CN103677958A - Virtualization cluster resource scheduling method and device - Google Patents
Virtualization cluster resource scheduling method and device Download PDFInfo
- Publication number
- CN103677958A CN103677958A CN201310689439.5A CN201310689439A CN103677958A CN 103677958 A CN103677958 A CN 103677958A CN 201310689439 A CN201310689439 A CN 201310689439A CN 103677958 A CN103677958 A CN 103677958A
- Authority
- CN
- China
- Prior art keywords
- resource
- computing node
- virtual machine
- node
- virtual
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
The invention discloses a virtualization cluster resource scheduling method which includes the steps that a management node simulates the optimal balanced state of a current virtualization cluster; resource using conditions of the virtualization cluster under the optimal balanced state are evaluated; if the evaluation result is that resources are appropriately used, load balanced scheduling is carried out; if the evaluation result is that the resources are overused, power-on scheduling is carried out; if the evaluation result is that the resources are insufficiently used, power-off scheduling is carried out. According to the embodiment, the corresponding management node is further provided. According to the technical scheme, scheduling evaluations are carried out based on judgment on the whole state of the virtualization cluster, scheduling suggestions can be rapidly and accurately given, the problem that load balancing and a DPS give opposite scheduling suggestions in the same scheduling of a DRS can be further solved, the virtualization cluster accordingly achieves the stable state rapidly, and the technical problems that in an existing DRS method, the virtual machine migration time number is increased, the scheduling time is prolonged, and the migration cost is improved are solved.
Description
Technical field
The present invention relates to communication technical field, be specifically related to a kind of resource regulating method and device of virtual cluster.
Background technology
Server virtualization technology is the gordian technique based on infrastructure layer in cloud computing.This technology is virtual by physical server is carried out, and realizes at many virtual machines of separate unit physical machine deploy (Virtual Machine, VM), improves the resource utilization of server, reduces use cost.Virtual cluster is that many virtualized servers are consisted of to an organic whole, carry out unified management, by Intel Virtualization Technology, by the abstract resource pool forming for the various resources such as storage, calculating, network of physical resource, the mode by on-demand application resource provides virtual machine to user.
A key property of virtual cluster is DRS(Dynamic Resources Scheduling, dynamic resource scheduling), based on this characteristic, cluster management system can regularly be monitored the resource utilization of each physical machine and virtual machine, according to resource distribution situation, utilize thermophoresis to adjust the distribution of virtual machine in physical machine, thereby load balancing and the cluster realized in cluster wide are integrated, when improving the resource utilization of each physical machine, guarantee that each physical machine all bears suitable load.
The DRS of virtual cluster generally comprises load balancing and DPM(Distributed power management, distributed power supply management) two processes.Load balancing, for the load of each physical machine of balance, is carried out the part virtual machine (vm) migration in high capacity physical machine on other physical machine.DPM continues the resource requirement in the virtual cluster of monitoring, and during poor efficiency, resource requirement reduces, and operating load will be incorporated in several physical machine, closes other no physical machine to reduce power consumption; During high usage, resource requirement increases, can buttoned-up physical machine is again online to guarantee to meet service class.
The focus of load balancing and DPM is different, and whether load balancing is paid close attention to the load of each physical machine in cluster balanced, and DPM pays close attention to the service condition of the whole resource of cluster.In prior art, DRS method is normally first carried out load balancing, then carries out DPM.But the suggestion that the difference of focus can make load balancing and DPM produce runs in the opposite direction, one takes turns not only load balancing but also DPM in scheduling, can increase migration number of times, and virtual machine is moved repeatedly at short notice, extends scheduling time, increases moving costs.Example as shown in Figure 1, virtual cluster comprises three computing node host1, host2 and host3, in host1, operation has virtual machine VM1-VM4, and in host2, operation has virtual machine VM5-VM6, while carrying out DRS scheduling, load balancing can provide the suggestion of vm4 being moved to host2, DPM can provide the suggestion that host3 is powered on and vm3 and vm4 are moved to host3, so vm4 can move twice at short notice.
Summary of the invention
The embodiment of the present invention provides a kind of resource regulating method and device of virtual cluster, to solve existing DRS method, can increase virtual machine (vm) migration number of times, extends scheduling time, increases the technical matters of moving costs.
First aspect present invention provides a kind of resource regulating method of virtual cluster, described virtual cluster comprises management node and at least one computing node, and described computing node comprises hardware layer, operates in the host on described hardware layer and operates at least one virtual machine on described host; Described method comprises: the optimum equalization state of the current described virtual cluster of described management node simulation; Resource service condition to described virtual cluster under optimum equalization state is assessed; If assessment result is resource, use suitably, carry out load balance scheduling; If assessment result is resource, overuse, scheduling powers on; If assessment result is resource, use deficiency, carry out lower electricity scheduling.
In the possible implementation of the first, the current optimum equalization state of the described management node described virtual cluster of simulation comprises: by a virtual machine (vm) migration on the highest computing node of resource use amount to the minimum computing node of resource use amount, and repeat this step, until do not have virtual machine to move, think that described virtual cluster reaches optimum equalization state.
In the possible implementation of the second, the current optimum equalization state of the described management node described virtual cluster of simulation comprises: first, by a virtual machine (vm) migration on the highest computing node of resource use amount to the minimum computing node of resource use amount, and repeat this step, until do not have virtual machine to move; Then, take the virtual machine that resource is maximum by one on the highest computing node of resource use amount, and take the minimum virtual machine of resource, mutually exchange migration on the minimum computing node of resource use amount, and repeat this step, until do not have the virtual machine can exchange migration; Think that described virtual cluster reaches optimum equalization state.
In the third possible implementation, the current optimum equalization state of the described management node described virtual cluster of simulation comprises: all virtual machines in described virtual cluster are sorted according to the size of its resource use amount; According to resource use amount order from big to small, be that each virtual machine finds the Optimal calculation node that can place, make after each virtual machine is placed to searched out Optimal calculation node, the equilibrium state of whole virtual cluster is optimum.
Any in conjunction with the first of first aspect or first aspect in the third possible implementation, in the 4th kind of possible implementation, described to described virtual cluster the resource service condition under optimum equalization state assess and comprise: the quantity of determining focus computing node and cold spot computing node under optimum equalization state, described focus computing node refers to that resource utilization surpasses the high computing node that carries utilization factor empirical value, and described cold spot computing node refers to that resource utilization is lower than the computing node of low year utilization factor empirical value; If the quantity of focus computing node and cold spot computing node is all zero, judge the resource use of described virtual cluster under optimum equalization state suitably; If focus computing node quantity is greater than zero, judge that the resource of described virtual cluster under optimum equalization state overuses; If cold spot computing node quantity is greater than zero, judge that the resource use of described virtual cluster under optimum equalization state is not enough.
The third or the 4th kind of possible implementation in conjunction with first aspect, in the 5th kind of possible implementation, describedly find in the process of the Optimal calculation node that can place for each virtual machine, if the Optimal calculation node that has virtual machine not search out can to place, judges that the resource of described virtual cluster under optimum equalization state overuses.
A kind of management node of second aspect present invention, for virtual cluster, described virtual cluster comprises described management node and at least one computing node, and described computing node comprises hardware layer, operates in the host on described hardware layer and operates at least one virtual machine on described host; Described management node comprises: analog module, for simulating the optimum equalization state of current described virtual cluster; Evaluation module, for to described virtual cluster the resource service condition under optimum equalization state assess; Scheduler module, if be that resource is used suitably for the assessment result of described evaluation module, carries out load balance scheduling; If the assessment result of described evaluation module is resource, overuse, scheduling powers on; If the assessment result of described evaluation module is resource, use deficiency, carry out lower electricity scheduling.
In the possible implementation of the first, described analog module comprises: the first analogue unit, for by a virtual machine (vm) migration on the highest computing node of resource use amount to the minimum computing node of resource use amount, and repeat this step, until do not have virtual machine to move.
In the possible implementation of the second, described analog module comprises: the first analogue unit, for by a virtual machine (vm) migration on the highest computing node of resource use amount to the minimum computing node of resource use amount, and repeat this step, until do not have virtual machine to move; The second analogue unit, for taking the virtual machine that resource is maximum by one on the highest computing node of resource use amount, and take the minimum virtual machine of resource, mutually exchange migration on the minimum computing node of resource use amount, and repeat this step, until do not have the virtual machine can exchange migration.
In the third possible implementation, described analog module comprises: the 3rd analogue unit, for all virtual machines of described virtual cluster are sorted according to the size of its resource use amount; According to resource use amount order from big to small, be that each virtual machine finds the Optimal calculation node that can place, make after each virtual machine is placed to searched out Optimal calculation node, the equilibrium state of whole virtual cluster is optimum.
Any in conjunction with the first of second aspect or second aspect in the third possible implementation, described evaluation module comprises: determining unit, quantity for focus computing node and cold spot computing node under definite optimum equalization state, described focus computing node refers to that resource utilization surpasses the high computing node that carries utilization factor empirical value, and described cold spot computing node refers to that resource utilization is lower than the computing node of low year utilization factor empirical value; Judging unit, if be all zero for the quantity of focus computing node and cold spot computing node, judges the resource use of described virtual cluster under optimum equalization state suitably; If focus computing node quantity is greater than zero, judge that the resource of described virtual cluster under optimum equalization state overuses; If cold spot computing node quantity is greater than zero, judge that the resource use of described virtual cluster under optimum equalization state is not enough.
The third or the 4th kind of possible implementation in conjunction with second aspect, in the 5th kind of possible implementation, described the 3rd analogue unit is found in the process of the Optimal calculation node that can place for each virtual machine, if the Optimal calculation node that has virtual machine not search out can to place, the resource of the described virtual cluster of described evaluation module judgement under optimum equalization state used not enough.
The embodiment of the present invention adopts carries out optimum equalization state simulation to virtual cluster, resource service condition under assessment optimum equalization state, according to assessment result, decide and carry out load balancing or power on or lower electric technical scheme of dispatching, based on the judgement of virtual cluster integrality is dispatched to assessment, can provide more accurately sooner scheduling suggestion, can also avoid load balancing and DPM in DRS same taken turns scheduling to provide contrary scheduling suggestion, thereby make virtual cluster reach fast steady state (SS), solve existing DRS method and can increase virtual machine (vm) migration number of times, extend scheduling time, increase the technical matters of moving costs.
Accompanying drawing explanation
Fig. 1 carries out the schematic diagram of DRS to virtual cluster in prior art;
Fig. 2 is the logical organization schematic diagram of virtual cluster;
Fig. 3 is the process flow diagram of the resource regulating method of the virtual cluster that provides of the embodiment of the present invention;
Fig. 4 is the schematic diagram of the management node that provides of the embodiment of the present invention;
Fig. 5 is the schematic diagram of the management node that provides of another embodiment of the present invention.
Embodiment
The embodiment of the present invention provides a kind of resource regulating method and device of virtual cluster, and can solve existing DRS method can increase virtual machine (vm) migration number of times, extends scheduling time, increases the technical matters of moving costs.In order to make those skilled in the art person understand better the present invention program, below in conjunction with the accompanying drawing in the embodiment of the present invention, technical scheme in the embodiment of the present invention is clearly and completely described, obviously, described embodiment is only the embodiment of a part of the present invention, rather than whole embodiment.Embodiment based in the present invention, those of ordinary skills, not making the every other embodiment obtaining under creative work prerequisite, should belong to the scope of protection of the invention.
To facilitate understanding of the present embodiment of the invention, first at this, introduce the several key elements that can introduce in embodiment of the present invention description;
Virtual machine: can simulate one or many virtual computing machines by software virtual machine on a physical computer, and these virtual machines carry out work just as real computing machine, can installing operating system and application program on virtual machine, virtual machine is addressable network resource also.For the application program of moving in virtual machine, similarly be in real computing machine, to carry out work.
Hardware layer: the hardware platform of virtualized environment operation.Wherein, hardware layer can comprise multiple hardwares, for example the hardware layer of certain computing node can comprise processor (for example CPU) and storer (for example internal memory), can also comprise network interface card, storer etc. high speed/low speed I/O (I/O, Input/Output) equipment, and there is the miscellaneous equipment of particular procedure function, as input and output memory management unit (IOMMU, Input/Output Memory Management Unit), wherein IOMMU can be used for the conversion of virtual machine physical address and Host physical address.
Host (Host): as administration and supervision authorities, in order to complete management, the distribution of hardware resource; For virtual machine presents virtual hardware platform; Realize scheduling and the isolation of virtual machine.Wherein, Host may be monitor of virtual machine (VMM, Virtual Machine Monitor); In addition, VMM and 1 franchise virtual machine coordinate sometimes, and both are in conjunction with forming Host.Wherein, virtual hardware platform provides various hardware resources to each virtual machine of operation on it, as virtual cpu, internal memory, virtual disk, Microsoft Loopback Adapter etc. are provided.Wherein, this virtual disk can corresponding Host a file or a logical block equipment.Virtual machine operates on the virtual hardware platform that Host is its preparation, the one or more virtual machines of the upper operation of Host.
The embodiment of the present invention provides a kind of resource regulating method of virtual cluster, and the method is applied to system of virtual cluster.
Please refer to Fig. 2, described system of virtual cluster comprises management node 210 and at least one computing node 220; Described computing node 220 comprises hardware layer, operates in the host (Host) on described hardware layer and operate at least one virtual machine on described Host.
Please refer to Fig. 3, described method comprises:
310, the optimum equalization state of the current virtual cluster of management node simulation.
In the present embodiment method, first by the optimum equalization state of the current whole virtual cluster of management node simulation, this simulation steps can be carried out by clocked flip, or is carried out by Event triggered, does not specifically limit.Management node supposes virtual machine equilibriums all in cluster to be assigned on all operating computing nodes in analog computation process, the operating load that makes each computing node is identical or approaching to such an extent as to can not further optimize, and thinks that simulation has obtained optimum equalization state.
320, to described virtual cluster, the resource service condition under optimum equalization state is assessed.
Described resource service condition can comprise central processing unit (central processing unit, CPU), internal memory (MEMORY, MEM), the storage IO(Input/Output of computing node, input and output), the service condition of network I/O etc.Take herein this resource service condition comprises that the service condition of CPU and internal memory is example, the CPU capacity of supposing a computing node is 10GHz, memory size is 10GB, and cpu resource and memory source that every the virtual machine moving on this computing node on average takies in a period of time are in the past respectively 2GHz and 2GB; If operation has 3 such virtual machines on this computing node, the CPU that the resource use amount of this computing node is 6GHz and the internal memory of 6GB, resource utilization is 60%.In like manner, can obtain resource use amount and the resource utilization of whole virtual cluster.Can set in advance a threshold value, for example 45% to 75%, by the resource utilization of virtual cluster and this threshold value are compared, assess the resource service condition of virtual cluster under optimum equalization state.If the resource utilization of virtual cluster under optimum equalization state in threshold range, is evaluated as resource and uses suitably; If exceeded the highest threshold values of threshold range as 75%, be evaluated as resource and overuse; If, be evaluated as resource and use not enough as 45% lower than the minimum threshold values of threshold range.
If 330 assessment results are resource, use suitably, carry out load balance scheduling; If assessment result is resource, overuse, scheduling powers on; If assessment result is resource, use deficiency, carry out lower electricity scheduling.
According to assessment result, be suitable, excessive or not enough, dispatch accordingly respectively.For example, if assessment result is resource, use suitably, enter load balance scheduling process, produce virtual machine (vm) migration suggestion; If being resource, assessment result overuses, explanation need to be opened new computing node, enter the scheduling process that powers on, produce the suggestion that powers on and also produce virtual machine (vm) migration suggestion to the part virtual machine (vm) migration on the heavier computing node of other load is arrived to the computing node of this new unlatching simultaneously; If assessment result is resource to be used not enoughly, explanation can closed portion computing node, enters lower electric scheduling process, produce that lower electricity suggestion produces virtual machine (vm) migration suggestion simultaneously in case by the virtual machine (vm) migration on this computing node to be closed to other computing node.By carrying out the suggestion of above-mentioned virtual machine (vm) migration, power on suggestion or lower electricity suggestion etc., complete dynamic resource scheduling.
Above, the embodiment of the present invention provides a kind of resource regulating method of virtual cluster, the method adopts carries out optimum equalization state simulation to virtual cluster, resource service condition under assessment optimum equalization state, according to assessment result, decide and carry out load balancing or power on or lower electric technical scheme of dispatching, based on the judgement of virtual cluster integrality is dispatched to assessment, can provide more accurately sooner scheduling suggestion, can also avoid load balancing and DPM in DRS same taken turns scheduling to provide contrary scheduling suggestion, thereby make virtual cluster reach fast steady state (SS), solve existing DRS method and can increase virtual machine (vm) migration number of times, extend scheduling time, increase the technical matters of moving costs.
In a kind of embodiment, the simulation steps of the optimum equalization state of the current described virtual cluster of the simulation of management node described in above-mentioned 310 specifically can comprise:
A virtual machine (vm) migration on the highest computing node of resource use amount, to the minimum computing node of resource use amount, and is repeated to this step, until do not have virtual machine to move, think that described virtual cluster reaches optimum equalization state.
Or, can comprise:
First, a virtual machine (vm) migration on the highest computing node of resource use amount, to the minimum computing node of resource use amount, and is repeated to this step, until do not have virtual machine to move; Then, take the virtual machine that resource is maximum by one on the highest computing node of resource use amount, and take the minimum virtual machine of resource, mutually exchange migration on the minimum computing node of resource use amount, and repeat this step, until do not have the virtual machine can exchange migration; Think that described virtual cluster reaches optimum equalization state.
In concrete application, can comprise:
A1, obtains every virtual machine in virtual cluster resource of for example on average taking in 5 minutes of a period of time in the past, that is, resource mean allocation value, for example, every virtual machine takies cpu resource 2GHz and committed memory resource 2GB.Meanwhile, obtain the average utilization avgU of resource in virtual cluster, that is, whole ratio of resources in the resource that all virtual machines take and virtual cluster, for example 60%.
A2, searches resource occupation amount or the highest computing node of utilization factor in virtual cluster, uses host
irepresent.Can be from host
ioptimum virtual machine vm of middle selection
ii, that is, take the virtual machine that stock number is maximum, make, by this virtual machine vm
iiafter moving out, host
iresource utilization approach average utilization avgU most.
A3 is virtual machine vm
iiselect Optimal calculation node, for example host
j, this computing node can be resource use amount or the minimum computing node of utilization factor in virtual cluster.By virtual machine vm
iimove to host
j, can make the equilibrium degree of virtual cluster lower, that is, more balanced.Wherein, when selecting Optimal calculation node, can further consider the constraint conditions such as placement location rule between virtual machine and computing node.
By repeating above-mentioned steps A2 and A3, can make the equilibrium degree of virtual cluster constantly reduce, when not having virtual machine to move, can think, virtual cluster has reached optimum equalization state.Preferably, after steps A 3, may further include:
A4, selects a computing node the computing node from resource utilization higher than avgU, for example resource utilization or the highest computing node of resource use amount represent with srcHost herein; Meanwhile, also from resource utilization, the computing node lower than avgU, select a computing node, for example resource utilization or the minimum computing node of resource use amount represent with dstHost herein; Then, from srcHost, select one to take the maximum or more virtual machine (vm) migration of resource to dstHost, from dstHost, select one to take the minimum or less virtual machine (vm) migration of resource to srcHost; By constantly repeating this exchange migration, can make the equilibrium degree of virtual cluster further reduce, when there is no virtual machine can exchange migration time, can think, virtual cluster has reached optimum equalization state.
Accordingly, the appraisal procedure that the resource service condition under optimum equalization state is assessed to described virtual cluster in above-mentioned 320 specifically can comprise:
B1, obtains following parameter: every virtual machine in virtual cluster is the resource that for example on average takies in 5 minutes of a period of time in the past, that is, and and resource mean allocation value; And, high utilization factor empirical value highestUtil, low year utilization factor empirical value lowestUtil, focus computing node quantity HotHostNum, the cold spot computing node quantity ColdHostNum of carrying.Wherein, highestUtil and lowestUtil can rule of thumb arrange, or determine according to historical experience by self study mode; The initial value of HotHostNum and ColdHostNum is 0.Described focus computing node refers to that resource utilization surpasses the high computing node that carries utilization factor empirical value, and described cold spot computing node refers to that resource utilization is lower than the computing node of low year utilization factor empirical value.
B2, determines the quantity of focus computing node and cold spot computing node under optimum equalization state.The cluster resource of take comprises that CPU and memory source are example, and the method for determining specifically comprises:
Successively by the cpu busy percentage cpuUtil of each computing node
iwith memory usage memUtil
icarry utilization factor empirical value highestUtil with height and compare, if cpuUtil
i-highestUtil>0 or memUtil
i-highestUtil>0, thinks that this computing node is focus computing node, so make HotHostNum++ make HotHostNum=1; All computing node judgement is complete, obtains the quantity of focus computing node.
Similarly, successively by the cpu busy percentage cpuUtil of each computing node
iwith memory usage memUtil
icompare with low year utilization factor empirical value lowestUtil, if lowestUtil-cpuUtil
i>0 and lowestUtil-memUtil
i>0, thinks that this computing node is cold spot computing node, so make ColdHostNum++ make ColdHostNum=1; All computing node judgement is complete, obtains the quantity of cold spot computing node.
B3, resource service condition to virtual cluster is assessed and is divided three classes: if the quantity ColdHostNum of the quantity HotHostNum of focus computing node and cold spot computing node is zero, judge the resource use of described virtual cluster under optimum equalization state suitably; If focus computing node quantity HotHostNum is greater than zero, judge that the resource of described virtual cluster under optimum equalization state overuses; If cold spot computing node quantity ColdHostNum is greater than zero, judge that the resource use of described virtual cluster under optimum equalization state is not enough.
By adopting above-mentioned steps, can simply realize the resource service condition under optimum equalization state of virtual cluster is assessed and classified, follow-up, according to classification situation, carry out corresponding scheduling.
In another kind of embodiment, the simulation steps of the optimum equalization state of the current described virtual cluster of the simulation of management node described in above-mentioned 310 specifically can comprise:
C0, obtains the resource service condition of each virtual machine in virtual cluster and every resource situation that computing node can be used;
C1, sorts all virtual machines in described virtual cluster according to the size of its resource use amount;
C2, according to resource use amount order from big to small, is that each virtual machine finds the Optimal calculation node that can place, makes after each virtual machine is placed to searched out Optimal calculation node, and the equilibrium state of whole virtual cluster is optimum;
According to the operation of C2, successively each virtual machine is placed on the Optimal calculation node searching out separately, thereby makes described virtual cluster reach optimum equalization state.
Accordingly, the appraisal procedure that in above-mentioned 320, the resource service condition under optimum equalization state is assessed to described virtual cluster can also comprise: in the process of the Optimal calculation node that can place for each virtual machine searching described in C2, if the Optimal calculation node that has virtual machine not search out can to place, can judge that the resource of described virtual cluster under optimum equalization state overuses.
If all virtual machines have all found and have been placed on Optimal calculation node separately, described appraisal procedure specifically can be identical with a upper embodiment,, comprises above-mentioned B1, B2 and B3 etc. that is.
By an application scenarios example, said method is described further below:
Supposing has 4 computing nodes in virtual cluster, the upper operation of computing node pm1 has 5 virtual machine vm1-vm5, the upper operation of computing node pm2 has powered on operation but in light condition, computing node pm4 is in power-down state of 5 virtual machine vm6-vm10, computing node pm3.As shown in table 1, wherein each computing node all has the cpu resource of 10GHz and the internal memory of 10GB (MEM) resource, and each virtual machine is the resource that for example on average takies in 5 minutes of a period of time in the past,, resource mean allocation value is to take cpu resource 2GHz and committed memory resource 2GB.
| ? | CPU | MEM |
| Computing node capacity | 10GHz | 10GB |
| Virtual machine specification | 2GHz | 2GB |
Simulation steps to the optimum equalization state of virtual cluster comprises:
D1, obtain the resource mean allocation value of every virtual machine as shown in Table 1, and computational resource utilization factor is as follows: the cpu busy percentage of computing node pm1 and pm2 is that 100%, MEM utilization factor is 100%, and the CPU of computing node pm3 and MEM utilization factor are all 0%.Now computing node pm1 and pm2 are the focus computing node of " falseness ", and pm3 is the cold spot computing node of " falseness ".The all computing node average utilizations of cluster avgU=(100%+100%+0)/3=66.7%.
D2, from the highest computing node of resource utilization for example pm1 select for example vm1 of a virtual machine, vm1 is moved out.
D3, simulation is by the virtual machine vm1 for example pm3 of Optimal calculation node that moves into, and after moving into, the CPU of computing node pm1 and MEM utilization factor are all 80%, and the CPU of computing node pm3 and MEM utilization factor are all 20%, and the CPU of computing node pm2 and MEM utilization factor are also 100%;
D4, the result based on step D3, repeats D2-D3, from computing node pm2, select virtual machine vm6 simulation to move on computing node pm3, after moving into, the CPU of computing node pm1 and pm2 and MEM utilization factor are all 80%, and the CPU of computing node pm3 and MEM utilization factor are all 40%;
D5, the result based on step D4, repeats D2-D3, from computing node pm1, select virtual machine vm2 simulation to move on computing node pm3, after moving into, the CPU of computing node pm1 and pm3 and MEM utilization factor are all 60%, and the CPU of computing node pm2 and MEM utilization factor are all 80%.Now, think that virtual cluster has reached optimum equalization state.
Accordingly, the appraisal procedure that the resource service condition under optimum equalization state is assessed to virtual cluster can comprise:
E1, supposes the high utilization factor empirical value highestUtil=75% of carrying, low year utilization factor empirical value lowestUtil=45%, the quantity HotHostNum=0 of focus computing node, the quantity ColdHostNum=0 of cold spot computing node.
E2, for computing node pm2, cpuUtil
2-highestUtil>0, memUtil
2-highestUtil>0, so HotHostNum=HotHostNum+1=1; For computing node pm1 and pm3, cpuUtil
i-highestUtil<0, memUtil
i-highestUtil<0, lowestUtil – cpuUtil
i<0, lowestUtil – memUtil
i<0; Therefore final HotHostNum=1, ColdHostNum=0.
E3, assesses the resource service condition of virtual cluster: because HotHostNum>0 overuses so be judged as resource, so, need to first solve hot issue, so epicycle DRS dispatches power on assessment and scheduling.
Enter and power on after assessment, input parameter adopts the current real resource data of virtual cluster, but not simulation data later, input parameter is that the cpu busy percentage of computing node pm1 and pm2 is 100%, MEM utilization factor is 100%, and the CPU of computing node pm3 and MEM utilization factor are all 0%.Power on after scheduling, the computing node pm4 operation that powers on, part virtual machine is migrated to pm3 and pm4 is upper moves.
Above, to assess after virtual cluster, enter and power on assessment and be scheduling to example, technical solution of the present invention is illustrated.
In another kind of application scenarios example, supposing has 2 virtual machine vm1～vm2 on computing node pm1, computing node pm2 is unloaded, on computing node pm3 and pm4, respectively there are a virtual machine vm3 and vm4, simulate after virtual cluster optimum equalization state, on pm1～pm4, respectively have a virtual machine, the CPU of each computing node and MEM utilization factor are all 20%.Carry out, after the assessment of resource service condition, calculating ColdHostNum=4, therefore, finally enter lower electricity assessment and scheduling.Lower electricity assessment and scheduling are based on current real resources data, be on pm1, to have 2 virtual machines, pm2 is unloaded, on pm3 and pm4, respectively there is a virtual machine, start to carry out, trial is electricity under 4 candidate's computing nodes, finally can 2 computing node of lower electricity success, and 4 virtual machines finally operate on another two computing nodes.
To sum up, the embodiment of the present invention provides a kind of resource regulating method of virtual cluster, the method adopts carries out optimum equalization state simulation to virtual cluster, resource service condition under assessment optimum equalization state, according to assessment result, decide and carry out load balancing or power on or lower electric technical scheme of dispatching, based on the judgement of virtual cluster integrality is dispatched to assessment, can provide more accurately sooner scheduling suggestion, can also avoid load balancing and DPM in DRS same taken turns scheduling to provide contrary scheduling suggestion, thereby make virtual cluster reach fast steady state (SS), solve existing DRS method and can increase virtual machine (vm) migration number of times, extend scheduling time, increase the technical matters of moving costs.
Please refer to Fig. 4, the embodiment of the present invention provides a kind of management node, for virtual cluster.
Described virtual cluster comprises described management node and at least one computing node, and described computing node comprises hardware layer, operates in the host on described hardware layer and operates at least one virtual machine on described host;
Described management node comprises:
In a kind of embodiment, described analog module 410 can comprise: the first analogue unit, for by a virtual machine (vm) migration on the highest computing node of resource use amount to the minimum computing node of resource use amount, and repeat this step, until do not have virtual machine to move.
In another kind of embodiment, described analog module 410 can comprise: the first analogue unit, for by a virtual machine (vm) migration on the highest computing node of resource use amount to the minimum computing node of resource use amount, and repeat this step, until do not have virtual machine to move; The second analogue unit, for taking the virtual machine that resource is maximum by one on the highest computing node of resource use amount, and take the minimum virtual machine of resource, mutually exchange migration on the minimum computing node of resource use amount, and repeat this step, until do not have the virtual machine can exchange migration.
In another embodiment, described analog module 410 can comprise: the 3rd analogue unit, for all virtual machines of described virtual cluster are sorted according to the size of its resource use amount; According to resource use amount order from big to small, be that each virtual machine finds the Optimal calculation node that can place, make after each virtual machine is placed to searched out Optimal calculation node, the equilibrium state of whole virtual cluster is optimum.
In a kind of embodiment, described evaluation module 420 can comprise: determining unit, quantity for focus computing node and cold spot computing node under definite optimum equalization state, described focus computing node refers to that resource utilization surpasses the high computing node that carries utilization factor empirical value, and described cold spot computing node refers to that resource utilization is lower than the computing node of low year utilization factor empirical value; Judging unit, if be all zero for the quantity of focus computing node and cold spot computing node, judges the resource use of described virtual cluster under optimum equalization state suitably; If focus computing node quantity is greater than zero, judge that the resource of described virtual cluster under optimum equalization state overuses; If cold spot computing node quantity is greater than zero, judge that the resource use of described virtual cluster under optimum equalization state is not enough.
Optionally, if described analog module 410 comprises the 3rd analogue unit, described the 3rd analogue unit is found in the process of the Optimal calculation node that can place for each virtual machine, if the Optimal calculation node that has virtual machine not search out can to place, described evaluation module 420 can judge that the resource of described virtual cluster under optimum equalization state used not enough.
To sum up, the embodiment of the present invention provides a kind of management node, for virtual cluster is carried out to scheduling of resource, employing is carried out optimum equalization state simulation to virtual cluster, resource service condition under assessment optimum equalization state, according to assessment result, decide and carry out load balancing or power on or lower electric technical scheme of dispatching, based on the judgement of virtual cluster integrality is dispatched to assessment, can provide more accurately sooner scheduling suggestion, can also avoid load balancing and DPM in DRS same taken turns scheduling to provide contrary scheduling suggestion, thereby make virtual cluster reach fast steady state (SS), solve existing DRS method and can increase virtual machine (vm) migration number of times, extend scheduling time, increase the technical matters of moving costs.
The embodiment of the present invention also provides a kind of computer-readable storage medium, and this computer-readable storage medium can have program stored therein, and this program comprises the part or all of step of the resource regulating method of the virtual cluster of recording in said method embodiment while carrying out.
Please refer to Fig. 5, the embodiment of the present invention also provides a kind of management node, for virtual cluster.
Described virtual cluster comprises described management node and at least one computing node, and described computing node comprises hardware layer, operates in the host on described hardware layer and operates at least one virtual machine on described host; Described management node comprises: in input media 510, output unit 520, storer 530 and the processor 540(network equipment, the quantity of processor can be one or more, and the processor of take in Fig. 5 is example).In some embodiments of the invention, input media 510, output unit 520, storer 530 and processor 540 can be connected by bus or alternate manner, wherein, in Fig. 5 to be connected to example by bus.
Wherein, processor 540 is carried out following steps: the optimum equalization state of simulating current described virtual cluster; Resource service condition to described virtual cluster under optimum equalization state is assessed; If assessment result is resource, use suitably, carry out load balance scheduling; If assessment result is resource, overuse, scheduling powers on; If assessment result is resource, use deficiency, carry out lower electricity scheduling.
In some embodiments of the invention, the current optimum equalization state of the processor 540 described virtual cluster of simulation can comprise: by a virtual machine (vm) migration on the highest computing node of resource use amount to the minimum computing node of resource use amount, and repeat this step, until do not have virtual machine to move, think that described virtual cluster reaches optimum equalization state.
In some embodiments of the invention, the current optimum equalization state of the processor 540 described virtual cluster of simulation can comprise: first, by a virtual machine (vm) migration on the highest computing node of resource use amount to the minimum computing node of resource use amount, and repeat this step, until do not have virtual machine to move; Then, take the virtual machine that resource is maximum by one on the highest computing node of resource use amount, and take the minimum virtual machine of resource, mutually exchange migration on the minimum computing node of resource use amount, and repeat this step, until do not have the virtual machine can exchange migration; Think that described virtual cluster reaches optimum equalization state.
In some embodiments of the invention, the current optimum equalization state of the processor 540 described virtual cluster of simulation can comprise: all virtual machines in described virtual cluster are sorted according to the size of its resource use amount; According to resource use amount order from big to small, be that each virtual machine finds the Optimal calculation node that can place, make after each virtual machine is placed to searched out Optimal calculation node, the equilibrium state of whole virtual cluster is optimum.
In some embodiments of the invention, the resource service condition of 540 pairs of described virtual clusters of processor under optimum equalization state assessed and comprised: the quantity of determining focus computing node and cold spot computing node under optimum equalization state, described focus computing node refers to that resource utilization surpasses the high computing node that carries utilization factor empirical value, and described cold spot computing node refers to that resource utilization is lower than the computing node of low year utilization factor empirical value; If the quantity of focus computing node and cold spot computing node is all zero, judge the resource use of described virtual cluster under optimum equalization state suitably; If focus computing node quantity is greater than zero, judge that the resource of described virtual cluster under optimum equalization state overuses; If cold spot computing node quantity is greater than zero, judge that the resource use of described virtual cluster under optimum equalization state is not enough.
In some embodiments of the invention, processor 540 is found in the process of the Optimal calculation node that can place for each virtual machine, if the Optimal calculation node that has virtual machine not search out can to place, judges that the resource of described virtual cluster under optimum equalization state overuses.
To sum up, the embodiment of the present invention provides a kind of management node, for virtual cluster is carried out to scheduling of resource, employing is carried out optimum equalization state simulation to virtual cluster, resource service condition under assessment optimum equalization state, according to assessment result, decide and carry out load balancing or power on or lower electric technical scheme of dispatching, based on the judgement of virtual cluster integrality is dispatched to assessment, can provide more accurately sooner scheduling suggestion, can also avoid load balancing and DPM in DRS same taken turns scheduling to provide contrary scheduling suggestion, thereby make virtual cluster reach fast steady state (SS), solve existing DRS method and can increase virtual machine (vm) migration number of times, extend scheduling time, increase the technical matters of moving costs.
The function of each functional module that is appreciated that the management node of the embodiment of the present invention can be according to the method specific implementation in said method embodiment, and its specific implementation process can, with reference to the associated description in said method embodiment, repeat no more herein.
In the above-described embodiments, the description of each embodiment is all emphasized particularly on different fields, in certain embodiment, there is no the part of detailed description, can be referring to the associated description of other embodiment.
It should be noted that, for aforesaid each embodiment of the method, for simple description, therefore it is all expressed as to a series of combination of actions, but those skilled in the art should know, the present invention is not subject to the restriction of described sequence of movement, because according to the present invention, some step can adopt other order or carry out simultaneously.Secondly, those skilled in the art also should know, the embodiment described in instructions all belongs to preferred embodiment, and related action and module might not be that the present invention is necessary.
One of ordinary skill in the art will appreciate that all or part of step in the whole bag of tricks of above-described embodiment can complete by hardware, also can complete by the relevant hardware of programmed instruction, this program can be stored in a computer-readable recording medium, and storage medium can comprise: ROM (read-only memory), random-access memory, disk or CD etc.
Resource regulating method and the device of the virtual the cluster above embodiment of the present invention being provided are described in detail, but the explanation of above embodiment is just understood method of the present invention and core concept thereof for helping, and should not be construed as limitation of the present invention.In the technical scope that those skilled in the art disclose in the present invention, the variation that can expect easily or replacement, within all should being encompassed in protection scope of the present invention.
Claims (12)
1. the resource regulating method of a virtual cluster, it is characterized in that, described virtual cluster comprises management node and at least one computing node, and described computing node comprises hardware layer, operates in the host on described hardware layer and operates at least one virtual machine on described host; Described method comprises:
The optimum equalization state of the current described virtual cluster of described management node simulation;
Resource service condition to described virtual cluster under optimum equalization state is assessed;
If assessment result is resource, use suitably, carry out load balance scheduling;
If assessment result is resource, overuse, scheduling powers on;
If assessment result is resource, use deficiency, carry out lower electricity scheduling.
2. method according to claim 1, is characterized in that, the current optimum equalization state of the described management node described virtual cluster of simulation comprises:
A virtual machine (vm) migration on the highest computing node of resource use amount, to the minimum computing node of resource use amount, and is repeated to this step, until do not have virtual machine to move, think that described virtual cluster reaches optimum equalization state.
3. method according to claim 1, is characterized in that, the current optimum equalization state of the described management node described virtual cluster of simulation comprises:
First, a virtual machine (vm) migration on the highest computing node of resource use amount, to the minimum computing node of resource use amount, and is repeated to this step, until do not have virtual machine to move;
Then, take the virtual machine that resource is maximum by one on the highest computing node of resource use amount, and take the minimum virtual machine of resource, mutually exchange migration on the minimum computing node of resource use amount, and repeat this step, until do not have the virtual machine can exchange migration; Think that described virtual cluster reaches optimum equalization state.
4. method according to claim 1, is characterized in that, the current optimum equalization state of the described management node described virtual cluster of simulation comprises:
All virtual machines in described virtual cluster are sorted according to the size of its resource use amount;
According to resource use amount order from big to small, be that each virtual machine finds the Optimal calculation node that can place, make after each virtual machine is placed to searched out Optimal calculation node, the equilibrium state of whole virtual cluster is optimum.
5. according to the method described in any one in claim 1-4, it is characterized in that, described to described virtual cluster the resource service condition under optimum equalization state assess and comprise:
Determine the quantity of focus computing node and cold spot computing node under optimum equalization state, described focus computing node refers to that resource utilization surpasses the high computing node that carries utilization factor empirical value, and described cold spot computing node refers to that resource utilization is lower than the computing node of low year utilization factor empirical value;
If the quantity of focus computing node and cold spot computing node is all zero, judge the resource use of described virtual cluster under optimum equalization state suitably;
If focus computing node quantity is greater than zero, judge that the resource of described virtual cluster under optimum equalization state overuses;
If cold spot computing node quantity is greater than zero, judge that the resource use of described virtual cluster under optimum equalization state is not enough.
6. according to the method described in claim 4 or 5, it is characterized in that:
Describedly find in the process of the Optimal calculation node that can place for each virtual machine, if the Optimal calculation node that has virtual machine not search out can to place judges that the resource of described virtual cluster under optimum equalization state overuses.
7. a management node, it is characterized in that, for virtual cluster, described virtual cluster comprises described management node and at least one computing node, and described computing node comprises hardware layer, operates in the host on described hardware layer and operates at least one virtual machine on described host; Described management node comprises:
Analog module, for simulating the optimum equalization state of current described virtual cluster;
Evaluation module, for to described virtual cluster the resource service condition under optimum equalization state assess;
Scheduler module, if be that resource is used suitably for the assessment result of described evaluation module, carries out load balance scheduling; If the assessment result of described evaluation module is resource, overuse, scheduling powers on; If the assessment result of described evaluation module is resource, use deficiency, carry out lower electricity scheduling.
8. management node according to claim 7, is characterized in that, described analog module comprises:
The first analogue unit, for by a virtual machine (vm) migration on the highest computing node of resource use amount to the minimum computing node of resource use amount, and repeat this step, until do not have virtual machine to move.
9. management node according to claim 7, is characterized in that, described analog module comprises:
The first analogue unit, for by a virtual machine (vm) migration on the highest computing node of resource use amount to the minimum computing node of resource use amount, and repeat this step, until do not have virtual machine to move;
The second analogue unit, for taking the virtual machine that resource is maximum by one on the highest computing node of resource use amount, and take the minimum virtual machine of resource, mutually exchange migration on the minimum computing node of resource use amount, and repeat this step, until do not have the virtual machine can exchange migration.
10. management node according to claim 7, is characterized in that, described analog module comprises:
The 3rd analogue unit, for sorting all virtual machines of described virtual cluster according to the size of its resource use amount; According to resource use amount order from big to small, be that each virtual machine finds the Optimal calculation node that can place, make after each virtual machine is placed to searched out Optimal calculation node, the equilibrium state of whole virtual cluster is optimum.
11. according to arbitrary described management node in claim 7 to 10, it is characterized in that, described evaluation module comprises:
Determining unit, quantity for focus computing node and cold spot computing node under definite optimum equalization state, described focus computing node refers to that resource utilization surpasses the high computing node that carries utilization factor empirical value, and described cold spot computing node refers to that resource utilization is lower than the computing node of low year utilization factor empirical value;
Judging unit, if be all zero for the quantity of focus computing node and cold spot computing node, judges the resource use of described virtual cluster under optimum equalization state suitably; If focus computing node quantity is greater than zero, judge that the resource of described virtual cluster under optimum equalization state overuses; If cold spot computing node quantity is greater than zero, judge that the resource use of described virtual cluster under optimum equalization state is not enough.
12. according to the management node described in claim 10 or 11, it is characterized in that:
Described the 3rd analogue unit is found in the process of the Optimal calculation node that can place for each virtual machine, if the Optimal calculation node that has virtual machine not search out can to place, the resource of the described virtual cluster of described evaluation module judgement under optimum equalization state used not enough.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310689439.5A CN103677958B (en) | 2013-12-13 | 2013-12-13 | The resource regulating method and device of a kind of virtual cluster |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310689439.5A CN103677958B (en) | 2013-12-13 | 2013-12-13 | The resource regulating method and device of a kind of virtual cluster |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103677958A true CN103677958A (en) | 2014-03-26 |
| CN103677958B CN103677958B (en) | 2017-06-20 |
Family
ID=50315615
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310689439.5A Expired - Fee Related CN103677958B (en) | 2013-12-13 | 2013-12-13 | The resource regulating method and device of a kind of virtual cluster |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103677958B (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104184813A (en) * | 2014-08-20 | 2014-12-03 | 杭州华为数字技术有限公司 | Load balancing method of virtual machines, related equipment and trunking system |
| CN104391735A (en) * | 2014-11-14 | 2015-03-04 | 深信服网络科技（深圳）有限公司 | Method and system for scheduling virtual machines in virtualized all-in-one machine cluster |
| CN104484220A (en) * | 2014-11-28 | 2015-04-01 | 杭州华为数字技术有限公司 | Method and device for dispatching dynamic resources of virtual cluster |
| CN105187512A (en) * | 2015-08-13 | 2015-12-23 | 航天恒星科技有限公司 | Method and system for load balancing of virtual machine clusters |
| CN106125888A (en) * | 2016-06-23 | 2016-11-16 | 河海大学常州校区 | The efficient power-economizing method of the utilization of resources based on virtual machine (vm) migration in cloud data center |
| CN106161539A (en) * | 2015-04-12 | 2016-11-23 | 北京典赞科技有限公司 | Schedule creating energy conservation optimizing method based on the fictitious host computer of ARM server |
| CN106445636A (en) * | 2016-09-28 | 2017-02-22 | 郑州云海信息技术有限公司 | Dynamic resource scheduling algorithm under PAAS platform |
| CN106817243A (en) * | 2015-12-01 | 2017-06-09 | 广达电脑股份有限公司 | Management system and management method for server resources |
| CN106959884A (en) * | 2017-03-23 | 2017-07-18 | 联想(北京)有限公司 | A kind of data processing method, electronic equipment and cluster |
| CN107197039A (en) * | 2017-06-30 | 2017-09-22 | 郑州云海信息技术有限公司 | A kind of PAAS platform service bag distribution methods and system based on CDN |
| CN107566457A (en) * | 2017-08-09 | 2018-01-09 | 天脉聚源（北京）科技有限公司 | A kind of method and device of balanced node device load |
| CN109656711A (en) * | 2018-10-30 | 2019-04-19 | 山东超越数控电子股份有限公司 | A kind of dynamic power supply management method and system based on server cluster |
| CN111158909A (en) * | 2019-12-27 | 2020-05-15 | 中国联合网络通信集团有限公司 | Cluster resource allocation processing method, device, equipment and storage medium |
| CN111367677A (en) * | 2020-03-31 | 2020-07-03 | 中国工商银行股份有限公司 | Cluster resource allocation method and system |
| CN113032102A (en) * | 2021-04-07 | 2021-06-25 | 广州虎牙科技有限公司 | Resource rescheduling method, device, equipment and medium |
| CN115373862A (en) * | 2022-10-26 | 2022-11-22 | 安超云软件有限公司 | Dynamic resource scheduling method, system and storage medium based on data center |
| CN116974700B (en) * | 2023-08-16 | 2024-04-09 | 北京志凌海纳科技有限公司 | Method, system, equipment and storage medium for realizing dynamic balance of resources |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100070784A1 (en) * | 2008-09-15 | 2010-03-18 | Vmware, Inc. | Reducing Power Consumption in a Server Cluster |
| CN102388380A (en) * | 2009-04-08 | 2012-03-21 | 微软公司 | Optimized virtual machine migration mechanism |
| CN102662754A (en) * | 2012-04-20 | 2012-09-12 | 浙江大学 | Multi-field supportable virtual machine dispatching device and multi-field supportable virtual machine dispatching method |
| US20130145364A1 (en) * | 2011-12-01 | 2013-06-06 | Tunghai University | Method of dynamic resource allocation for a virtual machine cluster |
| CN103440160A (en) * | 2013-08-15 | 2013-12-11 | 华为技术有限公司 | Virtual machine recovering method and virtual machine migration method , device and system |
-
2013
- 2013-12-13 CN CN201310689439.5A patent/CN103677958B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100070784A1 (en) * | 2008-09-15 | 2010-03-18 | Vmware, Inc. | Reducing Power Consumption in a Server Cluster |
| CN102388380A (en) * | 2009-04-08 | 2012-03-21 | 微软公司 | Optimized virtual machine migration mechanism |
| US20130145364A1 (en) * | 2011-12-01 | 2013-06-06 | Tunghai University | Method of dynamic resource allocation for a virtual machine cluster |
| CN102662754A (en) * | 2012-04-20 | 2012-09-12 | 浙江大学 | Multi-field supportable virtual machine dispatching device and multi-field supportable virtual machine dispatching method |
| CN103440160A (en) * | 2013-08-15 | 2013-12-11 | 华为技术有限公司 | Virtual machine recovering method and virtual machine migration method , device and system |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104184813B (en) * | 2014-08-20 | 2018-03-09 | 杭州华为数字技术有限公司 | The load-balancing method and relevant device and group system of virtual machine |
| CN104184813A (en) * | 2014-08-20 | 2014-12-03 | 杭州华为数字技术有限公司 | Load balancing method of virtual machines, related equipment and trunking system |
| CN104391735A (en) * | 2014-11-14 | 2015-03-04 | 深信服网络科技（深圳）有限公司 | Method and system for scheduling virtual machines in virtualized all-in-one machine cluster |
| CN104391735B (en) * | 2014-11-14 | 2018-11-06 | 深信服网络科技（深圳）有限公司 | Virtualize dispatching method of virtual machine and system in all-in-one machine cluster |
| CN104484220A (en) * | 2014-11-28 | 2015-04-01 | 杭州华为数字技术有限公司 | Method and device for dispatching dynamic resources of virtual cluster |
| CN104484220B (en) * | 2014-11-28 | 2017-12-22 | 杭州华为数字技术有限公司 | The method and device of the dynamic resource scheduling of virtual cluster |
| CN106161539A (en) * | 2015-04-12 | 2016-11-23 | 北京典赞科技有限公司 | Schedule creating energy conservation optimizing method based on the fictitious host computer of ARM server |
| CN105187512A (en) * | 2015-08-13 | 2015-12-23 | 航天恒星科技有限公司 | Method and system for load balancing of virtual machine clusters |
| CN106817243A (en) * | 2015-12-01 | 2017-06-09 | 广达电脑股份有限公司 | Management system and management method for server resources |
| CN106125888A (en) * | 2016-06-23 | 2016-11-16 | 河海大学常州校区 | The efficient power-economizing method of the utilization of resources based on virtual machine (vm) migration in cloud data center |
| CN106125888B (en) * | 2016-06-23 | 2018-12-25 | 河海大学常州校区 | The efficient power-economizing method of the utilization of resources in cloud data center based on virtual machine (vm) migration |
| CN106445636A (en) * | 2016-09-28 | 2017-02-22 | 郑州云海信息技术有限公司 | Dynamic resource scheduling algorithm under PAAS platform |
| CN106445636B (en) * | 2016-09-28 | 2019-08-02 | 郑州云海信息技术有限公司 | A kind of dynamic resource scheduling algorithm under PAAS platform |
| CN106959884A (en) * | 2017-03-23 | 2017-07-18 | 联想(北京)有限公司 | A kind of data processing method, electronic equipment and cluster |
| CN107197039A (en) * | 2017-06-30 | 2017-09-22 | 郑州云海信息技术有限公司 | A kind of PAAS platform service bag distribution methods and system based on CDN |
| CN107197039B (en) * | 2017-06-30 | 2019-08-09 | 郑州云海信息技术有限公司 | A kind of PAAS platform service packet distribution method and system based on CDN |
| CN107566457A (en) * | 2017-08-09 | 2018-01-09 | 天脉聚源（北京）科技有限公司 | A kind of method and device of balanced node device load |
| CN109656711A (en) * | 2018-10-30 | 2019-04-19 | 山东超越数控电子股份有限公司 | A kind of dynamic power supply management method and system based on server cluster |
| CN111158909A (en) * | 2019-12-27 | 2020-05-15 | 中国联合网络通信集团有限公司 | Cluster resource allocation processing method, device, equipment and storage medium |
| CN111158909B (en) * | 2019-12-27 | 2023-07-25 | 中国联合网络通信集团有限公司 | Cluster resource allocation processing method, device, equipment and storage medium |
| CN111367677A (en) * | 2020-03-31 | 2020-07-03 | 中国工商银行股份有限公司 | Cluster resource allocation method and system |
| CN113032102A (en) * | 2021-04-07 | 2021-06-25 | 广州虎牙科技有限公司 | Resource rescheduling method, device, equipment and medium |
| CN113032102B (en) * | 2021-04-07 | 2024-04-19 | 广州虎牙科技有限公司 | Resource rescheduling method, device, equipment and medium |
| CN115373862A (en) * | 2022-10-26 | 2022-11-22 | 安超云软件有限公司 | Dynamic resource scheduling method, system and storage medium based on data center |
| CN116974700B (en) * | 2023-08-16 | 2024-04-09 | 北京志凌海纳科技有限公司 | Method, system, equipment and storage medium for realizing dynamic balance of resources |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103677958B (en) | 2017-06-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103677958A (en) | Virtualization cluster resource scheduling method and device | |
| Xu et al. | Dynamic deployment of virtual machines in cloud computing using multi-objective optimization | |
| CN104270416B (en) | Control method for equalizing load and management node | |
| CN105302630B (en) | A kind of dynamic adjusting method and its system of virtual machine | |
| CN104184813A (en) | Load balancing method of virtual machines, related equipment and trunking system | |
| Riahi et al. | A multi-objective decision support framework for virtual machine placement in cloud data centers: a real case study | |
| CN102591443A (en) | Method, device and system for integrating virtual clusters | |
| CN111966453A (en) | Load balancing method, system, equipment and storage medium | |
| Deng et al. | A data and task co-scheduling algorithm for scientific cloud workflows | |
| CN105373432A (en) | Cloud computing resource scheduling method based on virtual resource state prediction | |
| Mazidi et al. | Autonomic resource provisioning for multilayer cloud applications with K‐nearest neighbor resource scaling and priority‐based resource allocation | |
| CN109447264A (en) | Virtual machine under cloud computing environment based on VHAM-R model places genetic optimization method | |
| CN104461673A (en) | Virtual machine transferring judgment method and device | |
| CN105389196A (en) | Method for processing seismic data by utilizing virtualization technology | |
| Guo et al. | Multi-objective optimization for data placement strategy in cloud computing | |
| Wang et al. | Gmpr: a two-phase heuristic algorithm for virtual machine placement in large-scale cloud data centers | |
| Monteiro et al. | Virtualized Web server cluster self-configuration to optimize resource and power use | |
| CN116647560A (en) | Method, device, equipment and medium for coordinated optimization control of Internet of things computer clusters | |
| Tang | Load balancing optimization in cloud computing based on task scheduling | |
| Wang et al. | An Efficient Load Prediction-Driven Scheduling Strategy Model in Container Cloud | |
| TWI425421B (en) | Load Balancing Design Method of Virtual Machine in Cloud Computing Environment | |
| CN104008011A (en) | Method for balancing resource load of entity machine in cluster environment | |
| Rudra et al. | Models and algorithms for energy conservation in internet of things | |
| CN103218249B (en) | A kind of virtual cluster control method and equipment, system of virtual cluster | |
| Abdullahi et al. | Variable neighborhood search-based symbiotic organisms search algorithm for energy-efficient scheduling of virtual machine in cloud data center |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170620 Termination date: 20181213 |