CN107168805A - A kind of resource regulating method based on virtual machine - Google Patents
A kind of resource regulating method based on virtual machine Download PDFInfo
- Publication number
- CN107168805A CN107168805A CN201710463471.XA CN201710463471A CN107168805A CN 107168805 A CN107168805 A CN 107168805A CN 201710463471 A CN201710463471 A CN 201710463471A CN 107168805 A CN107168805 A CN 107168805A
- Authority
- CN
- China
- Prior art keywords
- data center
- virtual machine
- resource
- task requests
- dispatch processor
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
- G06F9/505—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals considering the load
-
- 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
- G06F9/46—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
-
- 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
- G06F2209/00—Indexing scheme relating to G06F9/00
- G06F2209/50—Indexing scheme relating to G06F9/50
- G06F2209/5022—Workload threshold
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2209/00—Indexing scheme relating to G06F9/00
- G06F2209/50—Indexing scheme relating to G06F9/50
- G06F2209/508—Monitor
Abstract
The present invention proposes a kind of resource regulating method based on virtual machine, and methods described is used to be scheduled resource under cloud computing environment, and methods described includes:(1) dispatch processor receives the virtual machine task requests of user, and virtual machine task requests are transmitted to some data center by dispatch processor according to the load condition of each current data center;In cloud computing environment include several data centers, each data center by several physics units into;Each physical machine includes several virtual machines;(2) data center finds a physical machine for the virtual machine task requests, allocation result is returned into dispatch processor after virtual machine task requests are connected to;(3) dispatch processor receives the virtual machine task requests allocation result that data center sends, and is sent to user.Resource regulating method proposed by the present invention based on virtual machine, improves the level of resources utilization, reduces task processing time.
Description
Technical field
The present invention relates to field of cloud calculation, and in particular to a kind of resource regulating method based on virtual machine.
Background technology
Cloud computing is a current popular technical term, is counted as after personal computer is changed, internet is changed
Third time IT tides afterwards, just generate tremendous influence in the few years.The giants such as Google, Microsoft and IBM all before
Go to the follow-up research for promoting cloud computing technology with scale at an unprecedented rate, development and popularize, it is increasing academic living
Dynamic to bring cloud computing into schedule, people also change to the understanding of computer and network.It is envisioned that the life in our future
Also therefore can occur deep change with working method.
Under traditional mode, personal or enterprise customer needs to buy the infrastructure such as set of software and hardware and bandwidth ability structure
Build a set of IT system for possessing complete function.But for a user, the calculating such as CPU, hard disk and storage facilities and other software sheets
Body is not required in that.Just as running water, user is what is required is simply that water all beats a bite well without every household.With network
Popularization and the growth of bandwidth, are reached its maturity by the condition of network access computing resource non-indigenous, and this is just cloud computing technology
Generation and development lay the foundation.Cloud computing be it is a kind of network to provide dynamic telescopic virtualization by way of servicing
The computation schema of resource.Because the computing capability and infrastructure of cloud computing are distributed in network-side, held on high just as floating
Cloud is the same.User is when cloud computing service is received, it is not necessary to go to be concerned about that the cloud facility for providing it cloud service is in where, more
It need not be concerned about how cloud platform builds, how cloud program is run, so, it is referred to as " cloud ".
The service facility of cloud computing is not limited to by user terminal, it is meant that its scale and ability can increase without limitation to meet
Bigger demand.The core document published from Google in 2003 to 2006 Amazon EC2 (elastic calculation cloud) business
Industry application, then the Synaptic Hosting services released to Telecommunication giant AT&T, cloud computing is from cost-effective
Instrument develop into the pusher of profit.Nowadays, Google cloud computings oneself through possessing more than 100 ten thousand servers, the public affairs such as Amazon
Department also possesses hundreds of thousands platform server.The characteristics of resource utilization that cloud computing possessed is high is exactly taken a fancy to, each IT companies are not
It is disconnected to expand cloud computing scale so that the cost performance of cloud computing is constantly lifted in the form of the order of magnitude.With the diffusion of scale effect
And the exploitation and application of numerous new techniques, its cost performance even can reach the decades of times of traditional mode.
The fast development of cloud computing is using virtualization technology as support.It is used as the basic technology of cloud computing, virtualization
Technology provides more effective Resource Distribution and Schedule mode on microcosmic framework for cloud computing.For single physical main frame
Speech, resource can cause to waste too much.And the quantization unit of minimum is defined using virtual machine, when resource is inadequate, with virtual machine
Computing capability is added for unit, host number can be reduced when resource exceedes.Therefore, we create master using virtual machine
The basic processing unit of machine, then permutation and combination meets different demands for services, Ke Yirang in a different manner by these units
Resource obtains maximized application.The problems such as the server maintenance under cloud computing environment, load imbalance, the dynamic of virtual machine
Migrating technology is a kind of solution of efficient stable, ensure that service quality, meets user's request.The dynamic of virtual machine is moved
Shifting is to migrate the virtual machine on a server to another server in real time.In transition process, except the time is extremely short
Pause stage, virtual machine whole process does not interrupt the service of providing the user.Therefore, the dynamic migration of virtual machine becomes in cloud computing
Solve the effective ways for the problems such as load balancing, server maintenance and working environment are migrated.But at this stage, virtual machine is moved
Transposition degree is manual or semi-automatic realization, and dispatching efficiency is not high, influences the migration performance of virtual machine.In addition, virtual machine
Dynamic migration only carries out CPU, internal memory, network and sets the migration of the states such as each and I/O, and passes through SAN (storage between main frame
Area network), the mode shared storage device such as NAS (net-work-attached storage) is without data in magnetic disk
Migration.With the development of network technology and application demand, need to realize in the case where shared storage or wan environment can not be realized
Virtual machine total system real-time migration including data in magnetic disk.So, because the data in magnetic disk amount of virtual machine is big, data in magnetic disk migration
Just into the difficult point and emphasis of virtual machine (vm) migration.
In summary, virtual machine plays a key effect for the scheduling of resource under cloud computing environment.Exist in the prior art
The problem of scheduling of resource is inefficient.
The content of the invention
At least part of solution problems of the prior art, the present invention proposes a kind of scheduling of resource based on virtual machine
Method, methods described is used to be scheduled resource under cloud computing environment, and methods described includes:
(1) dispatch processor receives the virtual machine task requests of user, and dispatch processor is according to each current data center
Load condition, virtual machine task requests are transmitted to some data center;Included in cloud computing environment in several data
The heart, each data center by several physics units into;Each physical machine includes several virtual machines;
(2) data center finds a physical machine after virtual machine task requests are connected to for the virtual machine task requests,
Allocation result is returned into dispatch processor;
(3) dispatch processor receives the virtual machine task requests allocation result that data center sends, and is sent to user.
It is preferred that, dispatch processor is additionally operable to receive the resource updates information that data center sends, and updates resource information
Storehouse;When the information of some data center changes, data center can send the information after updating, scheduling to dispatch processor
After processor is received, the database of itself is updated, the uniformity with data center is kept, so as in distribution virtual machine
Used during task requests.
It is preferred that, dispatch processor is additionally operable to manage data center, includes the registration and deletion of data center;It is new when having
, it is necessary to which in dispatch processor registration, the information of the data center is added when data center needs to be added to the cloud computing environment
To the database of dispatch processor;When there is data center to need deletion, dispatch processor database by the data center
Corresponding information delete.
It is preferred that, the virtual machine task requests that user submits include at the beginning of virtual machine use between, it is the end time, virtual
The type of machine.
It is preferred that, dispatch processor described in step (1) is according to the load condition of each current data center, by virtual machine
Task requests are transmitted to some data center, specifically include:
Dispatch processor calculates the resource load value (XD of each data center1,…,XDi…,XDN), according to the value by small
Queued up to big order for data center;Wherein, XDiThe resource load value of i-th of data center is represented, N represents the cloud money
Data center's number under the environment of source;
First data center in data center's queue is selected, virtual machine task requests are forwarded to first number
According to center.
It is preferred that, first data Central Radical, according to virtual machine task requests, is user's distribution physical machine;If allotment
Reason machine fails, then the virtual machine task requests is forwarded into next data center in data center's queue, under
Physical machine is distributed for user in one data center, until the success of distribution physical machine or all data centers distribution physical machine are lost
Lose;
Update the surplus resources for the physical machine for being assigned to virtual machine task.
It is preferred that, the dispatch processor calculates the resource load value (XD of each data center1,…,XDi…,XDN), tool
Body includes:
VCPUj、VMEMj、VSTORjData center D is represented respectivelyiCPU sizes, memory size required for middle virtual machine j,
Storage size, M represents data center DiComprising all virtual machines quantity.
Resource regulating method proposed by the present invention based on virtual machine, improves the level of resources utilization, the processing of reduction task
Time.
Brief description of the drawings
Fig. 1 is a kind of flow chart of the resource regulating method based on virtual machine of the present invention;
Embodiment
Below in conjunction with the accompanying drawing of the present invention, technical scheme is clearly and completely described.Here will be detailed
Carefully exemplary embodiment is illustrated, its example is illustrated in the accompanying drawings.In the following description when referring to the accompanying drawings, unless otherwise table
Show, the same numbers in different accompanying drawings represent same or analogous key element.Embodiment party described in following exemplary embodiment
Formula does not represent all embodiments consistent with the present invention.On the contrary, they are only detailed with institute in such as appended claims
The example of the consistent apparatus and method of some aspects state, the present invention.
Referring to Fig. 1, the present invention proposes a kind of resource regulating method based on virtual machine, and methods described is used in cloud computing
Resource is scheduled under environment, methods described includes:
(1) dispatch processor receives the virtual machine task requests of user, and dispatch processor is according to each current data center
Load condition, virtual machine task requests are transmitted to some data center;Included in cloud computing environment in several data
The heart, each data center by several physics units into;Each physical machine includes several virtual machines;
(2) data center finds a physical machine after virtual machine task requests are connected to for the virtual machine task requests,
Allocation result is returned into dispatch processor;
(3) dispatch processor receives the virtual machine task requests allocation result that data center sends, and is sent to user.
Dispatch processor is additionally operable to receive the resource updates information that data center sends, and updates resource information bank;When certain
When the information of individual data center changes, data center can send the information after updating, dispatch processor to dispatch processor
After receiving, the database of itself is updated, the uniformity with data center is kept, to be asked in distribution virtual machine task
Used when asking.
Dispatch processor is additionally operable to manage data center, includes the registration and deletion of data center;When having in new data
, it is necessary to which in dispatch processor registration, the information of the data center is added into scheduling when the heart needs to be added to the cloud computing environment
The database of processor;When there is data center to need deletion, dispatch processor database by the corresponding of the data center
Information deletion.
User submit virtual machine task requests include virtual machine use at the beginning of between, the end time, the class of virtual machine
Type.
Dispatch processor described in step (1), please by virtual machine task according to the load condition of each current data center
Ask and be transmitted to some data center, specifically include:
Dispatch processor calculates the resource load value (XD of each data center1,…,XDi…,XDN), according to the value by small
Queued up to big order for data center;Wherein, XDiThe resource load value of i-th of data center is represented, N represents the cloud money
Data center's number under the environment of source;
First data center in data center's queue is selected, virtual machine task requests are forwarded to first number
According to center;
First data Central Radical, according to virtual machine task requests, is user's distribution physical machine;If distributing physical machine to lose
Lose, then the virtual machine task requests are forwarded to next data center in data center's queue, in next number
According to physical machine is distributed in center for user, until the success of distribution physical machine or the distribution physical machine failure of all data centers;
Update the surplus resources for the physical machine for being assigned to virtual machine task.
The dispatch processor calculates the resource load value (XD of each data center1,…,XDi…,XDN), specifically include:
VCPUj、VMEMj、VSTORjData center D is represented respectivelyiCPU sizes, memory size required for middle virtual machine j,
Storage size, M represents data center DiComprising all virtual machines quantity.
First data Central Radical, according to virtual machine task requests, is user's distribution physical machine, and one embodiment includes:
The type of virtual machine asked first according to user is arranged physical machine by the utilization rate ascending order of the type, described virtual
Machine type includes cpu type, type of memory, storage class;
Physical machine is divided into by multiple intervals according to the utilization rate of the type, each interval size can dynamically be set,
Then interval all physical machines where the minimum physical machine of utilization rate are found out, by the examination distribution of virtual machine task in the interval
In all physical machines, the cpu busy percentage of each physical machine after the interval physical machine distribution virtual machine task, internal memory profit are calculated respectively
With rate, the variance of space utilisation three obtains the load balancing value of each physical machine;
Choose the minimum physical machine of load balancing value to start to distribute virtual machine, as long as distributing the physics after the virtual machine task
The capacity of machine is not above threshold value, then into the distribution of work, otherwise takes the small physical machine of next load balancing value time to be allocated, such as
Really interval all physical machines can not all be distributed, then take out next interval physical machine and be allocated, until be allocated successfully for
Only.
Physical machine is divided into multiple intervals, each interval size can dynamically be set, every time minimum from utilization rate
Physical machine where interval be allocated, so can guarantee that and first assign the task to the small physical machine of utilization rate, it is ensured that whole number
It is more balanced according to center.Only find the run time that physical machine greatly reducing algorithm from an interval every time simultaneously.
The cpu busy percentage for calculating each physical machine after the interval physical machine distribution virtual machine task, memory usage,
The variance of space utilisation three, obtains the load balancing value of each physical machine, specifically includes:
CB=(AVG-CPU_U)2+(AVG-MEM_U)2+(AVG-STOR_U)2,
AVG=(CPU_U+MEM_U+STOR_U)/3
Wherein, CB represents the load balancing value of physical machine, and CPU_U, MEM_U, STOR_U represent the CPU profits of physical machine respectively
With rate, memory usage, space utilisation, AVG represent the cpu busy percentage of physical machine, memory usage, space utilisation it is flat
Average.
First data Central Radical, according to virtual machine task requests, is user's distribution physical machine, its another embodiment bag
Include:
Calculate all virtual machine tasks and the data received by first data center in a time window T
Resource load matrix in center between all physical machines;
According to received by the resource load matrix is first data center in one time window T
All virtual machine tasks distribute physical machine.
All virtual machine tasks in one time window T of the calculating received by first data center and should
Resource load matrix in data center between all physical machines, possess including:
Calculate each in all virtual machine tasks in a time window T received by first data center
Resource load V of the virtual machine task with respect to each physical machine in all physical machines in the data centerij,
Wherein, VijRepresent all virtual machine tasks received by first data center in a time window T
In i-th of virtual machine task with respect in the data center to the resource load of j physical machine,
Represent all received by first data center in a time window T
I-th of virtual machine task in virtual machine task is with respect to the cpu resource load in the data center to j physical machine;TCPU iTable
Show the cpu resource that i-th of virtual machine required by task is wanted, PCPU jRepresent the remaining available cpu resource of j-th of physical machine;
Represent the institute received by first data center in a time window T
There is i-th of virtual machine task in virtual machine task with respect to the memory source load in the data center to j physical machine;TMEM i
Represent the memory source that i-th of virtual machine required by task is wanted, PMEM jRepresent the remaining free memory resource of j-th of physical machine;
Represent the institute received by first data center in a time window T
There is i-th of virtual machine task in virtual machine task with respect to the storage resource load in the data center to j physical machine;TSTOR i
Represent the storage resource that i-th of virtual machine required by task is wanted, PSTOR jRepresent the remaining available storage resource of j-th of physical machine;
α, β, γ difference cpu resource load weights, memory source load weights, storage resource load weights, implement one
In example, α, β, γ difference value 0.5,0.3,0.2;
The resource load matrix is expressed as:
The matrix represents to have in a time window T m available physical machines, have n need to be scheduled it is virtual
Machine task requests, VijRepresent load value V of i-th of task in j-th of physical machine.
It is described to be received according to the resource load matrix by first data center in one time window T
All virtual machine tasks distribution physical machine arrived, is specifically included:
Find out the minimum V values V of each virtual machine taski, i.e., every a line for matrix finds out the minimum value of every a line;
Vi=min { Vi1,Vi2,…,Vim};
Then, the maximum V in these minimum values is found outmax,
Vmax=max { V1,V2,…,Vn};
Find out VmaxThe task i at place, the task is exactly first will being allocated for task, and it will be scheduled for this
The be expert at V of businessijIt is
VmaxLower j representated by j-th of physical machine on;That a line of task i in a matrix is then deleted, and updates square
L value of other tasks on j-th of node in battle array, i.e., all add VmaxIf (because other tasks will be in j-th of thing
Being performed on reason machine needs wait task i to perform, therefore their L values will add Vmax);Matrix update carries out next after finishing
The calculating of business scheduling, repeats above procedure, and dispatching sequence and the scheduling whereabouts of all tasks is finally determined.
A kind of resource regulating method based on virtual machine proposed by the present invention, is still further comprised:
Resource load situation to data center's physical machine current time t subsequent time t+1 is estimated;
According to physical machine physical machine current time t resource load and subsequent time t+1 resource load, selection needs are moved
Go out the physical machine and moved out virtual machine physical machine to be moved into of virtual machine.
The resource load situation of the subsequent time t+1 to data center's physical machine current time t is estimated, specifically
Including:
W (t+1)=μ W (t)+δ W (t-1)+ω W (t-2), wherein W (t+1), W (t), W (t-1), W (t-2) are represented respectively
The resource load of physical machine during moment t+1, t, t-1, t-2, μ, δ, ω be weights, in one embodiment respectively value 0.5,0.3,
0.2。
The selection needs to move out the physical machine of virtual machine, specifically includes:
When overload situations occur for physical machine and during low load situation, according to current state by the part in the physical machine or complete
Portion's virtual machine (vm) migration is run in other physical machines.
When physical machine is in overload, virtual machine is estimated into resource load amount by moment t+1 and carries out descending arrangement, choosing
The virtual machine for selecting Future load ranking prostatitis is moved out, simultaneously so that present physical machine disclosure satisfy that remaining virtual
Resource requirement at the time of machine during t+1;
When physical machine is in low load conditions, all virtual machines are all migrated out into present physical machine.
The physical machine overload situations include cpu busy percentage and are more than 80%, and the low load situation of physical machine is utilized including CPU
Rate is less than 20%
It is described to select virtual machine physical machine to be moved into of moving out, specifically include:
The resource information of all physical machines, Cong Zhongxuan in the resource requirement for the virtual machine moved into as needed and data center
Take in the physical machine of placement migration virtual machine.
Selection strategy is to estimate its moment t+1 workloads to all physical machines for meeting resources of virtual machine distribution requirement,
Resource requirement reduction queue is divided according to situation of estimating and resource requirement expands queue, purpose physics is determined by further screening
Machine, specific steps include:
(a) the resource requirement stroke in each physical machine is calculated, its value is all virtual machine moment t+1 on present physical machine
Workload estimate total amount and subtract current time workload total amount, it has reacted the variation tendency of resources of virtual machine demand;
(b) subtract current time resource load total amount to calculate unallocated stock number by physical machine total resources, screen
The difference gone out between unallocated stock number and resource requirement stroke i.e. expected residual stock number is more than zero physical machine list, shape
Into candidate physical machine list;
If (c) resource requirement stroke be negative value, add resource requirement reduce queue, and calculate surplus yield with
The difference of resource stroke is arranged in decreasing order.If on the occasion of then addition resource requirement expands queue, computational resource requirements stroke
Ratio with surplus yield is arranged queue by ascending order as the sacurity dispatching factor, and according to the factor value, if a side
Queue is sky, then directly selects the physical machine machine of the opposing party's ranking first as physical machine to be moved into, if being not sky, is entered
One step compares the first physical machine of two queues, and selection is one of to be used as physical machine to be moved into.
It is a kind of feedforward control based on the active control that resource load is estimated, passes through the future workload feelings of Prediction System
Condition analyzes the possibility that overload or low load occur for physical machine in advance, and carries out virtual machine integration according to algorithm, it is ensured that at system
In optimum state.Passive control based on real system status information is a kind of feedback control, passes through the system group such as monitoring unit
Part collects the real time information that resource is used in each virtual machine running and data center, the status information of physical machine as anti-
Feedback, readjusts deployment of the virtual machine in physical machine.
Know the fluctuation situation of resource load in advance by pre-estimating technology, scheduler portion can be allowed to implement more at leisure
Virtual machine migration policies, play and targetedly shift to an earlier date prevention effect;The reality of scheduling strategy can be known by feedback technique again
Border implementation status, virtual machine (vm) migration operation is implemented to the physical host in overload or low load conditions, is played and is corrected control in real time
Make and use.
The resource load situation of the subsequent time t+1 to data center's physical machine current time t is estimated, and enters one
Step includes:
Count user profile on present physical machine;
User is carried out in subsequent time t+1 required resource U (t+1) using heuristic or K- nearest neighbor algorithms pre-
Estimate, take U (t+1) and the larger resource load value as subsequent time t+1 present physical machines of W (t+1) intermediate value.
The present invention proposes the resource regulating method based on virtual machine, improves the level of resources utilization, the processing of reduction task
Time.
Those skilled in the art will readily occur to its of the present invention after considering specification and putting into practice invention disclosed herein
Its embodiment.The application be intended to the present invention any modification, purposes or adaptations, these modifications, purposes or
Person's adaptations follow the general principle of the present invention and including undocumented common knowledge in the art of the invention
Or conventional techniques.
It should be appreciated that the invention is not limited in the precision architecture for being described above and being shown in the drawings, and
And various modifications and changes can be being carried out without departing from the scope.The scope of the present invention is only limited by appended claim.
Claims (7)
1. a kind of resource regulating method based on virtual machine, methods described is used to be scheduled resource under cloud computing environment,
Methods described includes:
(1) dispatch processor receives the virtual machine task requests of user, and dispatch processor is negative according to each current data center
Virtual machine task requests are transmitted to some data center by load state;Several data centers are included in cloud computing environment, often
Individual data center by several physics units into;Each physical machine includes several virtual machines;
(2) data center finds a physical machine after virtual machine task requests are connected to for the virtual machine task requests, will divide
Dispatch processor is returned to result;
(3) dispatch processor receives the virtual machine task requests allocation result that data center sends, and is sent to user.
2. the resource regulating method as claimed in claim 1 based on virtual machine, wherein,
Dispatch processor is additionally operable to receive the resource updates information that data center sends, and updates resource information bank;When certain number
When being changed according to the information at center, data center can send the information after updating to dispatch processor, and dispatch processor is received
Afterwards, the database of itself is updated, the uniformity with data center is kept, so as to when distributing virtual machine task requests
Use.
3. the resource regulating method as claimed in claim 1 based on virtual machine, wherein,
Dispatch processor is additionally operable to manage data center, includes the registration and deletion of data center;When there is new data center to need
, it is necessary to which in dispatch processor registration, the information of the data center is added into dispatch deal when being added to the cloud computing environment
The database of device;When there is data center to need deletion, dispatch processor database by the corresponding information of the data center
Delete.
4. the resource regulating method as claimed in claim 1 based on virtual machine, wherein,
User submit virtual machine task requests include virtual machine use at the beginning of between, the end time, the type of virtual machine.
5. the resource regulating method as claimed in claim 1 based on virtual machine, wherein, dispatch processor described in step (1)
According to the load condition of each current data center, virtual machine task requests are transmitted to some data center, specifically included:
Dispatch processor calculates the resource load value (XD of each data center1,…,XDi…,XDN), it is ascending according to the value
Order be data center queue up;Wherein, XDiThe resource load value of i-th of data center is represented, N represents the cloud resource ring
Data center's number under border;
First data center in data center's queue is selected, virtual machine task requests are forwarded in first data
The heart.
6. the resource regulating method as claimed in claim 5 based on virtual machine, wherein,
First data Central Radical, according to virtual machine task requests, is user's distribution physical machine;If distributing physical machine failure,
The virtual machine task requests are forwarded to next data center in data center's queue, in next data center
In distribute physical machine for user, until the success of distribution physical machine or the distribution physical machine failure of all data centers;
Update the surplus resources for the physical machine for being assigned to virtual machine task.
7. the resource regulating method as claimed in claim 5 based on virtual machine, wherein, the dispatch processor is calculated per number
According to the resource load value (XD at center1,…,XDi…,XDN), specifically include:
<mrow>
<msub>
<mi>XD</mi>
<mi>i</mi>
</msub>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>j</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>M</mi>
</munderover>
<mrow>
<msub>
<mi>VCPU</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>+</mo>
<msub>
<mi>VMEM</mi>
<mi>j</mi>
</msub>
<mo>+</mo>
<msub>
<mi>VSTOR</mi>
<mi>j</mi>
</msub>
<mo>,</mo>
</mrow>
VCPUj、VMEMj、VSTORjData center D is represented respectivelyiCPU sizes, memory size required for middle virtual machine j, storage
Size, M represents data center DiComprising all virtual machines quantity.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710463471.XA CN107168805A (en) | 2017-06-19 | 2017-06-19 | A kind of resource regulating method based on virtual machine |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710463471.XA CN107168805A (en) | 2017-06-19 | 2017-06-19 | A kind of resource regulating method based on virtual machine |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107168805A true CN107168805A (en) | 2017-09-15 |
Family
ID=59819808
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710463471.XA Pending CN107168805A (en) | 2017-06-19 | 2017-06-19 | A kind of resource regulating method based on virtual machine |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107168805A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108279968A (en) * | 2017-12-29 | 2018-07-13 | 中国联合网络通信集团有限公司 | A kind of dispatching method and device of resources of virtual machine |
CN109743183A (en) * | 2019-01-07 | 2019-05-10 | 北京云基数技术有限公司 | A kind of cloud charging method and system |
CN110297693A (en) * | 2019-07-04 | 2019-10-01 | 北京伟杰东博信息科技有限公司 | A kind of method and its system of the distribution of distributed software task |
CN112732401A (en) * | 2020-12-29 | 2021-04-30 | 深圳前海微众银行股份有限公司 | Virtual machine resource allocation method, system, device and medium |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103577265A (en) * | 2012-07-25 | 2014-02-12 | 田文洪 | Method and device of offline energy-saving dispatching in cloud computing data center |
-
2017
- 2017-06-19 CN CN201710463471.XA patent/CN107168805A/en active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103577265A (en) * | 2012-07-25 | 2014-02-12 | 田文洪 | Method and device of offline energy-saving dispatching in cloud computing data center |
Non-Patent Citations (1)
Title |
---|
孙夏爽: ""多数据中心负载均衡调度的研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108279968A (en) * | 2017-12-29 | 2018-07-13 | 中国联合网络通信集团有限公司 | A kind of dispatching method and device of resources of virtual machine |
CN108279968B (en) * | 2017-12-29 | 2021-05-11 | 中国联合网络通信集团有限公司 | Virtual machine resource scheduling method and device |
CN109743183A (en) * | 2019-01-07 | 2019-05-10 | 北京云基数技术有限公司 | A kind of cloud charging method and system |
CN110297693A (en) * | 2019-07-04 | 2019-10-01 | 北京伟杰东博信息科技有限公司 | A kind of method and its system of the distribution of distributed software task |
CN110297693B (en) * | 2019-07-04 | 2020-07-28 | 北京伟杰东博信息科技有限公司 | Distributed software task allocation method and system |
CN112732401A (en) * | 2020-12-29 | 2021-04-30 | 深圳前海微众银行股份有限公司 | Virtual machine resource allocation method, system, device and medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107273185B (en) | Load balancing control method based on virtual machine | |
CN104657221B (en) | The more queue flood peak staggered regulation models and method of task based access control classification in a kind of cloud computing | |
CN109788046B (en) | Multi-strategy edge computing resource scheduling method based on improved bee colony algorithm | |
CN107273211A (en) | Data processing method based on virtual machine under a kind of cloud computing environment | |
Florence et al. | A load balancing model using firefly algorithm in cloud computing | |
CN107168805A (en) | A kind of resource regulating method based on virtual machine | |
Natesan et al. | Optimal task scheduling in the cloud environment using a mean grey wolf optimization algorithm | |
Joshi et al. | Load balancing in cloud computing: Challenges & issues | |
Ullah et al. | Task classification and scheduling based on K-means clustering for edge computing | |
CN104917839A (en) | Load balancing method for use in cloud computing environment | |
CN107562537A (en) | A kind of cloud computing method for scheduling task based on gravitation search | |
CN111831415A (en) | Multi-queue multi-cluster task scheduling method and system | |
CN112559122A (en) | Virtualization instance management and control method and system based on electric power special security and protection equipment | |
Khodar et al. | New scheduling approach for virtual machine resources in cloud computing based on genetic algorithm | |
Majumder et al. | Genetic algorithm-based two-tiered load balancing scheme for cloud data centers | |
Zaouch et al. | Load balancing for improved quality of service in the cloud | |
AU2021103249A4 (en) | A novel multi-level optimization for task scheduling and load balancing in cloud | |
Qaddoum et al. | Elastic neural network method for load prediction in cloud computing grid. | |
CN111124619B (en) | Container scheduling method for secondary scheduling | |
Ashalatha et al. | Dynamic load balancing methods for resource optimization in cloud computing environment | |
CN114090239A (en) | Model-based reinforcement learning edge resource scheduling method and device | |
KR102014246B1 (en) | Mesos process apparatus for unified management of resource and method for the same | |
Patni et al. | Distributed approach of load balancing in dynamic grid computing environment | |
Al Sallami | Load balancing in green cloud computation | |
Rahman et al. | Group based resource management and pricing model in cloud computing |
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: 20170915 |