CN107346264A - A kind of method, apparatus and server apparatus of virtual machine load balance scheduling - Google Patents
A kind of method, apparatus and server apparatus of virtual machine load balance scheduling Download PDFInfo
- Publication number
- CN107346264A CN107346264A CN201610294226.6A CN201610294226A CN107346264A CN 107346264 A CN107346264 A CN 107346264A CN 201610294226 A CN201610294226 A CN 201610294226A CN 107346264 A CN107346264 A CN 107346264A
- Authority
- CN
- China
- Prior art keywords
- resource
- virtual machine
- physical node
- factor
- resources
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5061—Partitioning or combining of resources
- G06F9/5077—Logical partitioning of resources; Management or configuration of virtualized resources
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application proposes a kind of method, apparatus and server apparatus of virtual machine load balance scheduling, is related to field of cloud computer technology, when creating each virtual machine, methods described includes：Obtain the every resource requirement for creating virtual machine；Obtain for creating all available physical nodes of virtual machine and every resource behaviour in service of each physical node；Utilization rate based on the weight for creating virtual machine each resource requirement and each resource of all available physical nodes, it is determined that meeting the physical node operation virtual machine of every resource requirement of virtual machine to be created.The method weighted by the vector based on resources of virtual machine demand realizes that balanced and cluster overall load is balanced between each node multiple resources in large-scale cloud environment.
Description
Technical field
The present invention relates to field of cloud computer technology, and in particular to a kind of method, apparatus of virtual machine load balance scheduling and
Server apparatus.
Background technology
Cloud computation data center is based primarily upon virtualization technology structure, by the way that the resources such as calculating, storage, network is virtual
Change, user is used various resources on demand, there is management automation, be easy to the advantages such as extension, resource utilization height.
Virtualization is the key technology of cloud computing, and the technology is real by being virtualized to physical server (host)
Can be with the multiple virtual machines of subordinate on existing separate unit physical node.The physical server of more virtualizations can form a virtualization collection
Group, physical resource therein is abstracted as the resource pool of the multiple resources such as calculating, storage, network by virtual cluster, by demand
The mode of application is supplied to user to use.
Scheduling of resource in virtual cluster is related to whether cluster resource can effectively use, and current dispatching method is main
It is divided into initial schedule and Real-Time Scheduling, the target of scheduling is broadly divided into cluster load balance or improves resource utilization.For example,
DRS (Distributed Resource Scheduler, distributed resource scheduling) technology, can the CPU based on node
The equilibrium assignment of (Central Processing Unit, central processing unit) resource or memory source progress initial resource, and
When cluster is run by the virtual machine (vm) migration on high capacity node to low-load node to realize load balancing.
The initial schedule of virtual machine load balancing mainly in the case where existing resource use environment is constant, newly creates
Virtual machine distributes resource, and ensures that cluster resource using equilibrium, reduces the wasting of resources as far as possible.Efficient initial schedule can subtract
The expense of Real-Time Scheduling during few virtual machine operation, Real-Time Scheduling then make up loaded when deficiency and the reply operation of initial schedule compared with
Big change.
Virtual machine load balancing initial schedule method has distributed the virtual machine of resource due to that can not transfer, and can not predict
Later virtual machine can not use the conventional loading strategy of cluster virtual machine scheduling and dynamic to advise the demand of resource quantity
Draw class method.At present main initial schedule method mainly consider single resource between node equilibrium (selection cpu resource, or
Memory source distributes resource using minimum node to create virtual machine), multi dimensional resource is on the whole between causing multiple nodes
Unbalanced and individual node on imbalance between multiple resources.For example, when carrying out load balancing based on memory source, it is likely that
Cause some nodes to exhaust disk resource due to being assigned to the virtual machine of multiple small internal memories of big disk so that the node residue compared with
More memory sources, but the new virtual machine that can not reallocate, cause the utilization of resources unbalanced and waste.
The content of the invention
The present invention is unfavorable for the equilibrium of cluster overall load to solve the initial schedule load balancing based on single resource, causes
The problem of wasting of resources, there is provided a kind of method, apparatus and server apparatus of virtual machine load balance scheduling, realize cloud environment
In between each node multiple resources balanced and cluster overall load it is balanced.
In order to realize foregoing invention purpose, the technical scheme that the present invention takes is as follows：
A kind of method of virtual machine load balance scheduling, when creating each virtual machine, methods described includes：
Obtain the every resource requirement for creating virtual machine；
Obtain for creating all available physical nodes of virtual machine and every resource behaviour in service of each physical node；
Utilization based on the weight for creating virtual machine each resource requirement and each resource of all available physical nodes
Rate, it is determined that meeting the physical node operation virtual machine of every resource requirement of virtual machine to be created.
Alternatively, obtain every resource requirement of establishment virtual machine also includes afterwards：
Calculate the vectorial factor of each resource requirement.
Alternatively, obtain every resource behaviour in service of each physical node also includes afterwards：
Calculate the vectorial factor of each resource of each physical node.
Optionally it is determined that meet every resource requirement of virtual machine to be created a physical node operation virtual machine it
It is preceding also to include：
According to the vectorial factor of each resource of each physical node and the weight of each resource, each physics is calculated
The utilization of resources total value of node.
Optionally it is determined that meet the physical node operation virtual machine of every resource requirement of virtual machine to be created
Including；
Determine the minimum physical node operation virtual machine of utilization of resources total value.
Alternatively, calculating the vectorial factor of each resource requirement includes：
Every resource requirement of the virtual machine of acquisition carries out to dimension is unified, obtain each resource requirement it is vectorial because
Son.
Alternatively, calculating the vectorial factor of each resource of each physical node includes：
Every resource behaviour in service of each physical node of acquisition is subjected to dimension unification, obtains each resource
The vectorial factor.
Alternatively, according to the vectorial factor of each resource of each physical node and the weight of each resource, calculate
The utilization of resources total value of each physical node includes：
For each physical node, will create the vectorial factor of each resource of virtual machine and the vectorial of total physical resource because
Son makees business, obtains the weight of each resource；
According to the result of the utilization rate weighted sum of each resource of physical node, the resource of each physical node is obtained
Utilize total value.
Alternatively, according to the result of the utilization rate weighted sum of each resource of physical node, each physics section is obtained
The utilization of resources total value of point includes：
The resources balance factor of each physical node is calculated, by the result of the weighted sum and the resources balance
The factor is accumulated, and obtains the utilization of resources total value of each physical node；The resources balance factor is that the physical node is distributing
After current virtual machine, the mean square deviation of every resource utilization discreet value.
In order to solve the above technical problems, the present invention also provides a kind of device of virtual machine load balance scheduling, including：
Virtual machine demand module, every resource requirement of virtual machine is created for obtaining；
Node uses module, for obtaining all available physical nodes for being used for creating virtual machine and each physical node
Every resource behaviour in service；
Creation module, for based on the every of the weight and all available physical nodes for creating virtual machine each resource requirement
The utilization rate of item resource, it is determined that meeting the physical node operation virtual machine of every resource requirement of virtual machine to be created.
Alternatively, the virtual machine demand module, is additionally operable to：
Calculate the vectorial factor of each resource requirement.
Alternatively, the node uses module, is additionally operable to：
Calculate the vectorial factor of each resource of each physical node.
Alternatively, the creation module is additionally operable to：
According to the vectorial factor of each resource of each physical node and the weight of each resource, each physics is calculated
The utilization of resources total value of node.
Alternatively, the creation module determines a physical node for meeting every resource requirement of virtual machine to be created
Operation virtual machine refers to；
Determine the minimum physical node operation virtual machine of utilization of resources total value.
Alternatively, the virtual machine demand module calculates the vectorial factor of each resource requirement and referred to：
Every resource requirement of the virtual machine of acquisition carries out to dimension is unified, obtain each resource requirement it is vectorial because
Son.
Alternatively, the node is calculated the vectorial factor of each resource of each physical node using module and referred to：
Every resource behaviour in service of each physical node of acquisition is subjected to dimension unification, obtains each resource
The vectorial factor.
Alternatively, the vectorial factor and described each resource of the creation module according to each resource of each physical node
Weight, the utilization of resources total value for calculating each physical node refers to：
For each physical node, will create the vectorial factor of each resource of virtual machine and the vectorial of total physical resource because
Son makees business, obtains the weight of each resource；
According to the result of the utilization rate weighted sum of each resource of physical node, the resource of each physical node is obtained
Utilize total value.
Alternatively, the creation module obtains according to the result of the utilization rate weighted sum of each resource of physical node
The utilization of resources total value for obtaining each physical node refers to：
The resources balance factor of each physical node is calculated, by the result of the weighted sum and the resources balance
The factor is accumulated, and obtains the utilization of resources total value of each physical node；The resources balance factor is that the physical node is distributing
After current virtual machine, the mean square deviation of every resource utilization discreet value.
The present invention also provides a kind of server apparatus of virtual machine load balance scheduling, including above-mentioned virtual machine load balancing
The device of scheduling.
Compared to the prior art the present invention, has the advantages that：
The present invention is unfavorable for the equilibrium of cluster overall load for the initial schedule load balancing based on single resource, causes to provide
The problem of source wastes, the method weighted by the vector based on resources of virtual machine demand realize each node in large-scale cloud environment
Balanced and cluster overall load is balanced between multiple resources.
Brief description of the drawings
Fig. 1 is the flow chart of the method for the virtual machine load balance scheduling of the embodiment of the present invention；
Fig. 2 is the structural representation of the device of the virtual machine load balance scheduling of the embodiment of the present invention.
Embodiment
To make the goal of the invention of the present invention, technical scheme and beneficial effect of greater clarity, below in conjunction with the accompanying drawings to this
The embodiment of invention is illustrated, it is necessary to illustrate, in the case where not conflicting, in the embodiment and embodiment in the application
Feature can mutually be combined.
As shown in figure 1, the embodiment of the present invention provides a kind of method of virtual machine load balance scheduling, each virtual machine is created
When, methods described includes：
S101, obtain the every resource requirement for creating virtual machine；
S102, obtain all available physical nodes for creating virtual machine and every resource of each physical node uses
Situation；
S103, based on the weight for creating virtual machine each resource requirement and each resource of all available physical nodes
Utilization rate, it is determined that meeting the physical node operation virtual machine of every resource requirement of virtual machine to be created.
The embodiment of the present invention supports the situation of the one or more virtual machines created, when creating multiple virtual machines, obtain
The list for the multiple virtual machines that need to be created, it is successively the virtual machine selection physical node in list by above-mentioned steps, for each row
Virtual machine in table needs to update physical resource behaviour in service after selecting physical node.
Wherein, the resource requirement of virtual machine includes following one or more in step S101：Cpu resource demand Ncpu,
Memory source demand Nmem, disk resource demand Ndisk, I/O resource requirements Nio, network bandwidth resources demand Nnet。
In step S102, each node resource behaviour in service includes following one or more：The cpu resource of node is total
Amount and usage amount, memory source total amount and usage amount, disk resource total amount and usage amount, I/O total resources and make
Dosage, network bandwidth resources total amount and usage amount.
The embodiment of the present invention is in the cloud computing environment based on virtualization technology, based on each resource requirement of establishment virtual machine
Weight and all available physical nodes each resource utilization rate so that between node load balancing (each node it
Between resource use it is balanced), (node cpu, internal memory/disk space, the resource such as disk and network I/O use load balancing in node
It is balanced).
Wherein, step S101 also includes after obtaining the every resource requirement for creating virtual machine：
Calculate the vectorial factor of each resource requirement.
Wherein, step S102 also includes after obtaining every resource behaviour in service of each physical node：
Calculate the vectorial factor of each resource of each physical node.
Wherein, the vectorial factor of acquisition is easy to calculating to compare to unify dimension.
Wherein, step S103 determines to meet that the physical node operation of every resource requirement of virtual machine to be created is empty
Also include before plan machine：
According to the vectorial factor of each resource of each physical node and the weight of each resource, each physics is calculated
The utilization of resources total value of node.
Using the means for weighting multi-C vector in the embodiment of the present invention, load balancing initial schedule method is realized, compensate for
Existing method is for the balanced deficiency of balanced and cluster overall load between each node multiple resources.
The setting principle of weight：Influence of such resource to overall utilization rate is more, then weight is bigger, so may insure
The virtual machine more to certain class resource requirement is more likely to be assigned on the relatively low physical node of such resource utilization.
Step S103 determines the physical node operation virtual machine for meeting every resource requirement of virtual machine to be created
Including；
Determine the minimum physical node operation virtual machine of utilization of resources total value.
Wherein, calculating the vectorial factor of each resource requirement includes：
Every resource requirement of the virtual machine of acquisition carries out to dimension is unified, obtain each resource requirement it is vectorial because
Son.
Wherein, calculating the vectorial factor of each resource of each physical node includes：
Every resource behaviour in service of each physical node of acquisition is subjected to dimension unification, obtains each resource
The vectorial factor.
Wherein, according to the vectorial factor of each resource of each physical node and the weight of each resource, calculate every
The utilization of resources total value of individual physical node includes：
For each physical node, will create the vectorial factor of each resource of virtual machine and the vectorial of total physical resource because
Son makees business, obtains the weight of each resource；
According to the result of the utilization rate weighted sum of each resource of physical node, the resource of each physical node is obtained
Utilize total value.
Wherein, according to the result of the utilization rate weighted sum of each resource of physical node, each physical node is obtained
Utilization of resources total value include：
The resources balance factor of each physical node is calculated, by the result of the weighted sum and the resources balance
The factor is accumulated, and obtains the utilization of resources total value of each physical node；The resources balance factor is that the physical node is distributing
After current virtual machine, the mean square deviation of every resource utilization discreet value.
Node comprehensive resources utilization rate is lower show to select the physical node can make clustered node between resource rate it is more equal
Weighing apparatus, the resources balance factor is smaller to show that distributing current virtual machine makes the use of the physical node items resource more balanced.Therefore provide
Source is selection optimal when currently carrying out virtual machine load balancing using the minimum physical node of total value.
The embodiment of the present invention come calculate node comprehensive resources utilization rate, can consider that node is each comprehensively using multi dimensional resource vector
Item resource uses, and improves the load balancing of cluster on the whole.The method calculating physical node for considering resource supply and demand weighting is comprehensive
Joint source utilization rate, resource can be distributed according to the different resource requirement of virtual machine and physical machine total resources, it is ensured that to certain class
The more virtual machine of resource requirement can be assigned on the idle relatively large number of physical node of such resource.Consider simultaneously between node
The balance that multiple resources use in the equilibrium of comprehensive resources utilization rate and node.
The application scenarios of the embodiment of the present invention include but is not limited to virtual cluster, server cluster, task scheduling system
In resource original allocation and system operation in resource redistribute.
As shown in Fig. 2 the embodiment of the present invention also provides a kind of device of virtual machine load balance scheduling, including：
Virtual machine demand module, every resource requirement of virtual machine is created for obtaining；
Node uses module, for obtaining all available physical nodes for being used for creating virtual machine and each physical node
Every resource behaviour in service；
Creation module, for based on the every of the weight and all available physical nodes for creating virtual machine each resource requirement
The utilization rate of item resource, it is determined that meeting the physical node operation virtual machine of every resource requirement of virtual machine to be created.
The virtual machine demand module, is additionally operable to：
Calculate the vectorial factor of each resource requirement.
The node uses module, is additionally operable to：
Calculate the vectorial factor of each resource of each physical node.
The creation module is additionally operable to：
According to the vectorial factor of each resource of each physical node and the weight of each resource, each physics is calculated
The utilization of resources total value of node.
The creation module determines to meet that the physical node operation of every resource requirement of virtual machine to be created is empty
Plan machine refers to；
Determine the minimum physical node operation virtual machine of utilization of resources total value.
The vectorial factor that the virtual machine demand module calculates each resource requirement refers to：
Every resource requirement of the virtual machine of acquisition carries out to dimension is unified, obtain each resource requirement it is vectorial because
Son.
The vectorial factor that the node calculates each resource of each physical node using module refers to：
Every resource behaviour in service of each physical node of acquisition is subjected to dimension unification, obtains each resource
The vectorial factor.
The creation module is according to the vectorial factor of each resource of each physical node and the power of each resource
Weight, the utilization of resources total value for calculating each physical node refer to：
For each physical node, will create the vectorial factor of each resource of virtual machine and the vectorial of total physical resource because
Son makees business, obtains the weight of each resource；
According to the result of the utilization rate weighted sum of each resource of physical node, the resource of each physical node is obtained
Utilize total value.
The creation module obtains each thing according to the result of the utilization rate weighted sum of each resource of physical node
The utilization of resources total value of reason node refers to：
The resources balance factor of each physical node is calculated, by the result of the weighted sum and the resources balance
The factor is accumulated, and obtains the utilization of resources total value of each physical node；The resources balance factor is that the physical node is distributing
After current virtual machine, the mean square deviation of every resource utilization discreet value.
The embodiment of the present invention also provides a kind of server apparatus of virtual machine load balance scheduling, including above-mentioned virtual machine is born
Carry the device of balance dispatching；The device of the virtual machine load balance scheduling includes：
Virtual machine demand module, every resource requirement of virtual machine is created for obtaining；
Node uses module, for obtaining all available physical nodes for being used for creating virtual machine and each physical node
Every resource behaviour in service；
Creation module, for based on the every of the weight and all available physical nodes for creating virtual machine each resource requirement
The utilization rate of item resource, it is determined that meeting the physical node operation virtual machine of every resource requirement of virtual machine to be created.
Embodiment one：
The present embodiment realizes virtual machine load balance scheduling using multi-C vector weighting, realizes that load is equal under cloud environment
Weighing apparatus.The cloud environment includes physical node, and virtual machine, the load equilibration scheduling method comprise the following steps：
The list for the one or more virtual machines that need to be created is obtained, is the virtual machine selection in list according to the following steps successively
Physical node, need to update physical resource behaviour in service after selecting physical node for the virtual machine in each list：
(1) the every resource requirement for creating virtual machine is obtained, and calculates each resource requirement vector factor of virtual machine Fvm；
(2) all available physical nodes for creating virtual machine and its every resource behaviour in service are obtained, and is calculated every
The vectorial factor F of each resource of individual physical nodestd；
(3) utilization rate of the weight based on each resource and every resource of each physical node, calculates each node
Utilization of resources total value, and select utilization of resources total value it is minimum physical node operation virtual machine.
Further, described step (1) comprises the following steps：
1.1st, requirements of the virtual machine to every resource are obtained from the request for creating virtual machine, the requirements are by user
Specified when creating virtual machine, embody and create the resource size that the virtual machine needs to take.If the virtual machine does not have to certain resource
Require, then using default value.As shown in Equation 1：
Nvm=(Ncpu,Nmem,Ndisk,Nio,Nnet) formula 1
The resource requirement value vector N of virtual machinevm, the vector value that is made up of all kinds of resource requirement values of virtual machine, wherein providing
Source requirements include cpu resource demand Ncpu, memory source demand Nmem, disk resource demand Ndisk, I/O resource requirements NioAnd net
Network bandwidth resources demand Nnet。
1.2nd, according to the resource requirement value vector N of virtual machinevmCalculate resources of virtual machine requirement vector factor Fvm, it is vectorial because
Sub- FvmPass through described value vector NvmUnified dimension is easy to calculating to compare.Calculation formula is as shown in Equation 2：
Wherein, FvmThe cpu vector factors including virtual machineThe internal memory vector factor The disk vector factorThe I/O vector factorsWith the network bandwidth vector factorThe vectorial factor F of resources of virtual machine demandvmRepresent relative size of the virtual machine to every resource requirement.
Further, described step (2) comprises the following steps：
2.1st, all available physical node listings (ID number of i.e. each physical node) that can be used for creating virtual machine are obtained,
And each node resource behaviour in service, resource behaviour in service includes the cpu resource total amount of node and usage amount, memory source are total
Amount and usage amount, disk resource total amount and usage amount, I/O total resources and usage amount, network bandwidth resources total amount and
Usage amount.
2.2nd, the vectorial factor F of physical resource is calculated according to all physics node resource total amountsstd, the vectorial factor passes through each
Physical resource unifies dimension and is easy to calculating to compare.Computational methods are as shown in Equation 3：
The physical resource vector factor is by every resource factor divided by cpu resource factor RcpuObtain, including the cpu resource factor
Rcpu, memory source factor Rmem, disk resource factor Rdisk, I/O resource factors RioWith network bandwidth resources factor Rnet。FstdBag
Include the cpu resource factorThe memory source factorThe disk resource factorI/O resource factorsWith the network bandwidth resources factorPhysical resource to
Measure factor FstdEach total resources relative size that expression physical node has, and resources of virtual machine requirement vector factor FvmWith thing
Manage resource vector factor FstdRatio can be used for embody weight of the virtual machine to a certain resource requirement, the weight can be used for
The calculate node physical resource vector factor, and then select optimal physical node.
Further, described step (3) comprises the following steps：
3.1st, the resource service condition based on each physical node obtained, the money of all available physical nodes is calculated successively
Source utilization rate.Every resource utilization includes：Cpu busy percentage Ucpu, memory usage Umem, disk utilization Udisk, I/O utilize
Rate UioWith network bandwidth utilization factor Unet.The utilization rate of every resource be corresponding resource usage amount divided by the resource it is total
Amount.
3.2nd, the comprehensive resources utilization rate U of each physical node is calculated successivelyweight.Comprehensive resources utilization rate UweightCalculating
Method as shown in Equation 4, for the weighted sum of the various resources of the physical node：
Wherein, weight can be preset corresponding to each resource, can also by currently create the resource requirement vector of virtual machine because
Sub- FvmWith total physical resource vector factor FstdRatio be worth to.FvmWith FstdRatio major embodiment physical resource quantity with
Resources of virtual machine demand corresponds to the relation between supply and demand of different type resource, and ratio is bigger to illustrate demand of the virtual machine to respective resources
More prominent, then when calculating the comprehensive resources utilization rate of physical node, the weight that such resource utilization obtains is bigger, to entirety
The influence of utilization rate is more, so may insure that the virtual machine more to certain class resource requirement is more likely to be assigned to such resource
On the relatively low physical node of utilization rate.The comprehensive resources utilization rate U of physical nodeweightIt is mainly used in instructing between clustered node
Load balancing, to reach the target of load balancing, can be current virtual machine select cluster in comprehensive resources utilization rate it is relatively low
Physical node is as host.
3.3rd, the resources balance factor L of each physical node is calculated successivelyall。LallComputational methods such as formula 5 and the institute of formula 6
Show：
That is resources balance factor LallIt is expressed as, each physical node is after current virtual machine is distributed, every resource utilization
The mean square deviation of discreet value.Utilize the mean square deviation (L of every resource utilization discreet valueall, by formula 5 and formula 6 calculate) body
The dispersion degree and degree of fluctuation of existing every resource utilization, whereinTable respectively
After showing that current virtual machine is assigned to present physical node, its CPU resource utilization, memory source utilization rate, disk resource utilize
The discreet value of rate, I/O resource utilizations, network bandwidth resources utilization rate.Resources balance factor Lall(each resource utilization it is equal
Variance) can represent certain physical node be assigned with it is current create virtual machine after each resource equilibrium level, i.e. LallSmaller expression should
Each resource of physical node, which uses, more to be balanced.
3.4th, each physical node utilization of resources total value S is calculated successively, and selects the minimum nodes of S as current virtual machine
Host.S computational methods are as shown in Equation 7：
S=Lall×UweightFormula 7
That is physical node utilization of resources total value S is node comprehensive resources utilization rate UweightWith its resources balance factor Lall
Product.UweightIt is lower show to select the physical node can make clustered node between resource rate it is more balanced, LallIt is smaller to show
Distribution current virtual machine makes the use of the physical node items resource more balanced.Therefore physical node minimum S is i.e. current is carried out
Optimal selection during virtual machine load balancing.
Although disclosed embodiment is as above, its content is only to facilitate understand the technical side of the present invention
Case and the embodiment used, are not intended to limit the present invention.Any those skilled in the art to which this invention pertains, not
On the premise of departing from disclosed core technology scheme, any modification can be made in the form and details of implementation and is become
Change, but the protection domain that the present invention is limited, the scope that must be still limited by appended claims are defined.
Claims (19)
- A kind of 1. method of virtual machine load balance scheduling, it is characterised in that when creating each virtual machine, methods described includes：Obtain the every resource requirement for creating virtual machine；Obtain for creating all available physical nodes of virtual machine and every resource behaviour in service of each physical node；Utilization rate based on the weight for creating virtual machine each resource requirement and each resource of all available physical nodes, really Surely the physical node operation virtual machine of every resource requirement of virtual machine to be created is met.
- 2. the method as described in claim 1, it is characterised in that：The every resource requirement for obtaining establishment virtual machine is also wrapped afterwards Include：Calculate the vectorial factor of each resource requirement.
- 3. method as claimed in claim 1 or 2, it is characterised in that：Obtain every resource behaviour in service of each physical node Also include afterwards：Calculate the vectorial factor of each resource of each physical node.
- 4. method as claimed in claim 3, it is characterised in that：It is determined that meet every resource requirement of virtual machine to be created Also include before one physical node operation virtual machine：According to the vectorial factor of each resource of each physical node and the weight of each resource, each physical node is calculated Utilization of resources total value.
- 5. method as claimed in claim 4, it is characterised in that it is determined that meet every resource requirement of virtual machine to be created One physical node operation virtual machine includes；Determine the minimum physical node operation virtual machine of utilization of resources total value.
- 6. method as claimed in claim 2, it is characterised in that calculating the vectorial factor of each resource requirement includes：Every resource requirement of the virtual machine of acquisition is subjected to dimension unification, obtains the vectorial factor of each resource requirement.
- 7. method as claimed in claim 3, it is characterised in that calculate the vectorial factor bag of each resource of each physical node Include：Every resource behaviour in service of each physical node of acquisition is subjected to dimension unification, obtains the vector of each resource The factor.
- 8. method as claimed in claim 4, it is characterised in that according to the vectorial factor of each resource of each physical node and The weight of each resource, calculating the utilization of resources total value of each physical node includes：For each physical node, the vectorial factor of each resource and the vectorial factor of total physical resource that will create virtual machine are made Business, obtain the weight of each resource；According to the result of the utilization rate weighted sum of each resource of physical node, the utilization of resources of each physical node is obtained Total value.
- 9. method as claimed in claim 4, it is characterised in that asked according to the weighting of the utilization rate of each resource of physical node The result of sum, obtaining the utilization of resources total value of each physical node includes：The resources balance factor of each physical node is calculated, by the result of the weighted sum and the resources balance factor Accumulate, obtain the utilization of resources total value of each physical node；The resources balance factor is that the physical node is current in distribution After virtual machine, the mean square deviation of every resource utilization discreet value.
- A kind of 10. device of virtual machine load balance scheduling, it is characterised in that including：Virtual machine demand module, every resource requirement of virtual machine is created for obtaining；Node uses module, and the items for all available physical nodes and each physical node for creating virtual machine are used for for obtaining Resource behaviour in service；Creation module, for each money based on the weight and all available physical nodes for creating virtual machine each resource requirement The utilization rate in source, it is determined that meeting the physical node operation virtual machine of every resource requirement of virtual machine to be created.
- 11. device as claimed in claim 10, it is characterised in that：The virtual machine demand module, is additionally operable to：Calculate the vectorial factor of each resource requirement.
- 12. the device as described in claim 10 or 11, it is characterised in that：The node uses module, is additionally operable to：Calculate the vectorial factor of each resource of each physical node.
- 13. device as claimed in claim 12, it is characterised in that：The creation module is additionally operable to：According to the vectorial factor of each resource of each physical node and the weight of each resource, each physical node is calculated Utilization of resources total value.
- 14. device as claimed in claim 13, it is characterised in that the creation module determines to meet virtual machine to be created The physical node operation virtual machine of every resource requirement refers to；Determine the minimum physical node operation virtual machine of utilization of resources total value.
- 15. device as claimed in claim 11, it is characterised in that the virtual machine demand module calculates each resource requirement The vectorial factor refers to：Every resource requirement of the virtual machine of acquisition is subjected to dimension unification, obtains the vectorial factor of each resource requirement.
- 16. device as claimed in claim 12, it is characterised in that the node calculates the every of each physical node using module The vectorial factor of item resource refers to：Every resource behaviour in service of each physical node of acquisition is subjected to dimension unification, obtains the vector of each resource The factor.
- 17. device as claimed in claim 13, it is characterised in that the creation module according to each physical node each money The vectorial factor in source and the weight of each resource, the utilization of resources total value for calculating each physical node refer to：For each physical node, the vectorial factor of each resource and the vectorial factor of total physical resource that will create virtual machine are made Business, obtain the weight of each resource；According to the result of the utilization rate weighted sum of each resource of physical node, the utilization of resources of each physical node is obtained Total value.
- 18. device as claimed in claim 13, it is characterised in that the creation module is according to each resource of the physical node Utilization rate weighted sum result, the utilization of resources total value for obtaining each physical node refers to：The resources balance factor of each physical node is calculated, by the result of the weighted sum and the resources balance factor Accumulate, obtain the utilization of resources total value of each physical node；The resources balance factor is that the physical node is current in distribution After virtual machine, the mean square deviation of every resource utilization discreet value.
- A kind of 19. server apparatus including such as claim 10 to claim 18 any one claim described device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610294226.6A CN107346264A (en) | 2016-05-05 | 2016-05-05 | A kind of method, apparatus and server apparatus of virtual machine load balance scheduling |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610294226.6A CN107346264A (en) | 2016-05-05 | 2016-05-05 | A kind of method, apparatus and server apparatus of virtual machine load balance scheduling |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107346264A true CN107346264A (en) | 2017-11-14 |
Family
ID=60254181
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610294226.6A Pending CN107346264A (en) | 2016-05-05 | 2016-05-05 | A kind of method, apparatus and server apparatus of virtual machine load balance scheduling |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107346264A (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108614726A (en) * | 2018-03-22 | 2018-10-02 | 华为技术有限公司 | Virtual machine creation method and device |
| CN108762891A (en) * | 2018-06-06 | 2018-11-06 | 郑州云海信息技术有限公司 | A kind of cloud platform resource regulating method and device |
| CN108874535A (en) * | 2018-05-14 | 2018-11-23 | 中国平安人寿保险股份有限公司 | A kind of task adjusting method, computer readable storage medium and terminal device |
| CN109067867A (en) * | 2018-07-30 | 2018-12-21 | 北京航空航天大学 | The virtualization container service elastic telescopic method of data-oriented central loading monitoring |
| CN109445684A (en) * | 2018-09-20 | 2019-03-08 | 华为技术有限公司 | The resource allocation methods and equipment of hard disk in a kind of distributed memory system |
| WO2019127492A1 (en) * | 2017-12-29 | 2019-07-04 | 华为技术有限公司 | Node flow ratio prediction method and device |
| CN110399200A (en) * | 2019-06-26 | 2019-11-01 | 苏州浪潮智能科技有限公司 | A kind of cloud platform resource regulating method and device |
| CN110795203A (en) * | 2018-08-03 | 2020-02-14 | 阿里巴巴集团控股有限公司 | Resource scheduling method, device and system and computing equipment |
| CN110795233A (en) * | 2019-09-18 | 2020-02-14 | 北京你财富计算机科技有限公司 | Distributed resource allocation method and device and electronic equipment |
| CN110825212A (en) * | 2019-10-17 | 2020-02-21 | 北京京东尚科信息技术有限公司 | Energy-saving scheduling method and device and computer storage medium |
| CN112148426A (en) * | 2020-08-25 | 2020-12-29 | 新华三大数据技术有限公司 | Bandwidth allocation method and device |
| CN113626282A (en) * | 2021-07-16 | 2021-11-09 | 济南浪潮数据技术有限公司 | Cloud computing physical node load monitoring method and device, terminal and storage medium |
| CN113905449A (en) * | 2021-09-30 | 2022-01-07 | 阿里巴巴达摩院(杭州)科技有限公司 | Computing resource scheduling method, system and equipment |
| CN114221962A (en) * | 2021-12-09 | 2022-03-22 | 兴业银行股份有限公司 | Cloud resource redistribution method and system based on peak utilization rate |
| WO2022257388A1 (en) * | 2021-06-08 | 2022-12-15 | 深圳前海微众银行股份有限公司 | Speed limiting method and apparatus for virtual machine, and device, storage medium and program |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102446115A (en) * | 2012-01-09 | 2012-05-09 | 浙江大学 | Dynamic deployment method for virtual machines |
| CN103713956A (en) * | 2014-01-06 | 2014-04-09 | 山东大学 | Method for intelligent weighing load balance in cloud computing virtualized management environment |
| CN103810023A (en) * | 2014-03-06 | 2014-05-21 | 中国科学院信息工程研究所 | Distributed application intelligent deployment method and system in cloud platform |
| CN103870314A (en) * | 2014-03-06 | 2014-06-18 | 中国科学院信息工程研究所 | Method and system for simultaneously operating different types of virtual machines by single node |
| CN104881325A (en) * | 2015-05-05 | 2015-09-02 | 中国联合网络通信集团有限公司 | Resource scheduling method and resource scheduling system |
| CN105279027A (en) * | 2015-11-19 | 2016-01-27 | 浪潮(北京)电子信息产业有限公司 | Virtual machine disposition method and apparatus |
-
2016
- 2016-05-05 CN CN201610294226.6A patent/CN107346264A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102446115A (en) * | 2012-01-09 | 2012-05-09 | 浙江大学 | Dynamic deployment method for virtual machines |
| CN103713956A (en) * | 2014-01-06 | 2014-04-09 | 山东大学 | Method for intelligent weighing load balance in cloud computing virtualized management environment |
| CN103810023A (en) * | 2014-03-06 | 2014-05-21 | 中国科学院信息工程研究所 | Distributed application intelligent deployment method and system in cloud platform |
| CN103870314A (en) * | 2014-03-06 | 2014-06-18 | 中国科学院信息工程研究所 | Method and system for simultaneously operating different types of virtual machines by single node |
| CN104881325A (en) * | 2015-05-05 | 2015-09-02 | 中国联合网络通信集团有限公司 | Resource scheduling method and resource scheduling system |
| CN105279027A (en) * | 2015-11-19 | 2016-01-27 | 浪潮(北京)电子信息产业有限公司 | Virtual machine disposition method and apparatus |
Cited By (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019127492A1 (en) * | 2017-12-29 | 2019-07-04 | 华为技术有限公司 | Node flow ratio prediction method and device |
| CN111527734B (en) * | 2017-12-29 | 2021-10-26 | 华为技术有限公司 | Node traffic ratio prediction method and device |
| CN111527734A (en) * | 2017-12-29 | 2020-08-11 | 华为技术有限公司 | Node traffic ratio prediction method and device |
| CN108614726B (en) * | 2018-03-22 | 2022-06-10 | 华为云计算技术有限公司 | Virtual machine creation method and device |
| US11960915B2 (en) | 2018-03-22 | 2024-04-16 | Huawei Cloud Computing Technologies Co., Ltd. | Method and apparatus for creating virtual machine based on parameter information of a virtual network interface card |
| CN108614726A (en) * | 2018-03-22 | 2018-10-02 | 华为技术有限公司 | Virtual machine creation method and device |
| CN108874535B (en) * | 2018-05-14 | 2022-06-10 | 中国平安人寿保险股份有限公司 | Task adjusting method, computer readable storage medium and terminal device |
| CN108874535A (en) * | 2018-05-14 | 2018-11-23 | 中国平安人寿保险股份有限公司 | A kind of task adjusting method, computer readable storage medium and terminal device |
| CN108762891A (en) * | 2018-06-06 | 2018-11-06 | 郑州云海信息技术有限公司 | A kind of cloud platform resource regulating method and device |
| CN109067867A (en) * | 2018-07-30 | 2018-12-21 | 北京航空航天大学 | The virtualization container service elastic telescopic method of data-oriented central loading monitoring |
| CN110795203A (en) * | 2018-08-03 | 2020-02-14 | 阿里巴巴集团控股有限公司 | Resource scheduling method, device and system and computing equipment |
| CN110795203B (en) * | 2018-08-03 | 2023-05-02 | 阿里巴巴集团控股有限公司 | Resource scheduling method, device, system and computing equipment |
| CN109445684B (en) * | 2018-09-20 | 2021-11-26 | 华为技术有限公司 | Resource allocation method and equipment for hard disk in distributed storage system |
| US11385814B2 (en) | 2018-09-20 | 2022-07-12 | Huawei Cloud Computing Technologies Co., Ltd. | Method and device for allocating resource of hard disk in distributed storage system |
| CN109445684A (en) * | 2018-09-20 | 2019-03-08 | 华为技术有限公司 | The resource allocation methods and equipment of hard disk in a kind of distributed memory system |
| CN110399200A (en) * | 2019-06-26 | 2019-11-01 | 苏州浪潮智能科技有限公司 | A kind of cloud platform resource regulating method and device |
| CN110795233A (en) * | 2019-09-18 | 2020-02-14 | 北京你财富计算机科技有限公司 | Distributed resource allocation method and device and electronic equipment |
| CN110795233B (en) * | 2019-09-18 | 2022-04-05 | 北京你财富计算机科技有限公司 | Distributed resource allocation method and device and electronic equipment |
| CN110825212A (en) * | 2019-10-17 | 2020-02-21 | 北京京东尚科信息技术有限公司 | Energy-saving scheduling method and device and computer storage medium |
| CN112148426A (en) * | 2020-08-25 | 2020-12-29 | 新华三大数据技术有限公司 | Bandwidth allocation method and device |
| WO2022257388A1 (en) * | 2021-06-08 | 2022-12-15 | 深圳前海微众银行股份有限公司 | Speed limiting method and apparatus for virtual machine, and device, storage medium and program |
| CN113626282A (en) * | 2021-07-16 | 2021-11-09 | 济南浪潮数据技术有限公司 | Cloud computing physical node load monitoring method and device, terminal and storage medium |
| CN113626282B (en) * | 2021-07-16 | 2023-12-22 | 济南浪潮数据技术有限公司 | Cloud computing physical node load monitoring method, device, terminal and storage medium |
| CN113905449B (en) * | 2021-09-30 | 2024-04-05 | 杭州阿里云飞天信息技术有限公司 | Computing resource scheduling method, system and equipment |
| CN113905449A (en) * | 2021-09-30 | 2022-01-07 | 阿里巴巴达摩院(杭州)科技有限公司 | Computing resource scheduling method, system and equipment |
| CN114221962A (en) * | 2021-12-09 | 2022-03-22 | 兴业银行股份有限公司 | Cloud resource redistribution method and system based on peak utilization rate |
| CN114221962B (en) * | 2021-12-09 | 2024-02-13 | 兴业银行股份有限公司 | Cloud resource reallocation method and system based on peak utilization rate |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107346264A (en) | A kind of method, apparatus and server apparatus of virtual machine load balance scheduling | |
| CN112783649B (en) | Cloud computing-oriented interaction-aware containerized micro-service resource scheduling method | |
| CN102724277B (en) | The method of live migration of virtual machine and deployment, server and group system | |
| CN105159762B (en) | Heuristic cloud computing method for scheduling task based on Greedy strategy | |
| CN104375897B (en) | Cloud computing resource scheduling method based on the unbalanced degree of minimum relative load | |
| CN103729246B (en) | Method and device for dispatching tasks | |
| CN110109756A (en) | A kind of network target range construction method, system and storage medium | |
| CN106020934A (en) | Optimized deploying method based on virtual cluster online migration | |
| CN104881325A (en) | Resource scheduling method and resource scheduling system | |
| CN103118124A (en) | Cloud computing load balancing method based on layering multiple agents | |
| CN101504620A (en) | Load balancing method, apparatus and system of virtual cluster system | |
| CN108108245B (en) | Hybrid scheduling method and system for cloud platform wide-node scientific workflow | |
| CN104679594B (en) | A kind of middleware distributed computing method | |
| CN105808341B (en) | A kind of methods, devices and systems of scheduling of resource | |
| CN102708003A (en) | Method for allocating resources under cloud platform | |
| Li et al. | An effective scheduling strategy based on hypergraph partition in geographically distributed datacenters | |
| CN114356543A (en) | Kubernetes-based multi-tenant machine learning task resource scheduling method | |
| CN106412124A (en) | Task allocation system and task allocation method for parallel ordering cloud service platform | |
| CN104917839A (en) | Load balancing method for use in cloud computing environment | |
| CN112559122A (en) | Virtualization instance management and control method and system based on electric power special security and protection equipment | |
| Li et al. | A wholistic optimization of containerized workflow scheduling and deployment in the cloud–edge environment | |
| Patra et al. | A randomized algorithm for load balancing in containerized cloud | |
| Kaur et al. | Equal distribution based load balancing technique for fog-based cloud computing | |
| Biswas et al. | A novel resource aware scheduling with multi-criteria for heterogeneous computing systems | |
| Wo et al. | Overbooking-based resource allocation in virtualized data center |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171114 |
|
| RJ01 | Rejection of invention patent application after publication |