CN107273181A - A kind of multilayer nest virtualization infrastructure and its method for allocating tasks - Google Patents
A kind of multilayer nest virtualization infrastructure and its method for allocating tasks Download PDFInfo
- Publication number
- CN107273181A CN107273181A CN201710405158.0A CN201710405158A CN107273181A CN 107273181 A CN107273181 A CN 107273181A CN 201710405158 A CN201710405158 A CN 201710405158A CN 107273181 A CN107273181 A CN 107273181A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msub
- task
- virtual machine
- nested
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
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
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
- G06F2009/45566—Nested virtual machines
-
- 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
Abstract
The invention belongs to field of cloud computer technology, disclose a kind of multilayer nest virtualization infrastructure and its method for allocating tasks, Hypervisor layers are one layer with the VME operating system layer run directly under its management by the multilayer nest virtualization infrastructure, and are represented with Li；Wherein i represents nested number of times, and L0 represents bottom physical hardware layer, and L1 represents VM layers of Regular corresponding with its of Hypervisor layers of Root；The multilayer nest virtualization infrastructure includes：Two layers of nested virtualization part and multilayer nest virtualization part.The present invention is by the improvement to existing nested virtualization structure and its method for allocating tasks, and multilayer nest virtualization infrastructure can efficiently reduce the generation of cloud resource fragment, can effectively improve the utilization ratio of cloud resource.Meanwhile, method for allocating tasks can reach the compromise of task treatment effeciency and cost on the basis of cloud resource utilization ratio is improved.
Description
Technical field
The invention belongs to field of cloud computer technology, more particularly to a kind of multilayer nest virtualization infrastructure and its task distribution side
Method.
Background technology
The technology attracted most attention as 21 century IT industry, the production and living that cloud computing gives people bring many facilities.
However, continuing to develop with cloud computing, cloud user's is on the increase, the deficiency of cloud computing conventional architectures and operation mode is also opened
Beginning shows.In the problems that cloud computing faces, cloud resource utilization rate is not enough, the excessive problem of cloud resource fragment is the most serious.
In the presence of resource isolation, the idling-resource piece for the Cloud Server leased as commodity will always be in the spare time before lease expires
Configuration state, and can not be recycled by cloud service provider.And being on the increase with the Cloud Server quantity leased, these
Scattered resource fragmentation will cause huge problem of resource waste.How efficiently using these scattered resource pieces become in order to
The significant challenge that cloud computing is faced.In many technologies, virtualization performance table in terms of hardware resource utilization is improved
Reveal color.And as the upgrading of traditional virtual, it is another that nested virtualization technology allows a VME operating system to run on
VME operating system；To solve the problems, such as that the Cloud Server low utilization of resources provides new solution direction.Except virtualization
Task allocation algorithms under technology, the cloud environment of a function admirable are also to solve the not enough key of cloud resource utilization rate.It is based on
The thought of load balancing, by reasonably designing, task allocation algorithms can improve the utilization rate of cloud resource in scheduling of resource aspect.
In summary, the problem of prior art is present be：Computing resource utilization rate deficiency, cloud resource are broken under cloud computing environment
Piece is excessive, causes the waste problem of resource, increases soft and hardware maintenance cost, cost of cloud service etc..
The content of the invention
The problem of existing for prior art, distributes the invention provides a kind of multilayer nest virtualization infrastructure and its task
Method.
The present invention is achieved in that a kind of multilayer nest virtualization infrastructure, and the multilayer nest virtualization infrastructure includes：
Two layers of nested virtualization broadcasts part and multilayer nest virtualizes part.
Further, there are multiple nested Hypervisor layers in the multilayer nest virtualization infrastructure；It is high-rise nested empty
Plan machine is defined as eps relative to the computing resource that the nested virtual machine of the second layer is additionally requiredi,j(i≥3)；Wherein i represents nesting level
The number of plies of place structure, j represents the virtual machine numbering of same nesting level；Simultaneously also by i-th layer of Hypervisor performance cost
It is mapped into epsi,jIn (i >=3)；Wherein idlei,jRepresent the idling-resource piece for the virtual machine that i-th layer of numbering is j.
Another object of the present invention is to provide a kind of method for allocating tasks using the multilayer nest virtualization infrastructure,
The expression formula of the method for allocating tasks is：
Wherein：
Set V={ v1,v2...,vk}：It is the set of a waiting task composition, the wherein number of task is k；
W=(vi)：Represent processing task viIt is expected that required processing time；
W=(vi,p)：Represent virtual machine P processing tasks viIt is expected that required processing time；
ts=(vi,p)：Represent virtual machine P processing tasks viThe time of beginning；
tf=(vi,p)：Represent virtual machine P processing tasks viThe time of end, meet below equation：
tf(vi, P) and=ts(vi, P) and+w (vi, P)；
prec(i,p)：Represent virtual machine P in processing task viBefore, the set of the task of required processing；If appointed in processing
Be engaged in viPreceding virtual machine P does not need task to be processed, prec (i, p)；And virtual machine P appointing in prec (i, p) has been handled
, will not start to process task v before businessi；
Avail=(vi,p)：Virtual machine P has been handled after the task in prec (i, p), and gets out execution task vi
Time：
Avail=(vi, p) it is time for having performed last task in prec (i, p)：
EFT=(vi, p)：Represent virtual machine P processing tasks viThe earliest end time, it is：
FRT(i)：Represent performing task viBefore, the Cloud Server is expected the lease overdue time, when having performed task vi
Afterwards, FRT (i) ' is obtained；
time：Represent execution task viIncreased server rental period, be：
Time=(FRT (i) '-FRT (i))；
cost(vi, Pj) show execution task viThe cost of required payment, is defined as：
cost(viPj)=time*mtu；
The cost rented in wherein mtu required for the Cloud Server under a certain configuration；
H={ HOST1, HOST2..., HOSTK}：Represent to have rented the set that Cloud Server is constituted.
Another object of the present invention is to provide a kind of cloud resource optimization method utilized, the cloud resource optimization side
Method includes：
(1) generation of nested virtual machine creating dispatch list
In multilayer nest virtualization, the hierachy number where highest nesting virtual machine should be the maximum for the n for meeting formula：
epsn＜ idle2,1, (n ＞ 2)；
Final remaining idling-resource is defined as：idle′1,1, the resource loss for obtaining multilayer nest virtualization infrastructure is total
With, and it is defined as wstn：
Wherein n represents the nested highest number of plies, and k represents the number of the L2 layers of virtual machine of standard of two layers of nested structure part；
Obtain optimizing the nested number of plies j of nested structure by following formula, and obtain nested virtual machine creating dispatch list：
Wherein：vmi,j：The virtual machine that i-th layer of numbering is j is represented, Cloud Server is defined as first layer virtual machine
vmi,j；
idlei,j：Represent the idling-resource piece for the virtual machine that i-th layer of numbering is j；Wherein idle1,1Represent Cloud Server
vm1,1Idle computing resources；
epsi,j(i≥3)：Represent that i-th layer is numbered the virtual machine for being j when performing some task；
min_resourcei：Represent that i-th layer of guest virtual machine wants minimum computing resource to handle some required by task,
And min_resourceiResource at least needed for operation VME operating system；
Rn：When representing nested n-layer, multilayer nest virtualization part takes idle1,1For carrying all nested virtual machines
Computing resource part.
(2) generation of schedule of tasks：
In task viDuring being allocated, under task scheduling figure at that time, its max [cost] and max [EFT (vi,
P) it is] fixed, finds the minimum value of formula, its corresponding virtual machine P is exactly to be best suitable for execution task viVirtual machine.When appoint
Be engaged in viFor last task to be allocated of this batch of task, that is, obtain the schedule of tasks of this batch of task.
Further, the generation of the virtual machine creating dispatch list includes：
First, the computing resource to rented Cloud Server is counted and analyzes the characteristic of waiting task；And according to institute
The information of collection, calculates the estimated idling-resource amount of Cloud Server；
After the estimated resource free quantity of Cloud Server is obtained, according to the nested virtualization structure relied on, generation is nested empty
Plan machine creates dispatch list.
The generation of the schedule of tasks includes：
First, all workable task processing virtual machines are traveled through；And count existing task processing virtual machine
Task disposition.
After the task disposition of task processing virtual machine is obtained, calculated and taken office according to institute's dependence task allocation algorithm
Business dispatch list.
Another object of the present invention is to provide a kind of Cloud Server of the application cloud resource optimization method.
Advantages of the present invention and good effect are：By changing to existing nested virtualization structure and its method for allocating tasks
Enter, multilayer nest virtualization infrastructure can efficiently reduce the generation of cloud resource fragment, the utilization of cloud resource can be effectively improved
Efficiency.Meanwhile, method for allocating tasks can reach task treatment effeciency and cost on the basis of cloud resource utilization ratio is improved
Compromise.The nested virtualization structure and task allocation algorithms of the present invention all has performance evaluation below.Due to existing virtualization
The limitation of technology, can only at most nesting third layer virtual machine, and the nested virtual machine performance loss of third layer is serious.So cannot
With experiment performance comparison, Zhi Nengyong are carried out come the multilayer nest virtualization infrastructure to the present invention with traditional nested virtualization structure
Theory analysis；Task allocation algorithms see attached list lattice with traditional heuristic task allocation algorithms and round-robin technique analysis.
Table 1：(EFT of this batch of task maximum comes the earliest time of the task allocation algorithms completion batch tasks carrying of the present invention
Embody) show medium, performance is inferior to traditional heuristic allocation algorithm better than loop distribution algorithm.
Table 2：Processing earliest time (the schedule of this batch task of the task allocation algorithms of the present invention for individual task
Length embodies) performance is inferior to traditional heuristic allocation algorithm better than loop distribution algorithm.
Table 3：The task allocation algorithms of the present invention are heuristic better than traditional for performing needed for batch tasks in terms of cost
Allocation algorithm and loop distribution algorithm.
Table 1
Table 2
Table 3
To sum up, task allocation algorithms of the invention can reach task processing effect on the basis of cloud resource utilization rate is improved
The compromise of rate and cost.
Brief description of the drawings
Fig. 1 is existing multilayer nest virtualization infrastructure schematic diagram provided in an embodiment of the present invention；
(a) two layers of nested virtualization structure in figure；(b) multilayer nest virtualization infrastructure.
Fig. 2 is cloud resource optimization method flow chart provided in an embodiment of the present invention.
Fig. 3 is task distribution example schematic diagram provided in an embodiment of the present invention；
In figure：(a) Cloud Server 1；(b) Cloud Server 2.
Fig. 4 is multilayer nest virtualization infrastructure schematic diagram provided in an embodiment of the present invention.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to embodiments, to the present invention
It is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not used to
Limit the present invention.
The present invention solves the problems such as computing resource utilization rate is not enough, cloud resource fragment is excessive under cloud computing environment.With reference to
Task distribution under cloud computing environment needs, the cloud resource optimization method virtualized using multilayer nest；Also, it is new comprising one
Nested virtualization structure and one be applied to nested cloud environment under method for allocating tasks；By to existing nested virtualization knot
The improvement of structure and its method for allocating tasks, multilayer nest virtualization infrastructure can efficiently reduce the generation of cloud resource fragment, can
Effectively improve the utilization ratio of cloud resource.Meanwhile, method for allocating tasks can reach on the basis of cloud resource utilization ratio is improved
To the compromise of task treatment effeciency and cost.
The application principle of the present invention is explained in detail below in conjunction with the accompanying drawings.
As shown in figure 1, multilayer nest virtualization infrastructure provided in an embodiment of the present invention is transported by Hypervisor layers and directly
VME operating system layer of the row under its management is referred to as one layer, and is represented with Li.Wherein i represents nested number of times, and
L0 represents bottom physical hardware layer, and L1 then represents VM layers of Regular corresponding with its of Hypervisor layers of Root.Two layers nested
The advantage of virtualization infrastructure is to only need to introduce L2Hypervisor, i.e., client Hypervisor layers, and all L2 virtual machines will
Directly it is controlled by it, greatly reduces the difficulty that the complexity and structure of structure are realized；And in multilayer nest virtualization infrastructure,
By recursive nestings, there are multiple nested Hypervisor layers in structure；High-rise nesting virtual machine is nested relative to the second layer
The computing resource that virtual machine is additionally required is defined as epsi,j(i≥3)；The number of plies of structure where wherein i represents nesting level, j is represented
The virtual machine numbering of same nesting level.I-th layer of Hypervisor performance cost is also mapped into eps simultaneouslyi,jIn (i >=3).
Wherein idlei,jRepresent the idling-resource piece for the virtual machine that i-th layer of numbering is j.
The application principle of the present invention is further described below in conjunction with the accompanying drawings.
As shown in figure 4, the multilayer nest virtualization infrastructure of design of the embodiment of the present invention includes：Two layers of nested virtualization part
Part is virtualized with multilayer nest.
There are multiple nested Hypervisor layers in the multilayer nest virtualization infrastructure；High-rise nesting virtual machine is relative
The computing resource being additionally required in the nested virtual machine of the second layer is defined as epsi,j(i≥3)；Wherein i represents structure where nesting level
The number of plies, j represent same nesting level virtual machine numbering；Also i-th layer of Hypervisor performance cost is mapped into simultaneously
epsi,jIn (i >=3)；Wherein idlei,jRepresent the idling-resource piece for the virtual machine that i-th layer of numbering is j.
(1) present invention includes a new multilayer nest virtualization based on the cloud resource optimization method that multilayer nest is virtualized
Structure and the task allocation algorithms under a new nested cloud environment.Task processing execution flow chart such as Fig. 2 under cloud environment
It is shown.By cooperating for designed nested virtualization structure and task allocation algorithms, cloud resource utilization ratio can be reached
Maximize.
The core of the cloud resource prioritization scheme of the present invention is the generation of two dispatch lists：
(1) nested virtual machine creating dispatch list
The generation of virtual machine creating dispatch list has following two steps：
First, the computing resource to rented Cloud Server is counted and analyzes the characteristic of waiting task, and such as processing should
Amount of computational resources that required by task is wanted etc..And according to collected information, calculate the estimated idling-resource amount of Cloud Server.
After the estimated resource free quantity of Cloud Server is obtained, according to the nested virtualization structure relied on, generation is nested empty
Plan machine creates dispatch list.
(2) task allocation schedule table
The generation of task allocation schedule table has following two steps：
Virtual machine can created and be used for the virtual machine for handling task, the Cloud Server do not rented also or gone back
The nested virtual machine not being created.If selected for the nested virtual machine not being created, now need embedding residing for the decision empty machine
Jacket layer.List processing is dispatched by the nested virtual machine creating generated above.
After for each task distribution task processing virtual machine, according to the task allocation algorithms relied on, task is generated
Allocation schedule table.
(2) nested virtualization structure design
It is defined as follows in the symbol used：
vmI, j：The virtual machine that i-th layer of numbering is j is represented, Cloud Server is defined as vm as first layer virtual machineI, j。
idleI, j：Represent the idling-resource piece for the virtual machine that i-th layer of numbering is j.Wherein idle1,1Represent Cloud Server
vm1, 1Idle computing resources.
epsI, j(i≥3)：Represent that i-th layer is numbered the virtual machine for being j when performing some task, exists to reach with the task
Identical treatment effeciency, required extra computing resource during the processing of second layer guest virtual machine.
min_resourcei：Represent that i-th layer of guest virtual machine wants minimum computing resource to handle some required by task,
And min_resourceiResource at least needed for operation VME operating system.If representing i-th layer of guest virtual machine is
With min_resourceiCreated for configuration, then it be called Li layers of nesting virtual machine of standard, and its idling-resource is referred to as
For idle_standard.From virtualization rudimentary knowledge, each layer of min_resource is different, and with layer
Number increase.Each layer same of idle_standard is also different.
Rn：When representing nested n-layer, multilayer nest virtualization part takes idle1,1For carrying all nested virtual machines
Computing resource part.
It is assumed that one layer of guest virtual machine is when performing the task of same type, their performance configuration is all min_
resourcei.Due to processing identical task, so their idling-resource should be equal, that is, following condition is met：
idle2, j=idle2, k, (j, k ＞ 0) (1-2)
In order to reach the saving of resource, in multilayer nest virtualization part, the hierachy number where highest nesting virtual machine,
It should be the maximum for the n for meeting formula (1-3)：
epsn＜ idle2,1, (n ＞ 2) (1-3)
Because each layer of eps is fixed, in order to reduce idle as far as possiblen,1, it is necessary to vmn,1It is that Ln layers of a standard is embedding
Set virtual machine, i.e. its configuration should be min_resource.Then, vmn-1,1Configuration be just determined, as min_resource
Required by task resource sum is handled with it.Such recurrence is gone down, the vm of multilayer nest part2,1Configuration RnJust it is determined.Finally
By idle1,1Remaining part is all for creating the nested virtual machine of L2 layers of standard of two layers of nested parts.Due to remaining resource
Number and not always min_resource2Multiple, therefore final remaining idling-resource is defined as by the present invention：idle'1,1。
Then, the resource loss summation of the multilayer nest virtualization infrastructure is obtained, and is defined as wstn：
Wherein n represents the nested highest number of plies, and k represents the number of the L2 layers of virtual machine of standard of two layers of nested structure part.
Meanwhile, in order to which the utilization ratio for improving bottom computing resource to greatest extent is maximized, it can be obtained by formula (1-5)
Optimize the nested number of plies j of nested structure：
According to the description of the multilayer nest virtualization infrastructure to upgrading, specific algorithm is as follows：
According to algorithm 1, the establishment dispatch list of virtual machine under a multilayer nest virtualization infrastructure can be obtained.When need wound
When building nested virtual machine, by this dispatch list, the computing resource that cloud user can be according to required for being handled task, suitable
The number of plies creates the nested virtual machine for meeting demand.When nested virtual machine all in this virtual machine creating dispatch list is all created
During completion, the cloud computing resources level of resources utilization is up to maximum.
(3) task allocation algorithms
Before task allocation algorithms are introduced, the symbol for needing to use in algorithm is defined first：
Set V={ v1,v2...,vk}：It is the set of a waiting task composition, the wherein number of task is k.
W=(vi)：Represent processing task viIt is expected that required processing time.
W=(vi,p)：Represent virtual machine P processing tasks viIt is expected that required processing time.
ts=(vi,p)：Represent virtual machine P processing tasks viThe time of beginning.
tf=(vi,p)：Represent virtual machine P processing tasks viThe time of end.It meets below equation：
tf(vi, P) and=ts, (vi, P) and+w (vi, P)；
prec(i,p)：Represent virtual machine P in processing task viBefore, the set of the task of required processing.If appointed in processing
Be engaged in viPreceding virtual machine P does not need task to be processed, prec (i, p).And virtual machine P appointing in prec (i, p) has been handled
, will not start to process task v before businessi。
Avail=(vi,p)：Virtual machine P has been handled after the task in prec (i, p), and gets out execution task vi
Time：
All tasks are all independent and can only be by the tasks of operating system serial operation；So avail=(vi,p)
The time of last task in prec (i, p) is as performed：
EFT=(vi,p)：Represent virtual machine P processing tasks viThe earliest end time.As：
EFT(vi, P) and=ts(vi, P) and+w (vi, P)；
FRT(i)：Represent performing task viBefore, the Cloud Server is expected the lease overdue time.When having performed task vi
Afterwards, FRT (i) ' is obtained.In scheme, only two kinds situations can cause FRT (i) '>FRT(i)：
If in the Cloud Server, for performing task viVirtual machine be not created also, after its establishment, existing fortune
Row then results in FRT (i) ' in the hydraulic performance decline of the virtual machine in the Cloud Server>FRT(i).
If EFT=(vi,p)>FRT (i), FRT (i) ' ← EFT=(vi,p)。
time：Represent execution task viIncreased server rental period, be：
Time=(FRT (i) '-FRT (i))；
cost(vi,Pj) show execution task viThe cost of required payment, is defined as：
cost(vi, Pj)=time*mtu；
Wherein mtu is the cost required for the Cloud Server under a certain configuration of rental.
H={ HOST1,HOST2...,HOSTK}:Represent to have rented the set that Cloud Server is constituted.When a certain Cloud Server
Rental period terminates/started, and the set can be rejected/added to system automatically.
On the basis of task treatment effeciency and service cost is considered, it is adaptable to the task distribution under nested cloud environment
Algorithm is as follows：
By formula (1-6), in task viDuring being allocated, under task scheduling figure at that time, its max
[cost] and max [EFT (vi, p)] and it is fixed, so only needing to find the minimum value of formula, its corresponding virtual machine P is exactly
It is best suitable for execution task viVirtual machine, selected virtual machine P can ensure task processing speed on the basis of simultaneously it is simultaneous
The expense of Gu task processing.According to the description to task allocation algorithms, task allocation algorithms are arranged as follows：
With reference to being explained in detail to the application effect for comparing the present invention.
In comparison model, L1 layers of virtual machine is Cloud Server.From the figure 3, it may be seen that two layers traditional of nested virtualization structure
Defect be：In the presence of resource isolation, the idling-resource piece of each nesting virtual machine can not be other virtual machine institutes
With, and will always be in idle state before lease expires.Compared to two layers nested virtualization structure, multilayer nest virtualization knot
Although structure can reduce the generation of resource fragmentation as far as possible, in the presence of resource expansion, eps generation can be caused, and
idle2,1＜ idlen,1.If idlen,1And the eps of each layeri,j(i >=3) sum is more than identical in two layers of nested virtualization structure
The idle of the virtual machine of quantity2,jSum, now the resource utilization situation of two layers of nested virtualization is obviously more empty than the multilayer nest
More preferably, now multilayer nest virtualization infrastructure will have no advantage in computing resource saving problem to say planization structure.Therefore, need
Design a performance more preferably, efficiency more excellent nested virtualization structure improves the resource utilization of cloud platform.
Existing cloud platform task allocation algorithms have greedy algorithm, Hadoop task allocation algorithms, the distribution of heuristic task
Algorithm etc..Greedy algorithm：Such as Min-min algorithms, Max-min algorithms, Sufferage algorithms.Wherein Min-min algorithms are followed
Be the oldest allocated of task and most fast processing, and easy first and difficult later principle is used to the priority of task, had the advantage that
Task waiting time can be reduced to greatest extent.Although oldest allocated and most fast processing of the Max-min algorithms also in compliance with task,
It is to be different from Min-min algorithms, the sequence of its task priority uses principle difficult at first and quite easy afterwards.And what Sufferage algorithms were used
It is then least disadvantage method, when task has prioritized contention, " weak tendency " task first distributes computing resource.Generally speaking, it is greedy
What algorithm considered is the speed of task processing.But pursuit computational efficiency simply, the load imbalance of calculate node will be caused
(processing pressure, which will focus on, calculates the strong node of performance).
Hadoop task allocation algorithms：According to the task processing requirements of different user, hadoop has different task distribution
Algorithm.Comparing classical has fifo algorithm, fair scheduling algorithm etc..Fifo algorithm, it was found from literal meaning, i.e., first
To task first handle.And it is the fair allocat principle of resource that equity dispatching rule, which is adhered to,.Such algorithm can cause substantial amounts of
The generation of resource fragmentation, and calculate node will the frequently switching between different task.
Heuristic task allocation algorithms.By carrying out priority ranking to task, and according to distribution experience, scheduling strategy etc.
Factor, such method can be quickly by each duty mapping to corresponding calculate node, and generates a work distribution chart.Compare
Famous heuritic approach has genetic algorithm and ant group algorithm etc..
According to different service QoS demands, the implementation strategy of task allocation algorithms is different.Loaded in view of Cloud Server
On the basis of balanced and service high availability, Hung proposes a task allocation algorithms and distributed to improve task in cloud environment
Efficiency.The algorithm considers the efficiency and cost problem of task processing simultaneously, and has reached the compromise of two kinds of factors.Pass through
Formula (1-1) minimum value is found, its corresponding task allocation schedule table is optimal task allocation schedule table.Wherein, V is represented
The set of tasks of waiting task composition, viRepresent the task of priority in set of tasks as i.H represents rented cloud service
The set of device composition, pkRepresent the Cloud Server numbered in Cloud Server set as k.EFT(vi,pk) represent task viIn cloud clothes
Be engaged in device pkIn estimated execution end time, corresponding cost (vi,pk) it is Cloud Server pkTask v is performediIt is expected that flower
Pin：
Although hung algorithm has reached the compromise of task treatment effeciency and cost, it can not solve appointing in Fig. 3
Business distribution condition.In the presence of cloud service charge mode, the nested virtual machine that runs in Cloud Server simultaneously need not be extra
Pay.In Cloud Server 1 in Fig. 3, OS1 and OS3 need not simultaneously pay, and be taken because the charge of Cloud Server 1 takes cloud by it
The time of business determines that as OS2 task handles the end time.Therefore, HostVM1 OS1 and HostVM2 OS1 can be protected
The service cost for demonstrate,proving processing task 7 is 0, so under the calculation formula of the algorithm policy, the conclusion drawn is：In processing task
When 7, HostVM1 OS1 and HostVM2 OS1 performances are consistent, and then task 7 can be randomly assigned to above-mentioned two by the algorithm
One in virtual machine.However, during actual task processing, the obvious performances of OS1 in HostVM2 are more preferable.Reason is task 7
When distributing to the OS1 in Cloud Server HostVM2, tasks carrying terminates to be expected to 7:04, and assign the task to cloud service
During OS1 in device HostVM1, now terminate can be 7 for tasks carrying:06.Obviously, the task allocation algorithms do not have fine solution
The problem.
The present invention by designed multilayer nest virtualization infrastructure with suitable for the task allocation algorithms under nested cloud environment
Cooperate, the utilization of cloud resource can be optimized simultaneously in software, hardware view.Relative to prior art, advantage master
There is following two aspect：
(1) nested virtualization structural advantage
When the nested virtual machine quantity of establishment is n, the resource loss of two layers traditional of nested virtualization structure isThe resource loss of traditional multilayer nest virtualization infrastructure is
The resource loss of multilayer nest virtualization infrastructure of the present invention is the wst of formula (1-4)x(2≤x≤y).Wherein x is the nested number of plies,
Y is determined to allow the nested highest number of plies by formula (1-3).Because nested virtualization structure can be created to each nested virtual machine
Build dispatch list calculating formula (1-4) wstx(2≤x≤y), wherein the situation comprising x=2 (is now that traditional two layers is nested
Virtualization infrastructure), therefore nested virtualization structure includes the performance of traditional two layers of nested virtualization structure.And at formula (1-5)
In the presence of, it is to obtain the nested virtual machine creating dispatch list that resource loss is minimum value, therefore nested virtualization structure is being carried
High computing resource utilization rate aspect of performance is better than two layers traditional of nested virtualization structure.And traditional multilayer nest virtualization knot
Structure is according to its nested virtual machine creating rule, and its nested number of plies is the quantity n of nested virtual machine.As (n+1)≤y, pass
The nested virtual machine creating dispatch list of the multilayer nest virtualization infrastructure generation of system is same virtual in multilayer nest virtualization infrastructure
Machine is created in the range of dispatch list traversal, so when nested virtualization structure include the property of traditional multilayer nest virtualization infrastructure
Energy.As (n+1) ＞ y, due to being unsatisfactory for formula (1-3), and idlen,1＞ idley,1, therefore traditional multilayer nest virtualization knot
Structure resource loss can be divided into two parts：2~y layers of nested virtual machine part virtual machine part nested with y+1~n+1.It is nested empty
The planization structure nesting highest number of plies takes y, then multilayer nest virtualizes partition losses than the 2 of traditional multilayer nest virtualization infrastructure
~y layers of nested virtual machine part have more idley,1.And under formula (1-3) constraint, traditional multilayer nest virtualization knot
Two layers nested virtualization part of the nested virtual machine parts of y+1~n+1 of structure but than the nested virtualization structure of the present invention has more
Due to idlen+1＞ idley,1, andSo the present invention is nested empty
Planization structure is improving computing resource utilization rate aspect of performance better than traditional multilayer nest virtualization knot in (n+1) ＞ y
Structure.To sum up, multilayer nest virtualization infrastructure of the present invention is nested better than traditional two layers empty in terms of computing resource utilization rate is improved
Planization part and multilayer nest virtualize part.
(2) task allocation algorithms advantage
The present invention is cloud environment nested virtualization structure, except that can reach task processing effect as Hung task allocation algorithms
The compromise of rate and cost, is also better than Hung task allocation algorithms in following several respects：
● computation complexity
The descriptions of task allocation algorithms of the present invention is understood, EFT (v in formula (1-6)i,P)≤max[EFT(vi, P)], cost
(vi, P)≤max [cost] and task processing cost cost for minute level cost.If the processing time of task to be allocated is x
Minute, the Cloud Server quantity rented is y, (cost (vi, P) and/max [cost])=1/x, now task distribution calculation of the present invention
Method is that the computation complexity of task distribution processing virtual machine is O (y*ex).Processing time and the cloud clothes of rental due to task
The quantity of business device is all constant, therefore task allocation algorithms proposed by the present invention are multiple for the calculating that individual task distributes processing virtual machine
Miscellaneous degree can be reduced to O (1).If the number of tasks of required processing be n, task allocation algorithms generate this batch of task allocation schedule table
Computation complexity be O (n).And in the task allocation algorithms that Hung is proposed, due to taking formula (1-6) can be made to obtain minimum value
Schedule of tasks, so needing to travel through all possible schedule of tasks.If the Cloud Server quantity rented is y,
The number of tasks of required processing is n, then can generate y*n schedule of tasks altogether.Understand, the task allocation algorithms pair that Hung is proposed
The computation complexity of one schedule of tasks calculating formula (1-1) is O (n).Now, to all schedule of tasks calculating formulas (1-1)
And it is O (n*y to obtain the computation complexity of minimum valuen).By the Cloud Server quantity rented is constant, therefore what Hung was proposed
The computation complexity of task allocation algorithms generation task allocation schedule table is O (2n).To sum up, task allocation algorithms of the present invention are in meter
The task allocation algorithms proposed in terms of calculating complexity better than Hung.
● algorithm performance
The task allocation algorithms that Hung is proposed can not solve Fig. 3 Task Allocation Problem.And task proposed by the present invention point
It is applied to solve the Task Allocation Problem with algorithm.From the figure 3, it may be seen that Host VM1 OS1 and Host VM2 OS1 can be protected
The service cost cost for demonstrate,proving processing task 7 is 0, now (cost (vi, P) and/max [cost])=0.But it is due to task of the present invention
Allocation algorithm introduces e, i.e., that the calculating of formula (1-6) reality is compared is EFT (vi,P)/max[EFT(vi,Pj)] part.Due to
The end time EFT of HostVM1 OS1 and HostVM2 OS1 processing tasks 7 is different, therefore the calculating acquired results of formula (1-6)
Difference, therefore task allocation algorithms of the present invention can solve Fig. 3 Task Allocation Problem, i.e., task 7 is distributed into Host VM2's
OS1 processing.
● algorithm customizes function
Not only in task treatment effeciency, the two factors reach compromise to task allocation algorithms with cost, compared to document Hung
The right of allocative decision can be changed according to self-demand by imparting to the also hommization of the task allocation algorithms of proposition cloud user, be
Cloud user provides algorithm and customizes function.By changing the value of the e in formula (1-6), algorithm can change Task Assigned Policy to effect
Rate stresses ratio with service cost.For example, when cloud user only consider task processing efficiency, and do not mind processing required by task branch
During the cost paid, it is only necessary to e is set into 1, now the customized task allocation algorithm is traditional heuristic task allocation algorithms.
Opposite, if cloud user thinks to reduce cost as far as possible, its need increases e value accordingly.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
Any modifications, equivalent substitutions and improvements made within refreshing and principle etc., should be included in the scope of the protection.
Claims (6)
1. a kind of multilayer nest virtualization infrastructure, it is characterised in that the multilayer nest virtualization infrastructure includes：Two layers nested empty
Planization part and multilayer nest virtualize part.
2. multilayer nest virtualization infrastructure as claimed in claim 1, it is characterised in that in the multilayer nest virtualization infrastructure
With multiple nested Hypervisor layers；High-rise nesting virtual machine is relative to the calculating that the nested virtual machine of the second layer is additionally required
Resource definition is epsi,j(i≥3)；The number of plies of structure where wherein i represents nesting level, j represents that the virtual machine of same nesting level is compiled
Number；I-th layer of Hypervisor performance cost is also mapped into eps simultaneouslyi,jIn (i >=3)；Wherein idlei,jRepresent i-th layer of volume
Number for j virtual machine idling-resource piece.
3. the method for allocating tasks of multilayer nest virtualization infrastructure described in a kind of utilization claim 1, it is characterised in that described
Business distribution method expression formula be：
<mrow>
<munder>
<munder>
<munder>
<mi>min</mi>
<mrow>
<msub>
<mi>v</mi>
<mi>i</mi>
</msub>
<mo>&Element;</mo>
<mi>V</mi>
</mrow>
</munder>
<mrow>
<mi>P</mi>
<mo>&Element;</mo>
<msub>
<mi>HOST</mi>
<mi>k</mi>
</msub>
<mo>,</mo>
<msub>
<mi>P</mi>
<mi>j</mi>
</msub>
<mo>&Element;</mo>
<msub>
<mi>HOST</mi>
<mi>q</mi>
</msub>
</mrow>
</munder>
<mrow>
<msub>
<mi>HOST</mi>
<mi>k</mi>
</msub>
<mo>,</mo>
<msub>
<mi>HOST</mi>
<mi>q</mi>
</msub>
<mo>&Element;</mo>
<mi>H</mi>
</mrow>
</munder>
<mrow>
<mo>&lsqb;</mo>
<mrow>
<mfrac>
<mrow>
<mi>E</mi>
<mi>F</mi>
<mi>T</mi>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>v</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<mi>P</mi>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>max</mi>
<mrow>
<mo>&lsqb;</mo>
<mrow>
<mi>E</mi>
<mi>F</mi>
<mi>T</mi>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>v</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>P</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
<mo>&rsqb;</mo>
</mrow>
</mrow>
</mfrac>
<mo>*</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<mi>cos</mi>
<mi>t</mi>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>v</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<mi>P</mi>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>max</mi>
<mrow>
<mo>&lsqb;</mo>
<mrow>
<mi>cos</mi>
<mi>t</mi>
</mrow>
<mo>&rsqb;</mo>
</mrow>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
</msup>
</mrow>
<mo>&rsqb;</mo>
</mrow>
<mo>;</mo>
</mrow>
Wherein：
Set V={ v1,v2...,vk}：It is the set of a waiting task composition, the wherein number of task is k；
W=(vi)：Represent processing task viIt is expected that required processing time；
W=(vi,p)：Represent virtual machine P processing tasks viIt is expected that required processing time；
ts=(vi,p)：Represent virtual machine P processing tasks viThe time of beginning；
tf=(vi,p)：Represent virtual machine P processing tasks viThe time of end, meet below equation：
tf(vi, P) and=ts(vi, P) and+w (vi, P)；
prec(i,p)：Represent virtual machine P in processing task viBefore, the set of the task of required processing；If in processing task vi
Preceding virtual machine P does not need task to be processed, prec (i, p)；And tasks of the virtual machine P in prec (i, p) has been handled it
Before, will not start to process task vi；
Avail=(vi,p)：Virtual machine P has been handled after the task in prec (i, p), and gets out execution task viWhen
Between：
<mrow>
<mi>a</mi>
<mi>v</mi>
<mi>a</mi>
<mi>i</mi>
<mi>l</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>v</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<mi>P</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<munder>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
</mrow>
<mrow>
<msub>
<mi>v</mi>
<mi>j</mi>
</msub>
<mo>&Element;</mo>
<mi>p</mi>
<mi>r</mi>
<mi>e</mi>
<mi>c</mi>
<mrow>
<mo>(</mo>
<mrow>
<mi>i</mi>
<mo>,</mo>
<mi>P</mi>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
</munder>
<mrow>
<mo>(</mo>
<msub>
<mi>t</mi>
<mi>f</mi>
</msub>
<mo>(</mo>
<mrow>
<msub>
<mi>v</mi>
<mi>j</mi>
</msub>
<mo>,</mo>
<mi>P</mi>
</mrow>
<mo>)</mo>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
Avail=(vi, p) it is time for having performed last task in prec (i, p)：
<mrow>
<mi>a</mi>
<mi>v</mi>
<mi>a</mi>
<mi>i</mi>
<mi>l</mi>
<mo>(</mo>
<mrow>
<msub>
<mi>v</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<mi>P</mi>
</mrow>
<mo>)</mo>
<mo>=</mo>
<msub>
<mi>t</mi>
<mi>f</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>v</mi>
<mrow>
<mo>(</mo>
<mrow>
<munder>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
</mrow>
<mrow>
<msub>
<mi>v</mi>
<mi>j</mi>
</msub>
<mo>&Element;</mo>
<mi>p</mi>
<mi>r</mi>
<mi>e</mi>
<mi>c</mi>
<mrow>
<mo>(</mo>
<mrow>
<mi>i</mi>
<mo>,</mo>
<mi>P</mi>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
</munder>
<mi>j</mi>
</mrow>
<mo>)</mo>
</mrow>
</msub>
<mo>,</mo>
<mi>P</mi>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
EFT=(vi,p)：Represent virtual machine P processing tasks viThe earliest end time, it is：
EFT(vi, P) and=ts(vi, P) and+w (vi, P)；
FRT(i)：Represent performing task viBefore, the Cloud Server is expected the lease overdue time, when having performed task viAfterwards, obtain
To FRT (i) '；
time：Represent execution task viIncreased server rental period, be：
Time=(FRT (i) '-FRT (i))；
cost(vi,Pj) show execution task viThe cost of required payment, is defined as：
cost(vi, Pj)=time*mtu；
The cost rented in wherein mtu required for the Cloud Server under a certain configuration；
H={ HOST1,HOST2...,HOSTK}:Represent to have rented the set that Cloud Server is constituted.
4. a kind of cloud resource optimization method of utilization claim 1,2 and 3, it is characterised in that the cloud resource optimization side
Method includes：
(1) generation of nested virtual machine creating dispatch list
In multilayer nest virtualization, the hierachy number where highest nesting virtual machine should be the maximum for the n for meeting formula：
epsn＜ idle2,1, (n ＞ 2)；
Final remaining idling-resource is defined as：idle'1,1, the resource loss summation of multilayer nest virtualization infrastructure is obtained, and
It is defined as wstn：
<mrow>
<msub>
<mi>wst</mi>
<mi>n</mi>
</msub>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>k</mi>
</msubsup>
<msub>
<mi>idle</mi>
<mrow>
<mn>2</mn>
<mo>,</mo>
<mi>i</mi>
</mrow>
</msub>
<mo>+</mo>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>q</mi>
<mo>=</mo>
<mn>3</mn>
</mrow>
<mi>n</mi>
</msubsup>
<msub>
<mi>eps</mi>
<mi>q</mi>
</msub>
<mo>+</mo>
<msub>
<mi>idle</mi>
<mrow>
<mi>n</mi>
<mo>,</mo>
<mn>1</mn>
</mrow>
</msub>
<mo>+</mo>
<msubsup>
<mi>idle</mi>
<mrow>
<mn>1</mn>
<mo>,</mo>
<mn>1</mn>
</mrow>
<mo>&prime;</mo>
</msubsup>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>n</mi>
<mo>&GreaterEqual;</mo>
<mn>3</mn>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>k</mi>
</msubsup>
<msub>
<mi>idle</mi>
<mrow>
<mn>2</mn>
<mo>,</mo>
<mi>i</mi>
</mrow>
</msub>
<mo>+</mo>
<msubsup>
<mi>idle</mi>
<mrow>
<mn>1</mn>
<mo>,</mo>
<mn>1</mn>
</mrow>
<mo>&prime;</mo>
</msubsup>
<mo>,</mo>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>n</mi>
<mo>=</mo>
<mn>2</mn>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>;</mo>
</mrow>
Wherein n represents the nested highest number of plies, and k represents the number of the L2 layers of virtual machine of standard of two layers of nested structure part；
Obtain optimizing the nested number of plies j of nested structure by following formula, and obtain nested virtual machine creating dispatch list：
<mrow>
<munder>
<mi>min</mi>
<mrow>
<mi>n</mi>
<mo>&GreaterEqual;</mo>
<mi>j</mi>
<mo>&GreaterEqual;</mo>
<mn>2</mn>
</mrow>
</munder>
<mrow>
<mo>&lsqb;</mo>
<mrow>
<msub>
<mi>wst</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>&rsqb;</mo>
</mrow>
<mo>;</mo>
</mrow>
Wherein：vmi,j：The virtual machine that i-th layer of numbering is j is represented, Cloud Server is defined as vm as first layer virtual machinei,j；
idlei,j：Represent the idling-resource piece for the virtual machine that i-th layer of numbering is j；Wherein idle1,1Represent Cloud Server vm1,1's
Idle computing resources；
epsi,j(i≥3)：Represent that i-th layer is numbered the virtual machine for being j when performing some task；
min_resourcei：Represent that i-th layer of guest virtual machine wants minimum computing resource to handle some required by task, and
min_resourceiResource at least needed for operation VME operating system；
Rn：When representing nested n-layer, multilayer nest virtualization part takes idle1,1Calculating for carrying all nested virtual machines
Resource part；
(2) generation of schedule of tasks：
<mrow>
<munder>
<munder>
<munder>
<mi>min</mi>
<mrow>
<msub>
<mi>v</mi>
<mi>i</mi>
</msub>
<mo>&Element;</mo>
<mi>V</mi>
</mrow>
</munder>
<mrow>
<mi>P</mi>
<mo>&Element;</mo>
<msub>
<mi>HOST</mi>
<mi>k</mi>
</msub>
<mo>,</mo>
<msub>
<mi>P</mi>
<mi>j</mi>
</msub>
<mo>&Element;</mo>
<msub>
<mi>HOST</mi>
<mi>q</mi>
</msub>
</mrow>
</munder>
<mrow>
<msub>
<mi>HOST</mi>
<mi>k</mi>
</msub>
<mo>,</mo>
<msub>
<mi>HOST</mi>
<mi>q</mi>
</msub>
<mo>&Element;</mo>
<mi>H</mi>
</mrow>
</munder>
<mrow>
<mo>&lsqb;</mo>
<mrow>
<mfrac>
<mrow>
<mi>E</mi>
<mi>F</mi>
<mi>T</mi>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>v</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<mi>P</mi>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>max</mi>
<mrow>
<mo>&lsqb;</mo>
<mrow>
<mi>E</mi>
<mi>F</mi>
<mi>T</mi>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>v</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<msub>
<mi>P</mi>
<mi>j</mi>
</msub>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
<mo>&rsqb;</mo>
</mrow>
</mrow>
</mfrac>
<mo>*</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<mi>cos</mi>
<mi>t</mi>
<mrow>
<mo>(</mo>
<mrow>
<msub>
<mi>v</mi>
<mi>i</mi>
</msub>
<mo>,</mo>
<mi>P</mi>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>max</mi>
<mrow>
<mo>&lsqb;</mo>
<mrow>
<mi>cos</mi>
<mi>t</mi>
</mrow>
<mo>&rsqb;</mo>
</mrow>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
</msup>
</mrow>
<mo>&rsqb;</mo>
</mrow>
<mo>;</mo>
</mrow>
In task viDuring being allocated, under task scheduling figure at that time, its max [cost] and max [EFT (vi,p)]
It is fixed, finds the minimum value of formula, its corresponding virtual machine P is exactly to be best suitable for execution task viVirtual machine.As task vi
For last task to be allocated of this batch of task, that is, obtain the schedule of tasks of this batch of task.
5. cloud resource optimization method as claimed in claim 4, it is characterised in that：
(1) generation of the virtual machine creating dispatch list includes：
First, the computing resource to rented Cloud Server is counted and analyzes the characteristic of waiting task；And according to collected
Information, calculate the estimated idling-resource amount of Cloud Server；
After the estimated resource free quantity of Cloud Server is obtained, according to the nested virtualization structure relied on, the nested virtual machine of generation
Create dispatch list.
(2) generation of the schedule of tasks includes：
First, all workable task processing virtual machines are traveled through；And count the task that existing task handles virtual machine
Disposition.
After the task disposition of task processing virtual machine is obtained, task is calculated according to institute's dependence task allocation algorithm and adjusted
Spend table.
6. the Cloud Server of cloud resource optimization method described in a kind of application claim 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710405158.0A CN107273181B (en) | 2017-05-31 | 2017-05-31 | Multilayer nested virtualization structure and task allocation method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710405158.0A CN107273181B (en) | 2017-05-31 | 2017-05-31 | Multilayer nested virtualization structure and task allocation method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107273181A true CN107273181A (en) | 2017-10-20 |
| CN107273181B CN107273181B (en) | 2021-01-22 |
Family
ID=60065359
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710405158.0A Active CN107273181B (en) | 2017-05-31 | 2017-05-31 | Multilayer nested virtualization structure and task allocation method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107273181B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102103517A (en) * | 2009-12-22 | 2011-06-22 | 英特尔公司 | Efficient nested virtualization |
| US20120131574A1 (en) * | 2010-11-23 | 2012-05-24 | Day Ii Michael D | Virtual machine data structures corresponding to nested virtualization levels |
| US20140053272A1 (en) * | 2012-08-20 | 2014-02-20 | Sandor Lukacs | Multilevel Introspection of Nested Virtual Machines |
| GB2502225B (en) * | 2011-02-17 | 2014-07-23 | Ibm | Multilevel support in a nested virtualization environment |
| CN105653372A (en) * | 2015-12-30 | 2016-06-08 | 中电科华云信息技术有限公司 | Cloud platform-based method for realizing multi-virtualization hybrid management and scheduling |
| US20160359889A1 (en) * | 2015-06-05 | 2016-12-08 | Cisco Technology, Inc. | Technologies for managing compromised sensors in virtualized environments |
| CN106648890A (en) * | 2016-12-06 | 2017-05-10 | 中国科学院重庆绿色智能技术研究院 | Cloud computing server resource on-line management method and system with energy consumption sensing function |
-
2017
- 2017-05-31 CN CN201710405158.0A patent/CN107273181B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102103517A (en) * | 2009-12-22 | 2011-06-22 | 英特尔公司 | Efficient nested virtualization |
| US20120131574A1 (en) * | 2010-11-23 | 2012-05-24 | Day Ii Michael D | Virtual machine data structures corresponding to nested virtualization levels |
| GB2502225B (en) * | 2011-02-17 | 2014-07-23 | Ibm | Multilevel support in a nested virtualization environment |
| US20140053272A1 (en) * | 2012-08-20 | 2014-02-20 | Sandor Lukacs | Multilevel Introspection of Nested Virtual Machines |
| US20160359889A1 (en) * | 2015-06-05 | 2016-12-08 | Cisco Technology, Inc. | Technologies for managing compromised sensors in virtualized environments |
| CN105653372A (en) * | 2015-12-30 | 2016-06-08 | 中电科华云信息技术有限公司 | Cloud platform-based method for realizing multi-virtualization hybrid management and scheduling |
| CN106648890A (en) * | 2016-12-06 | 2017-05-10 | 中国科学院重庆绿色智能技术研究院 | Cloud computing server resource on-line management method and system with energy consumption sensing function |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107273181B (en) | 2021-01-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104881325B (en) | A kind of resource regulating method and resource scheduling system | |
| Singh et al. | Task scheduling in cloud computing | |
| CN107491341B (en) | A kind of virtual machine distribution method based on particle group optimizing | |
| Xue et al. | An ACO-LB Algorithm for Task Scheduling in the Cloud Environment. | |
| Dhari et al. | An efficient load balancing scheme for cloud computing | |
| CN108182109B (en) | Workflow scheduling and data distribution method in cloud environment | |
| CN104331321B (en) | Cloud computing task scheduling method based on tabu search and load balancing | |
| CN103713956B (en) | Method for intelligent weighing load balance in cloud computing virtualized management environment | |
| CN103297499B (en) | Scheduling method and system based on cloud platform | |
| CN108182115A (en) | A kind of virtual machine load-balancing method under cloud environment | |
| US8527988B1 (en) | Proximity mapping of virtual-machine threads to processors | |
| CN103401939A (en) | Load balancing method adopting mixing scheduling strategy | |
| CN104023042B (en) | Cloud platform resource scheduling method | |
| CN105373426B (en) | A kind of car networking memory aware real time job dispatching method based on Hadoop | |
| CN104881322A (en) | Method and device for dispatching cluster resource based on packing model | |
| CN109240795A (en) | A kind of resource regulating method of the cloud computing resources pool model suitable for super fusion IT infrastructure | |
| CN105005503B (en) | Cloud computing load balancing method for scheduling task based on cellular automata | |
| Kaur et al. | Improved max-min scheduling algorithm | |
| CN107070965B (en) | Multi-workflow resource supply method under virtualized container resource | |
| CN104156505B (en) | A kind of Hadoop cluster job scheduling method and devices based on user behavior analysis | |
| CN110519386A (en) | Elastic resource supply method and device based on data clustering in cloud environment | |
| Gupta et al. | A theoretical comparison of job scheduling algorithms in cloud computing environment | |
| Dubey et al. | QoS driven task scheduling in cloud computing | |
| CN110958192B (en) | Virtual data center resource allocation system and method based on virtual switch | |
| CN105430074A (en) | Data dependency and access traffic based cloud data allocation storage optimization method and system |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |