system | |
CN105468435A (en) | NFV dynamic resource distribution method | |
CN106233276A (en) | The coordination access control of network-accessible block storage device | |
CN106452842B (en) | Network system based on network function virtualization intermediary system architecture | |
CN109684074A (en) | Physical machine resource allocation methods and terminal device | |
CN107346264A (en) | A kind of method, apparatus and server apparatus of virtual machine load balance scheduling | |
CN104503832B (en) | A kind of scheduling virtual machine system and method for fair and efficiency balance | |
CN104679594B (en) | A kind of middleware distributed computing method | |
CN105373432B (en) | A kind of cloud computing resource scheduling method based on virtual resource status predication | |
CN109783225B (en) | Tenant priority management method and system of multi-tenant big data platform | |
CN107864211B (en) | Cluster resource dispatching method and system | |
CN111970383B (en) | Multi-tenant sharing method, system and storage medium of data center network | |
CN110333945A (en) | A kind of dynamic load balancing method, system and terminal | |
CN106874115A (en) | A kind of resources of virtual machine distribution method and distributed virtual machine resource scheduling system | |
CN106713375A (en) | Method and device for allocating cloud resources | |
WO2020134133A1 (en) | Resource allocation method, substation, and computer-readable storage medium | |
CN108701132A (en) | Resource management system and method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
EXSB | Decision made by sipo to initiate substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |