CN108369524A - The technology of collaboration migration for virtual machine - Google Patents
The technology of collaboration migration for virtual machine Download PDFInfo
- Publication number
- CN108369524A CN108369524A CN201580085487.0A CN201580085487A CN108369524A CN 108369524 A CN108369524 A CN 108369524A CN 201580085487 A CN201580085487 A CN 201580085487A CN 108369524 A CN108369524 A CN 108369524A
- Authority
- CN
- China
- Prior art keywords
- scene
- storage page
- destination node
- copy
- iteration
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000013508 migration Methods 0.000 title claims abstract description 145
- 230000005012 migration Effects 0.000 title claims abstract description 115
- 238000005516 engineering process Methods 0.000 title abstract description 6
- 230000006399 behavior Effects 0.000 claims abstract description 71
- 230000001617 migratory effect Effects 0.000 claims abstract description 70
- 238000003860 storage Methods 0.000 claims description 214
- 238000000034 method Methods 0.000 claims description 92
- 230000015654 memory Effects 0.000 claims description 53
- 230000008569 process Effects 0.000 claims description 48
- 230000004044 response Effects 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 11
- 235000015170 shellfish Nutrition 0.000 claims description 5
- 239000000306 component Substances 0.000 description 68
- 238000009826 distribution Methods 0.000 description 13
- 238000004891 communication Methods 0.000 description 11
- 230000008859 change Effects 0.000 description 8
- 230000009471 action Effects 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 239000004065 semiconductor Substances 0.000 description 4
- 238000004590 computer program Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000003068 static effect Effects 0.000 description 3
- 230000001052 transient effect Effects 0.000 description 3
- 230000000712 assembly Effects 0.000 description 2
- 238000000429 assembly Methods 0.000 description 2
- 239000003990 capacitor Substances 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 230000014509 gene expression Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000013468 resource allocation Methods 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- FMFKNGWZEQOWNK-UHFFFAOYSA-N 1-butoxypropan-2-yl 2-(2,4,5-trichlorophenoxy)propanoate Chemical compound CCCCOCC(C)OC(=O)C(C)OC1=CC(Cl)=C(Cl)C=C1Cl FMFKNGWZEQOWNK-UHFFFAOYSA-N 0.000 description 1
- 230000018199 S phase Effects 0.000 description 1
- 241000700605 Viruses Species 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000007667 floating Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000012533 medium component Substances 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 229920000642 polymer Polymers 0.000 description 1
- 230000002028 premature Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 229910052710 silicon Inorganic materials 0.000 description 1
- 239000010703 silicon Substances 0.000 description 1
- 229910052814 silicon oxide Inorganic materials 0.000 description 1
- -1 silicon oxide nitride Chemical class 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000001550 time effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5083—Techniques for rebalancing the load in a distributed system
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
- G06F2009/4557—Distribution of virtual machine instances; Migration and load balancing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
- G06F2009/45575—Starting, stopping, suspending or resuming virtual machine instances
Abstract
Example includes to be used for virtual machine（VM）The technology of migration.Example include the VM migratory behaviours of one or more VM in multiple VM for being responsible for by source node of adjustment for scene move to destination node so that the multiple VM can in the substantially the same time scene migration final stage.Final stage may include stopping and the last copy stage.
Description
Technical field
Example described herein relates generally to the virtual machine between nodes（VM）Migration.
Background technology
For being responsible for by node/server（host）Virtual machine（VM）Scene migration be for system（Such as data
Center）It can realize the important feature that fault-tolerant ability, flexible resource management or Dynamic workload rebalance.Scene migration can
Including the VM being responsible for by source node is moved to destination node by the network connection between source node and destination node.It moves
The scene of being considered to be is moved, because applying of being carrying out by the VM migrated can be after during largely scene transit time
It is continuous to be executed by VM.Executing can be only before copying residual state information to destination node from source node by briefly temporary
Stop, so that VM can restore to execute application in destination node.
Description of the drawings
Fig. 1 illustrates the virtual machine (vm) migration for system.
Fig. 2 illustrates the work integrated mode of example first.
Fig. 3 illustrates example first scheme.
Fig. 4 illustrates example prognostic chart.
Fig. 5 illustrates example alternative plan.
Fig. 6 illustrates Example third-party case.
Fig. 7 illustrates example fourth program.
Fig. 8 illustrates the work integrated mode of example second.
Fig. 9 illustrates the example block diagram of equipment.
Figure 10 illustrates the example of logic flow.
Figure 11 illustrates the example of storage medium.
Figure 12 illustrates example calculations platform.
Specific implementation mode
It, can during the most of time migrated at the scene by the application being carrying out by VM as contemplated in the disclosure
Continuation is executed by VM, so can be considered to be from source node/server to the scene migration of the VM of destination node/server existing
.Most of scene migration of VM can include the vm health information of the memory used by VM when execution is applied.Cause
This, scene migration is usually directed to the process in two stages.First stage can be the pre-copy stage, and it includes by initial memory
（Such as the first iteration）And the change memory for remaining iteration（Such as dirty storage page）It is copied to from source node
Destination node, while VM is still carrying out application or VM and is still just running on the source node.The pre-copy stage can continue, directly
It is brought down below threshold value to the dirty storage page of residue in source node.Then second stage can be off and finally copy the stage,
It stops or pauses at the VM at source node, by residual state information（Such as remaining dirty storage page and/or processor state,
Input/output state）It copies destination node to, and then restores the VM at destination node.The VM shapes in two stages
The copy of state information can be carried out by the network connection maintained between source node and destination node.
Second（Stop and finally copies）The time it takes amount is important in stage, because applying in this period
It is not just executed by VM inside.To be temporarily unresponsive executing any network service that the when applied just is providing.First
（Pre-copy）The time it takes amount is also important in stage, because this stage is to the total time for completing scene migration
There can be maximum time to influence.In addition, the computing resource of scene migration consumption relatively high amount, therefore saved in source node or destination
The performance of the other VM run on point could be severely affected.
When VM execute it is one or more in application, can be with the memory operation Ji Guan of VM for the VM significant challenges migrated
Connection.If the rate for keeping storage page dirty is more than the rate for migrating distributed network bandwidth for VM, may spend not
Acceptable long-time stops and the execution of one or more applications in last copy stage to pause at, because of a large amount of number
According to may still keep being copied into destination node from source node.This unacceptable long-time is problematic to VM migrations
, and migration may be caused to fail.
Network function virtualizes（NFV）VM migrations in environment propose additional challenges.In some NFV in use, service
Device or node usually supervisor are arranged to support virtual network function（VNF）The VM of chain.This is referred to alternatively as service chaining use.Service chaining
Interdepending between the VNF supported in use communicates between VM may be present.For example, the first VM can be used for virtual firewall with
Support the first VNF.The output of first VNF can be fed to for virtual router to support the 2nd VM of the 2nd VNF.Second
The output of VNF can be fed to for virtual virus scanner to support the 3rd VM, etc. of the 3rd VNF.It, will for this example
The first VM of the first VNF is supported to move to destination or destination node and may cause for comprising by being maintained at the corresponding of source node
Second and the 3rd VM supported second and the 3rd VNF service chaining in other VNF service close.In addition, only migration
The 3rd VM for holding the 3rd VNF may cause not come freely still in the corresponding first and second VM institutes of source node by the 3rd VNF
Inappropriate functional situation caused by the output of the first and second VNF in the service chaining of support.
In some instances, it is arranged to support that the parallel or collaboration migration of the VM of the VNF of service chaining can be for above
The solution for the migration challenge mentioned.However, collaboration migration may have additional challenges.First, the transition process of each VM
It can be independent and be difficult to control.Second, determine when to start to stop and finally copy rank for a set of or one group of VM
Section.It is to be noted, that once the VM for the VNF being included in service chaining is supported to enter stopping and last copy stage, then support to wrap
That set for the VNF being contained in service chaining or other VM of that group also stop.As a result, when the service of the VNF supported is closed
Between can enter from the first VM stopping and the last copy stage continue until that the last one VM in that set or that group has completed to move
Move past journey（Restore in target/destination node）.How third ensured that each VM can be in the approximately or substantially upper identical time
Reaching stopping and last copy stage may be also difficult with minimizing or reducing the service shut-in time.It is challenged for these,
Need example described herein.
Fig. 1 illustrates the VM migrations of example system 100.In some instances, as shown in fig. 1, system 100 includes and can lead to
Cross source node/server 110 that network 140 is communicatively coupled with destination node/server 120.Source node/server
110 and destination node/server 120 may be disposed to be responsible for multiple VM.For example, source node/server 110 can be responsible for VM
112-1,112-2,112-3 are to 112-n, wherein " n " is greater than 3 any whole positive integer.Destination node/server 120 is also
It can be responsible for the multiple VM to be migrated from source node/server 110.Supervisor may include providing in corresponding source node/service
It is that device 110 or destination node/server 120 maintain or to its addressable composition physical resource, such as processor, storage
Device, storage device or Internet resources（It is not shown）.Source node/server 110 and destination node/server 120 all may include
Corresponding migration manager 114 and 124 is to promote the migrations of VM between these nodes.In addition, in some instances, system 100
It can be arranged to provide infrastructures as service (IaaS), platform as service (PaaS) or software as service
(SaaS) part for data center.
In some instances, VM 112-1 to 112-n can be comprised in one group of VM, be arranged to support to be used as service chaining
Multiple VNF of 101 part.VM 112-1,112-2,112-3 and VM 122-n may be able to carry out corresponding one or more
A application（App）111-1,111-2,111-3 and 111-n support the relevant work of multiple VNF to load to carry out.App 111-1、
Corresponding state information 113-1,113-2,113-3 and 113-n of 111-2,111-3 and 111-n can reflect for execute these one
It is a or more to apply to carry out corresponding VM 112-1,112-2, VM 112-3 that support that the relevant work of multiple VNF loads
With the current state of VM 122-n.For these examples, status information 113-1 may include storage page 115-1 and operation information
117-1 is performed simultaneously App 111-1 to carry out and supports at least one VNF in multiple VNF to reflect the current state of VM 112-1
Workload.In addition, status information 113-1,113-2,113-2 and 113-n may include shared memory page 119-1.It is shared
Storage page 119-1 can be individually performed with by VM 112-1 to 112-n to support as the multiple of a part for service chaining 101
Shared data between VNFde applications is associated.In other words, shared memory page 119-1 may include for App 111-1
It is carried out to 111-n and supports the input needed for the relevant work load of multiple VNF or output data.
According to some examples, the workload as multiple VNF of a part for service chaining 101 is supported to can be used for and to can
Including one or more clients of the data center of system 100 provide the network service associated by IaaS, PaaS or SaaS
Service chaining.The network service supported by the VM 112-1 to 112-n of service chaining 101 may include, but are not limited to data bank network clothes
Business, website host web service, route network service, electronic mail network service or virus scan network service.For to institute
The performance requirement for stating one or more clients offer IaaS, PaaS or SaaS may include meeting one or more Service Qualities
Measure (QoS) criterion, service level agreement（SLA）And/or RAS（Reliability, availability and serviceability）It is required that.
In some instances, in the logic and/or feature of source node/server 110（Such as migration manager 114）It may
VM 112-1 included in service chaining 101 to the scenes 112-n can be moved to destination node/server 120.These VM
Collaboration migration may be in response to source node/server 110 terminate close to life cycle or can begin at show it is too early therefore
The sign of barrier（Such as QoS criterion or SLA cannot be met when being responsible for VM 112-1 to 112-n）Instruction.Life cycle terminates
Or these instructions of premature failure can result in the need for moving to VM 112-1 to 112-n from the collaboration of source node/server 110
Destination node/server 120, while to the network service of support（Such as VNF）With influence as small as possible.Example is unlimited
In for these reasons that the VM in service chaining is moved to another node/server from a node/server scene.For
The other examples reason of scene migration is susceptible to by the displosure.
According to some examples, migration manager 114 may include corresponding for executing it based on corresponding VM 112-1 to 112-n
App 111-1 to 11-n are to carry out relevant work load（Such as to support multiple VNF）To determine for VM 112-1 to 112-n
The integrated mode that works independently logic and/or feature.The identified integrated mode that works independently can be at least partially based on to VM
The available composition physical resources of 112-1 to 112-n（Such as processor, memory, storage device or Internet resources）Distribution
Resource.The logic and/or feature of migration manager 114 also based on work integrated mode determining respectively and can be based on distributing to VM
112-1 to 112-n migrates 130 network bandwidth to predict for VM 112-1 to the scenes 112-n to be migrated 130 for scene
To the migratory behaviour of destination node/server 120.Prediction algorithm may include：Based on individually distributing to each VM for existing
The network bandwidth of field migration migrated for 130 periods by dirty storage at the scene to determine for each VM in VM 112-1 to 112-n
Device page copy is to destination node/server 120 until the dirty memory of residue for the corresponding VM in VM 112-1 to 112-n
Can need how much to copy iteration until the number of thresholds that page is brought down below remaining dirty storage page（Such as complete the pre-copy stage）.
In some instances, migration manager 114 also may include logic and/or feature to be moved based on the VM individually predicted
It migrates to copy dirty storage page to destination node/service during VM 112-1 to 112-n migrate 130 at the scene to determine
Device 120 is brought down below remaining dirty storage page until the dirty storage page of residue for the corresponding VM in VM 112-1 to 112-n
Number of thresholds until required relative time amount.
According to some examples, as described in more detail below, the logic and/or feature of migration manager 114 are also possible to energy
It is enough to be adjusted to that these VM is made to migrate at the scene of VM 112-1 to 112-n for the VM migratory behaviours of VM 112-1 to 112-n
The substantially the same time of period copies remaining dirty storage page to destination node/service from source node/server 110
Device 120（Such as start to stop and the last copy stage）.Adjustment may include, but are not limited to：It controls or causes to individually distributing to VM
Each VM is controlled or is caused to individually distributing to VM for the adjustment of the network bandwidth of scene migration in 112-1 to 112-n
112-1 to 112-n executes its corresponding App 111-1 to 111-n to carry out the adjustment of the process resource of relevant work load, or
The selective VM from VM 112-1 to 112-n is set to enter the waiting iteration for other VM, this may take longer for coming
Complete its corresponding pre-copy stage.
In some instances, as shown in fig. 1, scene migration 130 can be passed through by network 140 in source node/server
110 network interface 116 and then by being route in the network interface 126 of destination node/server 120.It is right
In these examples, network 140 can be a part for the internal network for the data center that may include system 100.In some examples
In, carry out free source node/server 110 to maintain the scene for be completed within an acceptable amount of time by network 140
The network bandwidth of migration 130 or distribution to some amount of its available limited amount available network bandwidth.The network of the distribution
The some or all of of bandwidth are preassigned for supporting that VM is migrated or some or all of of the network bandwidth of the distribution can
With from being not included in service chaining 101（It is not shown）In and other VM for being responsible for by source node/server 110 borrow.It is borrowed
Network bandwidth can be borrowed to be completed at least up to scene migration 130.
According to some examples, it to be copied into the number of thresholds of the dirty storage page of residue of destination node/server 120
Between being based on source node/server 110 when closed in threshold value（Such as it completes to stop and finally copy the stage）It is saved using by source
Distributed network bandwidth of the distribution of point/server 110 for scene migration VM 112-1 to 112-n will come from storage page 115-
1 to the 115-n dirty storage page of residue, the dirty storage page of residue from shared memory page 119-1 and included in operation
At least processor and input/output state in information 117-1 to 117-n copy the energy of destination node/server 120 to
Power.Shut-in time threshold value can based on for the VM 112-1 to 112-n included in service chaining 101 in source node/server 110
The requirement for stopping and restoring in destination node/server 120 in given time period.These VM are stopped and closed
Closing the requirement restored in destination node/server 120 in time threshold can be arranged for meeting one or more QoS standards
Then, SLA and/or RAS requirements.Such as, it is desirable that it could dictate that less than several milliseconds（a couple milliseconds）Shut-in time
Threshold value.
Fig. 2 illustrates example work integrated mode 200.In some instances, work integrated mode 200 may include for by being used for
The VM 112-1 to 112-n that source node/server 110 as shown in Figure 1 of system 100 is responsible for are come the work that is individually determined
Integrated mode.For these examples, the work integrated mode being individually determined can be individually performed accordingly based on corresponding VM 112-1 to 112-n
Using 111-1 to 111-n to carry out relevant work load（Such as to support multiple VNF）.Included in work integrated mode 200
Each work integrated mode can track the dirty storage that each VM is used on given time based on the dirty pattern of usage log is collected
The quantity of device page it is writeable（Memory）Work integrated mode.Carry out the dirty pattern of usage log for each VM to can be used for be happened at
Dirty storage page is tracked during previous ones during the scene migration of each VM.In other words, because dirty storage page is just from source
Node/server copy produces that generated during this period or iteration new is dirty to deposit to destination node/server
Reservoir page.When given VM writes to given storage page, writing to the storage page for giving VM can be arranged in the dirty pattern of daily record
Protection, and by data structure（Such as bitmap, hash table, journal buffer or page change log recording）It is set to indicate that given deposit
Reservoir page is in fault time（Such as VM logs off virtualization）Dirty situation.After being write to given storage page, removal pair
In the write-protect of given storage page.Data structure can be regular（Such as every 10 milliseconds）It is verified to determine for giving VM
Dirty storage page total quantity.
In some instances, as shown in Figure 2, for the integrated mode 200 that works, the quantity of the dirty storage page when starting
In initial burst after, the rate that dirty storage page generates tends to the work integrated mode of the determination for each VM a bit
Steadily.According to some examples, for the dirty storage page of the work integrated mode of the given determination in work integrated mode 200
Generation be described using example equation 1：
(1) D=f (t)
For example equation 1, D indicates the dirty storage page generated, and f (t) indicates monotonically increasing function.Therefore, finally,
Being supplied to VM all can be from 0 for executing all memories for the application for carrying out the workload with work integrated mode 200
The storage page that dirty storage page changes to essentially all supply is all dirty.
In some instances, can be assumed：D=f (t) for the integrated mode that works can VM transition process at the scene
Period keeps constant.Therefore, the work integrated mode with D=f (t) tracked during previous ones can be with for current iteration
It is identical.Even if workload can float during given 24 hours one day, but may need re-sampling or tracking workload with
Determine the work integrated mode of reflection floating workload.For example, tracking every 30 minutes or can per hour occur, what D determined
=f (t) will be suitable for migrating given VM to use.For example, if workload was for 24 hours one day first part's phases
Second part compared with 24 hours one day be it is high, then and therefore can be for the more dirty storage page of each grey iterative generation, can
The scene migration of given VM can be needed to illustrate this increase of the rate generated with dirty storage page.
Fig. 3 illustrates exemplary scenario 300.In some instances, scheme 300 can be described migrates row for the VM of scene migration
For example, it is dirty it includes being generated when VM of the copy among the VM in service chaining moves to destination node by scene
Multiple copy iteration that storage page may need.For example, included in the service chaining 101 being responsible for by source node/server 110
In the executable applications of VM 112-2, while being migrated to destination node/server 120, moved as scene shown in Fig. 1
Move 130 part.For these examples, it is supplied to all storage pages of VM 112-2 that can all be indicated by " R ".In Fig. 3
It is shown, the beginning for the first iteration of scheme 300, at least part in R storage page or it is all be considered as dirty, such as
By example equation（2）D0Represented by=R.In other words, according to example equation（2）And as shown in Figure 3, in R storage page
At least part or all destination node/server 120 can be copied into during the first iteration.
According to some examples, example equation can be used in the period for completing the first iteration（3）To determine：
(3) T0=D0/W
For example equation（3）, W can indicate to be used for that VM 112-2 are moved to the distribution of destination node/server 120
Network bandwidth（Such as unit is megabyte per second (MBps)）.
In the beginning of secondary iteration, newly-generated dirty storage page is generated by the VM 112-2 for executing App 111-2, and
Carrying out workload during T0 can be by example equation（4）It indicates：
(4) D1=f (T0)
The period for copying the dirty storage pages of D1 can be by example equation（5）It indicates：
(5) T1=D1/W
Therefore, when q iteration starts, the quantity of dirty storage page can be by example equation（6）It indicates, wherein " q " is appointing for ＞ 1
What positive integer：
(6) Dq=f (Tq-1)
The period for copying the dirty storage pages of Dq can be by example equation（7）It indicates：
(7) Tq=Dq/W
In some instances, M can indicate the number of thresholds in the 110 remaining dirty storage page of residue of source node/server, can
The end in triggering pre-copy stage and the beginning for stopping and finally copying the stage, stop and the last copy stage includes to stop at source
VM 112-2 of node/server 110 and then by the dirty memory of residue of memory 115-2 and shared memory 119-1
Page and operational status information 117-2 copy destination node/server 120 to.For these examples, equation（8）Indicate surplus
The quantity of remaining dirty storage page is brought down below the condition of convergence of M：
Therefore, the quantity of the dirty storage page of residue at convergence can be indicated by Dc, and example equation（9）Dc<M instructions are remaining
The quantity of dirty storage page has fallen to less than number of thresholds M.
It can be by example equation in the period for stopping and copied during the last copy stage Dc（10）It indicates：
(10) TS=(Dc+SI) /W
For example equation（10）, what SI expressions were exited included in the time stopped in source node/server 110 in VM 112-2
Operational status information in the operational status information 117-2 of VM 112-2.
According to some examples, predicted time 310 as shown in Figure 3 indicates that remaining dirty storage page is brought down below number of threshold values
Measure the time quantum of M.As shown in Figure 3, this includes period T0、T1To TqThe sum of.Predicted time 320, as shown in Figure 3, instruction
It completes to close or stop and the last copy stage is to terminate to move to VM 112-2 the time of destination node/server 120.
As shown in Figure 3, this includes period TS。
In some instances, threshold value M can be based on VM 112-2 and stop in source node/server 110 and based on by institute
The network bandwidth W of distribution in the shut-in time threshold value of scene migration VM 112-2 destination node/server 120 again
The ability for starting or restoring.
According to some examples, the network bandwidth W of all distribution can be another from being responsible for by source node/server 110
VM is borrowed.In other examples, the first part of the network bandwidth W distributed may include that the pre- scene that remains for migrates VM 112-1
To the pre-assigned network bandwidth of 112-n, and second part may include from being responsible for by source node/server 110 and
The borrowed network bandwidth for not cooperateing with other VM of migration to borrow with VM 112-1 to 112-n.
In some instances, shut-in time threshold value can be based on for its in VM 112-2 and VM 112-1 to 112-n
Its VM source node/server 110 stop and in given time period destination node/server 120 restarting or
The requirement of recovery.For these examples, which can be arranged for meeting one or more QoS criterion, SLA is required and/
Or RAS requirements.
It is true using the scheme 300 for VM 112-2 and VM 112-1 and 112-3 to 112-n according to some examples
The fixed migratory behaviour individually predicted can be used for determining VM to migrate at the scene during 130 copying dirty storage page to destination section
The dirty storage page of residue of corresponding VM in for VM 112-1 to 112-n of point/server 120 is brought down below remaining dirty storage
Required relative time amount until the number of thresholds of device page.As described in more detail below, those relative times can be used for by
In the substantially the same time during being adjusted to migrate 130 at the scene for the practical VM migratory behaviours of VM 112-1 to 112-n
The remaining dirty storage page of copy.
Fig. 4 illustrates example prognostic chart 400.In some instances, prognostic chart 400 can show to move based on the scene for VM
Move the predicted time of the M number for being brought down below remaining memory page using how many distributed network bandwidths.Prognostic chart 400 is for example
It can be based on the example equation for using various different values for distributed network bandwidth（1）Extremely（9）Use, and also held based on VM
One or more applications of row have the workload for the work integrated mode for determining D=f (t) to carry out.
As shown in Figure 4, for prognostic chart 400, in convergence below in 5 seconds（It is brought down below the time of M）Seem until right
At least 200 MBps are assigned in the migration of VM just to occur.In addition, in the distributed network bandwidth of 800 MBps or more, do not show
With the more associated apparent time benefits of bandwidth of distribution.
According to some examples, prognostic chart 400 can be used for determining VM migratory behaviours for giving VM and for given
A variety of different distributed network bandwidths of determined work integrated mode.It can be for each of being responsible for by source node/server
VM generate similar to prognostic chart 400 independent prognostic chart, with based on distribute to VM for scene move to destination node
Network bandwidth determines that one group of scene VM for that will be included in service chaining moves to the independent of destination source node/server
VM migratory behaviours.
In some instances, prognostic chart 400 can be additionally used in determine for by VM from source node/server migration to destination
Node/server will need how many distributed network bandwidths.For example, if being currently allocated for the Netowrk tape of primary scene migration
Wide is that 200 MBps and QoS, SLA and/or RAS require to be arranged to 0.5 second threshold value to be brought down below " M ", then prognostic chart 400
Indicate the distributed network bandwidth of 600 MBps of needs approximation.To which for this example, additional 400 MBps can be assigned to
With 400 associated VM of prognostic chart, required to meet QoS, SLA and/or RAS.
Fig. 5 illustrates exemplary scenario 500.In some instances, scheme 500 depicts sometimes during iteration q, for
It migrates under given network bandwidth that each VM is distributed and the process resource for distributing to each VM, VM 112-1 to 112-n exist
Dirty storage page is copied to destination node during the migration of scene（Such as destination server/node 120）Until for VM
Needed for until the number of thresholds that the dirty storage page of residue of corresponding VM is brought down below remaining dirty storage page in 112-1 to 112-n
The relative time amount wanted.In other words, in iteration 1 to during q-1, VM is in the pre-copy stage, and the condition of convergence is for this
A little VM occur.For these examples, by distributing the network bandwidth (Wx) distributed for the migration of VM 122-x, in migration tube
The logic and/or feature for managing device 114 will may be adjusted to make this for the practical VM migratory behaviours of VM 112-1 to 112-n
A little VM are in approximation or the substantially the same time completes each iteration in pre-copy stage.For example, as shown in Figure 5, base
It is the network bandwidth W of VM 112-x distribution in the predictive behavior of each VM by Fig. 4i-xIt is taken turns in i so that VM 112-1 to 112-
N will respectively complete iteration, and reach boundary 510,520,530 and 540 in the substantially the same time.As in substantially phase
With time reach these boundaries as a result, VM 112-1 to 112-n can also enter stopping and most in the substantially the same time
After copy the stage.This can cause network service for being supported by the VM 112-1 to 112-n included in service chaining 101 or
The minimum or relatively small service shut-in time of VNF.
It in some instances, can be based on the visitor/main memory size for distributing to each VM and for VM 112-1 to 112-n
Collaboration migrate distributed network bandwidth to estimate the iteration 1 to contain each VM in VM 112-1 to 112-n
Duration.It is assumed that being W for cooperateing with the overall network bandwidth of migration0, then for T0Iteration 1 duration can be used example
Equation（11）To determine：
According to some examples, the network bandwidth W of each VM is individually distributed to for iteration 11Then each VM will be used
Example equation（12）To determine：
In some instances, for 1 subsequent iteration of iteration, example equation 11 and 12 can be used to realize similar approach.Example
Equation（13）Depict subsequent iteration：
For example equation（13）, D1-VM122-xIt is for memory dirty in previous round VM 122-x.
Fig. 6 illustrates exemplary scenario 600.In some instances, scheme 600 is similar to scheme 500, depicts VM 112-1
Dirty storage page is copied to the relative time amount of destination node during being migrated at the scene to 112-n.However, for scheme
600, it may cause to move the practical VM for VM 112-1 to 112-n in the logic and/or feature of migration manager 114
Migrate for coarseness adjustment so that these VM only reach boundary 610 in the substantially the same time.Coarseness is adjusted in determination
When each VM completes to may require less accuracy in copy iteration aspect.
According to some examples, as shown in Figure 6, VM 112-3 can complete it and copy iteration, and quickly than other VM
Reach the number of thresholds of remaining dirty storage page.For these examples, waits for iteration 620 that can be inserted in the pre- of VM 112-3 and copy
In the shellfish stage, so that VM 112-3 iterations 2, until all residue VM complete their corresponding pre-copy stage.It waits for
The quantity of iteration can be indicated that wherein p indicates any whole positive integer more than 1 by " p " in figure 6.Although Fig. 6 shows VM
112-1,112-2 and 112-n, but it respectively has equal q-1 iteration.In other examples, at least some of these VM exist
The iteration of different number can be passed through before reaching boundary 610.
In some instances, coarseness adjustment can make VM 112-1 to 112-n substantially together with waiting iteration 620 is inserted into
The identical time enters stopping and last copy stage.Similar to scheme 500, this can cause for by being included in service chaining 101
In minimum or relatively small service shut-in times of network service or VNF for being supported of VM 112-1 to 112-n.However,
In some examples, when gross migration time for scheme 600 for the gross migration time of scheme 500 than that can spend relatively longer
Between.This longer time can be by adding multiple waiting iteration rather than distributed resource is to cause in pre-copy rank between VM
The copy iteration that is essentially equal during section and cause.
Fig. 7 illustrates exemplary scenario 700.In some instances, scheme 700 is similar to scheme 500 and 600, depicts VM
Dirty storage page is copied to the relative time amount of destination node by 112-1 to 112-n during can migrating at the scene.Scheme 700
The combination or mixing that can indicate scheme 500 and 600 lead to practical VM migratory behaviour tune of the VM 112-1 to 112-n by them
It is made into at least some iteration for making these VM complete the pre-copy stage in the substantially the same time.For example, as shown in Figure 7,
Iteration 1 can be completed in the substantially the same time, and therefore, and VM 112-1 to 112-n can be reached in the substantially the same time
To boundary 710.
In some instances, make VM 112-1 to 112-n the substantially the same time reach boundary 710 adjustment it
Afterwards, the logic at migration manager 114 and/or feature, which can return, returns to not adjusting or being previously assigned after iteration 1
Process resource and/or network bandwidth.As shown in Figure 7, both VM 112-1 and VM 112-2 quickly reach surplus than other VM
The number of thresholds of remaining dirty storage page.VM 112-1 are shown as reaching number of thresholds after iteration 2, and VM 112-2 are shown
Go out to reach number of thresholds after iteration 3.For these examples, waits for iteration 630 that can be inserted in the pre- of VM 112-1 and copy
In the shellfish stage, so that VM 112-1 iterations 2, until VM 112-3 and 112-n complete their corresponding pre-copy stage.Deng
Wait for that iteration can be also inserted in the pre-copy stage of VM 112-2 so that VM 112-2 iterations 3, until VM 112-3 and
112-n completes their corresponding pre-copy stage.It is to be noted, that the value of " p " or " q " can be equal or different.In other words, VM
112-1 and 112-2 can be Jing Guo identical or different quantity p wait for iteration.VM 112-3 and 112-n can also reach boundary 720
Q iteration Jing Guo identical or different quantity before.
According to some examples, VM 112-1 to 112-n can be substantially or the approximate identical time reaches boundary
720, as shown in Figure 7.To the knot of the Mixed adjustment as the network bandwidth to being distributed at least some copy iteration
Fruit, VM 112-1 to 112-n can enter stopping and last copy stage in the substantially the same time.Similar to 500 He of scheme
600, these Mixed adjustments can cause the network for being supported by the VM 112-1 to 112-n included in service chaining 101 to take
The minimum or relatively small service shut-in time of business or VNF.However, in some instances, being used for the gross migration time of scheme 700
The relatively longer time can be spent than the gross migration time for scheme 500.This longer time can be multiple etc. by adding
Wait for iteration rather than distributed resource causes to cause the copy iteration being essentially equal during the pre-copy stage between VM.
However, since iteration 1 is usually directed to relatively long copy iteration compared to subsequent copy iteration, for gross migration
The time effects of time may be much smaller compared to scheme 600 for scheme 700.This can be for scheme 700 compared to side
Case 600 needs the less result for waiting for iteration.
Fig. 8 illustrates example work integrated mode 800.In some instances, as shown in Figure 8, work integrated mode 800 includes
The first work integrated mode for VM 112-3（Allocation）, it is included in work integrated mode 200 shown in Figure 2
Identical work integrated mode.For these examples, work integrated mode 800 includes the second work integrated mode for VM 112-3（Contracting
Deduction is matched）, show to be adjusted to disappear for executing the process resource of one or more applications if distributing to given VM
Subtract and generates the rate of dirty storage page and how may then influence work integrated mode.
According to some examples, VM 112-3 can make the process resource adjustment that it is distributed so that the speed that dirty storage page generates
Rate cuts down half by approximation.For these examples, this reduction can be indicated based on the migratory behaviour of the prediction for VM 112-3
VM 112-3, which have, completes its relative time amount for copying iteration and reaching the number of thresholds of remaining dirty storage page, at least
It is two double-lengths of VM 112-1,112-2 or VM 112-n.It adjusts process resource and reduces 50% so that the speed that dirty storage page generates
Rate, which slows down, can reduce the time that VM 112-3 complete its copy iteration, and therefore, VM 112-3 can be with VM 112-1,112-
2 or 112-n reaches the number of thresholds of remaining dirty storage page in the substantially the same time.
As shown in Figure 8, have for reducing assignment integrated mode and reach about 25000 after 20 seconds and dirty deposit
Reservoir page vs. reaches the curve of about 25000 dirty storage pages after 10 seconds before the distribution adjusted.
In some instances, other than adjustment process resource or without adjusting process resource, it also can adjust and individually divide
The network bandwidth of dispensing VM 112-1 to 112-n.For these examples, network bandwidth can be for changing with comparatively faster copy
The VM in generation is reduced and/or the VM with relatively slow copy iteration is increased, so that VM 112-1 to 112-n are existing
In the substantially the same remaining dirty storage page of time copy during the migration of field.These adjustment can for example face figure in realization
It is carried out while scheme 500,600 or 700 described in 5-7.
Fig. 9 illustrates the example block diagram for equipment 900.Although the equipment 900 being shown in FIG. 9 has in a certain topology
Have an element of limited quantity, but can be appreciated that, equipment 900 may include more or fewer elements in alternate topologies, such as
Given realization is desired.
According to some examples, equipment 900 can be by being arranged to be responsible for the circuit of source node/server maintenance of multiple VM
920 are supported.Circuit 920 may be disposed to execute the module or component 922-a that one or more softwares or firmware are realized.It is worth
It is noted that " a " is intended that the variable for indicating any positive integer with " b " with " c " and similar identifier used herein.From
And for example, if realizing the value of setting a=4, the software of component 922-a or the complete or collected works of firmware may include component 922-1,922-
2,922-3 or 922-4.In the example context without being limited thereto of presentation, and the different variables used in the whole text can indicate it is identical or
Different integer values.In addition, these " components " can be stored in the software/firmware in computer-readable medium, and although
Component is shown in Figure 9 for discrete blocks, but these components are not confined in completely different computer-readable medium component by this
Storage device（Such as single memory etc.）.
According to some examples, circuit 920 may include for realizing promote VM from source node/server to destination node/
The logic of the collaboration migration of server and/or the processor or processor circuit of feature（Such as migration manager 114）.As above
It is previously mentioned, circuit 920 can be in source node/server（Such as source node/server 110）A part for the circuit at place,
It may include process cores or element.Including the circuit of one or more process cores can be any various commercially available places
Device is managed, includes (but not limited to)：AMD Athlon, Duron and Opteron processors;ARM application, it is embedded or
Safe processor;IBM and Motorola DragonBall and PowerPC processors;At IBM and Sony units
Manage device;Intel®、Atom®、Celeron® Core (2) Duo®、Core i3、Core i5、Core i7、Itanium
, Pentium, Xeon, Xeon Phi, XScale processor；And similar processor.According to some examples, electricity
Road 920 also may include application-specific integrated circuit（ASIC）, and at least some component 922-a can be implemented as the hardware member of ASIC
Part.
According to some examples, equipment 900 may include schema component 922-1.Schema component 922-1 can be executed by circuit 920
To determine that the integrated mode that works independently of the VM for being responsible for by source node, the integrated mode that works independently are individually performed based on corresponding VM
At least one application is to carry out relevant work load.For these examples, schema component 922-1 may be in response to migration request 905
And work integrated mode is determined based on the information included in pattern information 910, the instruction of information 910 is when corresponding VM is in given processing
Corresponding VM generates dirty storage page when one or more applications being individually performed under resource allocation to carry out relevant work load
Respective rate.The integrated mode that works independently, which can be comprised in, to be maintained to the addressable data structures of schema component 922-1（Such as look into
Look for table（LUT））In work integrated mode 924-a in.
In some instances, equipment 900 also may include prediction component 922-2.Prediction component 922-2 can be held by circuit 920
Row is with based on the determined work integrated mode determined respectively by schema component 922-1（Such as included in work integrated mode 924-a
In）And it is predicted based on distributing to VM and moving to the network bandwidth of destination node for scene for migrating the scenes VM
To the independent VM migratory behaviours of destination node.For these examples, prediction component 922-2, which may have access to, is included in work integrated mode
Information in 924-a, distribution 924-b, threshold value 924-c and QoS/SLA 924-d is moved with predicting the independent VM for scene migration
Migrate for.Similar to work integrated mode 924-a, the information being included in distribution 924-b, threshold value 924-c and QoS/SLA 924-d
It can be maintained to the addressable data structures of prediction component 922-2（Such as LUT）In.In addition, for these examples, QoS/
SLA information 915 may include that threshold value 924-c and/or the information included in QoS/SLA 924-d is arranged.
In some instances, independent VM migratory behaviours can be predicted in prediction component 922-2 so that by schema component 922-1 institutes
Determining work integrated mode can be used for based on individually distribute to each VM for scene migration network bandwidth come determine for
Each VM in VM copies dirty storage page to destination node until for the corresponding VM's in VM during migrating at the scene
Need how much to copy iteration until the number of thresholds that remaining dirty storage page is brought down below remaining dirty storage page.
According to some examples, equipment 900 also may include time component 922-3.Time component 922-3 can be held by circuit 920
Row during VM is migrated at the scene to determine based on the VM migratory behaviours individually predicted to copy dirty storage page to destination
Node is until the dirty storage page of residue for the corresponding VM in VM is brought down below the number of thresholds of remaining dirty storage page
Required relative time amount.For these examples, iteration timing 924-e（Such as it is stored in LUT）It may include that these are opposite
Time quantum, as the opposite copy iteration for each iteration in multiple copy iteration.
In some instances, equipment 900 also may include adaptive-migration component 922-4.Adaptive-migration component 922-4 can
It is executed by circuit 920 to be adjusted to make VM dirty to be deposited remaining during the scene migration of VM the practical VM migratory behaviours for being used for VM
Reservoir page copies destination node in the substantially the same time from source node.For these examples, adaptive-migration component
922-4 can be via self-adjusted block 924-f（Such as maintained in LUT）To track these adjustment.Redistributing 930 can wrap
Containing the adjustment to practical VM migratory behaviours, this includes the tune to individually distributing to each VM for the network bandwidth of scene migration
It is whole and/or make to distribute to VM and be adjusted with the process resource for carrying out relevant work load.Iteration 940 is waited for may include to practical VM
The adjustment of migratory behaviour, this includes that the VM comparatively fast copied is made to repeat given copy iteration to allow the VM copied more slowly copies remaining
Dirty storage page, and to copy their the dirty memory of residue during so that VM is migrated at the scene in the substantially the same time
Each of page.
Include herein is one group of logic flow of the exemplary method opinion for representing the novel aspect for executing disclosed framework
Journey.Although for the simplification of explanation, one or more methodology shown in this article are shown and described as a series of actions, this
Field technology personnel will understand and appreciate that methodology is not limited by the sequence acted.Some actions can be by difference according to this
Occur sequentially and/or simultaneously with other actions different from being illustrated and described herein.For example, it will be appreciated by those skilled in the art that and
It recognizes, methodology may alternatively be represented as a series of be mutually related state or events, such as in state diagram.And
And novelty is realized, it is not that the everything of illustration is required in methodology.
Logic flow can be realized with software, firmware and/or hardware.In software and firmware embodiments, logic flow can
By being stored at least one non-transitory computer-readable medium or machine readable media（Such as light, magnetic or semiconductor storage）
On computer executable instructions realize.In embodiment context without being limited thereto.
Figure 10 illustrates the example of logic flow 1000.Logic flow 1000 can be represented by described herein one or more
A logic, feature or device（Such as equipment 1000）Performed some or all of operations.More specifically, logic flow 1000
It can at least be realized by schema component 922-1, prediction component 922-2, time component 922-3 or adaptive-migration component 922-4.
According to some examples, logic flow 1000 can determine that the VM's for being responsible for by source node is independent in frame 1002
Work integrated mode, and at least one application is individually performed to carry out relevant work load based on corresponding VM in the integrated mode that works independently（Example
Such as under given processing resource allocation）.For these examples, schema component 922-1 can determine the integrated mode that works independently.
In some instances, logic flow 1000 frame 1004 can be based on the work integrated mode and base determined respectively
It is predicted in distributing to VM and moving to the network bandwidth of destination node for scene for the scenes VM to be moved to destination
The independent VM migratory behaviours of node.For these examples, the given process resource for distributing to VM is considered.In addition, showing for these
Example, prediction component 922-2 can predict the independent VM migratory behaviours for scene migration VM.
According to some examples, logic flow 1000 can be determined in frame 1006 based on the independent VM migratory behaviours predicted
VM copies dirty storage page to destination node until the dirty storage of residue for the corresponding VM in VM during migrating at the scene
Required relative time amount until the number of thresholds that device page is brought down below remaining dirty storage page.For these examples, time
Component 922-3 can determine relative time amount.
In some instances, the migration scheme for scene migration VM may be implemented to incite somebody to action in frame 1008 in logic flow 1000
Practical VM migratory behaviours for VM be adjusted to that VM is made to migrate at the scene during by remaining dirty storage page substantially the same
Time copies destination node to from source node.For these examples, adaptive-migration component 922-4 may be implemented for scene
Migrate the migration scheme of VM.
Figure 11 illustrates the example of storage medium 1100.Storage medium 1100 may include product.In some instances, it stores
Medium 1100 may include any non-transitory computer-readable medium or machine readable media, such as light, magnetic or semiconductor storage dress
It sets.Storage medium 1100 can store various types of computer executable instructions, such as finger for realizing logic flow 1000
It enables.Computer-readable or machine readable storage medium example may include any tangible medium that can store electronic data, packet
Containing volatile memory or nonvolatile memory, can be removed or non-removable memory, erasable or nonerasable memory,
Writeable or recordable memory, etc..The example of computer executable instructions may include the code of any suitable type, such as source
Code, the code of compiling, the code of interpretation, executable code, static code, dynamic code, the code of object-oriented, vision generation
Code and it is such.In example context without being limited thereto.
Figure 12 illustrates example calculations platform 1200.In some instances, as shown in Figure 12, computing platform 1200 can wrap
Containing processing component 1240, other platform assemblies 1250 or communication interface 1260.According to some examples, computing platform 1200 can save
It is realized in point/server.Node/server can be coupled to other node/servers by network, and can be packet
A part for the data center of node/server containing the multiple network connections for being arranged to supervisor VM.
According to some examples, processing component 1240 can perform the processing operation for equipment 900 and/or storage medium 1100
Or logic.Processing component 1240 may include various hardware elements, software element or combination.The example of hardware element can wrap
Containing device, logic device, component, processor, microprocessor, circuit, processor circuit, circuit element（Such as transistor, resistance
Device, capacitor, inductor etc.）, integrated circuit, application-specific integrated circuit（ASIC）, programmable logic device（PLD）, number letter
Number processor（DSP）, field programmable gate array（FPGA）, memory cell, logic gate, register, semiconductor device, core
Piece, microchip, chipset etc..The example of software element may include component software, program, application, computer program, using journey
Sequence, device driver, system program, software development procedures, machine program, operating system software, middleware, firmware, software mould
Block, routine, subroutine, function, method, regulation, software interface, application programming interfaces（API）, instruction set, calculation code, calculating
Machine code, code segment, computer code segments, word, value, symbol, or any combination thereof.Determine whether using hardware element and/or
Software element, which carrys out implementation example, to be changed according to any amount of factor, all computation rates as desired, power grade, heat-resisting
Property, process cycle budget, input data rate, output data rate, memory resource, data bus speed and other designs
Or performance constraints, it is such as desired for given example.
In some instances, other platform assemblies 1250 may include public computing element, such as one or more processing
Device, coprocessor, memory cell, chipset, controller, peripheral hardware, interface, oscillator, timing means, regards multi-core processor
Frequency card, audio card, multimedia input/output（I/O）Component（Such as digital display）, power supply etc..Memory cell
Example may include and be not limited to using one or more higher speed memory cells form various types of computers
Readable and machine readable storage medium, such as read-only memory（ROM）, random access memory（RAM）, dynamic ram（DRAM）、
Double data rate DRAM（DDRAM）, synchronous dram（SDRAM）, static state RAM（SRAM）, programming ROM（PROM）, it is erasable can
Programming ROM（EPROM）, electrically erasable ROM（EEPROM）, flash memory, such as ferroelectric polymer memory it is poly-
Close object memory, ovonic memory, phase transformation or ferroelectric memory, silicon oxide nitride oxide silicon（SONOS）Memory, magnetically or optically card,
Such as redundant array of independent disks（RAID）Apparatus array, the solid state memory device of driving（Such as USB storage）, solid-state driving
（SSD）And it is suitable for storing the storage medium of any other type of information.
In some instances, communication interface 1260 may include the logic and/or feature for supporting communication interface.For this
A little examples, communication interface 1260 may include being operated with by direct or network communication link according to various communication protocols or standard
Or one or more communication interfaces that channel is communicated.Direct communication can be via using in one or more industrial standards
（Including filial generation and deformation）（Those of be such as associated with PCIe specification）Described in communication protocol or standard and occur.Network is logical
Letter can be via using communication protocol or standard（That such as described in one or more ethernet standards promulgated by IEEE
A bit）And occur.For example, such ethernet standard may include IEEE 802.3.Network communication also can be according to one or more
A OpenFlow specifications（Such as OpenFlow hardware abstractions API specification）And occur.
As mentioned above, computing platform 1200 may be implemented in server/node of data center.Correspondingly, originally
Text description computing platform 1200 function and/or specific configuration can be comprised in the various embodiments of computing platform 1200 or
Person is omitted, as appropriate desired for server/node institute.
Discrete circuit, application-specific integrated circuit can be used in the component and feature of computing platform 1200（ASIC）, logic gate and/or
Any combinations of single-chip framework are realized.In addition, microcontroller, programmable logic battle array can be used in the feature of computing platform 1200
Row and/or microprocessor or any combinations above-mentioned are realized being suitble to appropriate place.It is to be noted, that hardware, firmware and/or software
Element may be collectively referred to herein as or be individually referred to as herein " logic " or " circuit ".
It should be appreciated that the example computing platform 1200 shown in the block diagram of Figure 12 can indicate the one of many potential realizations
A example functionally described.Correspondingly, the division of the block function of describing in the accompanying drawings, omit or comprising and do not imply that for real
Hardware component, circuit, software and/or the element of these existing functions be bound to be divided in embodiment, omit or comprising.
At least one exemplary one or more aspects can by be stored in indicate processor in various logic it is at least one
Representative instruction on machine readable media realizes that described instruction makes described when being read by machine, computing device or system
Machine, computing device or system make the logic for executing technique described herein.Such expression of referred to as " IP kernel " can quilt
It is stored on tangible machine readable media, and is supplied to various customers or manufacturing facility, to be loaded into described in actually manufacture
In the making machine of logic or processor.
Hardware element, software element or combination can be used to realize for various examples.In some instances, hardware member
Part may include device, component, processor, microprocessor, circuit, circuit element（Such as transistor, resistor, capacitor, inductance
Device etc.）, integrated circuit, application-specific integrated circuit（ASIC）, programmable logic device（PLD）, digital signal processor（DSP）、
Field programmable gate array（FPGA）, memory cell, logic gate, register, semiconductor device, chip, microchip, chipset
Etc..In some instances, software element may include component software, program, application, computer program, application program, system journey
Sequence, machine program, operating system software, middleware, firmware, software module, routine, subroutine, function, method, regulation, software
Interface, application programming interfaces（API）, instruction set, calculation code, computer code, code segment, computer code segments, word, value,
Symbol, or any combination thereof.Determining whether to carry out implementation example using hardware element and/or software element can be according to any amount of
Factor and change, all computation rates as desired, power grade, heat resistance, processing cycle budget, input data rate, output
Data rate, memory resource, data bus speed and other designs or performance constraints, such as given realization are desired.
Some examples may include product or at least one computer-readable medium.Computer-readable medium may include for depositing
Store up the non-transient storage media of logic.In some instances, non-transient storage media may include that the one of electronic data can be stored
Kind or more type computer readable storage medium, comprising volatile memory or nonvolatile memory, it is removable or
Non-removable memory, erasable or nonerasable memory, writeable or recordable memory etc..In some instances, it patrols
It collects and may include various software elements, such as component software, program, application, computer program, application program, system program, machine
Program, operating system software, middleware, firmware, software module, routine, subroutine, function, method, regulation, software interface,
API, instruction set, calculation code, computer code, code segment, computer code segments, word, value, symbol, or any combination thereof.
According to some examples, computer-readable medium may include the non-transient storage media for storing or preserving instruction,
Described instruction makes the machine, computing device or system execute according to described when being executed by machine, computing device or system
Exemplary method and/or operation.Instruction may include the code of any suitable type, such as code of source code, compiling, interpretation
Code, executable code, static code, dynamic code and such.Instruction can be according to for commanding machine, computing device
Or system executes pre-defined computer language, mode or the syntax of a certain function to realize.Any be suitble to can be used in instruction
Advanced, rudimentary, object-oriented, visual, compiling and/or the programming language of interpretation realize.
Statement " in one example " or " example " can be used to be described together with their derivative for some examples.These arts
Language is it is meant that specific features, structure or characteristic in conjunction with described in the example are comprised at least one example.In specification
In the phrase that occurs everywhere be not necessarily all referring to same example " in one example ".
Statement " coupling " and " connection " can be used to be described together with their derivative for some examples.These terms are not necessarily
It is intended as mutual synonym.For example, the description using term " connection " and/or " coupling " may indicate that two or more elements
It is in direct physical contact with each other or is in electrical contact.However, term " coupling " also means that two or more elements do not connect directly each other
It touches, but still intemperates with one another or interact.
Following example is related to the additional example of presently disclosed technology.
1. a kind of example apparatus of example, it may include circuit and schema component, the schema component by the circuit for being held
For row to determine the integrated mode that works independently of the VM for being responsible for by source node, it is mono- that the integrated mode that works independently is based on corresponding VM
At least one application is solely executed to carry out relevant work load.The equipment can further include prediction component, be used for by the electricity
Road is executed to move to destination section based on work integrated mode determining respectively and based on the VM is distributed to for scene
The network bandwidth of point predicts the independent VM migratory behaviours for the scenes VM to be moved to the destination node.It is described to set
Standby can further include time component, be used to be executed by the circuit described to determine based on the VM migratory behaviours individually predicted
VM migrate at the scene during by dirty storage page copy to the destination node until the corresponding VM for the VM residue
Required relative time amount until the number of thresholds that dirty storage page is brought down below remaining dirty storage page.The equipment can be also
Including adaptive-migration component, it is used to be executed to adjust the practical VM migratory behaviours for the VM so that institute by the circuit
VM is stated to copy remaining dirty storage page from the source node in the substantially the same time during the scene migration of the VM
To the destination node.
The equipment of example 2. as described in example 1, the number of thresholds can be based on remaining dirty storage page and at least locate
Reason device and input/output state copy the destination node to so that corresponding VM execution at least one application is to close
It moves to the network bandwidth of the destination node for live and carries out using distributing to the VM in time threshold
The relevant work load.
The equipment of example 3. as described in example 2, the number of thresholds can for the VM at least one of VM be not
With.
The equipment of example 4. as described in example 2, the shut-in time threshold value can be based on saving the VM in the source
Point stops and in given time period in the requirement of destination node restarting.The requirement can be arranged for meeting
One or more QoS criterion or SLA.
The equipment of example 5. as described in example 1, the prediction component can be based on being determined respectively by the schema component
Integrated mode work to predict that independent VM migratory behaviours, the work integrated mode are used to be based on individually to distribute to each VM for existing
Migration network bandwidth come determine each VM in the VM is migrated at the scene during copy dirty storage page to institute
Destination node is stated until the dirty storage page of residue for the corresponding VM in the VM is brought down below remaining dirty storage page
The number of thresholds needs how many copy iteration.
Equipment of the example 6. as described in example 5, for adjusting the practical VM migratory behaviours for the VM so that the VM
During migrating at the scene remaining dirty storage page is copied to the destination in the substantially the same time from the source node
The adaptive-migration component of node may include for making individually to distribute to each VM the network for scene migration
The adaptive-migration component that bandwidth is adjusted.The adaptive-migration component can also make individually to distribute to the VM to carry out
The process resource of the relevant work load is adjusted.For these examples, any of described adjustment or the two can make institute
VM is stated to reach the number of thresholds of remaining dirty memory page and start the remaining dirty memory of copy in the substantially the same time
Page.
Equipment of the example 7. as described in example 6, can further include the time component, can determine in the VM
Each VM copies dirty storage page to the destination node during migrating at the scene and needs multiple copy iteration.The equipment
Can further include the adaptive-migration component, making the network bandwidth for individually distributing to each VM be adjusted and/or
The process resource for individually distributing to the VM is set to be adjusted.For these examples, any of described adjustment or the two
It can make to complete each of the multiple copy iteration of VM copy iteration in substantially the same time quantum, with
During so that the VM is migrated at the scene remaining dirty storage page is copied to institute in the substantially the same time from the source node
State destination node.
Equipment of the example 8. as described in example 6 can further include the time component, to determine in the VM
Each VM migrate at the scene during copy dirty storage page to the destination node and need multiple copy iteration.Such as example
Equipment described in 6 can further include the adaptive-migration component, to make the network bandwidth for individually distributing to each VM
It is adjusted.The adaptive-migration component can also be such that the process resource for individually distributing to the VM is adjusted.For these
Example, any of described adjustment carried out by the adaptive-migration component or the two can make for described in the VM
The first copy iteration in multiple copy iteration is completed in substantially the same time quantum.The adaptive-migration component can be also
Make to recover to being previously assigned for network bandwidth and/or process resource after the first copy iteration.The adaptive-migration
When component can also make the first VM in the VM spend longer compared to the first VM in response to the 2nd VM in the VM
The area of a room repeats described first described subsequent at least part for copying iteration for copying iteration to complete then copy iteration.Institute
Stating adaptive-migration component can also make completed to stop the subsequent copy after the subsequent copy iteration by the 2nd VM
The repetition of the first VM of iteration, so that in the substantially the same time that the residue is dirty during the VM is migrated at the scene
Storage page copies the destination node to from the source node.
Equipment of the example 9. as described in example 6 can further include the time component, to determine in the VM
Each VM migrate at the scene during copy dirty storage page to the destination node and need multiple copy iteration.Such as example
Equipment described in 6 can further include the adaptive-migration component, to make the network bandwidth for individually distributing to each VM
It is adjusted and/or the process resource for individually distributing to the VM is made to be adjusted, any of described adjustment or the two are used
In make in the multiple copy iteration of the VM first copy iteration completed in substantially the same time quantum.Such as
Equipment described in example 6 can further include the schema component, to based on the network bandwidth and/or process resource adjusted
To determine second working independently integrated mode for the VM.Equipment as described in example 6 can further include the prediction component,
Its second independent VM migrated to the scene predicted for the VM based on the network bandwidth and/or process resource that are adjusted
Migratory behaviour.Equipment as described in example 6 can further include the time component, to based on the second independent VM predicted
Migratory behaviour copies dirty storage page to the destination node until being used for determine during the VM is migrated at the scene
Required for until the number of thresholds that the dirty storage page of residue for stating the corresponding VM in VM is brought down below remaining dirty storage page
Relative surplus time quantum.Equipment as described in example 6 can further include the adaptive-migration component, is used for adjust
Practical VM migratory behaviours of the VM during the subsequent copy iteration of the first copy iteration is so that the VM is moved at the scene
During shifting remaining dirty storage page is copied to the destination node in the substantially the same time from the source node.
Equipment of the example 10. as described in example 5, to adjust the practical VM migratory behaviours for the VM so that described
VM in the substantially the same time copies remaining dirty storage page to the purpose during migrating at the scene from the source node
The adaptive-migration component of ground node will during may include migrating each VM in the VM at the scene to determination
Dirty storage page copies the time component that the destination node needs multiple copy iteration to.For these examples, institute
The determination that adaptive-migration component responds are stated in multiple copy iteration can make the first VM in the VM in response in the VM
2nd VM completes the last copy iteration of the 2nd VM compared to the first VM amounts of taking longer time and repeats institute
At least part of the last copy iteration of the first VM is stated to reach the number of thresholds of remaining dirty storage page.It is described adaptive
Answer migration component that the last copy for stopping the first VM after the last copy iteration for completing the 2nd VM can also be made to change
The repetition of first VM in generation, so as to be deposited in the substantially the same time by the residue is dirty during the VM is migrated at the scene
Reservoir page copies the destination node to from the source node.
The equipment of example 11. as described in example 1, the VM may be disposed to support the multiple of the part as service chaining
VNF, the integrated mode that works independently are individually performed at least one application based on the corresponding VM and support the multiple VNF to carry out
Relevant work load.
The equipment of example 12. as described in example 1, the source node and the destination node may include being arranged to carry
For in the data center of IaaS, PaaS or SaaS.
The equipment of example 13. as described in example 1 can also include to be coupled to the circuit with presentation user's interface view
Digital display.
14. a kind of exemplary method of example, it may include determine the list of the VM for being responsible for by source node in processor circuit
Solely work integrated mode.At least one application can be individually performed to carry out relevant work in the integrated mode that works independently based on corresponding VM
Load.The method can further include being moved for scene based on work integrated mode determining respectively and based on the VM is distributed to
The network bandwidth of destination node is moved on to predict to move for the scenes VM to be moved to the independent VM of the destination node
Migrate for.The method can further include will during the VM is migrated at the scene to determine based on the independent VM migratory behaviours predicted
Dirty storage page copies the destination node to until the corresponding VM of the dirty storage page of residue to(for) the VM is brought down below
Required relative time amount until the number of thresholds of remaining dirty storage page.The method can further include realizing to be used for the VM
Scene migration migration scheme with during adjusting practical VM migratory behaviours for the VM so that the VM is migrated at the scene
Remaining dirty storage page is copied to the destination node in the substantially the same time from the source node.
Method of the example 15. as described in example 14, the number of thresholds can be based on by remaining dirty storage pages and at least
Processor and input/output state copy the destination node to so that corresponding VM execution at least one application is to close
It closes next real for the live network bandwidth for moving to the destination node using the VM is distributed in time threshold
The row relevant work load.
Method of the example 16. as described in example 15, the number of thresholds can for the VM at least one of VM be
Different.
Method of the example 17. as described in example 15, the shut-in time threshold value can be based on will be in the source for the VM
Node stops and is arranged for meeting in the requirement of destination node restarting, the requirement in given time period
One or more QoS criterion or SLA.
Method of the example 18. as described in example 14 predicts that independent VM migratory behaviours may include the working set determined respectively
Pattern, the work integrated mode be used for based on individually distribute to each VM for scene migration network bandwidth come determine for
Each VM in the VM copies dirty storage page to the destination node until for the VM during migrating at the scene
In the dirty storage page of residue of corresponding VM be brought down below the number of thresholds of remaining dirty storage page how many copies needed to change
Generation.
Method of the example 19. as described in example 18, realize adjusting practical VM migratory behaviours for the VM so that
The VM copies remaining dirty storage page from the source node in the substantially the same time described during migrating at the scene
The migration scheme of destination node may include the Netowrk tape that adjustment individually distributes to each VM for scene migration
Wide and/or adjustment individually distributes to the VM to carry out the process resource of the relevant work load, so that the VM reaches surplus
The number of thresholds of remaining dirty memory page and start the remaining dirty storage page of copy in the substantially the same time.
Method of the example 20. as described in example 19 can further include that determination migrates each VM in the VM at the scene
Period copies dirty storage page to the destination node and needs multiple copy iteration.The method can further include that adjustment is independent
It distributes to the network bandwidth of each VM and/or adjusts the process resource for individually distributing to the VM so that for institute
It states each of the multiple copy iteration of VM copy iteration to complete in substantially the same time quantum, so that the VM exists
Remaining dirty storage page is copied to the destination during the migration of scene from the source node in the substantially the same time to save
Point.
Method of the example 21. as described in example 19 can further include that determination migrates each VM in the VM at the scene
Period copies dirty storage page to the destination node and needs multiple copy iteration.The method can further include that adjustment is independent
It distributes to the network bandwidth of each VM and/or adjusts the process resource for individually distributing to the VM so that for institute
The in the multiple copy iteration of VM first copy iteration is stated to complete in substantially the same time quantum.The method can be also
It is included in after the first copy iteration and recovers to being previously assigned for network bandwidth and/or process resource.The method can be also
Including making the first VM in the VM in response to the 2nd VM in the VM compared to the first VM amounts of taking longer time
To complete then to copy described subsequent at least part for copying iteration of iteration and repetition the first copy iteration.The side
Method can further include stop the subsequent copy iteration after completing the subsequent copy iteration by the 2nd VM described the
The repetition of one VM so that during the VM is migrated at the scene in the substantially the same time by the dirty storage page of residue from institute
It states source node and copies the destination node to.
Method of the example 22. as described in example 19 can further include that determination migrates each VM in the VM at the scene
Period copies dirty storage page to the destination node and needs multiple copy iteration.The method can further include that adjustment is independent
It distributes to the network bandwidth of each VM and/or adjusts the process resource for individually distributing to the VM so that for institute
The in the multiple copy iteration of VM first copy iteration is stated to complete in substantially the same time quantum.The method can be also
Include that second working independently integrated mode for the VM is determined based on the network bandwidth and/or process resource that are adjusted.Institute
The method of stating can further include predicting the of the scene migration for the VM based on the network bandwidth and/or process resource that are adjusted
Two independent VM migratory behaviours.The method can further include determining that the VM exists based on the second independent VM migratory behaviours predicted
During the migration of scene dirty storage page is copied to the destination node until the residue for the corresponding VM in the VM is dirty
Required relative surplus time quantum until the number of thresholds that storage page is brought down below remaining dirty storage page.The side
Method can further include realizing the migration scheme of the scene migration for the VM, and the migration scheme adjustment exists for the VM
Practical VM migratory behaviours during the subsequent copy iteration of the first copy iteration, the practical VM migratory behaviours make the VM
During migrating at the scene remaining dirty storage page is copied to the destination in the substantially the same time from the source node
Node.
Method of the example 23. as described in example 18 is realized the migration scheme and is moved with the practical VM for adjusting for the VM
Migrate for so that during the VM is migrated at the scene in the substantially the same time by remaining dirty storage page from the source node
Copy to the destination node may include determining each VM in the VM is migrated at the scene during by dirty storage page
It copies the destination node to and needs multiple copy iteration.Realize that the migration scheme can further include first made in the VM
VM completes the 2nd VM most in response to the 2nd VM in the VM compared to the first VM amounts of taking longer time
It copies iteration afterwards and repeats at least part of the last copy iteration of the first VM to reach the institute of remaining dirty storage page
State number of thresholds.Realize the migration scheme can further include the 2nd SM complete the 2nd VM last copy iteration it
The repetition for stopping the first VM of the last copy iteration of the first VM afterwards, so that in base during the VM is migrated at the scene
The identical time copies the dirty storage page of residue to the destination node from the source node in sheet.
Method of the example 24. as described in example 14, the VM may be disposed to support the multiple of the part as service chaining
VNF.It is the multiple to carry out support that at least one application can be individually performed in the integrated mode that works independently based on the corresponding VM
The relevant work of VNF loads.
Method of the example 25. as described in example 14, the source node and the destination node may include being arranged to carry
For in the data center of IaaS, PaaS or SaaS.
26. example at least one machine readable media of example, it may include multiple instruction, described instruction may be in response to just by
System executes in computing platform and the system is made to execute the method according to any one of example 14 to 25.
A kind of 27. example apparatus of example, it may include the portion for executing the method as described in any one of example 14 to 25
Part.
28. example at least one machine readable media of example, may include multiple instruction, described instruction is in response to by system
Execute the integrated mode that works independently that the system can be made to determine the VM for being responsible for by source node.The integrated mode that works independently
At least one application can be individually performed based on corresponding VM to carry out relevant work load.Described instruction can also be such that the system is based on
The work integrated mode that determines respectively and based on distributing to the VM for the live network bandwidth for moving to destination node
To predict the independent VM migratory behaviours for the scenes VM to be moved to the destination node.Described instruction can also make described
System copies dirty storage page to institute to determine based on the independent VM migratory behaviours predicted during the VM is migrated at the scene
Destination node is stated until the corresponding VM of the dirty storage page of residue to(for) the VM is brought down below the threshold of remaining dirty storage page
Required relative time amount until value quantity.Described instruction can also make the system realize that the scene for the VM migrates
Migration scheme is to adjust practical VM migratory behaviours for the VM so that substantially the same during the VM is migrated at the scene
Time copy remaining dirty storage page to the destination node from the source node.
At least one machine readable media of the example 29. as described in example 28, the number of thresholds can be based on will be remaining dirty
Storage page and at least processor and input/output state copy to the destination node so as to corresponding VM execute described in extremely
A few application for scene in threshold value between when closed to move to the destination node using distributing to the VM
The network bandwidth loads to carry out the relevant work.
At least one machine readable media of the example 30. as described in example 29, the number of thresholds can for the VM it
At least one of VM be different.
At least one machine readable media of the example 31. as described in example 29, the shut-in time threshold value can be based on for
The VM will stop in the source node and in given time period in the requirement of destination node restarting, described
It is required that being arranged for meeting one or more QoS criterion or SLA.
At least one machine readable media of the example 32. as described in example 28, described instruction is for making the system prediction
Independent VM migratory behaviours may include that the work integrated mode determined respectively, the work integrated mode are used for every based on individually distributing to
A VM for the network bandwidth of scene migration come determine each VM in the VM is migrated at the scene during by dirty storage
Device page copy is to the destination node until the dirty storage page of residue for the corresponding VM in the VM is brought down below residue
The number of thresholds of dirty storage page needs how many copy iteration.
At least one machine readable media of the example 33. as described in example 32, for making the system realize adjusting
For the VM practical VM migratory behaviours so that the VM migrate at the scene during will be remaining dirty in the substantially the same time
The described instruction that storage page copies the migration scheme of the destination node to from the source node may include for making
The system call interception individually distributes to each VM and individually distributes to institute for the network bandwidth of scene migration and/or adjustment
VM is stated to carry out the process resource of the relevant work load, so that the VM reaches the number of thresholds of remaining dirty memory page
And start the described instruction of the remaining dirty storage page of copy in the substantially the same time.
At least one machine readable media of the example 34. as described in example 33, described instruction can further make the system
Determine during migrate at the scene for each VM in the VM that copy dirty storage page to destination node needs more
A copy iteration.Described instruction can also make the system call interception individually distribute to the network bandwidth and/or the adjustment of each VM
Individually distribute to the process resource of the VM so that change for each of the multiple copy iteration of VM copy
In generation, completes in substantially the same time quantum, so that will be remaining in the substantially the same time during the VM is migrated at the scene
Dirty storage page copies the destination node to from the source node.
At least one machine readable media of the example 35. as described in example 33, described instruction can further make the system
Determine during migrate at the scene for each VM in the VM that copy dirty storage page to destination node needs more
A copy iteration.Described instruction can also make the system call interception individually distribute to the network bandwidth and/or the adjustment of each VM
Individually distribute to the process resource of the VM so that change for the first copy in the multiple copy iteration of the VM
In generation, completes in substantially the same time quantum.Described instruction can also be such that the system is restored after the first copy iteration
To being previously assigned for network bandwidth and/or process resource.Described instruction can also make the system that the first VM in the VM be made to ring
The 2nd VM in VM described in Ying Yu completes then to copy iteration and repetition compared to the first VM amounts of taking longer time
At least part of the subsequent copy iteration of the first copy iteration.Described instruction can also make the system by described
2nd VM completes the repetition that the subsequent copy iteration stops the first VM of the subsequent copy iteration later, so that described
VM copies the dirty storage page of residue from the source node in the substantially the same time described during migrating at the scene
Destination node.
At least one machine readable media of the example 36. as described in example 33, described instruction can further make the system
Determine during migrate at the scene for each VM in the VM that copy dirty storage page to destination node needs more
A copy iteration.Described instruction can also make the system call interception individually distribute to the network bandwidth and/or the adjustment of each VM
Individually distribute to the process resource of the VM so that change for the first copy in the multiple copy iteration of the VM
In generation, completes in substantially the same time quantum.Described instruction can also make the system based on the network bandwidth adjusted and/or
Process resource determines second working independently integrated mode for the VM.Described instruction can also make the system be based on being adjusted
Network bandwidth and/or process resource come predict for the VM scene migration the second independent VM migratory behaviours.The finger
Order can also make the system will be dirty during the VM is migrated at the scene to determine based on the second independent VM migratory behaviours predicted
Storage page copies the destination node to until the dirty storage page of residue for the corresponding VM in the VM is brought down below
Required relative surplus time quantum until the number of thresholds of remaining dirty storage page.Described instruction can also make the system
Realize that the migration scheme of the scene migration for the VM, the migration scheme adjustment are copied for the VM described first
Practical VM migratory behaviours during the subsequent copy iteration of shellfish iteration, the practical VM migratory behaviours make the VM migrate at the scene
Period copies remaining dirty storage page to the destination node in the substantially the same time from the source node.
At least one machine readable media of the example 37. as described in example 32, for making the system realize the migration
Scheme is to adjust practical VM migratory behaviours for the VM so that during the VM is migrated at the scene when substantially the same
Between remaining dirty storage page is copied to from the source node destination node described instruction may include it is described for making
System determination copies dirty storage page to the destination node during being migrated at the scene for each VM in the VM and needs
Want the described instruction of multiple copy iteration.Described instruction can also make the system make the first VM in the VM in response to the VM
In the 2nd VM complete the last copy iteration of the 2nd VM compared to the first VM amounts of taking longer time and weigh
At least part of the last copy iteration of multiple first VM is to reach the number of thresholds of remaining dirty storage page.It is described
Instruction can also make the system stop the first VM after the last copy iteration that the 2nd SM completes the 2nd VM
Last copy iteration the first VM repetition so that in substantially the same time general during the VM migrate at the scene
The dirty storage page of residue copies the destination node to from the source node.
At least one machine readable media of the example 38. as described in example 28, the VM may be disposed to support as service
At least one application is individually performed with reality based on the corresponding VM in multiple VNF of a part for chain, the integrated mode that works independently
Row supports the relevant work load of the multiple VNF.
At least one machine readable media of the example 39. as described in example 28, the source node and the destination node
It may include in the data center for being arranged to provide IaaS, PaaS or SaaS.
1.72 (b) is saved to meet 37 C.F.R., it is desirable that abstract will allow it is emphasized that providing the abstract of the disclosure
Reader quickly finds out the nature of the disclosure of the art.Advocated by understanding, it will be not used in interpretation or limits claim
Scope or meaning.In addition, in previous embodiment, it can be seen that, various features are combined together in single example
In, to make the disclosure simplify.The example that the method for the disclosure is not interpreted as reflection prescription is needed than in each right
The intention for the more features of feature being expressly recited in it is required that.On the contrary, as appended claims reflect, the theme of invention exists
In fewer than single disclosed exemplary all features.Therefore, appended claims are hereby incorporated into specific implementation mode,
In each claim represent own as independent example.In appended claims, term "comprising" and " wherein "
It is used separately as the equivalent word of plain English of corresponding term " comprising " and " wherein ".Moreover, term " first ", " second ", " third "
Etc. be solely for marking, it is not intended to requirement numerically is applied to their object.
Although acting specific language description theme with to structure feature and/or methodology, it is to be understood that with
Attached theme defined in claims is not necessarily limited to special characteristic described above or action.On the contrary, described above specific
Feature and action are disclosed as the exemplary forms of realization claim.
Claims (25)
1. a kind of equipment, including：
Circuit；
Schema component, for being executed by the circuit to determine the virtual machine for being responsible for by source node（VM）Work independently
At least one application is individually performed to carry out relevant work load based on corresponding VM in integrated mode, the integrated mode that works independently；
Prediction component, for being executed with based on work integrated mode determining respectively and based on distributing to the VM by the circuit
The network bandwidth of destination node is moved to for scene to predict to save for the scenes VM to be moved to the destination
The independent VM migratory behaviours of point；
Time component, for being executed by the circuit to determine the VM based on the VM migratory behaviours individually predicted at the scene
During migration by dirty storage page copy to the destination node until the corresponding VM for the VM the dirty memory of residue
Required relative time amount until the number of thresholds that page is brought down below remaining dirty storage page；And
Adaptive-migration component, for being executed by the circuit to adjust the practical VM migratory behaviours for the VM so that described
VM copies remaining dirty storage page from the source node in the substantially the same time during the scene migration of the VM
The destination node.
2. equipment as described in claim 1, the number of thresholds be based on by remaining dirty storage page and at least processor and
Input/output state copy to the destination node so that corresponding VM is executed at least one application with threshold between when closed
It moves to the network bandwidth of the destination node for live and carries out the phase using distributing to the VM in value
Answer workload.
3. equipment as described in claim 1, including the prediction component, to be determined respectively based on by the schema component
Work integrated mode predict independent VM migratory behaviours, the work integrated mode be used for based on individually distribute to each VM for
Scene migration network bandwidth come determine each VM in the VM is migrated at the scene during dirty storage page is copied to
The destination node is brought down below remaining dirty storage page until the dirty storage page of residue for the corresponding VM in the VM
The number of thresholds need how many copy iteration.
4. equipment as claimed in claim 3 is used to adjust the practical VM migratory behaviours for the VM so that the VM is existing
During the migration of field remaining dirty storage page is copied to the destination node in the substantially the same time from the source node
The adaptive-migration component include the adaptive-migration component for performing the following operations：
Make individually to distribute to each VM to be adjusted for the network bandwidth of scene migration and/or make individually to distribute to described
VM is adjusted with the process resource for carrying out the relevant work load, and any of described adjustment or the two are for making the VM
Reach the number of thresholds of remaining dirty memory page and starts the remaining dirty storage page of copy in the substantially the same time.
5. equipment as claimed in claim 4, including：
The time component copies dirty storage page to during being migrated at the scene for each VM in the VM to determination
The destination node needs multiple copy iteration；And
The adaptive-migration component, to make the network bandwidth for individually distributing to each VM be adjusted and/or make individually
The process resource for distributing to the VM is adjusted, and any of described adjustment or the two are used to make the institute for the VM
It states each of multiple copy iteration copy iteration to complete in substantially the same time quantum, so that the VM is migrated at the scene
Period copies remaining dirty storage page to the destination node in the substantially the same time from the source node.
6. equipment as claimed in claim 4, including：
The time component copies dirty storage page to during being migrated at the scene for each VM in the VM to determination
The destination node needs multiple copy iteration；
The adaptive-migration component, to：
The network bandwidth for individually distributing to each VM is set to be adjusted and/or make the processing money for individually distributing to the VM
Source is adjusted, and any of described adjustment or the two are used to make first in the multiple copy iteration for the VM to copy
Shellfish iteration is completed in substantially the same time quantum；
Make to recover to being previously assigned for network bandwidth and/or process resource after the first copy iteration；
Make the first VM in the VM in response to the 2nd VM in the VM compared to the first VM amounts of taking longer time
To complete then to copy described subsequent at least part for copying iteration of iteration and repetition the first copy iteration；And
Make the first VM for stopping the subsequent copy iteration after completing the subsequent copy iteration by the 2nd VM
Repetition so that during the VM is migrated at the scene in the substantially the same time by the dirty storage page of residue from the source
Node copies the destination node to.
7. equipment as claimed in claim 4, including：
The time component copies dirty storage page to during being migrated at the scene for each VM in the VM to determination
The destination node needs multiple copy iteration；
The adaptive-migration component, to make the network bandwidth for individually distributing to each VM be adjusted and/or make individually
The process resource for distributing to the VM is adjusted, and any of described adjustment or the two are used to make the institute for the VM
The the first copy iteration stated in multiple copy iteration is completed in substantially the same time quantum；
The schema component, to determine second for the VM based on the network bandwidth and/or process resource that are adjusted
Work independently integrated mode；
The prediction component, to predict the scene for the VM based on the network bandwidth and/or process resource that are adjusted
Second independent VM migratory behaviours of migration；
The time component, to based on the second independent VM migratory behaviours predicted come during determining that the VM is migrated at the scene
Dirty storage page is copied to the destination node until the dirty storage page of residue for the corresponding VM in the VM is down to
Less than remaining dirty storage page the number of thresholds until required relative surplus time quantum；And
The adaptive-migration component, to adjust for the VM during the subsequent copy iteration of the first copy iteration
Practical VM migratory behaviours so that the VM migrate at the scene during the substantially the same time by remaining dirty storage page from
The source node copies the destination node to.
8. equipment as claimed in claim 3, including：To adjust the practical VM migratory behaviours for the VM so that described
VM in the substantially the same time copies remaining dirty storage page to the purpose during migrating at the scene from the source node
The adaptive-migration component of ground node includes will be dirty during being migrated at the scene for each VM in the VM to determination
Storage page copies the time component that the destination node needs multiple copy iteration, the adaptive-migration component to
In response to the determination of multiple copy iteration：
Make the first VM in the VM in response to the 2nd VM in the VM compared to the first VM amounts of taking longer time
At least part for the last copy iteration for repeating the first VM to complete the last copy iteration of the 2nd VM is to reach
To the number of thresholds of remaining dirty storage page；And
Make described the of the last copy iteration for stopping the first VM after the last copy iteration for completing the 2nd VM
The repetition of one VM so that during the VM is migrated at the scene in the substantially the same time by the dirty storage page of residue from institute
It states source node and copies the destination node to.
9. equipment as described in claim 1, including：Digital display is coupled to the circuit with presentation user's interface view.
10. a kind of method, including：
The virtual machine for being responsible for by source node is determined in processor circuit（VM）The integrated mode that works independently, the independent work
Make integrated mode and at least one application is individually performed to carry out relevant work load based on corresponding VM；
Based on the work integrated mode determined respectively and based on the VM is distributed to destination node is moved to for scene
Network bandwidth predicts the independent VM migratory behaviours for the scenes VM to be moved to the destination node；
Based on the independent VM migratory behaviours predicted dirty storage page is copied to institute during the VM is migrated at the scene to determine
Destination node is stated until the corresponding VM of the dirty storage page of residue to(for) the VM is brought down below the threshold of remaining dirty storage page
Required relative time amount until value quantity；And
Realize the migration scheme of the scene migration for the VM to adjust the practical VM migratory behaviours for the VM so that institute
It states and copies remaining dirty storage page to the mesh from the source node in the substantially the same time during VM is migrated at the scene
Ground node.
11. method as claimed in claim 10, the number of thresholds is based on remaining dirty storage page and at least processor
And input/output state copy to the destination node so that corresponding VM is executed at least one application between when closed
Described in being carried out using distributing to the VM for scene and moving to the network bandwidth of the destination node in threshold value
Relevant work loads.
12. method as claimed in claim 10 predicts that independent VM migratory behaviours include the work integrated mode determined respectively, institute
Work integrated mode is stated for determining for the VM based on individually distributing to network bandwidths of each VM for scene migration
In each VM migrate at the scene during copy dirty storage page to the destination node until for the phase in the VM
The number of thresholds that the dirty storage page of residue of VM is brought down below remaining dirty storage page is answered to need how many copy iteration.
13. method as claimed in claim 12 is realized to adjust the practical VM migratory behaviours for the VM so that described
VM in the substantially the same time copies remaining dirty storage page to the purpose during migrating at the scene from the source node
The migration scheme of ground node includes：
Adjustment individually distributes to each VM and individually distributes to the VM for the network bandwidth of scene migration and/or adjustment
To carry out the process resource of relevant work load so that the VM reach the number of thresholds of remaining dirty memory page and
Start the remaining dirty storage page of copy in the substantially the same time.
14. at least one machine readable media, including multiple instruction, described instruction make the system in response to being just executed by the system
System：
Determine the virtual machine for being responsible for by source node（VM）The integrated mode that works independently, the integrated mode that works independently is based on
At least one application is individually performed to carry out relevant work load in corresponding VM；
Based on the work integrated mode determined respectively and based on the VM is distributed to destination node is moved to for scene
Network bandwidth predicts the independent VM migratory behaviours for the scenes VM to be moved to the destination node；
Based on the independent VM migratory behaviours predicted dirty storage page is copied to institute during the VM is migrated at the scene to determine
Destination node is stated until the corresponding VM of the dirty storage page of residue to(for) the VM is brought down below the threshold of remaining dirty storage page
Required relative time amount until value quantity；And
Realize the migration scheme of the scene migration for the VM to adjust the practical VM migratory behaviours for the VM so that institute
It states and copies remaining dirty storage page to the mesh from the source node in the substantially the same time during VM is migrated at the scene
Ground node.
15. at least one machine readable media as claimed in claim 14, the number of thresholds is based on remaining dirty memory
It is described at least one so that corresponding VM is executed that page and at least processor and input/output state copy the destination node to
Using to use the net for distributing to the VM for scene and moving to the destination node in threshold value between when closed
Network bandwidth loads to carry out the relevant work.
16. at least one machine readable media as claimed in claim 15, the number of thresholds among the VM extremely
A few VM is different.
17. at least one machine readable media as claimed in claim 15, including the shut-in time threshold value are based on for institute
The requirement that VM will stop and be restarted in the destination node in given time period in the source node is stated, it is described to want
It asks and is arranged for meeting one or more service quality (QoS) criterion or service level agreement（SLA）.
18. at least one machine readable media as claimed in claim 14, described instruction is for keeping the system prediction independent
VM migratory behaviours include the work integrated mode determined respectively, the work integrated mode be used for based on individually distribute to each VM with
For scene migration network bandwidth come determine each VM in the VM is migrated at the scene during dirty storage page is copied
Shellfish is to the destination node until the dirty storage page of residue for the corresponding VM in the VM is brought down below remaining dirty storage
The number of thresholds of device page needs how many copy iteration.
19. at least one machine readable media as claimed in claim 18, for making the system realize the migration scheme
To adjust practical VM migratory behaviours for the VM so that in substantially the same time general during the VM migrate at the scene
The described instruction that remaining dirty storage page copies the destination node to from the source node includes for making the system hold
The following instruction operated of row：
Adjustment individually distributes to each VM and individually distributes to the VM for the network bandwidth of scene migration and/or adjustment
To carry out the process resource of relevant work load so that the VM reach the number of thresholds of remaining dirty memory page and
Start the remaining dirty storage page of copy in the substantially the same time.
20. at least one machine readable media as claimed in claim 19, including described instruction, described further to make
System：
Determine that copying dirty storage page to the destination node during being migrated at the scene for each VM in the VM needs
Want multiple copy iteration；And
The processing that the network bandwidth of each VM is individually distributed in adjustment and/or the VM is individually distributed in adjustment provides
Source so that complete in substantially the same time quantum for each of the multiple copy iteration of VM copy iteration
At so as to which remaining dirty storage page is copied from the source node in the substantially the same time during the VM is migrated at the scene
To the destination node.
21. at least one machine readable media as claimed in claim 19, including described instruction, further to make the system
System：
Determine that copying dirty storage page to the destination node during being migrated at the scene for each VM in the VM needs
Want multiple copy iteration；
The processing that the network bandwidth of each VM is individually distributed in adjustment and/or the VM is individually distributed in adjustment provides
Source so that complete in substantially the same time quantum for the first copy iteration in the multiple copy iteration of the VM
At；
Being previously assigned for network bandwidth and/or process resource is recovered to after the first copy iteration；
Make the first VM in the VM in response to the 2nd VM in the VM compared to the first VM amounts of taking longer time
To complete then to copy described subsequent at least part for copying iteration of iteration and repetition the first copy iteration；And
Stop the first VM's of the subsequent copy iteration after completing the subsequent copy iteration by the 2nd VM
It repeats, so as to save the dirty storage page of residue from the source in the substantially the same time during the VM is migrated at the scene
Point copies the destination node to.
22. at least one machine readable media as claimed in claim 19, including described instruction, further to make the system
System：
Determine that copying dirty storage page to the destination node during being migrated at the scene for each VM in the VM needs
Want multiple copy iteration；
The processing that the network bandwidth of each VM is individually distributed in adjustment and/or the VM is individually distributed in adjustment provides
Source so that complete in substantially the same time quantum for the first copy iteration in the multiple copy iteration of the VM
At；
Second working independently integrated mode for the VM is determined based on the network bandwidth and/or process resource that are adjusted；
Predict that the second independent VM of the scene migration for the VM is moved based on the network bandwidth and/or process resource that are adjusted
Migrate for；
Dirty storage page is copied during the VM is migrated at the scene to determine based on the second independent VM migratory behaviours predicted
To the destination node until the dirty storage page of residue for the corresponding VM in the VM is brought down below remaining dirty memory
Required relative surplus time quantum until the number of thresholds of page；And
Realize the migration scheme of the scene migration for the VM, migration scheme adjustment is for the VM described the
Practical VM migratory behaviours during the subsequent copy iteration of one copy iteration, the practical VM migratory behaviours make the VM at the scene
During migration remaining dirty storage page is copied to the destination node in the substantially the same time from the source node.
23. at least one machine readable media as claimed in claim 19, to make the system realize the migration scheme
To adjust practical VM migratory behaviours for the VM so that in substantially the same time general during the VM migrate at the scene
The described instruction that remaining dirty storage page copies the destination node to from the source node includes keeping the system true
It is fixed each VM in the VM migrate at the scene during to copy dirty storage page to the destination node needs multiple
Copy the described instruction of iteration；
Make the first VM in the VM in response to the 2nd VM in the VM compared to the first VM amounts of taking longer time
At least part for the last copy iteration for repeating the first VM to complete the last copy iteration of the 2nd VM is to reach
To the number of thresholds of remaining dirty storage page；And
Stop the last copy iteration of the first VM after the last copy iteration that the 2nd SM completes the 2nd VM
The first VM repetition so that in the substantially the same time by the dirty storage of residue during the VM is migrated at the scene
Device page copies the destination node to from the source node.
24. at least one machine readable media as claimed in claim 14, including the VM are arranged to support to be used as service chaining
A part multiple virtual network functions（VNF）, the integrated mode that works independently is individually performed at least based on the corresponding VM
One application supports the relevant work of the multiple VNF to load to carry out.
25. at least one machine readable media as claimed in claim 14, including it is included in the work that is arranged to provide infrastructures
For service (IaaS), platform as service (PaaS) or software as the source node in the data center of service (SaaS)
With the destination node.
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2015/098025 WO2017106997A1 (en) | 2015-12-21 | 2015-12-21 | Techniques for co-migration of virtual machines |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108369524A true CN108369524A (en) | 2018-08-03 |
Family
ID=59088742
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201580085487.0A Pending CN108369524A (en) | 2015-12-21 | 2015-12-21 | The technology of collaboration migration for virtual machine |
Country Status (3)
| Country | Link |
|---|---|
| EP (1) | EP3394746A4 (en) |
| CN (1) | CN108369524A (en) |
| WO (1) | WO2017106997A1 (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109347716A (en) * | 2018-09-25 | 2019-02-15 | 中国联合网络通信集团有限公司 | The instantiation method and device of consumer VNF |
| CN109347661A (en) * | 2018-09-25 | 2019-02-15 | 中国联合网络通信集团有限公司 | The instantiation method and device of consumer VNF |
| CN110769272A (en) * | 2019-11-19 | 2020-02-07 | 深圳市网心科技有限公司 | Node adjusting method and related equipment thereof |
| CN111352695A (en) * | 2018-12-24 | 2020-06-30 | 中国电信股份有限公司 | Virtual machine migration method and device and computer readable storage medium |
| WO2021218504A1 (en) * | 2020-04-28 | 2021-11-04 | International Business Machines Corporation | Transactional memory based memory page de-duplication |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11899957B2 (en) | 2020-05-19 | 2024-02-13 | EMC IP Holding Company LLC | Cost-optimized true zero recovery time objective for multiple applications |
| US11836512B2 (en) | 2020-05-19 | 2023-12-05 | EMC IP Holding Company LLC | Virtual machine replication strategy based on predicted application failures |
| US11797400B2 (en) * | 2020-05-19 | 2023-10-24 | EMC IP Holding Company LLC | Cost-optimized true zero recovery time objective for multiple applications based on interdependent applications |
| US11934283B2 (en) | 2020-05-19 | 2024-03-19 | EMC IP Holding Company LLC | Cost-optimized true zero recovery time objective for multiple applications using failure domains |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2329370A1 (en) * | 2008-09-15 | 2011-06-08 | International Business Machines Corporation | Securing live migration of a virtual machine within a service landscape |
| US20110264788A1 (en) * | 2010-04-23 | 2011-10-27 | Glauber Costa | Mechanism for Guaranteeing Deterministic Bounded Tunable Downtime for Live Migration of Virtual Machines Over Reliable Channels |
| CN103365704A (en) * | 2012-03-26 | 2013-10-23 | 中国移动通信集团公司 | Memory pre-copying method in virtual machine migration, device executing memory pre-copying method and system |
| CN103955399A (en) * | 2014-04-30 | 2014-07-30 | 华为技术有限公司 | Migrating method and device for virtual machine, as well as physical host |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100115095A1 (en) * | 2008-10-31 | 2010-05-06 | Xiaoyun Zhu | Automatically managing resources among nodes |
| US8214829B2 (en) * | 2009-01-15 | 2012-07-03 | International Business Machines Corporation | Techniques for placing applications in heterogeneous virtualized systems while minimizing power and migration cost |
| US8689211B2 (en) * | 2009-05-25 | 2014-04-01 | International Business Machines Corporation | Live migration of virtual machines in a computing environment |
| US9052949B2 (en) * | 2012-06-21 | 2015-06-09 | Vmware, Inc. | Scheduling a processor to support efficient migration of a virtual machine |
| JPWO2014010213A1 (en) * | 2012-07-11 | 2016-06-20 | 日本電気株式会社 | Non-transitory computer-readable medium storing migration system, migration method, and control program |
| US9558005B2 (en) * | 2014-05-19 | 2017-01-31 | Intel Corporation | Reliable and deterministic live migration of virtual machines |
-
2015
- 2015-12-21 CN CN201580085487.0A patent/CN108369524A/en active Pending
- 2015-12-21 WO PCT/CN2015/098025 patent/WO2017106997A1/en unknown
- 2015-12-21 EP EP15910997.4A patent/EP3394746A4/en not_active Withdrawn
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2329370A1 (en) * | 2008-09-15 | 2011-06-08 | International Business Machines Corporation | Securing live migration of a virtual machine within a service landscape |
| US20110264788A1 (en) * | 2010-04-23 | 2011-10-27 | Glauber Costa | Mechanism for Guaranteeing Deterministic Bounded Tunable Downtime for Live Migration of Virtual Machines Over Reliable Channels |
| CN103365704A (en) * | 2012-03-26 | 2013-10-23 | 中国移动通信集团公司 | Memory pre-copying method in virtual machine migration, device executing memory pre-copying method and system |
| CN103955399A (en) * | 2014-04-30 | 2014-07-30 | 华为技术有限公司 | Migrating method and device for virtual machine, as well as physical host |
Non-Patent Citations (2)
| Title |
|---|
| JIE ZHENG等: ""Pacer: A Progress Management System for Live Virtual Machine Migration in Cloud Computing"", 《IEEE TRANSACTIONS ON NETWORK AND SERVICE MANAGEMENT》 * |
| JIE ZHEN等: ""COMMA: Coordinating the Migration of Multitier Applications"", 《VEE "14: PROCEEDINGS OF THE 10TH ACM SIGPLAN/SIGOPS INTERNATIONAL CONFERENCE ON VIRTUAL EXECUTION ENVIRONMENTS》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109347716A (en) * | 2018-09-25 | 2019-02-15 | 中国联合网络通信集团有限公司 | The instantiation method and device of consumer VNF |
| CN109347661A (en) * | 2018-09-25 | 2019-02-15 | 中国联合网络通信集团有限公司 | The instantiation method and device of consumer VNF |
| CN109347716B (en) * | 2018-09-25 | 2021-05-11 | 中国联合网络通信集团有限公司 | Instantiation method and device of consumer VNF |
| CN111352695A (en) * | 2018-12-24 | 2020-06-30 | 中国电信股份有限公司 | Virtual machine migration method and device and computer readable storage medium |
| CN111352695B (en) * | 2018-12-24 | 2023-09-29 | 中国电信股份有限公司 | Virtual machine migration method, device and computer readable storage medium |
| CN110769272A (en) * | 2019-11-19 | 2020-02-07 | 深圳市网心科技有限公司 | Node adjusting method and related equipment thereof |
| WO2021218504A1 (en) * | 2020-04-28 | 2021-11-04 | International Business Machines Corporation | Transactional memory based memory page de-duplication |
| GB2610511A (en) * | 2020-04-28 | 2023-03-08 | Ibm | Transactional memory based memory page de-duplication |
Also Published As
| Publication number | Publication date |
|---|---|
| EP3394746A1 (en) | 2018-10-31 |
| WO2017106997A1 (en) | 2017-06-29 |
| EP3394746A4 (en) | 2019-08-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108369524A (en) | The technology of collaboration migration for virtual machine | |
| CN107924328B (en) | Technique for selecting virtual machine for migration | |
| CN107735767B (en) | Apparatus and method for virtual machine migration | |
| US11614893B2 (en) | Optimizing storage device access based on latency | |
| US11281538B2 (en) | Systems and methods for checkpointing in a fault tolerant system | |
| US20210182190A1 (en) | Intelligent die aware storage device scheduler | |
| US10853139B2 (en) | Dynamic workload management based on predictive modeling and recommendation engine for storage systems | |
| JP5594664B2 (en) | System and method for storage tiering and migration techniques based on quality of service | |
| CN102821158B (en) | A kind of method and cloud system realizing virtual machine (vm) migration | |
| CN106201659B (en) | A kind of method and host of live migration of virtual machine | |
| CN107247619B (en) | Live migration of virtual machine method, apparatus, system, storage medium and equipment | |
| CN109690487A (en) | System and method for executing the real-time migration of software container | |
| CN105808159A (en) | Storage operation interrupt | |
| CN107077366B (en) | Method and apparatus for checkpointing/passing between primary and secondary virtual machines | |
| CN107615244A (en) | The technology of one or more containers is run on a virtual machine | |
| US11157355B2 (en) | Management of foreground and background processes in a storage controller | |
| US20220357891A1 (en) | Efficient Read By Reconstruction | |
| CN109844714B (en) | System and method for allocating input/output bandwidth in a storage system | |
| CN106959893A (en) | Accelerator, EMS memory management process and data handling system for accelerator | |
| CN105940386A (en) | Migrating data between memories | |
| US10264064B1 (en) | Systems and methods for performing data replication in distributed cluster environments | |
| US20140164323A1 (en) | Synchronous/Asynchronous Storage System | |
| US9727252B2 (en) | Methods and systems for optimal snapshot distribution within a protection schedule | |
| CN105045541B (en) | A kind of incremental snapshot method and device | |
| US11768763B2 (en) | Flash secure erase |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180803 |