CN104484233B - A kind of resource allocation methods - Google Patents
A kind of resource allocation methods Download PDFInfo
- Publication number
- CN104484233B CN104484233B CN201410602590.5A CN201410602590A CN104484233B CN 104484233 B CN104484233 B CN 104484233B CN 201410602590 A CN201410602590 A CN 201410602590A CN 104484233 B CN104484233 B CN 104484233B
- Authority
- CN
- China
- Prior art keywords
- resource
- tasks
- parallel processing
- redundancy
- alternative
- 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.)
- Active
Links
Landscapes
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses a kind of resource allocation methods.Firstly generate multiple alternative tasks in parallel processing schemes corresponding with pending task, it is determined that after the reference runnability of each alternative tasks in parallel processing scheme, feasible tasks in parallel processing scheme is screened with reference to the current resource parameters of default requirement of real-time and base station resource pond, then it is one or more physical implementations each feasible tasks in parallel processing scheme filtered out to be distinguished into correspondence mappings, and the resource and multiple attribute decision making (MADM) strategy that are consumed according to each physical implementation determine optimal physical implementation.So that final Resource Allocation Formula matches with current environment, ensure that scheme is that resource pool current available resource can be supported and can meet requirement of real-time.
Description
Technical field
The present invention relates to communication technical field, more particularly to a kind of resource allocation methods.
Background technology
Traditional communication equipment is often that a manufacturer provides a whole set of solution, system maintenance or upgrading dependence
It is high.And as energy resources anxiety, global mobile communication network operator faced increasingly serious cost pressure in recent years.Mostly
Number mainstream carrier generally possesses the network of multiple different communication standards, to ensure the service quality of network, it is necessary to dispose a large amount of
Base station to solve the problems, such as the network coverage.But site and the relative scarcity of machine room resource, the demand shape with a large amount of base station deployments
Into the contradiction for being difficult to coordinate.And due to the keen competition of Mobile Communications Market, single user average income increasess slowly or even declined,
The " profit " ability of operator does not improve therewith, and this will cause the compression of networking and equipment purchase investment.Continue for industry
Profit and long-run development consider that Mobile Communication Industry circle proposes solve this problem by changing the access network architecture.
However, inventor is realizing that equipment is found during being multiplexed by changing the access network architecture, existing base
Stand resource pool platform architecture generally use general processor, multinomial processing task is scheduled by operating system.With existing void
Resource allocation visual angle in planization resource scheduling system is different, and processing/computing resource, storage are not only related in base station resource pond
The distribution of the multi dimensional resources such as resource, it is also necessary to consider the requirement of real-time of Base-Band Processing.To ensure requirement of real-time, due to interface
The traffic is larger, and the scheduling of resource between server is limited.And conventional baseband solution is different from, base station resource pond should
Suitable physical resource is selected to complete Base-Band Processing process for Current resource pool environment.Therefore, although current currently existing scheme
In have been directed to the dispatching distribution of resource pool multi dimensional resource, but still there are the following problems:It is not known in implementation method and considers real-time
It is required that the resource of resource pool multiserver is considered as equal resource, it is that the scheduling of resource between multiserver and server is not domestic-investment
Source scheduling is distinguished, and can be related to compared with multi-interface communication amount yet with the scheduling of resource between server so as to influence the reality of task processing
Shi Xing.In addition, existing method generally use fix baseband processing arrangements, consider scheduling of resource how with current environment phase
Adapt to.
The content of the invention
The invention provides a kind of resource allocation methods, to solve in the prior art, resource distribution mode real-time is not
Height, inflexible problem, this method include:
Multiple alternative tasks in parallel processing schemes corresponding with pending task are generated, at the alternative tasks in parallel
Reason scheme is the set of parallel processing plan corresponding with the pending task;
The reference runnability of each alternative tasks in parallel processing scheme is determined, the reference runnability is described
Base station resource pond individually performs the runnability of each alternative tasks in parallel processing scheme;
According to the current resource parameters of default requirement of real-time, the base station resource pond and each alternative task simultaneously
The reference runnability of rowization processing scheme, screened from the alternative tasks in parallel processing scheme at feasible tasks in parallel
Reason scheme;The resource parameters include current available resources, run time amount of redundancy and resource consumption amount of redundancy, the fortune
Row time redundancy amount is additionally required by the base station resource pond when performing each alternative tasks in parallel processing scheme to disappear
The percentage of the run time of consumption, the resource consumption amount of redundancy are that the base station resource pond is performing each alternative task simultaneously
The percentage of the resource of consumption is additionally required during rowization processing scheme;
It is one or more physics realization sides by each feasible tasks in parallel processing scheme difference correspondence mappings filtered out
Case, physical implementation are that feasible tasks in parallel processing scheme is being distributed to the currently available physics of the base station resource pond
The set formed after resource;
The resource and multiple attribute decision making (MADM) strategy consumed according to each physical implementation determines that optimal physical is realized
Scheme, and carry out resource allocation according to the optimal physical implementation.
As can be seen here, by applying above technical scheme, multiple alternative tasks corresponding with pending task are firstly generated
Parallelization processing scheme, it is determined that after reference the runnability of each alternative tasks in parallel processing scheme, with reference to default real-time
Property require and the current resource parameters of base station resource pond screen feasible tasks in parallel processing scheme, then by filter out it is each can
Row tasks in parallel processing scheme difference correspondence mappings are one or more physical implementations, and according to each physics realization
The resource and multiple attribute decision making (MADM) strategy that scheme is consumed determine optimal physical implementation.So that final Resource Allocation Formula
Matched with current environment, ensure that scheme is that resource pool current available resource can be supported and can meet requirement of real-time.
Brief description of the drawings
Fig. 1 is a kind of schematic flow sheet of resource allocation methods proposed by the present invention.
Embodiment
In order to solve the problems, such as proposed in background technology, the present invention proposes a kind of physical layer resources of base station resource pond
Distribution method, parallelization processing is carried out to physical layer base band calculating task first, form a variety of base band Parallelization Scheme compositions
Scheme collection merges statistical-reference runnability, then according to requirement of real-time and Current resource pool environment from alternative set
It is middle to determine feasible alternative collection, the scheme that alternative is concentrated finally is mapped as physical implementation, forms implementation
Set, then the implementation matched based on multiple attributive decision making method selection with resource pool current environment.
As shown in figure 1, be a kind of physical layer resources distribution method schematic flow sheet of base station resource pond proposed by the present invention,
This method is applied in base station resource pond, and the base station resource pond is connected as BBU with the RRU by optical transport network, bag
Include following steps:
S101, the multiple alternative tasks in parallel processing schemes corresponding with pending task of generation, the alternative task is simultaneously
Rowization processing scheme is the set of parallel processing plan corresponding with the pending task.
For pending task, the present invention carries out parallelization processing to physical layer base band calculating task first, is formed a variety of
The scheme collection of base band Parallelization Scheme composition merges statistical-reference runnability.Specifically, this step by tasks in parallel by institute
It is each parallel processing plan to state pending Task-decomposing processing, using the set of each parallel processing plan as described alternative
Business parallelization processing scheme.
S102, the reference runnability of each alternative tasks in parallel processing scheme is determined, it is described to refer to runnability
The runnability of each alternative tasks in parallel processing scheme is individually performed for the base station resource pond.
S103, according to current resource parameters of default requirement of real-time, the base station resource pond and each described alternative
The reference runnability of tasks in parallel processing scheme, feasible task is screened simultaneously from the alternative tasks in parallel processing scheme
Rowization processing scheme.
Based on S101 and S102 result, the present invention is according to resource pool current idle disposal ability, idling-resource
Position distribution, the situation of task scheduling migration the amount of redundancy of run time and resource consumption is set, and according to reference to maneuverability
Alternative feasible under the present circumstances can be selected from alternative set with amount of redundancy.Wherein, the resource parameters bag
Current available resources, run time amount of redundancy and resource consumption amount of redundancy are included, the run time amount of redundancy is the base
Resource pool of standing is additionally required the percentage of the run time of consumption when performing each alternative tasks in parallel processing scheme,
The institute when performing each alternatively tasks in parallel processing scheme is extra for the base station resource pond for the resource consumption amount of redundancy
Need the percentage of resource consumed.
In addition, before this step, the present invention will also be according to the current ambient parameter setting of the base station resource pond
Run time amount of redundancy and the resource consumption amount of redundancy, the ambient parameter include the base station resource pond current free time
Disposal ability, idling-resource position distribution, task scheduling migration situation, it is specific to set rule as follows:
If the free processing capacity is larger, less run time amount of redundancy and resource consumption amount of redundancy are set;
If the scheduling of the pending task need to perform according to load balancing, less run time amount of redundancy is set;
If the task scheduling migration situation is more frequent, less run time amount of redundancy is set, and set compared with
Big resource consumption amount of redundancy;
If the idling-resource position distribution is more concentrated, less run time amount of redundancy is set;
If the idling-resource position distribution is more scattered, larger run time amount of redundancy is set.
S104, it is that one or more physics are real by each feasible tasks in parallel processing scheme difference correspondence mappings filtered out
Existing scheme, physical implementation are feasible tasks in parallel processing scheme to be distributed to the base station resource pond currently available
The set formed after physical resource.
Result based on S103, this step obtain the reference of alternative tasks in parallel processing scheme to be screened first
Reference run time and reference resources consumption included in runnability;If the reference run time is according to the fortune
Still less than the requirement of real-time after row time redundancy amount increase, and reference resources consumption is superfluous according to the resource consumption
Still less than the available resources after surplus increase, then distributed the alternative resource allocative decision as the feasible alternative resource
Scheme.
S105, the resource and multiple attribute decision making (MADM) strategy consumed according to each physical implementation determine optimal physical
Implementation, and carry out resource allocation according to the optimal physical implementation.
After S101-S104 is finished, the scheme that alternative is concentrated is mapped as physical implementation, shape by the present invention
Into implementation set, then the implementation matched based on multiple attributive decision making method selection with resource pool current environment.Specifically,
This step can disappear the idle available processes resource after performing each physical implementation, free time with storage resource and scheme
Input attribute of the interface resource of consumption as scheme, each physical implementation is determined according to the multiple attribute decision making (MADM) strategy
Synthesized attribute value, and using synthesized attribute value highest physical implementation as the optimal physical implementation.
It is pointed out that in order to reduce the processing load of this step, before this step, can be according to default multiple
Interface resource consumption level, a pair physical implementation corresponding with each feasible tasks in parallel processing scheme screen,
Retain the physical implementation of interface resource consumption level minimum in each feasible tasks in parallel processing scheme.
In order to which the technological thought of the present invention is expanded on further, in conjunction with specific application scenarios, to the technical side of the present invention
Case illustrates.The specific embodiment is divided into following three steps:
Step 1:Parallelization processing is carried out to physical layer base band calculating task based on load balancing principle, forms a variety of base band
The scheme collection of Parallelization Scheme composition merges statistical-reference runnability
The physical layer base band signal of wireless communication system is handled by multiple calculating tasks or Algorithm constitution, wherein including some
Computation-intensive algorithm, such as signal detection, channel decoding, the computationally intensive of them, complexity are high.Physical layer algorithm it is another
Feature is that have clear and definite order execution relation between calculating task, i.e., the output of one calculating task will be used as next calculating task
Input.If not carrying out parallelization processing to computation-intensive algorithm, it causes as the bottleneck of physical layer Base-Band Processing
Calculating task elapsed time is excessive, can not meet real-time demand.
Therefore it is to meet system real time requirement, it is necessary to which by algorithm, especially computation-intensive algorithm is split with reality
Existing parallelization processing.Because a certain calculating task or algorithm there can be a variety of parallelization processing modes, it is made up of multiple calculating tasks
Physical layer baseband processing arrangements also have a variety of parallel structures.Using different parallel structures design scheme its consumption it is each
Class resource is different, such as computing resource, storage resource, interface resource, and the time of operation is also different.The physics in base station resource pond
The selection of layered scheme needs to consider the factors such as requirement of real-time, current idle resource situation to adapt to resource pool environmental change, institute
It is alternative so that kinds of schemes formation scheme set can be provided.
