CN106155782A - Virtual machine migration method and device - Google Patents

Virtual machine migration method and device Download PDF

Info

Publication number
CN106155782A
CN106155782A CN201510165952.3A CN201510165952A CN106155782A CN 106155782 A CN106155782 A CN 106155782A CN 201510165952 A CN201510165952 A CN 201510165952A CN 106155782 A CN106155782 A CN 106155782A
Authority
CN
China
Prior art keywords
physical server
virtual machine
model
timeslice
ratio
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
Application number
CN201510165952.3A
Other languages
Chinese (zh)
Other versions
CN106155782B (en
Inventor
高瞻
杨杰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Lenovo Beijing Ltd
Original Assignee
Lenovo Beijing Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Lenovo Beijing Ltd filed Critical Lenovo Beijing Ltd
Priority to CN201510165952.3A priority Critical patent/CN106155782B/en
Publication of CN106155782A publication Critical patent/CN106155782A/en
Application granted granted Critical
Publication of CN106155782B publication Critical patent/CN106155782B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Abstract

The embodiment of the invention discloses a kind of virtual machine migration method and device, pre-set the timeslice when virtual machine of same virtual cpu core number runs on the physical server of different model and share ratio transformational relation, when carrying out virtual machine (vm) migration, share ratio transformational relation by above-mentioned timeslice and determine candidate physical type of server, and the second timeslice that virtual machine to be migrated is when running on the physical server of candidate physical server model shares ratio, it is then determined that purpose physical server, the model of purpose physical server belongs to described candidate physical server model, the CPU core calculation of purpose physical server is more than or equal to the virtual cpu core number of described virtual machine to be migrated;By in virtual machine (vm) migration to be migrated to purpose physical server, and ratio of the timeslice of virtual machine to be migrated being shared is adjusted to second timeslice corresponding with the model of purpose physical server and shares ratio, reduce the difference of Practical Calculation performance before and after virtual machine (vm) migration.

Description

Virtual machine migration method and device
Technical field
The present invention relates to field of computer technology, more particularly, it relates to a kind of virtual machine migration method and Device.
Background technology
Intel Virtualization Technology is one of key technology of cloud computing.One physical server is virtualized into multiple stage Logic virtual machine, is possible not only to be greatly promoted cloud computing environment and calculates the utilization ratio of resource and save energy consumption, The dynamic migration of Intel Virtualization Technology offer simultaneously, scheduling of resource so that the load of cloud computing service obtains height Effect management and extension, so that the service of cloud computing has more elasticity and motility.
Virtual machine (vm) migration refers to the resource distribution needing to adjust virtual machine between physical server, or, When certain physical server needs repairing, virtual machine is moved to another from a physical server The process of platform physical server.
But, owing to there may be performance difference between physical server, virtual machine is at different physical services When migrating between device, the Practical Calculation performance before and after it migrates also can there are differences.This species diversity can be straight Connect and have influence on the service application that user operates in virtual machine, it is impossible to ensure the grade of service of cloud computing service Agreement (service level agreement, SLA).Therefore, the reality before and after virtual machine (vm) migration how is reduced The difference of border calculated performance becomes problem demanding prompt solution.
Summary of the invention
It is an object of the invention to provide a kind of virtual machine migration method and device, before reducing virtual machine (vm) migration After the difference of Practical Calculation performance.
For achieving the above object, the technical scheme is that
A kind of virtual machine migration method, including:
Obtain the virtual machine to be migrated very first time sheet on the physical server of source and share ratio;
Ratio is shared based on described very first time sheet, the model of source physical server, and preset same The timeslice when virtual machine of virtual cpu core number runs on the physical server of different model shares ratio Example transformational relation determines that candidate physical server model and virtual machine to be migrated are in candidate physical server-type Number physical server on the second timeslice when running share ratio, wherein, described second timeslice is altogether Ratio of enjoying is less than or equal to 1;
Determine that purpose physical server, the model of described purpose physical server belong to described candidate physical clothes Business type number, the CPU core calculation of described purpose physical server is more than or equal to described virtual machine to be migrated Virtual cpu core number;
By in described virtual machine (vm) migration to be migrated to described purpose physical server, and by described void to be migrated Plan machine timeslice in described purpose physical server is shared ratio and is adjusted to and described purpose physical services The second timeslice that the model of device is corresponding shares ratio.
Said method, it is preferred that the virtual machine of described same virtual cpu core number is at the thing of different model Timeslice when running on reason server is shared ratio transformational relation and is included:
Virtual cpu core number is the virtual machine of m computational when running on the physical server of j model Can be equivalent to described virtual cpu core number be the virtual machine of m on the physical server of i model with the time Sheet shares ratio when being calculated performance during 1 operation, and described virtual cpu core number is that the virtual machine of m exists Run required timeslice on the physical server of described j model sharing ratio is tij
Said method, it is preferred that described share ratio, source physical server based on described very first time sheet Model, and the virtual machine of preset same virtual cpu core number is at the physical server of different model Timeslice during upper operation is shared ratio transformational relation and is determined candidate physical server model, and to be migrated The second timeslice when virtual machine runs on the physical server of candidate physical server model shares ratio Including:
If it is not 1 that described very first time sheet shares ratio, then according to the model of described source physical server, institute State very first time sheet and share ratio, and the virtual machine of described preset same virtual cpu core number is not Timeslice when running on the physical server of same model shares ratio transformational relation, determines that transition physics takes Business type number;Wherein, described virtual machine to be migrated is at the physical services of described transition physical server model With the physics of physical server that timeslice ratio is 1 use described transition physical server model on device Calculated performance during CPU time slice, with described virtual machine to be migrated on the physical server of described source with institute State very first time sheet and share computational when ratio uses the physical cpu timeslice of described source physical server Can be identical;
According to described transition physical server model, and described preset same virtual cpu core number Timeslice when virtual machine runs on the physical server of different model is shared ratio transformational relation and is determined time Election server model, and virtual machine to be migrated transports on the physical server of candidate physical server model The second timeslice during row shares ratio, and wherein, described second timeslice shares ratio less than or equal to 1.
Said method, it is preferred that the virtual machine being determined by described same virtual cpu core number exists Timeslice when running on the physical server of different model shares ratio transformational relation, including:
With timeslice, virtual machine that virtual cpu core number is m is shared ratio respectively is 1 to operate in i type Number physical server and j model physical server on;
On the physical server of described i model described virtual cpu core number be m virtual machine in run Different types of live load, carries out test to the virtual machine that described virtual cpu core number is m and obtains institute State the first calculated performance parameter Ci of the virtual machine that virtual cpu core number is m;
On the physical server of described j model described virtual cpu core number be m virtual machine in run Described different types of live load, tests the virtual machine that described virtual cpu core number is m The second calculated performance parameter Cj to the virtual machine that described virtual cpu core number is m;
If described Ci < Cj, then to the virtual machine physics in j model that described virtual cpu core number is m The timeslice that can take the physical cpu in the physical server of described j model when running on server is shared Ratio is adjusted, until the difference of Ci Yu Cj is in default scope, then and tijFor the difference of Ci and Cj in advance If time in the range of, described virtual cpu core number is that the virtual machine of m is on the physical server of j model The timeslice that can take the physical cpu in the physical server of described j model during operation shares ratio;
If Ci > Cj, then by the virtual machine physical services in i model that described virtual cpu core number is m The timeslice that can take the physical cpu in the physical server of described i model when running on device shares ratio It is adjusted, until the difference of Ci Yu Cj is in default scope, then tijFor the difference of Ci and Cj default In the range of time, described virtual cpu core number is that the virtual machine of m runs on the physical server of i model Time can take the timeslice of the physical cpu in the physical server of described i model and share the inverse of ratio;
If Ci=Cj, then tij=1.
Said method, it is preferred that determined that by calculating the virtual machine of described same virtual cpu core number exists Timeslice when running on the physical server of different model shares ratio transformational relation, including:
Calculating described virtual cpu core number is that the virtual machine of m is when running on the physical server of i model The first calculated performance parameter Ci;
Calculating described virtual cpu core number is that the virtual machine of m is when running on the physical server of j model The second calculated performance parameter Cj;
Wherein, tij=Ci/Cj.
A kind of virtual machine (vm) migration device, described device includes:
Acquisition module, shares for obtaining the virtual machine to be migrated very first time sheet on the physical server of source Ratio;
First determines module, for sharing ratio, the type of source physical server based on described very first time sheet Number, and the virtual machine of preset same virtual cpu core number transports on the physical server of different model Timeslice during row is shared ratio transformational relation and is determined candidate physical server model and virtual machine to be migrated The second timeslice when running on the physical server of candidate physical server model shares ratio, wherein, Described second timeslice shares ratio less than or equal to 1;
Second determines module, is used for determining purpose physical server, the model of described purpose physical server Belong to described candidate physical server model, the CPU core calculation of described purpose physical server more than or etc. Virtual cpu core number in described virtual machine to be migrated;
Transferring module, for by described virtual machine (vm) migration to be migrated to described purpose physical server, and The described virtual machine to be migrated timeslice in described purpose physical server is shared ratio be adjusted to and institute The second timeslice stating the model of purpose physical server corresponding shares ratio.
Said apparatus, it is preferred that the virtual machine of described same virtual cpu core number is at the thing of different model Timeslice when running on reason server is shared ratio transformational relation and is included:
Virtual cpu core number is the virtual machine of m computational when running on the physical server of j model Can be equivalent to described virtual cpu core number be the virtual machine of m on the physical server of i model with the time Sheet shares ratio when being calculated performance during 1 operation, and described virtual cpu core number is that the virtual machine of m exists Run required timeslice on the physical server of described j model sharing ratio is tij
Said apparatus, it is preferred that described first determines that module includes:
First determines unit, if sharing ratio for described very first time sheet is not 1, then according to described source thing The model of reason server, described very first time sheet shares ratio, and described preset same virtual cpu The timeslice when virtual machine of core number runs on the physical server of different model shares ratio conversion pass System, determines transition physical server model;Wherein, described virtual machine to be migrated takes at described transition physics It is 1 use described transition physical server model with timeslice ratio on the physical server of business type number Calculated performance during the physical cpu timeslice of physical server, with described virtual machine to be migrated in described source Share ratio with described very first time sheet on physical server and use the physical cpu of described source physical server Calculated performance during timeslice is identical;
Second determines unit, for according to described transition physical server model and described preset same The timeslice when virtual machine of one virtual cpu core number runs on the physical server of different model is shared Ratio transformational relation determines candidate server model, and virtual machine to be migrated is in candidate physical server-type Number physical server on the second timeslice when running share ratio, wherein, described second timeslice is altogether Ratio of enjoying is less than or equal to 1.
Said apparatus, it is preferred that described device also includes:
First transformational relation determines module, for being determined by described same virtual cpu core number Timeslice when virtual machine runs on the physical server of different model shares ratio transformational relation, including:
First test cell, for sharing ratio when the virtual machine that virtual cpu core number is m with timeslice Be on 1 physical server operating in i model described virtual cpu core number be m virtual machine in run During different types of live load, the virtual machine that described virtual cpu core number is m is carried out test and obtains Described virtual cpu core number is the first calculated performance parameter Ci of the virtual machine of m;
Second test cell, for sharing ratio when the virtual machine that virtual cpu core number is m with timeslice Be on 1 physical server operating in j model described virtual cpu core number be m virtual machine in run During described different types of live load, the virtual machine that described virtual cpu core number is m is tested Obtain the second calculated performance parameter Cj of the virtual machine that described virtual cpu core number is m;
3rd determines unit, if for described Ci < Cj, is then the void of m to described virtual cpu core number Plan machine can take the physics in the physical server of described j model when running on the physical server of j model The timeslice of CPU is shared ratio and is adjusted, until the difference of Ci Yu Cj is in default scope, the most really Determine tijDuring for the difference of Ci and Cj in default scope, described virtual cpu core number is the virtual machine of m The physical server of j model can take when running the physical cpu in the physical server of described j model Timeslice share ratio;
4th determines unit, if for Ci > Cj, is then the virtual machine of m by described virtual cpu core number The physical server of i model can take when running the physical cpu in the physical server of described i model Timeslice share ratio and be adjusted, until the difference of Ci Yu Cj is in default scope, it is determined that tij During for the difference of Ci and Cj in default scope, described virtual cpu core number is that the virtual machine of m is at i Physical cpu in the physical server of described ji model can be taken when running on the physical server of model Timeslice shares the inverse of ratio;
5th determines unit, if for Ci=Cj, it is determined that tij=1.
Said apparatus, it is preferred that described device also includes:
Second transformational relation determines module, for determining described same virtual cpu core number by calculating Timeslice when virtual machine runs on the physical server of different model shares ratio transformational relation, including:
First computing unit, is that the virtual machine of m is in i model for calculating described virtual cpu core number The first calculated performance parameter Ci when running on physical server;
Second computing unit, is that the virtual machine of m is in j model for calculating described virtual cpu core number The second calculated performance parameter Cj when running on physical server;
6th determines unit, is used for determining tij: tij=Ci/Cj.
By above scheme, a kind of virtual machine migration method of the application offer and device, pre-set same The timeslice when virtual machine of one virtual cpu core number runs on the physical server of different model is shared Ratio transformational relation, when carrying out virtual machine (vm) migration, shares ratio transformational relation by above-mentioned timeslice, And the very first time sheet that virtual machine to be migrated is on the physical server of source shares ratio, determine candidate physical Type of server, and virtual machine to be migrated runs on the physical server of candidate physical server model Time the second timeslice share ratio, it is then determined that purpose physical server, the type of purpose physical server Number belong to described candidate physical server model, the CPU core calculation of described purpose physical server more than or Virtual cpu core number equal to described virtual machine to be migrated;By virtual machine (vm) migration to be migrated to purpose physics In server, and ratio of the virtual machine to be migrated timeslice in described purpose physical server being shared is adjusted Whole share ratio for the second timeslice corresponding with the model of described purpose physical server, reduce virtual The difference of the Practical Calculation performance before and after machine migration.
Accompanying drawing explanation
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to reality Execute the required accompanying drawing used in example or description of the prior art to be briefly described, it should be apparent that below, Accompanying drawing in description is only some embodiments of the present invention, for those of ordinary skill in the art, On the premise of not paying creative work, it is also possible to obtain other accompanying drawing according to these accompanying drawings.
A kind of flowchart of the virtual machine migration method that Fig. 1 provides for the embodiment of the present application;
A kind of structural representation of the virtual machine (vm) migration device that Fig. 2 provides for the embodiment of the present application;
The first a kind of structural representation determining module that Fig. 3 provides for the embodiment of the present application;
Fig. 4 determines a kind of structural representation of module for the first transformational relation that the embodiment of the present application provides;
Fig. 5 determines a kind of structural representation of module for the second transformational relation that the embodiment of the present application provides.
Term " first " in specification and claims and above-mentioned accompanying drawing, " second ", " the 3rd " (if present) such as " the 4 " is for distinguishing similar part, the most suitable without being used for describing Sequence or precedence.Should be appreciated that the data of so use can be exchanged in the appropriate case, in order to here The embodiments herein described can be implemented with the order in addition to here illustrating.
Detailed description of the invention
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out Clearly and completely describe, it is clear that described embodiment is only a part of embodiment of the present invention, and It is not all, of embodiment.Based on the embodiment in the present invention, those of ordinary skill in the art are not doing Go out the every other embodiment obtained under creative work premise, broadly fall into the scope of protection of the invention.
The one referring to the virtual machine migration method that Fig. 1, Fig. 1 provide for the embodiment of the present application realizes flow process Figure, may include that
Step S11: obtain the virtual machine to be migrated very first time sheet on the physical server of source and share ratio;
Creating in the physical server of virtual machine, a physical cpu both can process virtual cpu thread, Other thread (non-virtual CPU line journey) can also be processed.Assume that a physical cpu processes one simultaneously The virtual cpu thread thread1 of virtual machine and non-virtual CPU line journey thread2, the two thread Time timesharing use physical cpu.If it is 1:1 (i.e. thread1 that the time of physical cpu is shared ratio by them Timeslice to share ratio be 1), the dispatching cycle of thread is t, then thread1 uses the physics of t time After CPU, then being used the physical cpu of t time by thread2, then thread1 re-uses the t time Physical cpu, circulate like this;If the time of physical cpu is shared ratio by thread1 and thread2 For 1:2 (i.e. it is 1/2 that the timeslice of thread1 shares ratio), then thread1 uses the physics of t time After CPU, thread2 can use the physical cpu of 2t time, and then thread1 uses the physics of t time CPU, circulates like this.
For ease of narration, in the embodiment of the present invention, by virtual machine to be migrated on the physical server of source time Between sheet share ratio and be designated as very first time sheet and share ratio.
Step S12: share ratio based on described very first time sheet, the model of source physical server, and in advance The time when virtual machine of the same virtual cpu core number put runs on the physical server of different model Sheet is shared ratio transformational relation and is determined that candidate physical server model and virtual machine to be migrated are in candidate physical The second timeslice when running on the physical server of server model shares ratio, wherein, and described second Timeslice shares ratio less than or equal to 1;
The computing capability of virtual machine depends on the number of virtual cpu and the performance of physical server, general feelings Condition next one virtual cpu core correspond to a physical core or the hyperthread logic core of physical cpu.
In the embodiment of the present invention, pre-build that the virtual machine of same virtual cpu core number is in different model Physical server on timeslice when running share ratio transformational relation.
The virtual cpu core number assuming virtual machine is m, and this virtual cpu core number is the virtual machine of m On the physical server of the first model with timeslice share ratio for s1 run time calculated performance, quite It is that the virtual machine of m is shared with timeslice on the physical server of the second model in this virtual cpu core number Ratio is calculated performance during s2 operation, then, in the embodiment of the present invention, virtual cpu core number is m's Virtual machine on the physical server of the first model and when running on the physical server of the second model time Between sheet share ratio transformational relation and be: virtual cpu core number is the virtual machine physics in the first model of m Calculated performance when sharing the operation of ratio s1 with timeslice on server is equivalent to this virtual cpu core number and is The virtual machine of m the second model physical server run time calculated performance time, this virtual cpu core It is s2 that timeslice when the virtual machine that number is m runs on the physical server of the second model shares ratio.
In the embodiment of the present invention, by the same void corresponding with the core number of the CPU of virtual machine to be migrated The timeslice when virtual machine of plan CPU core calculation runs on the physical server of different model shares ratio Transformational relation determines candidate physical type of server.
Wherein, owing to virtual machine based on same virtual cpu core number is at the physical server of different model Upper timeslice when running shares the physical server model that ratio transformational relation determines may be had multiple, and Calculated performance when virtual machine to be migrated runs on the physical server of these models is equivalent to void to be migrated When plan machine shares calculated performance when ratio is run with very first time sheet on the physical server of source, to be migrated The timeslice of virtual machine shares ratio may be more than 1, it is also possible to less than or equal to 1, and the present invention implements In example, determine timeslice when virtual machine to be migrated runs on physical server share ratio less than or etc. Physical server model in 1 is candidate physical server model.
Step S13: determine that purpose physical server, the model of described purpose physical server belong to described time Select physical server model, the CPU core calculation of described purpose physical server be more than or equal to described in wait to move Move the virtual cpu core number of virtual machine;
In the embodiment of the present invention, from the physical server that model is candidate physical type number, determine a thing The CPU core calculation of reason server is more than or equal to the thing of the virtual cpu core number of described virtual machine to be migrated Physical server for the purpose of reason server.
If not finding purpose physical server, illustrate that virtual machine to be migrated cannot be carried out migrating.
Step S14: by described virtual machine (vm) migration to be migrated to described purpose physical server, and by described The virtual machine to be migrated timeslice in described purpose physical server is shared ratio and is adjusted to and described purpose The second timeslice that the model of physical server is corresponding shares ratio.
By in virtual machine (vm) migration to be migrated to purpose physical server, and virtual machine to be migrated is taken in purpose It is that the second timeslice shares ratio that timeslice in business device shares ratio setting, wherein, and virtual machine to be migrated The calculated performance shared when ratio is run with the second timeslice on purpose physical server is equivalent to be migrated Virtual machine shares calculated performance when ratio is run with very first time sheet on the physical server of source.
A kind of virtual machine migration method that the embodiment of the present invention provides, pre-sets same virtual cpu core The timeslice when virtual machine of number runs on the physical server of different model shares ratio transformational relation, When carrying out virtual machine (vm) migration, share ratio transformational relation by above-mentioned timeslice, and to be migrated virtual Machine very first time sheet on the physical server of source shares ratio, determines candidate physical type of server, with And the second timeslice that virtual machine to be migrated is when running on the physical server of candidate physical server model The ratio of sharing, it is then determined that purpose physical server, the model of purpose physical server belongs to described candidate Physical server model, the CPU core calculation of described purpose physical server is more than or equal to described to be migrated The virtual cpu core number of virtual machine;By in virtual machine (vm) migration to be migrated to purpose physical server, and will The virtual machine to be migrated timeslice in described purpose physical server is shared ratio and is adjusted to and described purpose The second timeslice that the model of physical server is corresponding shares ratio, reduces the reality before and after virtual machine (vm) migration The difference of border calculated performance.
Optionally, the virtual machine of described same virtual cpu core number is on the physical server of different model Timeslice during operation is shared ratio transformational relation and be may include that
Virtual cpu core number is the virtual machine of m computational when running on the physical server of j model Can be equivalent to described virtual cpu core number be the virtual machine of m on the physical server of i model with the time Sheet shares ratio when being calculated performance during 1 operation, and described virtual cpu core number is that the virtual machine of m exists Run required timeslice on the physical server of described j model sharing ratio is tij
Generally, when creating a virtual machine on a physical server, this virtual machine is at physical services Timeslice on device is shared ratio and is initialized as 1.
For more convenient understanding, in the embodiment of the present invention, matrix T can be usedmCharacterizing virtual cpu core number is The virtual machine of m is transported on the physical server of different model (by the different model of different digital representation) Timeslice during row shares ratio transformational relation, and this matrix is as follows:
T m = t 11 t 12 . . . t 1 N t 21 t 22 . . . t 2 N . . . t N 1 t N 2 . . . t NN
Wherein, tij(1≤i≤N, 1≤j≤N) represents that virtual cpu core number is that the virtual machine of m is in j model Physical server on calculated performance when running to be equivalent to virtual cpu core number be that the virtual machine of m is at i On the physical server of model with timeslice share ratio be calculated performance during 1 operation time, virtual cpu Core number is that the virtual machine of m timeslice when running on the physical server of j model shares ratio.Also That is, virtual cpu core number is that the virtual machine of m is total to timeslice on the physical server of i model The ratio of enjoying is that calculated performance during 1 operation is equivalent to virtual machine that virtual cpu core number is m in j model Physical server on share ratio as t with timesliceijCalculated performance during operation.
Optionally, described share ratio based on described very first time sheet, the model of source physical server, with And the virtual machine of preset same virtual cpu core number is when running on the physical server of different model Timeslice is shared ratio transformational relation and is determined candidate physical server model, and virtual machine to be migrated is being waited Select the second timeslice when running on the physical server of physical server model to share ratio may include that
If it is 1 that very first time sheet shares ratio, then can be directly by searching Tm, find out all candidate physical Type of server k, wherein, trk≤ 1, r are the model of source, virtual machine place to be migrated physical server, k∈{1,2,…,N}.In the embodiment of the present invention, if the model of purpose physical server is k1, then with described mesh The second timeslice corresponding to the model of physical server to share ratio be trk1
If it is not 1 that described very first time sheet shares ratio, then according to the model of described source physical server, institute State very first time sheet and share ratio, and the virtual machine of described preset same virtual cpu core number is not Timeslice when running on the physical server of same model shares ratio transformational relation, determines that transition physics takes Business type h;Wherein, described virtual machine to be migrated is at the physical services of described transition physical server model With the physics of physical server that timeslice ratio is 1 use described transition physical server model on device Calculated performance during CPU time slice, with described virtual machine to be migrated on the physical server of described source with institute State very first time sheet and share computational when ratio uses the physical cpu timeslice of described source physical server Can be identical;
Concrete, t can be passed throughhr=s finds transition physical server model h, and wherein, s is void to be migrated Very first time sheet when plan machine runs on the physical server of source shares ratio, and r is virtual machine place to be migrated The model of source physical server, h ∈ 1,2 ..., N}.
Sharing ratio when very first time sheet is not 1, illustrates that virtual machine to be migrated is from another physical server On move on this source physical server.In the embodiment of the present invention, if very first time sheet shares ratio not Be 1, then the virtual machine first passing through same virtual cpu core number is transported on the physical server of different model Timeslice during row is shared ratio transformational relation and is found the timeslice of virtual machine to be migrated to share ratio when being 1 Physical server model, i.e. transition physical server model, then, then by transition physical server Model, and when the virtual machine of same virtual cpu core number runs on the physical server of different model Timeslice share ratio transformational relation and find candidate physical server.
According to described transition physical server model, and described preset same virtual cpu core number Timeslice when virtual machine runs on the physical server of different model is shared ratio transformational relation and is determined time Select physical server model, and virtual machine to be migrated is at the physical server of candidate physical server model Upper run time the second timeslice share ratio, wherein, described second timeslice share ratio less than or etc. In 1.
Concrete, candidate physical server model k can be determined by transitional services type h, wherein, thk≤ 1, h are transition physical server model, k ∈ 1,2 ..., N}.Now, if purpose physical server Model is k1, then it is t that the second corresponding with the model of described purpose physical server timeslice shares ratiohk1
Optionally, the virtual machine of described same virtual cpu core number can be determined by different shaped Number physical server on timeslice when running share ratio transformational relation, specifically include:
With timeslice, virtual machine that virtual cpu core number is m is shared ratio respectively is 1 to operate in i type Number physical server and j model physical server on;When specifically testing,
On the physical server of described i model, in the virtual machine that described virtual cpu core number is m Run different types of live load, the virtual machine that described virtual cpu core number is m is tested The first calculated performance parameter Ci to the virtual machine that described virtual cpu core number is m;
On the physical server of described j model, in the virtual machine that described virtual cpu core number is m Run described different types of live load, the virtual machine that described virtual cpu core number is m is surveyed Examination obtains the second calculated performance parameter Cj of the virtual machine that described virtual cpu core number is m;
Above-mentioned different type may include that CPU intensive type, memory-intensive, disk I/O intensive with And the combination in any type of these types.
SPEC ((Standard Performance Evaluation Corporation, systematic function can be utilized Assessment test) etc. benchmark test instrument test virtual machine calculated performance.
Calculated performance parameter can characterize with the instruction number (i.e. instruction number/second) of average execution per second.
If described Ci < Cj, then to the virtual machine physics in j model that described virtual cpu core number is m The timeslice that can take the physical cpu in the physical server of described j model when running on server is shared Ratio is adjusted, until the difference of Ci Yu Cj is in default scope, then and tijFor the difference of Ci and Cj in advance If time in the range of, described virtual cpu core number is that the virtual machine of m is on the physical server of j model The timeslice that can take the physical cpu in the physical server of described j model during operation shares ratio;
If the first calculated performance parameter is less than the second calculated performance parameter, then find virtual machine in j model Process corresponding when running in physical machine, and adjust the physical cpu of the j model that this process can take Timeslice ratio, until the difference of Ci Yu Cj is in default scope.When the difference of Ci Yu Cj is at default model When enclosing interior, virtual cpu core number is that the virtual machine of m can account for when running on the physical server of j model Sharing ratio by the timeslice of the physical cpu in the physical server of described j model is tij
If Ci > Cj, then by the virtual machine physical services in i model that described virtual cpu core number is m The timeslice that can take the physical cpu in the physical server of described i model when running on device shares ratio It is adjusted, until the difference of Ci Yu Cj is in default scope, then tijFor the difference of Ci and Cj default In the range of time, described virtual cpu core number is that the virtual machine of m runs on the physical server of i model Time can take the timeslice of the physical cpu in the physical server of described i model and share the inverse of ratio;
If the first calculated performance parameter is more than the second calculated performance parameter, then find virtual machine in i model Process corresponding when running in physical machine, and adjust the physical cpu of the i model that this process can take Timeslice ratio, until the difference of Ci Yu Cj is in default scope.When the difference of Ci Yu Cj is at default model When enclosing interior, virtual cpu core number is that the virtual machine of m can account for when running on the physical server of i model The inverse sharing ratio by the timeslice of the physical cpu in the physical server of described i model is tij
If Ci=Cj, then tij=1.
Optionally, it is also possible to determine that the virtual machine of described same virtual cpu core number is in difference by calculating Timeslice when running on the physical server of model shares ratio transformational relation, specifically includes:
Calculating described virtual cpu core number is that the virtual machine of m is when running on the physical server of i model The first calculated performance parameter Ci;
Calculating described virtual cpu core number is that the virtual machine of m is when running on the physical server of j model The second calculated performance parameter Cj;
Wherein, tij=Ci/Cj.
In the embodiment of the present invention, the instruction number x of the CPU execution per second of physical server, base can be obtained In virtual machine monitor (Virtual Machine Monitor, VMM) scheduling physical cpu to virtual machine process Algorithm to calculate the physical cpu time distributing to virtual machine process the most per second be T, then virtual machine Calculated performance parameter is x/T.
Corresponding with embodiment of the method, the embodiment of the present invention also provides for a kind of virtual machine (vm) migration device, this A kind of structural representation of the virtual machine (vm) migration device that bright embodiment provides is as in figure 2 it is shown, may include that
Acquisition module 21, first determines module 22, and second determines module 23 and transferring module 24;Wherein,
Acquisition module 21 is shared for obtaining the virtual machine to be migrated very first time sheet on the physical server of source Ratio;
Creating in the physical server of virtual machine, a physical cpu both can process virtual cpu thread, Other thread (non-virtual CPU line journey) can also be processed.Assume that a physical cpu processes one simultaneously The virtual cpu thread thread1 of virtual machine and non-virtual CPU line journey thread2, the two thread Time timesharing use physical cpu.If it is 1:1 (i.e. thread1 that the time of physical cpu is shared ratio by them Timeslice to share ratio be 1), the dispatching cycle of thread is t, then thread1 uses the physics of t time After CPU, then being used the physical cpu of t time by thread2, then thread1 re-uses the t time Physical cpu, circulate like this;If the time of physical cpu is shared ratio by thread1 and thread2 For 1:2 (i.e. it is 1/2 that the timeslice of thread1 shares ratio), then thread1 uses the physics of t time After CPU, thread2 can use the physical cpu of 2t time, and then thread1 uses the physics of t time CPU, circulates like this.
For ease of narration, in the embodiment of the present invention, by virtual machine to be migrated on the physical server of source time Between sheet share ratio and be designated as very first time sheet and share ratio.
First determines that module 22 is for sharing ratio, the type of source physical server based on described very first time sheet Number, and the virtual machine of preset same virtual cpu core number transports on the physical server of different model Timeslice during row is shared ratio transformational relation and is determined candidate physical server model and virtual machine to be migrated The second timeslice when running on the physical server of candidate physical server model shares ratio, wherein, Described second timeslice shares ratio less than or equal to 1;
The computing capability of virtual machine depends on the number of virtual cpu and the performance of physical server, general feelings Condition next one virtual cpu core correspond to a physical core or the hyperthread logic core of physical cpu.
In the embodiment of the present invention, pre-build that the virtual machine of same virtual cpu core number is in different model Physical server on timeslice when running share ratio transformational relation.
The virtual cpu core number assuming virtual machine is m, and this virtual cpu core number is the virtual machine of m On the physical server of the first model with timeslice share ratio for s1 run time calculated performance, quite It is that the virtual machine of m is shared with timeslice on the physical server of the second model in this virtual cpu core number Ratio is calculated performance during s2 operation, then, in the embodiment of the present invention, virtual cpu core number is m's Virtual machine on the physical server of the first model and when running on the physical server of the second model time Between sheet share ratio transformational relation and be: virtual cpu core number is the virtual machine physics in the first model of m Calculated performance when sharing the operation of ratio s1 with timeslice on server is equivalent to this virtual cpu core number and is The virtual machine of m the second model physical server run time calculated performance time, this virtual cpu core It is s2 that timeslice when the virtual machine that number is m runs on the physical server of the second model shares ratio.
In the embodiment of the present invention, by the same void corresponding with the core number of the CPU of virtual machine to be migrated The timeslice when virtual machine of plan CPU core calculation runs on the physical server of different model shares ratio Transformational relation determines candidate physical type of server.
Wherein, owing to virtual machine based on same virtual cpu core number is at the physical server of different model Upper timeslice when running shares the physical server model that ratio transformational relation determines may be had multiple, and Calculated performance when virtual machine to be migrated runs on the physical server of these models is equivalent to void to be migrated When plan machine shares calculated performance when ratio is run with very first time sheet on the physical server of source, to be migrated The timeslice of virtual machine shares ratio may be more than 1, it is also possible to less than or equal to 1, and the present invention implements In example, determine timeslice when virtual machine to be migrated runs on physical server share ratio less than or etc. Physical server model in 1 is candidate physical server model.
Second determine module 23 for determining purpose physical server, the model of described purpose physical server Belong to described candidate physical server model, the CPU core calculation of described purpose physical server more than or etc. Virtual cpu core number in described virtual machine to be migrated;
In the embodiment of the present invention, from the physical server that model is candidate physical type number, determine a thing The CPU core calculation of reason server is more than or equal to the thing of the virtual cpu core number of described virtual machine to be migrated Physical server for the purpose of reason server.
If not finding purpose physical server, illustrate that virtual machine to be migrated cannot be carried out migrating.
Transferring module 24 for by described virtual machine (vm) migration to be migrated to described purpose physical server, and The described virtual machine to be migrated timeslice in described purpose physical server is shared ratio be adjusted to and institute The second timeslice stating the model of purpose physical server corresponding shares ratio.
By in virtual machine (vm) migration to be migrated to purpose physical server, and virtual machine to be migrated is taken in purpose It is that the second timeslice shares ratio that timeslice in business device shares ratio setting, wherein, and virtual machine to be migrated The calculated performance shared when ratio is run with the second timeslice on purpose physical server is equivalent to be migrated Virtual machine shares calculated performance when ratio is run with very first time sheet on the physical server of source.
A kind of virtual machine (vm) migration device that the embodiment of the present invention provides, pre-sets same virtual cpu core The timeslice when virtual machine of number runs on the physical server of different model shares ratio transformational relation, When carrying out virtual machine (vm) migration, share ratio transformational relation by above-mentioned timeslice, and to be migrated virtual Machine very first time sheet on the physical server of source shares ratio, determines candidate physical type of server, with And the second timeslice that virtual machine to be migrated is when running on the physical server of candidate physical server model The ratio of sharing, it is then determined that purpose physical server, the model of purpose physical server belongs to described candidate Physical server model, the CPU core calculation of described purpose physical server is more than or equal to described to be migrated The virtual cpu core number of virtual machine;By in virtual machine (vm) migration to be migrated to purpose physical server, and will The virtual machine to be migrated timeslice in described purpose physical server is shared ratio and is adjusted to and described purpose The second timeslice that the model of physical server is corresponding shares ratio, reduces the reality before and after virtual machine (vm) migration The difference of border calculated performance.
Optionally, the virtual machine of described same virtual cpu core number is on the physical server of different model Timeslice during operation is shared ratio transformational relation and be may include that
Virtual cpu core number is the virtual machine of m computational when running on the physical server of j model Can be equivalent to described virtual cpu core number be the virtual machine of m on the physical server of i model with the time Sheet shares ratio when being calculated performance during 1 operation, and described virtual cpu core number is that the virtual machine of m exists Run required timeslice on the physical server of described j model sharing ratio is tij
Generally, when creating a virtual machine on a physical server, this virtual machine is at physical services Timeslice on device is shared ratio and is initialized as 1.
For more convenient understanding, in the embodiment of the present invention, matrix T can be usedmCharacterizing virtual cpu core number is The virtual machine of m is transported on the physical server of different model (by the different model of different digital representation) Timeslice during row shares ratio transformational relation, and this matrix is as follows:
T m = t 11 t 12 . . . t 1 N t 21 t 22 . . . t 2 N . . . t N 1 t N 2 . . . t NN
Wherein, tij(1≤i≤N, 1≤j≤N) represents that virtual cpu core number is that the virtual machine of m is in j model Physical server on calculated performance when running to be equivalent to virtual cpu core number be that the virtual machine of m is at i On the physical server of model with timeslice share ratio be calculated performance during 1 operation time, virtual cpu Core number is that the virtual machine of m timeslice when running on the physical server of j model shares ratio.Also That is, virtual cpu core number is that the virtual machine of m is total to timeslice on the physical server of i model The ratio of enjoying is that calculated performance during 1 operation is equivalent to virtual machine that virtual cpu core number is m in j model Physical server on share ratio as t with timesliceijCalculated performance during operation.
Optionally, first determines that a kind of structural representation of module 22 is as it is shown on figure 3, may include that
First determines that unit 31 and second determines unit 32;Wherein,
If first to determine that unit 31 shares ratio for described very first time sheet be not 1, then according to described source The model of physical server, described very first time sheet shares ratio, and described preset same virtual cpu The timeslice when virtual machine of core number runs on the physical server of different model shares ratio conversion pass System, determines transition physical server model;Wherein, described virtual machine to be migrated takes at described transition physics It is 1 use described transition physical server model with timeslice ratio on the physical server of business type number Calculated performance during the physical cpu timeslice of physical server, with described virtual machine to be migrated in described source Share ratio with described very first time sheet on physical server and use the physical cpu of described source physical server Calculated performance during timeslice is identical;
Concrete, t can be passed throughhr=s finds transition physical server model h, and wherein, s is void to be migrated Very first time sheet when plan machine runs on the physical server of source shares ratio, and r is virtual machine place to be migrated The model of source physical server, h ∈ 1,2 ..., N}.
Sharing ratio when very first time sheet is not 1, illustrates that virtual machine to be migrated is from another physical server On move on this source physical server.In the embodiment of the present invention, if very first time sheet shares ratio not Be 1, then the virtual machine first passing through same virtual cpu core number is transported on the physical server of different model Timeslice during row is shared ratio transformational relation and is found the timeslice of virtual machine to be migrated to share ratio when being 1 Physical server model, i.e. transition physical server model, then, then by transition physical server Model, and when the virtual machine of same virtual cpu core number runs on the physical server of different model Timeslice share ratio transformational relation and find candidate physical server.
First determines that unit 31 can be also used for, if it is 1 that very first time sheet shares ratio, then and can be direct By searching Tm, find out all candidate physical type of servers k, wherein, trk≤ 1, r are to be migrated virtual The model of source, machine place physical server, k ∈ 1,2 ..., N}.In the embodiment of the present invention, if purpose physics clothes The model of business device is k1, then the second corresponding with the model of described purpose physical server timeslice shares ratio Example is trk1
Second determines that unit 32 is for according to described transition physical server model and described preset same The timeslice when virtual machine of one virtual cpu core number runs on the physical server of different model is shared Ratio transformational relation determines candidate server model, and virtual machine to be migrated is in candidate physical server-type Number physical server on the second timeslice when running share ratio, wherein, described second timeslice is altogether Ratio of enjoying is less than or equal to 1.
Concrete, candidate physical server model k can be determined by transitional services type h, wherein, thk≤ 1, h are transition physical server model, k ∈ 1,2 ..., N}.Now, if purpose physical server Model is k1, then it is t that the second corresponding with the model of described purpose physical server timeslice shares ratiohk1
Optionally, on the basis of embodiment illustrated in fig. 2, the virtual machine (vm) migration that the embodiment of the present invention provides Device can also include:
First transformational relation determines module, for being determined by described same virtual cpu core number Timeslice when virtual machine runs on the physical server of different model shares ratio transformational relation.
Concrete, the first transformational relation that the embodiment of the present invention provides determines a kind of structural representation of module As shown in Figure 4, may include that
First test cell 41, for sharing with timeslice when the virtual machine that virtual cpu core number is m Ratio is on 1 physical server operating in i model, is the virtual of m at described virtual cpu core number When machine runs different types of live load, the virtual machine that described virtual cpu core number is m is carried out Test obtains the first calculated performance parameter Ci of the virtual machine that described virtual cpu core number is m;
Second test cell 42, for sharing with timeslice when the virtual machine that virtual cpu core number is m Ratio is on 1 physical server operating in j model, is the virtual of m at described virtual cpu core number When machine runs described different types of live load, it is the virtual machine of m to described virtual cpu core number Carry out testing the second calculated performance parameter Cj obtaining the virtual machine that described virtual cpu core number is m;
Above-mentioned different type may include that CPU intensive type, memory-intensive, disk I/O intensive with And the combination in any type of these types.
SPEC ((Standard Performance Evaluation Corporation, systematic function can be utilized Assessment test) etc. benchmark test instrument test virtual machine calculated performance.
Calculated performance parameter can characterize with the instruction number (i.e. instruction number/second) of average execution per second.
3rd determines unit 43, if for described Ci < Cj, is then m's to described virtual cpu core number Virtual machine can take the thing in the physical server of described j model when running on the physical server of j model The timeslice of reason CPU is shared ratio and is adjusted, until the difference of Ci Yu Cj is in default scope, then Determine tijDuring for the difference of Ci and Cj in default scope, described virtual cpu core number is the virtual of m Machine can take the physics in the physical server of described j model when running on the physical server of j model The timeslice of CPU shares ratio;
If the first calculated performance parameter is less than the second calculated performance parameter, then find virtual machine in j model Process corresponding when running in physical machine, and adjust the physical cpu of the j model that this process can take Timeslice ratio, until the difference of Ci Yu Cj is in default scope.When the difference of Ci Yu Cj is at default model When enclosing interior, virtual cpu core number is that the virtual machine of m can account for when running on the physical server of j model Sharing ratio by the timeslice of the physical cpu in the physical server of described j model is tij
4th determines unit 44, if for Ci > Cj, is then the virtual of m by described virtual cpu core number Machine can take the physics in the physical server of described i model when running on the physical server of i model The timeslice of CPU is shared ratio and is adjusted, until the difference of Ci Yu Cj is in default scope, the most really Determine tijDuring for the difference of Ci and Cj in default scope, described virtual cpu core number is the virtual machine of m The physical server of i model can take when running the physical cpu in the physical server of described ji model Timeslice share the inverse of ratio;
If the first calculated performance parameter is more than the second calculated performance parameter, then find virtual machine in i model Process corresponding when running in physical machine, and adjust the physical cpu of the i model that this process can take Timeslice ratio, until the difference of Ci Yu Cj is in default scope.When the difference of Ci Yu Cj is at default model When enclosing interior, virtual cpu core number is that the virtual machine of m can account for when running on the physical server of i model The inverse sharing ratio by the timeslice of the physical cpu in the physical server of described i model is tij
5th determines unit 45, if for Ci=Cj, it is determined that tij=1.
Optionally, on the basis of embodiment illustrated in fig. 2, the virtual machine (vm) migration that the embodiment of the present invention provides Device can also include:
Second transformational relation determines module, for determining described same virtual cpu core number by calculating Timeslice when virtual machine runs on the physical server of different model shares ratio transformational relation.
Concrete, the second transformational relation that the embodiment of the present invention provides determines a kind of structural representation of module As it is shown in figure 5, may include that
First computing unit 51, is that the virtual machine of m is in i model for calculating described virtual cpu core number Physical server on the first calculated performance parameter Ci when running;
Second computing unit 52, is that the virtual machine of m is in j model for calculating described virtual cpu core number Physical server on the second calculated performance parameter Cj when running;
6th determines unit 53, is used for determining tij: tij=Ci/Cj.
In the embodiment of the present invention, the instruction number x of the CPU execution per second of physical server, base can be obtained In virtual machine monitor (Virtual Machine Monitor, VMM) scheduling physical cpu to virtual machine process Algorithm to calculate the physical cpu time distributing to virtual machine process the most per second be T, then virtual machine Calculated performance parameter is x/T.
Described above to the disclosed embodiments, makes professional and technical personnel in the field be capable of or uses The present invention.Multiple amendment to these embodiments will be aobvious and easy for those skilled in the art See, generic principles defined herein can without departing from the spirit or scope of the present invention, Realize in other embodiments.Therefore, the present invention is not intended to be limited to the embodiments shown herein, And it is to fit to the widest scope consistent with principles disclosed herein and features of novelty.

Claims (10)

1. a virtual machine migration method, it is characterised in that including:
Obtain the virtual machine to be migrated very first time sheet on the physical server of source and share ratio;
Ratio is shared based on described very first time sheet, the model of source physical server, and preset same The timeslice when virtual machine of virtual cpu core number runs on the physical server of different model shares ratio Example transformational relation determines that candidate physical server model and virtual machine to be migrated are in candidate physical server-type Number physical server on the second timeslice when running share ratio, wherein, described second timeslice is altogether Ratio of enjoying is less than or equal to 1;
Determine that purpose physical server, the model of described purpose physical server belong to described candidate physical clothes Business type number, the CPU core calculation of described purpose physical server is more than or equal to described virtual machine to be migrated Virtual cpu core number;
By in described virtual machine (vm) migration to be migrated to described purpose physical server, and by described void to be migrated Plan machine timeslice in described purpose physical server is shared ratio and is adjusted to and described purpose physical services The second timeslice that the model of device is corresponding shares ratio.
Method the most according to claim 1, it is characterised in that described same virtual cpu core The timeslice when virtual machine of number runs on the physical server of different model shares ratio transformational relation bag Include:
Virtual cpu core number is the virtual machine of m computational when running on the physical server of j model Can be equivalent to described virtual cpu core number be the virtual machine of m on the physical server of i model with the time Sheet shares ratio when being calculated performance during 1 operation, and described virtual cpu core number is that the virtual machine of m exists Run required timeslice on the physical server of described j model sharing ratio is tij
Method the most according to claim 2, it is characterised in that described based on described very first time sheet The ratio of sharing, the model of source physical server, and the virtual machine of preset same virtual cpu core number Timeslice when running on the physical server of different model is shared ratio transformational relation and is determined candidate physical Server model, and virtual machine to be migrated runs on the physical server of candidate physical server model Time the second timeslice share ratio and include:
If it is not 1 that described very first time sheet shares ratio, then according to the model of described source physical server, institute State very first time sheet and share ratio, and the virtual machine of described preset same virtual cpu core number is not Timeslice when running on the physical server of same model shares ratio transformational relation, determines that transition physics takes Business type number;Wherein, described virtual machine to be migrated is at the physical services of described transition physical server model With the physics of physical server that timeslice ratio is 1 use described transition physical server model on device Calculated performance during CPU time slice, with described virtual machine to be migrated on the physical server of described source with institute State very first time sheet and share computational when ratio uses the physical cpu timeslice of described source physical server Can be identical;
According to described transition physical server model, and described preset same virtual cpu core number Timeslice when virtual machine runs on the physical server of different model is shared ratio transformational relation and is determined time Election server model, and virtual machine to be migrated transports on the physical server of candidate physical server model The second timeslice during row shares ratio, and wherein, described second timeslice shares ratio less than or equal to 1.
Method the most according to claim 2, it is characterised in that be determined by described same void The timeslice when virtual machine of plan CPU core calculation runs on the physical server of different model shares ratio Transformational relation, including:
With timeslice, virtual machine that virtual cpu core number is m is shared ratio respectively is 1 to operate in i type Number physical server and j model physical server on;
On the physical server of described i model described virtual cpu core number be m virtual machine in run Different types of live load, carries out test to the virtual machine that described virtual cpu core number is m and obtains institute State the first calculated performance parameter Ci of the virtual machine that virtual cpu core number is m;
On the physical server of described j model described virtual cpu core number be m virtual machine in run Described different types of live load, tests the virtual machine that described virtual cpu core number is m The second calculated performance parameter Cj to the virtual machine that described virtual cpu core number is m;
If described Ci < Cj, then to the virtual machine physics in j model that described virtual cpu core number is m The timeslice that can take the physical cpu in the physical server of described j model when running on server is shared Ratio is adjusted, until the difference of Ci Yu Cj is in default scope, then and tijFor the difference of Ci and Cj in advance If time in the range of, described virtual cpu core number is that the virtual machine of m is on the physical server of j model The timeslice that can take the physical cpu in the physical server of described j model during operation shares ratio;
If Ci > Cj, then by the virtual machine physical services in i model that described virtual cpu core number is m The timeslice that can take the physical cpu in the physical server of described i model when running on device shares ratio It is adjusted, until the difference of Ci Yu Cj is in default scope, then tijFor the difference of Ci and Cj default In the range of time, described virtual cpu core number is that the virtual machine of m runs on the physical server of i model Time can take the timeslice of the physical cpu in the physical server of described i model and share the inverse of ratio;
If Ci=Cj, then tij=1.
Method the most according to claim 2, it is characterised in that determine described same void by calculating The timeslice when virtual machine of plan CPU core calculation runs on the physical server of different model shares ratio Transformational relation, including:
Calculating described virtual cpu core number is that the virtual machine of m is when running on the physical server of i model The first calculated performance parameter Ci;
Calculating described virtual cpu core number is that the virtual machine of m is when running on the physical server of j model The second calculated performance parameter Cj;
Wherein, tij=Ci/Cj.
6. a virtual machine (vm) migration device, it is characterised in that described device includes:
Acquisition module, shares for obtaining the virtual machine to be migrated very first time sheet on the physical server of source Ratio;
First determines module, for sharing ratio, the type of source physical server based on described very first time sheet Number, and the virtual machine of preset same virtual cpu core number transports on the physical server of different model Timeslice during row is shared ratio transformational relation and is determined candidate physical server model and virtual machine to be migrated The second timeslice when running on the physical server of candidate physical server model shares ratio, wherein, Described second timeslice shares ratio less than or equal to 1;
Second determines module, is used for determining purpose physical server, the model of described purpose physical server Belong to described candidate physical server model, the CPU core calculation of described purpose physical server more than or etc. Virtual cpu core number in described virtual machine to be migrated;
Transferring module, for by described virtual machine (vm) migration to be migrated to described purpose physical server, and The described virtual machine to be migrated timeslice in described purpose physical server is shared ratio be adjusted to and institute The second timeslice stating the model of purpose physical server corresponding shares ratio.
Device the most according to claim 6, it is characterised in that described same virtual cpu core The timeslice when virtual machine of number runs on the physical server of different model shares ratio transformational relation bag Include:
Virtual cpu core number is the virtual machine of m computational when running on the physical server of j model Can be equivalent to described virtual cpu core number be the virtual machine of m on the physical server of i model with the time Sheet shares ratio when being calculated performance during 1 operation, and described virtual cpu core number is that the virtual machine of m exists Run required timeslice on the physical server of described j model sharing ratio is tij
Device the most according to claim 7, it is characterised in that described first determines that module includes:
First determines unit, if sharing ratio for described very first time sheet is not 1, then according to described source thing The model of reason server, described very first time sheet shares ratio, and described preset same virtual cpu The timeslice when virtual machine of core number runs on the physical server of different model shares ratio conversion pass System, determines transition physical server model;Wherein, described virtual machine to be migrated takes at described transition physics It is 1 use described transition physical server model with timeslice ratio on the physical server of business type number Calculated performance during the physical cpu timeslice of physical server, with described virtual machine to be migrated in described source Share ratio with described very first time sheet on physical server and use the physical cpu of described source physical server Calculated performance during timeslice is identical;
Second determines unit, for according to described transition physical server model and described preset same The timeslice when virtual machine of one virtual cpu core number runs on the physical server of different model is shared Ratio transformational relation determines candidate server model, and virtual machine to be migrated is in candidate physical server-type Number physical server on the second timeslice when running share ratio, wherein, described second timeslice is altogether Ratio of enjoying is less than or equal to 1.
Device the most according to claim 7, it is characterised in that described device also includes:
First transformational relation determines module, for being determined by described same virtual cpu core number Timeslice when virtual machine runs on the physical server of different model shares ratio transformational relation, including:
First test cell, for sharing ratio when the virtual machine that virtual cpu core number is m with timeslice Be on 1 physical server operating in i model described virtual cpu core number be m virtual machine in run During different types of live load, the virtual machine that described virtual cpu core number is m is carried out test and obtains Described virtual cpu core number is the first calculated performance parameter Ci of the virtual machine of m;
Second test cell, for sharing ratio when the virtual machine that virtual cpu core number is m with timeslice Be on 1 physical server operating in j model described virtual cpu core number be m virtual machine in run During described different types of live load, the virtual machine that described virtual cpu core number is m is tested Obtain the second calculated performance parameter Cj of the virtual machine that described virtual cpu core number is m;
3rd determines unit, if for described Ci < Cj, is then the void of m to described virtual cpu core number Plan machine can take the physics in the physical server of described j model when running on the physical server of j model The timeslice of CPU is shared ratio and is adjusted, until the difference of Ci Yu Cj is in default scope, the most really Determine tijDuring for the difference of Ci and Cj in default scope, described virtual cpu core number is the virtual machine of m The physical server of j model can take when running the physical cpu in the physical server of described j model Timeslice share ratio;
4th determines unit, if for Ci > Cj, is then the virtual machine of m by described virtual cpu core number The physical server of i model can take when running the physical cpu in the physical server of described i model Timeslice share ratio and be adjusted, until the difference of Ci Yu Cj is in default scope, it is determined that tij During for the difference of Ci and Cj in default scope, described virtual cpu core number is that the virtual machine of m is at i Physical cpu in the physical server of described ji model can be taken when running on the physical server of model Timeslice shares the inverse of ratio;
5th determines unit, if for Ci=Cj, it is determined that tij=1.
Device the most according to claim 7, it is characterised in that described device also includes:
Second transformational relation determines module, for determining described same virtual cpu core number by calculating Timeslice when virtual machine runs on the physical server of different model shares ratio transformational relation, including:
First computing unit, is that the virtual machine of m is in i model for calculating described virtual cpu core number The first calculated performance parameter Ci when running on physical server;
Second computing unit, is that the virtual machine of m is in j model for calculating described virtual cpu core number The second calculated performance parameter Cj when running on physical server;
6th determines unit, is used for determining tij: tij=Ci/Cj.
CN201510165952.3A 2015-04-09 2015-04-09 Virtual machine migration method and device Active CN106155782B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510165952.3A CN106155782B (en) 2015-04-09 2015-04-09 Virtual machine migration method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510165952.3A CN106155782B (en) 2015-04-09 2015-04-09 Virtual machine migration method and device

Publications (2)

Publication Number Publication Date
CN106155782A true CN106155782A (en) 2016-11-23
CN106155782B CN106155782B (en) 2019-06-25

Family

ID=57335916

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510165952.3A Active CN106155782B (en) 2015-04-09 2015-04-09 Virtual machine migration method and device

Country Status (1)

Country Link
CN (1) CN106155782B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108874534A (en) * 2018-04-18 2018-11-23 华为技术有限公司 A kind of dispatching method and device of PCPU resource

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7203944B1 (en) * 2003-07-09 2007-04-10 Veritas Operating Corporation Migrating virtual machines among computer systems to balance load caused by virtual machines
US20090106409A1 (en) * 2007-10-18 2009-04-23 Fujitsu Limited Method, apparatus and recording medium for migrating a virtual machine
CN102253857A (en) * 2011-06-24 2011-11-23 华中科技大学 Xen virtual machine scheduling control method in multi-core environment
CN103294546A (en) * 2013-04-03 2013-09-11 华中科技大学 Multi-dimensional resource performance interference aware on-line virtual machine migration method and system
US20130346973A1 (en) * 2012-06-25 2013-12-26 Fujitsu Limited Management server, and virtual machine move control method
CN104156255A (en) * 2014-07-31 2014-11-19 华为技术有限公司 Virtual machine migration method, virtual machine migration device and source physical host

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7203944B1 (en) * 2003-07-09 2007-04-10 Veritas Operating Corporation Migrating virtual machines among computer systems to balance load caused by virtual machines
US20090106409A1 (en) * 2007-10-18 2009-04-23 Fujitsu Limited Method, apparatus and recording medium for migrating a virtual machine
CN102253857A (en) * 2011-06-24 2011-11-23 华中科技大学 Xen virtual machine scheduling control method in multi-core environment
US20130346973A1 (en) * 2012-06-25 2013-12-26 Fujitsu Limited Management server, and virtual machine move control method
CN103294546A (en) * 2013-04-03 2013-09-11 华中科技大学 Multi-dimensional resource performance interference aware on-line virtual machine migration method and system
CN104156255A (en) * 2014-07-31 2014-11-19 华为技术有限公司 Virtual machine migration method, virtual machine migration device and source physical host

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
姜雅楠: "云计算环境下虚拟机实时迁移及负载均衡研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108874534A (en) * 2018-04-18 2018-11-23 华为技术有限公司 A kind of dispatching method and device of PCPU resource
CN108874534B (en) * 2018-04-18 2022-09-16 华为技术有限公司 Method and device for scheduling PCPU (physical packet Unit) resources

Also Published As

Publication number Publication date
CN106155782B (en) 2019-06-25

Similar Documents

Publication Publication Date Title
US10176004B2 (en) Workload-aware load balancing to minimize scheduled downtime during maintenance of host or hypervisor of a virtualized computing system
CN105843683B (en) Method, system and equipment for the distribution of dynamic optimization platform resource
WO2018099299A1 (en) Graphic data processing method, device and system
US9009707B2 (en) Sharing reconfigurable computing devices between workloads
CN106354616B (en) Monitor the method, apparatus and high performance computing system of application execution performance
US11157302B2 (en) Idle processor management in virtualized systems via paravirtualization
CN104615480A (en) Virtual processor scheduling method based on NUMA high-performance network processor loads
WO2015032201A1 (en) Virtual machine placing method and device
DE102021123338A1 (en) PREDICTIVE CONTROL USING ONE OR MORE NEURAL NETWORKS
Ghribi Energy efficient resource allocation in cloud computing environments
Ahmadi et al. Performance evaluation of server virtualization in data center applications
CN107301088A (en) A kind of method and apparatus for managing virtual machine batch migration
Kim et al. Fair-share scheduling in single-ISA asymmetric multicore architecture via scaled virtual runtime and load redistribution
US20170140839A1 (en) Address translation stimuli generation for post-silicon functional validation
CN106155782A (en) Virtual machine migration method and device
Ismail et al. CPU and memory performance analysis on dynamic and dedicated resource allocation using XenServer in Data Center environment
CN107741868B (en) Setting method and device for Basic Input Output System (BIOS)
US20230195489A1 (en) Pluggable diagnostic tool for telco ran troubleshooting
Raj et al. Efficient resource allocation in resource provisioning policies over resource cloud communication paradigm
US11507437B2 (en) Deploying multiple different applications into a single short-lived container along with a master runtime
US9524193B1 (en) Transparent virtualized operating system
Cabrera et al. Energy efficient dynamic load balancing over multigpu heterogeneous systems
Muñoz et al. How to run scientific applications with dirac in federated hybrid clouds
CN107391228A (en) A kind of virtual machine migration management and device
Prades et al. Made-to-measure GPUs on virtual machines with rCUDA

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant