CN107203255A - Power-economizing method and device are migrated in a kind of network function virtualized environment - Google Patents
Power-economizing method and device are migrated in a kind of network function virtualized environment Download PDFInfo
- Publication number
- CN107203255A CN107203255A CN201610166392.8A CN201610166392A CN107203255A CN 107203255 A CN107203255 A CN 107203255A CN 201610166392 A CN201610166392 A CN 201610166392A CN 107203255 A CN107203255 A CN 107203255A
- Authority
- CN
- China
- Prior art keywords
- migration
- nfv
- physical
- virtual machine
- migrated
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/26—Power supply means, e.g. regulation thereof
- G06F1/32—Means for saving power
- G06F1/3203—Power management, i.e. event-based initiation of a power-saving mode
- G06F1/3234—Power saving characterised by the action undertaken
- G06F1/329—Power saving characterised by the action undertaken by task scheduling
-
- 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
- G06F9/5088—Techniques for rebalancing the load in a distributed system involving task migration
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Debugging And Monitoring (AREA)
Abstract
The embodiment of the present invention, which is disclosed, migrates power-economizing method and device in a kind of network function virtualized environment, be related to data center's energy-saving distribution field.Methods described step：1) the divisible technology of capacity is used, physical machine PM cpu busy percentage bound T is seth、Tl, monitor CPU, the internal memory for the PM for thering is NFV to apply, network, if reach migration threshold values, computation migration probability；Frequent migration is avoided, last time transit time is checked to the virtual machine VM of migration, if adjacent transit time twice is less than certain value, is not migrated；2) load is less than T in 1)lPM, judge that VM thereon is transportable no, if feasible, migration；Otherwise do not migrate；If 3) 1) in have load TcBeyond ThPM, selection thereon VM migrate, make the T of PM after migratione≤Th, and Th‑TeIt is minimum；If 4) NFV application life cycles terminate, correspondence VM is closed；5) dormancy is carried out after migrating；6) PM is selected with matching way at first, and then NFV clusters is saved；7) after having migrated, the information such as PM CPU, internal memory, network are updated.
Description
Technical field
The present invention relates to data center's energy-saving distribution technical field, more particularly to a kind of network function virtualization (NFV) ring
Migration power-economizing method and device under border.
Background technology
Network Functions Virtualization (NFV, network function virtualization) refer to virtual by IT
Change technology, various network software functions are carried using the volume server, memory and interchanger of industrywide standard
Technical standard.NFV will realize the flexible loading of software, realize in the spirit of each position such as data center, network node and user terminal
Deployment configuration living, so as to accelerate the speed of network design and adjustment, reduces the complexity of service deployment, improves the network equipment
The functions such as unitized, generalization, suitability.
NFV framework part mainly includes the simple frameworks of NFV (Fig. 1) and its frame of reference (Fig. 2).Simple framework is NFV
Most classical is also a most simplified Organization Chart, and frame of reference is the refinement that certain level is carried out to simple framework, including is connect
The introducing of mouth and the refinement of NFV management and layout.
Fig. 1 is a figure most simple and classical in ETSI NFV standard architectures, it can be seen that NFV include NFVI, VNFs,
Several major parts such as NFV Management and Orchestration (MANO)：
NFVI(Network Functions Virtualization Infrastructure)：NFVI is exactly cloud computing
Infrastructure layer in structure, it is by the related CPU/ internal memories/hard disk/Internet resources all-round virtualization of hardware.
VNF (Virtualized Network Function, virtual network function)：Realized as a pure software
Network function, can operate on NFVI, correspondence conventional physical network function.The resource that they see down into entirely quilt
" empty resource " after virtualization software closing is hiding, is podium level plus software layer in cloud computing.In wherein original telecommunication apparatus
Podium level can correspond to the podium level of cloud computing, the software layer of operation layer correspondence cloud computing.From the perspective of installation and deployment,
VNF be by a virtual machine (VM) or multiple Imaginary Mechanisms into.From the perspective of software application development business, VNF is exactly portion
The software affixed one's name in the VM of one or more interconnections is realized.
NFV MANO (NFV Management and Orchestration, NFV management and layout)：It is responsible for NFVI
Software and hardware resources life cycle management and layout, and life cycle management and layout to VNFs.NFV MANO emphasis
It is concerned with virtual management tasks all under NFV frameworks.
NFV frame of references focus on to describe the change that the network of network operator occurs during it carries out NFV transformations
Change.Fig. 2 NFV frame of references define the Primary Reference point between different functional module and module.Some of functional modules
It is already present in current deployment.And it is other the need for subsequently supplement, to support virtualization process and follow-up operation.
These functional blocks have：
Virtualised Network Function(VNF)：The network function of virtualization.
Element Management System(EMS)：Network management system.One or more VNF can be managed.It can make
With former network management system unified management virtualization and non-virtualized network element.
NFV Infrastructure(NFVI)：It is NFV infrastructure layers.
Virtualised Infrastructure Managers:Virtualized infrastructure is managed.Infrastructure layer manufacturer
The infrastructure layer managing system of offer, is responsible for carrying out physical hardware virtual resources unified management, monitoring, optimization.
Such as, OpenStack.
VNF Managers：It is responsible for VNF life cycle management.One VNF Manager can manage one or more
VNF.Note, do not refer to the service management of network element on EMS here, and referring to provide EMS and VNF includes deployment/dilatation/contracting
Hold/the automatic capability such as offline.
Orchestrator：Composer, is responsible for basic resource and the layout of upper layer software (applications) resource of NFV infrastructure layer
And management, realize network service on the basis of NFV infrastructure layer.According to the demand of business, required for adjusting each VNF
How much is resource, and VNF etc. is migrated between each rack, computer room, region, is full automatic core competence.
OSS/BSS：Business support system (BSS) and OSS (OSS).Need to reduce to greatest extent to existing
OSS/BSS influence.In order to adapt to NFV trend, OSS/BSS will support to operate on cloud computing platform in itself, at the same support and
VNF Manager and Orchestrator above-mentioned Orchestrator, VNF Manager (s) of intercommunication, Virtualized
This three part of Infrastructure Manager (s), has collectively constituted NFV Management and Orchestration
(NFV MANO)。
The content of the invention
The technical problem to be solved in the present invention is：The power cost risen steadily, the energy supply of worsening shortages, green ring
Protect the society triggered and economic pressures force global enterprise all in research power-saving technology.In ETSI GS NFV standards (004
V1.1.1 (2013-10)) in, propose requirement for the efficient energy consumptions of NFV.But also lack specific solution.The present invention
The dispatching method and device for the virtual machine energy-conservation migration being directed under a kind of network function virtualization (NFV) environment.
In order to solve the above technical problems, in a first aspect, virtualizing (NFV) the embodiments of the invention provide a kind of network function
The method for saving migration, methods described includes following seven step：
1) using capacity divisible technology (so that virtual machine whole volume is proportional to corresponding physical machine whole volume,
Illustrated hereinafter in application example), and the cpu busy percentage upper limit T of each physical host (PM) is sethAnd lower limitMonitoring peace
Cpu busy percentage equipped with the NFV physical machines applied, memory usage, network load, by computation migration probability function, judges
Whether dynamic migration threshold value is reached；In order to avoid frequent migration occurs, is increased to the virtual machine of migration the time of previous migration
Stamp, if less than one value (5 minutes) of the adjacent time migrated twice, without migration；
If 2) step 1) in find load be less than lower limitPhysical host, computation migration probability function
Wherein f (x, α, β) is distributed for Beta：
The maximum physical machine of selection migration probability is migrated；
Using the whole moving methods of dormancy formula, first judge whether that all virtual machines thereon are all transportable, if feasible, hold
Row migration, the 1st physical machine that the virtual machine selection migrated out can be received is re-assigned to other physical hosts；If can not be whole
Migration, then abandon this migration, and using the method energy-conservation for turning down physical machine dominant frequency；
If 3) step 1) middle discovery load TcBeyond upper limit ThPhysical host, computation migration probability function
Wherein f (x, α, β) is distributed for Beta：
The maximum physical machine of selection migration probability is migrated；
Then select thereon a virtual machine migrated, method be so that migration after the physical host utilization rate be Tc,(Tc
≤Th), and (Th-Tc) it is minimum, i.e., less than equal to the utilization rate upper limit, and and the utilization rate upper limit difference it is minimum；
If 4) NFV application life cycles terminate, correspondence virtual machine (or cluster) is closed；
5) migration finishes progress sleep operation to save；
6) physical machine is selected according to First-Fit (matching way at first), physical machine is operated；And then can be right
NFV clusters are saved；
7) after the completion of migrating, the corresponding cpu busy percentage of each physical host, the monitoring such as memory usage and network load are updated
Information.
In 1) planting possible implementation the, the cpu busy percentage of the monitoring current time All hosts server is interior
Utilization rate is deposited, network load judges whether to reach dynamic migration threshold value.
It is described to reach transition condition when NFV clusters load too low in 2) planting possible implementation the, pass through energy-conservation
Migration algorithm, is migrated.
It is described to reach transition condition when NFV clusters load too high in 3) planting possible implementation the, pass through energy-conservation
Migration algorithm, is migrated.
In 4) planting possible implementation the, if the NFV application life cycles terminate, correspondence should be closed virtual
Machine (or cluster).
In 5) planting possible implementation the, methods described includes needing to judge whether want sleep operation after migrating.
In 6) planting possible implementation the, methods described, which also includes transition process, to be needed to operate physical machine,
And then NFV clusters can be saved.
In 7) planting possible implementation the, methods described also includes virtual machine and uses the shared storages such as NFS；Migrate
Cheng Hou, updates the corresponding cpu busy percentage of each physical host, the monitoring information such as memory usage and network load.
In 8) planting possible implementation the, methods described also includes NFV applications CPU and internal memory, the feelings of the network bandwidth
Condition can be monitored and state can be returned into scheduler module.
In 9) planting possible implementation the, methods described also includes only one of which or multiple same on a virtual machine
Class NFV is applied.
In 10) planting possible implementation the, methods described also includes distributing to all void on a physical server
The upper limit of CPU (or internal memory, the storage IO, network I/O etc.) upper limits no more than the offer of physical server of plan machine.
In 10) planting possible implementation the, when methods described is additionally included in computation migration probability function, pass through measurement
The cpu busy percentage of single physical machine, the average and migration probability for calculating cpu busy percentage is compared, and judges whether migration.
Second aspect, the embodiments of the invention provide a kind of network function virtualization (NFV) energy-conservation moving apparatus, the dress
Putting three big modules includes：
Monitoring modular, for monitoring each physical machine and cluster cpu busy percentage, memory usage, network load condition；
Computing module, for according to above-mentioned monitoring information, comprehensive consideration information, according to above-mentioned moving method, to be calculated and moved
The source virtual machine and target physical machine of shifting；
Transferring module, for the virtual machine for needing to migrate according to the calculating message scheduling.
According to second aspect, in the first possible implementation, the monitoring modular：Each physics is monitored in real time
Machine and cluster cpu busy percentage, memory usage, network load condition provide foundation for migration.
According to second aspect, in second of possible implementation, the computing module：
(a) migrated during low load, all virtual machines on the physical host of low load are all migrated so as to dormancy or the thing is closed
Reason machine；(b) migration when overloading so that the load utilization of the main frame after migration：Less than equal to the utilization rate upper limit, and with profit
It is minimum with the difference of the rate upper limit, so ensure to maximally utilize existing physical host to reduce migration simultaneously；(c) physical host
Break down, all migrate virtual machine thereon.
According to second aspect, in the third possible implementation, the computing module：
Configuration to physical machine and selectable virtual machine uses the divisible configuring technical of capacity, it is ensured that minimum in distribution
Change idle capacity fragment so that used physical machine sum is minimized.
According to second aspect, in the third possible implementation, the transferring module：
Physical machine utilization rate bound is set, and both using minimal number of server, what reduction frequent migration was brought again opens
Sell, the measure such as this process main frame simultaneously too low including the use of dormancy utilization rate, the main frame for waking up dormancy is saved with further dynamic
Energy.
The third aspect, the embodiments of the invention provide the method for reducing power consumption that a kind of network function virtualizes (NFV) system,
Characterized in that, described network function virtualization (NFV) system uses any possible realization of first aspect or first aspect
Method described in mode is migrated and saved.
Brief description of the drawings
Fig. 1 is network function virtualization (NFV) group system configuration diagram of an embodiment of the present invention；
Fig. 2 is network function virtualization (NFV) group system configuration diagram of another embodiment of the invention；
Fig. 3 is that the network function of an embodiment of the present invention virtualizes the flow of (NFV) group system energy-saving scheduling method
Figure；
Fig. 4 be an embodiment of the present invention network function virtualization (NFV) group system energy-saving distribution algorithm with it is original
The schematic diagram of architecture combined；
Fig. 5 is that the network function of an embodiment of the present invention virtualizes the module of (NFV) system cluster system call device
Structural representation；
Embodiment
Below according to drawings and examples, the embodiment to the present invention is described in further detail.Implement below
Example is used to illustrate the present invention, but is not intended to limit the scope of the present invention.
It will be understood by those skilled in the art that in the method for various embodiments of the present invention, the sequence number size of each step is not
Mean the priority of execution sequence, the execution sequence of each step should be determined with its function and internal logic, without tackling the present invention
The implementation process of specific embodiment constitutes any limit.
Embodiment one：
As shown in figure 4, the embodiment of the present invention additionally provides a kind of network function virtualization (NFV) group system energy-saving distribution
Algorithm frame schematic diagram：
The NFV architecture frameworks that figure displaying ESTI recommends, are below corresponding a kind of description of application scenarios under the framework：
1) NFV user submits task (such as backup tasks) to system；
2) NFV providers return to the information such as the time needed for backup scheduling to client；
3) other dependent events are triggered；
4) after accomplishing a task, the next item down task is continued；
5) after ending task, server is closed.
The key for realizing above scene is then intelligent scheduler, the target to realize optimization NFV resource deployments.This intelligence
Energy scheduler will determine the void under current state (storage, network, computing resource state, power consumption state etc.) by data analysis
Plan machine is disposed, and resource is selected using resource management, and meet business constraints.
Embodiment two：
As shown in figure 5, the embodiment of the present invention additionally provides a kind of network function virtualization (NFV) system cluster system call
The structural representation of device, the device 500 includes：
Monitoring modular 510, for monitoring each physical machine and cluster cpu busy percentage, memory usage, network load feelings
Condition.
Computing module 520, for according to above-mentioned monitoring information, comprehensive consideration CPU, the resource information such as internal memory and network to be pressed
The source virtual machine and target physical machine of migration are calculated according to the moving method.
One specific embodiment is as follows：
1) whether each physical host inspects periodically cpu busy percentage in threshold values up and down(lower limit threshold values) and Th(upper limit valve
Value) between；
If 2) cpu busy percentage is less than lower limitWhen, the virtual machine of the autonomous asynchronous decision selection of server thereon is migrated,
Triggering scheduling system migration operation；If cpu busy percentage is higher than ThWhen, migrated according to overload moving method；
Transferring module 530, for the virtual machine for needing to migrate according to the calculating message scheduling.
Virtual machine (VM) is migrated：
1) first to physical server and the selectable divisible configuring technical of virtual machine capacity：As shown in Table A -1 and A-2
Several typical configuring conditions, the relation between different virtual machine capacity configuration is proportional, for example, VM 1-1,1-2,1-3, (interior
Deposit, CPU, store) capacity is overall into 1：4：8 relation, accounts for the 1/16,1/4,1/2 of physical machine PM-1 types respectively.Capacity can divide
Cut configuration to ensure, in distribution, to minimize idle capacity fragment so that used physical machine sum is minimized.
2) allocation algorithm is using minimum total energy consumption NFV algorithms：Distributed with reference to centralised allocation and distributed ad-hoc
Advantage；If physical machine isomorphism during distribution task, because using the divisible configuring technical of capacity, then first assignable machine is selected
(First-Fit)；If physical machine isomery, physical server increase energy consumption being allocated at least after selection distribution, target is
Using minimal number of physical machine, in order to reduce computation complexity, current portions can be only selected to meet the physical machine of distributive condition
It is allocated；
3) asynchronous migration：Server utilization bound is set, and each physical host is asynchronously carried out according to methods described
Migration, both using minimal number of server, reduces the expense that frequent migration is brought, this process is simultaneously including the use of dormancy profit again
With measures such as the too low main frame of rate, the main frames for waking up dormancy with further dynamic energy-saving.
4) three kinds of situations of migration point：(a) migrated during low load, it is provable to maximize energy-conservation, it is necessary to the physics of low load
All virtual machines on main frame all migrate so as to dormancy or close the physical machine；(b) migration when overloading so that being somebody's turn to do after migration
The load utilization of main frame：Less than equal to the utilization rate upper limit, and and the utilization rate upper limit difference it is minimum, it is so same to reduce migration
When again ensure maximally utilize existing physical host；(c) physical host breaks down, and all migrates virtual machine thereon.
5) in order to avoid frequent migration occurs, increase the virtual machine of migration the timestamp of previous migration, if it is adjacent twice
The time of migration is less than a value (5 minutes), then without migration.
Several typical virtual machine configuration examples of Table A -1
| VM Type | CPU(Compute Units) | Memory | Storage | PM Type |
| Type1-1 | 1(1core*1unit) | 1.875GB | 211.25GB | Type1 |
| Type1-2 | 4(2cores*2units) | 7.5GB | 845GB | Type1 |
| Type1-3 | 8(4cores*2units) | 15GB | 1690GB | Type1 |
| Type2-1 | 6.5(2cores*3.25units) | 17.1GB | 422.5GB | Type2 |
| Type2-2 | 13(4cores*3.25units) | 34.2GB | 845GB | Type2 |
| Type2-3 | 26(8cores*3.25units) | 68.4GB | 1690GB | Type2 |
| Type3-1 | 5(2cores*2.5units) | 1.875GB | 422.5GB | Type3 |
| Type3-2 | 20(8cores*2.5units) | 7.5GB | 1690GB | Type3 |
The typical physical server configuration example of three kinds of Table A -2
| PM Type | CPU(Compute Units) | Memory | Storage |
| Type1 | 16(16*1、4*4、2*8) | 30GB | 3380GB |
| Type2 | 52(8*6.5、4*13、2*26) | 136GB | 3380GB |
| Type3 | 40(8*5、2*20) | 14GB | 3380GB |
Embodiment three：
Various embodiments of the present invention are further illustrated by specific embodiment.
1st, it is now assumed that there is the Type1 types physical machine 3 of Table A -2, and there are 6 void of Type1-2 types virtual machine of Table A -1
Plan machine, 4 virtual machines of Type1-1 types virtual machine；And the threshold values up and down of cluster energy-saving distribution, its lower limit threshold values are setFor
0.2 and upper limit threshold values ThFor 0.9；And function parameter is migrated accordinglyWherein f (x, α, β) is
(wherein x is virtual machine cpu utilization rates, and 3) α, β is set to for Beta distributions：
The state of current cluster such as following table：
The resource utilization that 1-No. 3 physical machines can be calculated from upper table is respectively：75% (3/4), 75% (3/4),
18.75% (1/8), it is found that the cpu busy percentage of No. 3 physical machines is less than lower limit threshold valuesFunction is migrated according to lower limit probability,
And then the virtual machine in No. 3 physical machines is migrated, its migration probability is calculated respectively, because physical machine is isomorphism, according to
First-Fit modes are distributed；Calculate the resource utilization after the First virtual machine (vm) migration of No. 3 machines to No. 1 physical machine first,
Now the resource utilization of No. 1 physical machine is 81.25%, less than upper limit threshold values Th, and migration probability is larger, it is possible to move
Move；Same continues to move to No. 1 physical machine by the 2nd virtual machine of No. 3 physical machines, the now utilization of resources of No. 1 physical machine
Rate is 87.5%, still is below upper limit threshold values Th, but if the 3rd virtual machine is continued to move to No. 1 physical machine, then its
Resource utilization is 93.75%, higher than upper limit threshold values Th, so by the 3rd virtual machine (vm) migration to No. 2 physical machines, now No. 2 things
The resource utilization of reason machine be 81.25%, meet above and below threshold values requirement；Virtual machine is all moved out in such No. 3 physical machines,
Resting state is placed in, to save energy consumption；State such as following table after migration：
2nd, it is now assumed that there is the Type1 types physical machine 3 of Table A -2, and there are 7 void of Type1-2 types virtual machine of Table A -1
Plan machine, 5 virtual machines of Type1-1 types virtual machine；And the threshold values up and down of cluster energy-saving distribution, its lower limit threshold values are setFor
0.2 and upper limit threshold values ThFor 0.9；And function parameter is migrated accordinglyWherein f (x, α, β) is
(wherein x is virtual machine cpu utilization rates, and 3) α, β is set to for Beta distributions：
The state of current cluster such as following table：
The resource utilization that 1-No. 3 physical machines can be calculated from upper table is respectively：93.75% (15/16), 75%
(3/4), 37.5% (3/8), find No. 1 physical machine is higher than lower limit threshold values Th=0.9, function is migrated according to Upper Probability, entered
And the virtual machine in No. 1 physical machine is migrated；First by selection migration Type1-1 types virtual machine, Type1-2 type virtual machines
The virtual machine of which kind of model, according to the principle of upper limit virtual machine (vm) migration, if after 1 Type1-1 type of migration, No. 1 physical machine
Resource utilization is 87.5%；Migrate after 1 Type1-2 type, the resource utilization of No. 1 physical machine is 68.75%, and all meeting will
Ask, and the utilization rate for migrating resource after 1 Type1-1 type is higher and meets upper limit transition condition, so selection migration Type1-1
Type；Then selection target physical machine, is selected according to First-Fit modes, is moved to after No. 2 physical machines, and its utilization rate is
81.25%, and migration probability is larger, meets condition, so selection moves to No. 2 physical machines.State such as following table after migration：
One of ordinary skill in the art will appreciate that realize all or part of flow in above-described embodiment method, being can be with
The hardware of correlation is instructed to complete by computer program, described program can be stored in an embodied on computer readable storage and be situated between
In matter, the program is upon execution, it may include such as the flow of the embodiment of above-mentioned each method.Wherein, described storage medium can be
Magnetic disc, CD, read-only memory (Read-Only Memory, ROM) or random access memory (Random Access
Memory, RAM) etc..
The foregoing is only a specific embodiment of the invention, but protection scope of the present invention is not limited thereto, any
Those familiar with the art the invention discloses technical scope in, the change or replacement that can be readily occurred in, all should
It is included within the scope of the present invention.Therefore, protection scope of the present invention should be defined by scope of the claims.
Claims (9)
1. the migration power-economizing method under a kind of network function virtualized environment, it is characterised in that methods described includes step：
1) using the divisible technology of capacity (so that virtual machine whole volume is proportional to corresponding physical machine whole volume, hereinafter
Application example in illustrate), and the cpu busy percentage upper limit T of each physical host (PM) is sethWith lower limit TlMonitoring is installed
There is the cpu busy percentage of the physical machine of NFV applications, memory usage, network load judges whether to reach dynamic migration threshold value,
And computation migration probability function；In order to avoid frequent migration occurs, the timestamp of previous migration is increased to the virtual machine of migration, if
The adjacent time migrated twice is less than a value (such as 5 minutes), then without migration；
If 2) step 1) in find load be less than lower limit TlPhysical host, first judge whether that all virtual machines thereon can all be moved
Move, if feasible, perform migration, the 1st physical machine that the virtual machine selection migrated out can be received is re-assigned to other physics
Main frame；If can not all migrate, this migration is abandoned；
If 3) step 1) middle discovery load TcBeyond upper limit ThPhysical host, then select thereon a virtual machine migrated, make
The physical host utilization rate is T after must migratingc,(Tc≤Th), and (Th-Tc) it is minimum, i.e., less than equal to the utilization rate upper limit, and with profit
It is minimum with the difference of the rate upper limit；
If 4) NFV application life cycles terminate, correspondence virtual machine (or cluster) is closed；
5) migration finishes progress sleep operation to save；
6) virtual machine migrated out according to migration probability function and First-Fit (matching way at first) distribution, to physical machine
Operated；And then NFV clusters can be saved；
7) after the completion of migrating, the corresponding cpu busy percentage of each physical host, the monitoring letter such as memory usage and network load are updated
Breath.
2. according to the method described in claim 1, it is characterised in that the CPU for obtaining current time All hosts server is utilized
Rate, memory usage, network bandwidth load, and computation migration probability function, judge whether to reach that dynamic is moved according to migration models
Whether sliding door limit value and life cycle terminate；If NFV clusters load too low is too high and reach transition condition or life
End cycle is ordered, by saving migration algorithm, dynamic migration of virtual machine is carried out.
3. according to the method described in claim 1, it is characterised in that whether needs will carry out sleep operation after judging migration, such as
Fruit NFV application life cycles terminate, then should close and correspond to virtual machine (or cluster), underlying virtual software support dynamic migration,
And transition process needs to operate physical machine, and then NFV clusters can be saved.
4. according to the method described in claim 1, it is characterised in that NFV applications CPU and internal memory, the situation of the network bandwidth can be with
It is monitored and state can be returned to scheduler module, and distribute to all virtual machines on a physical server CPU (or
Internal memory, storage IO, network I/O etc.) upper limit no more than the offer of physical server the upper limit.
5. according to the method described in claim 1, it is characterised in that only one of which or multiple same class NFV on a virtual machine
Using.
6. according to the method described in claim 1, it is characterised in that, can be to each migration in order to avoid frequent migration switches
Virtual machine increase the timestamp of previous migration, if the adjacent time migrated twice is less than a value (such as 5 minutes), do not move
Move, wait similar method.
7. method according to any one of claim 1 to 4, it is characterised in that in computation migration probability function, pass through
The cpu busy percentage of single physical machine is measured, the average and migration probability for calculating cpu busy percentage are compared, and judge whether migration.
8. the migration energy saver under a kind of network function virtualized environment, it is characterised in that described device includes：
Monitoring modular, for monitoring each physical machine and cluster cpu busy percentage, memory usage, network load condition；
Computing module, for according to above-mentioned monitoring information, comprehensive consideration information, according to above-mentioned moving method, to calculate migration
Source virtual machine and target physical machine；Configuration to physical machine and selectable virtual machine uses the divisible configuring technical of capacity,
Ensure to minimize idle capacity fragment in distribution so that used physical machine sum is minimized；
(a) migrated during low load, all virtual machines on the physical host of low load are all migrated so as to dormancy or the physical machine is closed；
(b) migration when overloading so that the load utilization of the main frame after migration：Less than equal to the utilization rate upper limit, and with profit
It is minimum with the difference of the rate upper limit, so ensure to maximally utilize existing physical host to reduce migration simultaneously；
(c) physical host breaks down, and all migrates virtual machine thereon.
Transferring module, for the virtual machine for needing to migrate according to the calculating message scheduling, sets physical machine utilization rate bound,
Both minimal number of server is used, the expense that frequent migration is brought is reduced again, this process is simultaneously including the use of dormancy utilization rate
The measures such as too low main frame, the main frame of wake-up dormancy are with further dynamic energy-saving.
9. the migration energy conserving system under a kind of network function virtualization (NFV) environment, it is characterised in that including in claim 8
Described dispatching device；A kind of energy-conservation moving method under network function virtualization (NFV) environment, it is characterised in that the NFV
Method any one of group system usage right requirement 1 to 7 is scheduled.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610166392.8A CN107203255A (en) | 2016-03-20 | 2016-03-20 | Power-economizing method and device are migrated in a kind of network function virtualized environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610166392.8A CN107203255A (en) | 2016-03-20 | 2016-03-20 | Power-economizing method and device are migrated in a kind of network function virtualized environment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107203255A true CN107203255A (en) | 2017-09-26 |
Family
ID=59904629
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610166392.8A Pending CN107203255A (en) | 2016-03-20 | 2016-03-20 | Power-economizing method and device are migrated in a kind of network function virtualized environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107203255A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108429815A (en) * | 2018-03-23 | 2018-08-21 | 重庆邮电大学 | Dynamic resource scheduling method based on OpenStack |
| CN108540405A (en) * | 2017-12-18 | 2018-09-14 | 清华大学 | Internet resources moving method and device |
| CN108664116A (en) * | 2018-04-27 | 2018-10-16 | 北京邮电大学 | Adaptive electricity saving method, device and the cpu controller of network function virtualization |
| CN109002348A (en) * | 2018-07-26 | 2018-12-14 | 郑州云海信息技术有限公司 | Load-balancing method and device in a kind of virtualization system |
| CN109062668A (en) * | 2018-08-01 | 2018-12-21 | 重庆邮电大学 | A kind of virtual network function moving method of the multipriority based on 5G access network |
| CN110275758A (en) * | 2019-05-09 | 2019-09-24 | 重庆邮电大学 | A kind of virtual network function intelligence moving method |
| CN111488053A (en) * | 2020-04-17 | 2020-08-04 | 苏州浪潮智能科技有限公司 | Power consumption adjusting method and device of network function virtualization system |
| CN112416530A (en) * | 2020-12-08 | 2021-02-26 | 西藏宁算科技集团有限公司 | Method and device for flexibly managing cluster physical machine nodes and electronic equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102096461A (en) * | 2011-01-13 | 2011-06-15 | 浙江大学 | Energy-saving method of cloud data center based on virtual machine migration and load perception integration |
| CN104679594A (en) * | 2015-03-19 | 2015-06-03 | 成都艺辰德迅科技有限公司 | Middleware distributed calculating method |
| CN104850459A (en) * | 2015-05-22 | 2015-08-19 | 中国联合网络通信集团有限公司 | Virtual machine migrating method and apparatus |
| CN104866375A (en) * | 2015-05-22 | 2015-08-26 | 中国联合网络通信集团有限公司 | Virtual machine migration method and apparatus |
| CN104881316A (en) * | 2015-05-22 | 2015-09-02 | 中国联合网络通信集团有限公司 | Virtual machine transferring method and device |
| CN104881325A (en) * | 2015-05-05 | 2015-09-02 | 中国联合网络通信集团有限公司 | Resource scheduling method and resource scheduling system |
-
2016
- 2016-03-20 CN CN201610166392.8A patent/CN107203255A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102096461A (en) * | 2011-01-13 | 2011-06-15 | 浙江大学 | Energy-saving method of cloud data center based on virtual machine migration and load perception integration |
| CN104679594A (en) * | 2015-03-19 | 2015-06-03 | 成都艺辰德迅科技有限公司 | Middleware distributed calculating method |
| CN104881325A (en) * | 2015-05-05 | 2015-09-02 | 中国联合网络通信集团有限公司 | Resource scheduling method and resource scheduling system |
| CN104850459A (en) * | 2015-05-22 | 2015-08-19 | 中国联合网络通信集团有限公司 | Virtual machine migrating method and apparatus |
| CN104866375A (en) * | 2015-05-22 | 2015-08-26 | 中国联合网络通信集团有限公司 | Virtual machine migration method and apparatus |
| CN104881316A (en) * | 2015-05-22 | 2015-09-02 | 中国联合网络通信集团有限公司 | Virtual machine transferring method and device |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108540405A (en) * | 2017-12-18 | 2018-09-14 | 清华大学 | Internet resources moving method and device |
| CN108540405B (en) * | 2017-12-18 | 2020-04-07 | 清华大学 | Network resource migration method and device |
| CN108429815A (en) * | 2018-03-23 | 2018-08-21 | 重庆邮电大学 | Dynamic resource scheduling method based on OpenStack |
| CN108664116B (en) * | 2018-04-27 | 2020-03-27 | 北京邮电大学 | Self-adaptive power saving method and device for network function virtualization and CPU controller |
| CN108664116A (en) * | 2018-04-27 | 2018-10-16 | 北京邮电大学 | Adaptive electricity saving method, device and the cpu controller of network function virtualization |
| CN109002348A (en) * | 2018-07-26 | 2018-12-14 | 郑州云海信息技术有限公司 | Load-balancing method and device in a kind of virtualization system |
| CN109062668A (en) * | 2018-08-01 | 2018-12-21 | 重庆邮电大学 | A kind of virtual network function moving method of the multipriority based on 5G access network |
| CN109062668B (en) * | 2018-08-01 | 2021-10-08 | 重庆邮电大学 | Multi-priority virtual network function migration method based on 5G access network |
| CN110275758A (en) * | 2019-05-09 | 2019-09-24 | 重庆邮电大学 | A kind of virtual network function intelligence moving method |
| CN110275758B (en) * | 2019-05-09 | 2022-09-30 | 重庆邮电大学 | Intelligent migration method for virtual network function |
| CN111488053A (en) * | 2020-04-17 | 2020-08-04 | 苏州浪潮智能科技有限公司 | Power consumption adjusting method and device of network function virtualization system |
| CN111488053B (en) * | 2020-04-17 | 2023-02-28 | 苏州浪潮智能科技有限公司 | Power consumption adjusting method and device of network function virtualization system |
| CN112416530A (en) * | 2020-12-08 | 2021-02-26 | 西藏宁算科技集团有限公司 | Method and device for flexibly managing cluster physical machine nodes and electronic equipment |
| CN112416530B (en) * | 2020-12-08 | 2023-12-22 | 西藏宁算科技集团有限公司 | Method and device for elastically managing cluster physical machine nodes and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107203255A (en) | Power-economizing method and device are migrated in a kind of network function virtualized environment | |
| Feller et al. | Energy-aware ant colony based workload placement in clouds | |
| US11570937B2 (en) | Methods and apparatus to control power delivery based on predicted power utilization in a data center | |
| US10776149B2 (en) | Methods and apparatus to adjust energy requirements in a data center | |
| CN101488098B (en) | Multi-core computing resource management system based on virtual computing technology | |
| CN105302630B (en) | A kind of dynamic adjusting method and its system of virtual machine | |
| CN102662750A (en) | Virtual machine resource optimal control method and control system based on elastic virtual machine pool | |
| US11206193B2 (en) | Method and system for provisioning resources in cloud computing | |
| Khan et al. | Energy, performance and cost efficient cloud datacentres: A survey | |
| CN102521014B (en) | Deploying method and deploying device for virtual machine | |
| CN102759984A (en) | Power supply and performance management system for virtualization server cluster | |
| CN101231552A (en) | Regulating power consumption | |
| CN104503832B (en) | A kind of scheduling virtual machine system and method for fair and efficiency balance | |
| CN103365700A (en) | Cloud computing virtualization environment-oriented resource monitoring and adjustment system | |
| Sampaio et al. | Towards high-available and energy-efficient virtual computing environments in the cloud | |
| Jin et al. | Energy-efficient task scheduling for CPU-intensive streaming jobs on Hadoop | |
| US20220214917A1 (en) | Method and system for optimizing rack server resources | |
| Fernández-Cerero et al. | Energy policies for data-center monolithic schedulers | |
| Li et al. | Dynamic virtual machine consolidation algorithm based on balancing energy consumption and quality of service | |
| CN107122235A (en) | Public infrastructure resource regulating method based on application priority | |
| CN105930202B (en) | A kind of virtual machine migration method of three threshold values | |
| Feller et al. | State of the art of power saving in clusters and results from the EDF case study | |
| CN106293000B (en) | A kind of virtual machine storage subsystem power-economizing method towards cloud environment | |
| Lu et al. | Synchronous Dislocation Scheduling Quantum Algorithm Optimization in Virtual Private Cloud Computing Environment | |
| Dong et al. | Energy efficient virtual machine consolidation in mobile media cloud |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170926 |