Step 1.1 carries out paralell design to physical layer Base-Band Processing, forms alternative set
, can to ensure that polycaryon processor is fully used because base station resource pond often uses general purpose processor platform
To consider to carry out Task-decomposing to calculating task especially computation-intensive task based on the principle of load balancing, pass through flowing water
Line, the mode of tasks in parallel design a variety of parallel processing plans.These parallel processing plans constitute alternative set.
Step 1.2 obtains the reference runnability of alternative
When referring to reference to runnability without other baseband processing arrangements in base station resource pond, individually perform at a certain base band
The runnability of reason scheme.The business of Base-Band Processing is different according to as defined in agreement, and the data volume of respective handling is also different.For
Different pieces of information amount, some scheme, which also has, a variety of refers to runnability.
Include run time, all kinds of resource situations of consumption of each scheme with reference to runnability.All kinds of moneys wherein consumed
Source refers to the logical resource consumed when calculating task performs, including processing/computing resource, storage resource.
Can online under first carry out alternative in advance and refer to runnability to obtain.According to descending with reference to run time
By ranking alternatives, and record other runnabilities.Following table provides record alternative and its example with reference to runnability.
Table 1 records alternative and its example with reference to runnability
Step 2:Feasible alternative side is determined from alternative set according to requirement of real-time and Current resource pool environment
Case collection
There are a large amount of cells, the data of user handling in base station resource pond, when the data of certain user reach base station resource pond
When, used baseband processing arrangements should be selected according to Current resource pool environment.And the operation that the scheme chosen is consumed
Time, all kinds of resource situations are because there is the reference performance counted under the influence and line that other schemes, data processing are brought difference.
Therefore can not be directly by the foundation with reference to runnability alternatively.Because carrying out practically performance is by resource pool environment shadow
Ring, it is necessary to consider current environment influence factor.
Step 2.1 sets the amount of redundancy of run time and resource consumption according to base station resource pond current environment
The base station resource pond current environment factor for needing to consider sets the amount of redundancy of run time and resource consumption, including:
Current idle disposal ability, the position distribution of idling-resource, the situation of task scheduling migration.Wherein current idle disposal ability can
To be represented by current open ended portfolio and the ratio of the open ended total traffic of resource pool.Task scheduling, migration are respectively
Refer to pending calculating task to be assigned to processor and carry out calculating task to physical resource according to resource pool current environment
Remap.The position distribution of idling-resource can be concentrated/degree of scatter measurement by idling-resource.
Represent run time and resource consumption amount of redundancy respectively with percent value a%, b%.Amount of redundancy is worked as according to resource pool
Preceding environmental factor and experience are configured.If current idle disposal ability is larger, can be set less run time and
Resource consumption amount of redundancy.If the scheduling reference fairness of task, load balancing principle perform, when should set less operation
Between amount of redundancy.It is typically due to when run time is unable to reach requirement of real-time or more idle resource, it is necessary to carry out task immigration.
If task immigration is more frequent, less run time amount of redundancy and larger resource consumption amount of redundancy should be set.If
There is more idling-resource integrated distribution, then less run time amount of redundancy can be set, if idling-resource distribution is scattered, it is meant that
Scattered resource coordinating is worked the interface communication amount for expending more, cause real-time to be affected, then need it is settable compared with
Big run time amount of redundancy.
Step 2.2. is feasible under the present circumstances according to being selected with reference to runnability and amount of redundancy from alternative set
Alternative
The resource that feasible alternative needs should be that resource pool current available resource can be supported and can meet real-time
It is required that.Therefore the selection of feasible alternative needs to consider current environment.
The reference run time combination amount of redundancy of each scheme in alternative set is compared with requirement of real-time first
Compared with selecting the scheme in accordance with requirement of real-time.Judgement mode is as follows:With reference to run time * (1+a%)<Requirement of real-time, then sentence
For feasible alternative.
Then the reference resources consumption combination resource of each scheme in the alternative set for meeting requirement of real-time is disappeared
Amount of redundancy is consumed compared with current available resource, and selection meets the scheme of available resources support.Judgement mode is as follows:With reference to money
Source consumption * (1+b%)<Current available resource, then it is judged to feasible alternative.Wherein processing/computing resource, storage resource need
Calculate respectively.
Finally all qualified feasible alternatives form feasible alternative collection.
Step 3:Alternative is mapped as physical implementation, implementation set is formed, based on multiple attribute decision making (MADM) side
The implementation that method selection matches with resource pool current environment
This step will select final implementation from feasible alternative and be mapped on physical resource.Physical resource
Processing/computing resource, storage resource including reality, interface resource.The run time and the resource of consumption of usual scheme are lances
Shield.Selection scheme should make the resource of consumption as few as possible when mapping in the case where meeting requirement of real-time.
Alternative is mapped as physical implementation by step 3.1, forms implementation set
Alternative is mapped as into physical implementation to refer to consume resource according to the alternative reference of first step acquisition,
The calculating task of alternative is distributed in currently available physical resource.Logical resource is only accounted for reference to consumption resource,
That is the consumption of processing/computing resource, storage resource, consideration processing/computing resource, storage are needed when carrying out physical resource mapping
Distribution of the resource in resource pool.If alternative is cooperateed with using multiple processors completes a certain calculating task, and multiple processing
Device distribution farther out (as being distributed on different server) if need to consume more interface communication visit capacities.Therefore physics is being carried out
During resource impact, it is also necessary to consider the size of the consumption of interface resource, i.e. interface communication amount.Processing/computing resource, storage resource
A variety of situations are distributed with resource pool, even if same scheme consumption identical processing/computing resource, storage resource, corresponding
Interface resource may also be different.Therefore an alternative can correspond to multiple physical implementations.
The state distribution that resource pool tackles wherein each physical resource is monitored, and understands available processing/calculating money in real time
Source, the position of storage resource.When carrying out physical resource mapping, the logical resource lookup selection needed according to alternative is available
Physical resource, and the interface resource for completing project task is counted, form physical implementation set corresponding to alternative.Following table
Give the example of physical implementation consumption resource.
The example of physical implementation consumption resource corresponding to the alternative of table 2
When the idling-resource of resource pool is more, available physical resources and its position distribution situation are more, cause alternative side
Physical implementation is more corresponding to case, judges overlong time.Therefore the physical resource of resource pool can be distributed to consume
Interface resource divided rank.The interface communication amount consumed during two physical resource co-ordinations is larger, then they belong to higher
Interface resource grade.Answer first selection processing/computing resource, storage money in the physical resource of relatively low same interface resource grade
Source.If the physical resource of inferior grade can not meet to require, available resources are searched into the physical resource of greater degree.Such as
Because base station resource pond generally comprises multiple servers, the processor on different server cooperates, and can consume larger amount of
The indirect port communications of server.Therefore the process resource of different server belongs to the physical resource of higher interface resource grade, and same
Process resource in one server belongs to the physical resource of relatively low interface resource grade.It can be searched first in server available
Resource, if not enough, re-mapped in the resource between server.All possible implementation need not be enumerated.
The implementation that step 3.2 is matched based on multiple attributive decision making method selection with resource pool current environment
Using the implementation of the selection matching current environment from feasible implementation set as Multiple Attribute Decision Problems.
The scheme for wishing to choose consumes less interface resource in the case where reaching requirement of real-time and makes the processing money of remaining free time
Source, storage resource are more.
If the implementation obtained by previous step integrates as X=(x1,x2...xn), wherein xiRepresent i-th of implementation, i
=1,2...n, n are implementation number.Property set P=(p1,p2,p3), p1Refer to idle available processing after using the program to provide
Source, the reference that the program is subtracted by Current resource pond available processes stock number consume the acquisition of process resource amount, p2Finger uses the party
Idle available storage resource amount after case, the reference consumption storage money of the program can be subtracted with storage resource amount by Current resource pond
Source amount obtains, p3Refer to the interface resource amount of consumption.p1,p2It is profit evaluation model attribute, p3It is cost type attribute.
For a certain implementation xi, by attribute pjEstimated, obtain xiOn pjProperty value αij, form decision-making square
Battle array Γ=(αij)n×3, it is as follows:
p1 | p2 | p3 | |
x1 | α11 | α12 | α13 |
x2 | α21 | α22 | α23 |
… | |||
xn | αn1 | αn2 | αn3 |
Above-mentioned matrix is standardized to obtain Standard Process R=(rij)n×m, wherein for p1,p2Property value αij,j
∈ 1,2, orderFor p3Property value αi3, orderJ=3.
Normalization matrix is further obtained,Wherein
Calculate pjThe comentropy of output,J=1,2,3.
Calculate weight vectors, β=(β1,β2,β3), wherein
FoundationNumerical procedure xiSynthesized attribute value zi(β) (i=1,2 ..., n).Utilize zi(β) is to scheme
It is ranked up and chooses maximum as final implementation.
Comentropy is employed in previous step as weight, can also rule of thumb set weight beta=(β1,β2,β3), makeZ is calculated according to thisi(β) chooses final implementation.
By applying above technical scheme, multiple alternative tasks in parallelization processing corresponding with pending task are firstly generated
Scheme, it is determined that after the reference runnability of each alternative tasks in parallel processing scheme, with reference to default requirement of real-time and base
The current resource parameters of resource pool of standing screen feasible tasks in parallel processing scheme, each feasible tasks in parallel that then will be filtered out
It is one or more physical implementations to change processing scheme difference correspondence mappings, and is consumed according to each physical implementation
Resource and multiple attribute decision making (MADM) strategy determine optimal physical implementation.So that final Resource Allocation Formula and current environment
Matching, ensure that scheme is that resource pool current available resource can be supported and can meet requirement of real-time.
Through the above description of the embodiments, those skilled in the art can be understood that the present invention can lead to
Hardware realization is crossed, the mode of necessary general hardware platform can also be added by software to realize.Based on such understanding, this hair
Bright technical scheme can be embodied in the form of software product, and the software product can be stored in a non-volatile memories
In medium (can be CD-ROM, USB flash disk, mobile hard disk etc.), including some instructions are causing a computer equipment (can be
Personal computer, server, or network equipment etc.) perform method described in each implement scene of the present invention.
It will be appreciated by those skilled in the art that accompanying drawing is a schematic diagram for being preferable to carry out scene, module in accompanying drawing or
Flow is not necessarily implemented necessary to the present invention.
It will be appreciated by those skilled in the art that the module in device in implement scene can be described according to implement scene into
Row is distributed in the device of implement scene, can also carry out one or more dresses that respective change is disposed other than this implement scene
In putting.The module of above-mentioned implement scene can be merged into a module, can also be further split into multiple submodule.
The invention described above sequence number is for illustration only, does not represent the quality of implement scene.
Disclosed above is only several specific implementation scenes of the present invention, and still, the present invention is not limited to this, Ren Heben
What the technical staff in field can think change should all fall into protection scope of the present invention.
Claims (6)
1. a kind of resource allocation methods, it is characterised in that methods described is applied in base station resource pond, and the base station resource pond is made
It is connected for BBU with RRU by optical transport network, this method includes:
Generate multiple alternative tasks in parallel processing schemes corresponding with pending task, the alternative tasks in parallelization processing side
Case is the set of parallel processing plan corresponding with the pending task;
The reference runnability of each alternative tasks in parallel processing scheme is determined, the reference runnability is the base station
Resource pool individually performs the runnability of each alternative tasks in parallel processing scheme;
According to the current resource parameters of default requirement of real-time, the base station resource pond and each alternative tasks in parallel
The reference runnability of processing scheme, feasible tasks in parallelization processing side is screened from the alternative tasks in parallel processing scheme
Case;The resource parameters include current available resources, run time amount of redundancy and resource consumption amount of redundancy, during the operation
Between amount of redundancy be additionally required consumption when performing each alternative tasks in parallel processing scheme by the base station resource pond
The percentage of run time, the resource consumption amount of redundancy are that the base station resource pond is performing each alternative tasks in parallel
The percentage of the resource of consumption is additionally required during processing scheme;
It is one or more physical implementations by each feasible tasks in parallel processing scheme difference correspondence mappings filtered out, thing
Reason implementation is that feasible tasks in parallel processing scheme is being distributed to the currently available physical resource of the base station resource pond
The set formed afterwards;
The resource and multiple attribute decision making (MADM) strategy consumed according to each physical implementation determines optimal physical implementation,
And carry out resource allocation according to the optimal physical implementation.
2. the method as described in claim 1, it is characterised in that the multiple alternative tasks in parallel corresponding with pending task of generation
Change processing scheme, be specially:
It is each parallel processing plan to be handled the pending Task-decomposing by tasks in parallel, by each parallel processing plan
Set as the alternative tasks in parallel processing scheme.
3. the method as described in claim 1, it is characterised in that according to default requirement of real-time, the base station resource pond
Current resource parameters and reference the runnability of each alternative tasks in parallel processing scheme from the alternatively task simultaneously
Before feasible tasks in parallel processing scheme is screened in rowization processing scheme, in addition to:
Set the run time amount of redundancy and the resource consumption superfluous according to the current ambient parameter of the base station resource pond
Surplus, the ambient parameter include the current free processing capacity of the base station resource pond, idling-resource position distribution, task and adjusted
Spend migration situation;Wherein,
If the free processing capacity is larger, less run time amount of redundancy and resource consumption amount of redundancy are set;
If the scheduling of the pending task need to perform according to load balancing, less run time amount of redundancy is set;
If the task scheduling migration situation is more frequent, less run time amount of redundancy is set, and sets larger
Resource consumption amount of redundancy;
If the idling-resource position distribution is more concentrated, less run time amount of redundancy is set;
If the idling-resource position distribution is more scattered, larger run time amount of redundancy is set.
4. method as claimed in claim 3, it is characterised in that worked as according to default requirement of real-time, the base station resource pond
Preceding resource parameters and the reference runnability of each alternative resource allocative decision are carried out at feasible alternative tasks in parallel
The screening of reason scheme, it is specially:
Obtain reference run time included in the reference runnability of alternative tasks in parallel processing scheme to be screened with
And reference resources consumption;
If it is described with reference to run time after according to the run time amount of redundancy increase still less than the requirement of real-time, and institute
Reference resources consumption is stated after according to the resource consumption amount of redundancy increase still less than the available resources, then by the alternative money
Source allocative decision is as the feasible alternative resource allocative decision.
5. the method as described in claim 1, it is characterised in that the resource includes process resource, storage resource and interface
Resource, the resource and multiple attribute decision making (MADM) strategy consumed according to each physical implementation determine optimal physical realization side
Case, it is specially:
The idle available processes resource after each physical implementation will be performed, the free time can be disappeared with storage resource and the program
Input attribute of the interface resource of consumption as scheme, each physical implementation is determined according to the multiple attribute decision making (MADM) strategy
Synthesized attribute value, and using synthesized attribute value highest physical implementation as the optimal physical implementation.
6. method as claimed in claim 5, it is characterised in that the resource consumed according to each physical implementation with
And before multiple attribute decision making (MADM) strategy determines optimal physical implementation, in addition to:
According to default multiple interface resource consumption levels, pair physics corresponding with each feasible tasks in parallel processing scheme
Implementation is screened, and retains the thing of interface resource consumption level minimum in each feasible tasks in parallel processing scheme
Manage implementation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410602590.5A CN104484233B (en) | 2014-10-31 | 2014-10-31 | A kind of resource allocation methods |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410602590.5A CN104484233B (en) | 2014-10-31 | 2014-10-31 | A kind of resource allocation methods |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104484233A CN104484233A (en) | 2015-04-01 |
CN104484233B true CN104484233B (en) | 2017-12-05 |
Family
ID=52758777
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410602590.5A Active CN104484233B (en) | 2014-10-31 | 2014-10-31 | A kind of resource allocation methods |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104484233B (en) |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106155807A (en) * | 2015-04-15 | 2016-11-23 | 阿里巴巴集团控股有限公司 | A kind of method and apparatus realizing scheduling of resource |
CN106155977B (en) * | 2015-06-03 | 2019-07-30 | 上海红神信息技术有限公司 | Restructural method for managing resource and system in reconfigurable system |
CN112449021B (en) * | 2016-01-27 | 2023-10-27 | 创新先进技术有限公司 | Internet resource screening method and device |
CN107239337A (en) * | 2016-03-28 | 2017-10-10 | 北京智梵网络科技有限公司 | The distribution of virtual resources and dispatching method and system |
CN107025137B (en) * | 2016-11-24 | 2020-11-17 | 创新先进技术有限公司 | Resource query method and device |
CN108628667A (en) * | 2017-03-16 | 2018-10-09 | 哈尔滨英赛克信息技术有限公司 | A kind of heuristic dynamic task scheduling system and its implementation based on multiple attribute decision making (MADM) |
CN109831321B (en) * | 2019-01-04 | 2021-02-09 | 东南大学 | RDMA application transmission parameter self-adaptive selection method in data center |
CN113992666B (en) * | 2021-09-18 | 2023-12-29 | 阿里巴巴(中国)有限公司 | Cloud resource allocation method, system, equipment and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101119314A (en) * | 2007-09-14 | 2008-02-06 | 中国科学院计算技术研究所 | Multimode terminal service stream control system and device and method |
CN102780765A (en) * | 2012-06-27 | 2012-11-14 | 浙江大学 | Cloud manufacturing service resource match and combination method based on performance fusion |
CN103841647A (en) * | 2014-03-07 | 2014-06-04 | 北京邮电大学 | Multi-granularity telescopic packaging device and method based on base station resource pool |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8416746B2 (en) * | 2008-07-07 | 2013-04-09 | Apple Inc. | Medium access control for wireless systems |
-
2014
- 2014-10-31 CN CN201410602590.5A patent/CN104484233B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101119314A (en) * | 2007-09-14 | 2008-02-06 | 中国科学院计算技术研究所 | Multimode terminal service stream control system and device and method |
CN102780765A (en) * | 2012-06-27 | 2012-11-14 | 浙江大学 | Cloud manufacturing service resource match and combination method based on performance fusion |
CN103841647A (en) * | 2014-03-07 | 2014-06-04 | 北京邮电大学 | Multi-granularity telescopic packaging device and method based on base station resource pool |
Non-Patent Citations (1)
Title |
---|
未来移动通信基站体系结构-定性理论、方法与实践;王勇;《中国博士学位论文全文数据库 信息科技辑》;20101115;I136-111 * |
Also Published As
Publication number | Publication date |
---|---|
CN104484233A (en) | 2015-04-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104484233B (en) | A kind of resource allocation methods | |
US20200167184A1 (en) | System and Method for Optimizing Placements of Virtual Machines on Hypervisor Hosts | |
CN104038540B (en) | Method and system for automatically selecting application proxy server | |
CN102938790B (en) | Resource allocation methods in cloud computing system | |
CN107291545A (en) | The method for scheduling task and equipment of multi-user in computing cluster | |
CN103885831B (en) | The system of selection of virtual machine host machine and device | |
CN110418356A (en) | A kind of calculating task discharging method, device and computer readable storage medium | |
CN109784703A (en) | Business data processing method, device, computer equipment and storage medium | |
CN108901075A (en) | A kind of resource allocation methods based on GS algorithm | |
CN106534302A (en) | Multi-task demand service combination method and system | |
CN106202092A (en) | The method and system that data process | |
CN108132827A (en) | A kind of network slice method for mapping resource, relevant device and system | |
CN109889370A (en) | A kind of network equipment location determining method, device and computer readable storage medium | |
CN104581748A (en) | A method and device for identifying a scene in a wireless communication network | |
CN103077127B (en) | A kind of method and apparatus of specified data migrating objects | |
CN110297959A (en) | Date storage method, device, storage medium and Edge Server | |
CN112465371B (en) | Resource data distribution method, device and equipment | |
Satpathy et al. | CoMap: An efficient virtual network re-mapping strategy based on coalitional matching theory | |
CN108306912A (en) | Virtual network function management method and its device, network function virtualization system | |
CN104581832B (en) | A kind of realization method and system of data service shunting | |
CN112882805A (en) | Profit optimization scheduling method based on task resource constraint | |
CN103501509B (en) | A kind of method and device equalizing radio network controller load | |
CN106874108A (en) | Thin cloud is minimized in mobile cloud computing use number technology | |
CN109298914A (en) | A kind of Docker based on three-tier architecture and virtual machine initial placement method | |
CN114003389A (en) | Cloud platform resource allocation method and system based on joint selection and virtual machine placement |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |