CN106095586A - A kind of method for allocating tasks, Apparatus and system - Google Patents
A kind of method for allocating tasks, Apparatus and system Download PDFInfo
- Publication number
- CN106095586A CN106095586A CN201610474608.7A CN201610474608A CN106095586A CN 106095586 A CN106095586 A CN 106095586A CN 201610474608 A CN201610474608 A CN 201610474608A CN 106095586 A CN106095586 A CN 106095586A
- Authority
- CN
- China
- Prior art keywords
- node
- computing capability
- task
- main controlled
- calculating
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5083—Techniques for rebalancing the load in a distributed system
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
Abstract
The invention discloses a kind of method for allocating tasks, Apparatus and system.Described method includes: main controlled node receives task requests message, determines the computing capability that the task that described task requests is asked takies；Described main controlled node obtains the residue computing capability of each calculating node from this locality, according to the described computing capability taken and the residue computing capability of each calculating node, determines available calculating node；Described main controlled node is by described task distribution to described available calculating node, and can use the residue computing capability calculating node described in local update, it is thus achieved that the residue computing capability after described available calculating node updates.So, it is possible not only to eliminate polling mode and obtains each communication resource calculating node residue computing capability consumption, it is also possible to save the task distribution time, improve task allocative efficiency.
Description
Technical field
The present invention relates to data processing field, in particular it relates to a kind of method for allocating tasks, Apparatus and system.
Background technology
The Organization Chart of distributed system shown in Figure 1, distributed system is generally by a main controlled node and multiple calculating
Node forms.Wherein, after main controlled node receives task requests, can be according to the disposal ability of each calculating node, i.e. node
Surplus capacity, carries out task distribution.
Generally, main controlled node can obtain the surplus capacity of each calculating node before task is distributed by polling mode,
It is to say, main controlled node respectively with each calculating node communication, obtain the surplus capacity calculating node, and then according to this residue
Ability, completes task distribution between each calculating node.
So poll obtains the scheme of surplus capacity, not only can consume main controlled node and the communication money calculated between node
Source, along with the continuous accumulation of communication time between main controlled node and each calculating node, also results in task distribution overlong time, effect
The low inferior problem of rate.
Summary of the invention
It is an object of the invention to provide a kind of method for allocating tasks, Apparatus and system, obtain each in order to eliminate polling mode
The communication resource that individual calculating node residue computing capability consumes, and save the task distribution time, improve task allocative efficiency.
Embodiments providing a kind of method for allocating tasks, described method includes: main controlled node receives task requests
Message, determines the computing capability that the task that described task requests is asked takies；Described main controlled node obtains each meter from this locality
The residue computing capability of operator node, according to the described computing capability taken and the residue computing capability of each calculating node, determines
Available calculating node；Described main controlled node by the distribution of described task to described available calculating node, and can described in local update
Use tricks the residue computing capability of operator node, it is thus achieved that the residue computing capability after described available calculating node updates.
Alternatively, described main controlled node obtains the residue computing capability of each calculating node from this locality, including: described master control
Node obtains total computing capability of described calculating node, and described total computing capability is to be stored in described main controlled node during initial configuration
Local；Based on described total computing capability, described main controlled node is held according to task distribution condition and the task of described calculating node
Market condition, determines the residue computing capability of this calculating node.
Alternatively, described method also includes: described main controlled node receives the tasks carrying that described available calculating node returns
As a result, described task action result includes the current residual computing capability of described available calculating node；Described main controlled node judges
Whether the residue computing capability after described renewal is consistent with described current residual computing capability, if do not corresponded, then by described can
The residue computing capability of operator node of using tricks is adjusted to described current residual computing capability.
Embodiments providing a kind of method for allocating tasks, described method includes: the available node that calculates receives master control
The role assignment message that node sends, the task that described available calculating node is asked according to task requests by described main controlled node
The residue computing capability of each calculating node that the computing capability taken and described main controlled node this locality preserve determines；Described can
Operator node of using tricks performs the task of described main controlled node distribution, and after tasks carrying completes, returns to described main controlled node and appoints
Business performs result.
Alternatively, after tasks carrying completes, described method also includes: obtain the current residual of described available calculating node
Computing capability；And, described available calculating node to the task action result that described main controlled node returns include described currently
Residue computing capability.
Embodiments providing a kind of task allocation apparatus, described device is applied to main controlled node, described device bag
Include: first determines unit, be used for receiving task requests message, determine the calculating energy that the task that described task requests is asked takies
Power；Ability acquiring unit, for obtaining the residue computing capability of each calculating node from described main controlled node this locality；Second determines
Unit, for the computing capability taken described in basis and the residue computing capability of each calculating node, determines available calculating node；
Task allocation unit, for distributing described task to described available calculating node, and calculates joint available described in local update
The residue computing capability of point, it is thus achieved that the residue computing capability after described available calculating node updates.
Alternatively, described ability acquiring unit, for obtaining total computing capability of described calculating node；Based on described total
Calculation ability, according to task distribution condition and the tasks carrying situation of described calculating node, determines the residue meter of this calculating node
Calculation ability, described total computing capability is to be stored in described main controlled node this locality during initial configuration.
Alternatively, described device also includes: result receives unit, for receiving the task that described available calculating node returns
Performing result, described task action result includes the current residual computing capability of described available calculating node；Ability judging unit,
Whether the residue computing capability after judging described renewal is consistent with described current residual computing capability；Capacity adjusting unit,
For when described ability judging unit judges not correspond, it is described that the residue computing capability of described available calculating node is adjusted to
Current residual computing capability.
Embodiments providing a kind of task allocation apparatus, described device is applied to available calculating node, described dress
Putting and include: message reception units, for receiving the role assignment message that main controlled node sends, described available calculating node is by described
It is each that computing capability that the task that main controlled node is asked according to task requests takies and described main controlled node this locality preserve
The residue computing capability calculating node determines；Task executing units, for performing the task of described main controlled node distribution；Result is returned
Receipt unit, for after tasks carrying completes, returns task action result to described main controlled node.
Alternatively, described device also includes: ability acquiring unit, for after tasks carrying completes, obtains described available
Calculate the current residual computing capability of node；Described result returns unit, for returning tasks carrying knot to described main controlled node
Really, described task action result includes described current residual computing capability.
Embodiments providing a kind of task distribution system, described system includes that the task of being applied to main controlled node is divided
It is equipped with and puts, and be applied to the available task allocation apparatus calculating node.
In technical solution of the present invention, main controlled node is cached with the residue computing capability of each calculating node, therefore divides in task
Timing, can directly read from this locality, it is not necessary to as prior art needs to set up with each calculating node respectively to communicate, obtain in real time
The residue computing capability of each calculating node, so, is possible not only to eliminate polling mode and obtains each calculating node residue calculating
The communication resource that ability consumes, it is also possible to save the task distribution time, improve task allocative efficiency.
Other features and advantages of the present invention will be described in detail in detailed description of the invention part subsequently.
Accompanying drawing explanation
Accompanying drawing is used to provide a further understanding of the present invention, and constitutes the part of description, with following tool
Body embodiment is used for explaining the present invention together, but is not intended that limitation of the present invention.In the accompanying drawings:
Fig. 1 is the Organization Chart of distributed system；
Fig. 2 is the flow chart of main controlled node side of the present invention method for allocating tasks embodiment 1；
Fig. 3 is the flow chart of main controlled node side of the present invention method for allocating tasks embodiment 2；
Fig. 4 is that the present invention can be with the flow chart calculating node side method for allocating tasks；
Fig. 5 is the structural representation that the present invention is applied to the task allocation apparatus of main controlled node；
Fig. 6 is the structural representation that the present invention is applied to the available task allocation apparatus calculating node.
Detailed description of the invention
Below in conjunction with accompanying drawing, the detailed description of the invention of the present invention is described in detail.It should be appreciated that this place is retouched
The detailed description of the invention stated is merely to illustrate and explains the present invention, is not limited to the present invention.
See Fig. 2, it is shown that the flow chart of method for allocating tasks embodiment 1 of the present invention, the method is applied to main controlled node
Side, may include that
S101, main controlled node receives task requests message, determines the calculating that the task that described task requests is asked takies
Ability.
S102, described main controlled node obtains the residue computing capability of each calculating node, takies according to described from this locality
Computing capability and the residue computing capability of each calculating node, determine available calculating node.
The method for allocating tasks of the present invention, can apply in distributed system shown in Fig. 1.Main controlled node receives task
After request message, the following two aspect information of the acquisition that can be triggered:
(1) task that task requests is asked needs the computing capability taken, namely the distribution of this subtask needs consumption many
Few computing capability.
(2) the residue computing capability of multiple calculating nodes of main controlled node management, namely calculate the execution energy that node is current
Power.
As a kind of example, the residue computing capability calculating node can be obtained in the following manner: first, main controlled node
Obtaining the total computing capability calculating node, wherein, total computing capability is to be stored in main controlled node this locality during initial configuration.Secondly,
Based on total computing capability, main controlled node, according to calculating task distribution condition and the tasks carrying situation of node, determines this calculating
The residue computing capability of node.
It is to say, main controlled node can obtain, when carrying out initial configuration, the total computing capability calculating node；Additionally,
When main controlled node can also obtain the distribution of every subtask, the computing capability calculating node taken.So, main controlled node just can evidence
This determines the residue computing capability calculating node.It should be noted that when initial configuration by calculate node always calculate energy
Trying hard to keep, it is local to be stored in main controlled node, additionally aids and reduces the resource and time consumed when main controlled node obtains total computing capability.
For example, the total computing capability calculating node 1 is 8, and task distribution for the first time occupies 4 of calculating node 1
Computing capability, if task is assigned to calculate execution at node 1 for the first time, and the task of not receiving calculating node 1 return is held
Row result, i.e. task distribution condition is for distribute, and tasks carrying situation is for being not fully complete, then main controlled node can determine therefrom that out
The residue computing capability calculating node 1 is 4.If main controlled node receives calculates the task action result that node 1 returns, and not
Distributing new task to calculating node 1, i.e. task distribution condition is unallocated, tasks carrying situation is for complete, then master control saves
Point can determine therefrom that the residue computing capability calculating node 1 is 8.
S103, described main controlled node by the distribution of described task to described available calculating node, and can described in local update
Use tricks the residue computing capability of operator node, it is thus achieved that the residue computing capability after described available calculating node updates.
In the present invention program, according to the needs of reality application, the available node that calculates can be one, it is also possible to is multiple,
This can be not specifically limited by the present invention, and the computing capability that the available residue computing capability calculating node takies not less than task is
Can.
For example, the residue computing capability calculating node 1 is 4, and the residue computing capability calculating node 2 is 4, calculates joint
The residue computing capability of point 3 is 2, if the computing capability that task takies is 6, then, main controlled node can be true by calculating node 1 and 2
It is set to available calculating node, and in order to meet the task requirement to computing capability, does the distribution of following task: take calculating node 1
Whole remain computing capability, i.e. 4 computing capabilitys；Taking the some residual computing capability calculating node 2, i.e. 2 calculate energy
Power.
It should be noted that the residue calculating node that the present invention program is based primarily upon main controlled node local cache calculates energy
Power, carries out task distribution, in order to ensure the accuracy that task is distributed, often carries out a subtask distribution, and main controlled node needs basis
Task distribution condition, does renewal process to the residue computing capability of available calculating node, it is ensured that main controlled node local cache data
Accuracy and promptness.In above-mentioned example, after main controlled node carries out task distribution, after calculating the renewal of node 1, residue calculates
Ability is 0, and remaining computing capability after calculating the renewal of node 2 is 2.
To sum up, main controlled node is cached with the residue computing capability of each calculating node, therefore when task is distributed, can be direct
Read from this locality, it is not necessary to as prior art needs to set up with each calculating node respectively to communicate, obtain each calculating node in real time
Residue computing capability, so, be possible not only to eliminate that polling mode obtains that each calculates that node residue computing capability consumes is logical
Letter resource, it is also possible to save the task distribution time, improve task allocative efficiency.
Alternatively, in order to improve the accuracy of main controlled node local cache data further, present invention also offers task
Distribution method embodiment 2, specifically can be found in flow chart shown in Fig. 3, it is also possible to including:
S201, described main controlled node receives the task action result that described available calculating node returns, described tasks carrying
Result includes the current residual computing capability of described available calculating node.
S202, the residue computing capability after described main controlled node judges described renewal with described current residual computing capability is
No it is consistent；
S203, if do not corresponded, is then adjusted to described current residual by the residue computing capability of described available calculating node
Computing capability.
It is to say, the present invention also provides for a kind of Adjusted Option remaining computing capability.Specifically, calculating node can be by
The current residual computing capability of self feeds back to main controlled node, main controlled node adjust this locality according to current residual computing capability slow
Deposit, further ensure that data cached accuracy.It should be noted that in order to reduce main controlled node and calculate between node as far as possible
The communication resource that information is consumed alternately, current residual computing capability can be carried in task action result, feed back to master control
Node.
Seeing Fig. 4, it is shown that the flow chart of method for allocating tasks of the present invention, the method is applied to available calculating node side,
May include that
S301, the available node that calculates receives the role assignment message that main controlled node sends, and described available calculating node is by institute
State computing capability that main controlled node takies according to the task that task requests is asked and that described main controlled node this locality preserves is every
The residue computing capability of individual calculating node determines.
S302, described available calculating node performs the task of described main controlled node distribution, and after tasks carrying completes, to
Described main controlled node returns task action result.
It is to say, when main controlled node carries out task distribution, calculate node and communicate, in real time without setting up with main controlled node
Report the residue computing capability of this calculating node.Main controlled node can directly utilize the residue computing capability of its local cache, then
The computing capability taken in conjunction with required by task, carries out task distribution, that is, calculate node to directly receive appointing of main controlled node transmission
Business assignment messages, and perform being distributed of task as indicated.So, it is possible not only to eliminate polling mode and obtains each meter
The communication resource that operator node residue computing capability consumes, it is also possible to save the task distribution time, improve task allocative efficiency.
Alternatively, in order to improve the accuracy of main controlled node local cache data further, present invention also offers as follows
Task allocative decision.Specifically, the available node that calculates can obtain working as of described available calculating node after tasks carrying completes
Front residue computing capability, and the task action result including described current residual computing capability is returned to main controlled node.So, main
Control node just can be according to current residual computing capability, it may be judged whether need to adjust the data of its local cache, it is ensured that local cache
The accuracy of data, distributes ready for task next time.
Corresponding with method shown in Fig. 2, the embodiment of the present invention also provides for a kind of task allocation apparatus 400, and this device 400 should
For main controlled node.Schematic diagram shown in Figure 5, may include that
First determines unit 401, is used for receiving task requests message, determines that the task that described task requests is asked takies
Computing capability；
Ability acquiring unit 402, for obtaining the residue computing capability of each calculating node from described main controlled node this locality；
Second determines unit 403, and the residue for the computing capability taken described in basis and each calculating node calculates energy
Power, determines available calculating node；
Task allocation unit 404, for distributing described task to described available calculating node, and described in local update
The available residue computing capability calculating node, it is thus achieved that the residue computing capability after described available calculating node updates.
Alternatively, described ability acquiring unit, for obtaining total computing capability of described calculating node；Based on described total
Calculation ability, according to task distribution condition and the tasks carrying situation of described calculating node, determines the residue meter of this calculating node
Calculation ability, described total computing capability is to be stored in described main controlled node this locality during initial configuration.
Alternatively, described device also includes:
Result receives unit, for receiving the task action result that described available calculating node returns, described tasks carrying
Result includes the current residual computing capability of described available calculating node；
Ability judging unit, the residue computing capability after judging described renewal with described current residual computing capability is
No it is consistent；
Capacity adjusting unit, for when described ability judging unit judges not correspond, by described available calculating node
Residue computing capability is adjusted to described current residual computing capability.
Corresponding with method shown in Fig. 4, the embodiment of the present invention also provides for a kind of task allocation apparatus 500, and this device 500 should
For available calculating node.Schematic diagram shown in Figure 6, may include that
Message reception units 501, for receive main controlled node send role assignment message, described available calculating node by
Computing capability that the task that described main controlled node is asked according to task requests takies and described main controlled node this locality preserve
The residue computing capability of each calculating node determines；
Task executing units 502, for performing the task of described main controlled node distribution；
Result returns unit 503, for after tasks carrying completes, returns task action result to described main controlled node.
Alternatively, described device also includes:
Ability acquiring unit, for after tasks carrying completes, the current residual obtaining described available calculating node calculates
Ability；
Described result returns unit, for returning task action result, described task action result to described main controlled node
Include described current residual computing capability.
Accordingly, the embodiment of the present invention also provides for a kind of task distribution system, and this system can include appointing shown in Fig. 5
Business distributor 400, and the task allocation apparatus 500 shown in Fig. 6.
The preferred embodiment of the present invention is described in detail above in association with accompanying drawing, but, the present invention is not limited to above-mentioned reality
Execute the detail in mode, in the technology concept of the present invention, technical scheme can be carried out multiple letter
Monotropic type, these simple variant belong to protection scope of the present invention.
It is further to note that each the concrete technical characteristic described in above-mentioned detailed description of the invention, at not lance
In the case of shield, can be combined by any suitable means, in order to avoid unnecessary repetition, the present invention to various can
The compound mode of energy illustrates the most separately.
Additionally, combination in any can also be carried out between the various different embodiment of the present invention, as long as it is without prejudice to this
The thought of invention, it should be considered as content disclosed in this invention equally.
Claims (11)
1. a method for allocating tasks, it is characterised in that described method includes:
Main controlled node receives task requests message, determines the computing capability that the task that described task requests is asked takies；
Described main controlled node obtains the residue computing capability of each calculating node from this locality, according to the described computing capability taken and
The residue computing capability of each calculating node, determines available calculating node；
Described main controlled node is by described task distribution extremely described available calculating node, and calculates node available described in local update
Residue computing capability, it is thus achieved that the residue computing capability after described available calculating node updates.
Method the most according to claim 1, it is characterised in that described main controlled node obtains each calculating node from this locality
Residue computing capability, including:
Described main controlled node obtains total computing capability of described calculating node, and described total computing capability is to be stored in during initial configuration
Described main controlled node is local；
Based on described total computing capability, described main controlled node is according to the task distribution condition of described calculating node and tasks carrying
Situation, determines the residue computing capability of this calculating node.
Method the most according to claim 1 and 2, it is characterised in that described method also includes:
Described main controlled node receives the task action result that described available calculating node returns, and described task action result includes institute
State the available current residual computing capability calculating node；
Whether the residue computing capability after described main controlled node judges described renewal is consistent with described current residual computing capability, as
Fruit does not corresponds, then the residue computing capability of described available calculating node is adjusted to described current residual computing capability.
4. a method for allocating tasks, it is characterised in that described method includes:
The available node that calculates receives the role assignment message that main controlled node sends, and described available calculating node is by described main controlled node
Computing capability that being asked according to task requests of task takies and each calculating node that described main controlled node this locality preserves
Residue computing capability determine；
Described available calculating node performs the task of described main controlled node distribution, and after tasks carrying completes, to described master control
Node returns task action result.
Method the most according to claim 4, it is characterised in that after tasks carrying completes, described method also includes:
Obtain the current residual computing capability of described available calculating node；And
To the task action result that described main controlled node returns, described available calculating node includes that described current residual calculates energy
Power.
6. a task allocation apparatus, it is characterised in that described device is applied to main controlled node, described device includes:
First determines unit, is used for receiving task requests message, determines the calculating that the task that described task requests is asked takies
Ability；
Ability acquiring unit, for obtaining the residue computing capability of each calculating node from described main controlled node this locality；
Second determines unit, for the computing capability taken described in basis and the residue computing capability of each calculating node, determines
Available calculating node；
Task allocation unit, for distributing described task to described available calculating node, and can use tricks described in local update
The residue computing capability of operator node, it is thus achieved that the residue computing capability after described available calculating node updates.
Device the most according to claim 6, it is characterised in that described ability acquiring unit, is used for obtaining described calculating and saves
Total computing capability of point；Based on described total computing capability, according to task distribution condition and the tasks carrying of described calculating node
Situation, determines that the residue computing capability of this calculating node, described total computing capability are to be stored in described master control joint during initial configuration
Point this locality.
8. according to the device described in claim 6 or 7, it is characterised in that described device also includes:
Result receives unit, for receiving the task action result that described available calculating node returns, described task action result
Current residual computing capability including described available calculating node；
Ability judging unit, for judging the residue computing capability after described renewal and described current residual computing capability whether phase
Symbol；
Capacity adjusting unit, for when described ability judging unit judges not correspond, by the residue of described available calculating node
Computing capability is adjusted to described current residual computing capability.
9. a task allocation apparatus, it is characterised in that described device is applied to available calculating node, and described device includes:
Message reception units, for receiving the role assignment message that main controlled node sends, described available calculating node is by described master
Control computing capability and each meter of described main controlled node this locality preservation that node takies according to the task that task requests is asked
The residue computing capability of operator node determines；
Task executing units, for performing the task of described main controlled node distribution；
Result returns unit, for after tasks carrying completes, returns task action result to described main controlled node.
Device the most according to claim 9, it is characterised in that described device also includes:
Ability acquiring unit, for after tasks carrying completes, obtains the current residual computing capability of described available calculating node；
Described result returns unit, for returning task action result to described main controlled node, wraps in described task action result
Include described current residual computing capability.
11. 1 kinds of task distribution systems, it is characterised in that described system includes the task as described in any one of claim 6～8
Distributor, and the task allocation apparatus described in claim 9 or 10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610474608.7A CN106095586A (en) | 2016-06-23 | 2016-06-23 | A kind of method for allocating tasks, Apparatus and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610474608.7A CN106095586A (en) | 2016-06-23 | 2016-06-23 | A kind of method for allocating tasks, Apparatus and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106095586A true CN106095586A (en) | 2016-11-09 |
Family
ID=57252729
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610474608.7A Pending CN106095586A (en) | 2016-06-23 | 2016-06-23 | A kind of method for allocating tasks, Apparatus and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106095586A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106874120A (en) * | 2017-02-28 | 2017-06-20 | 联想(北京)有限公司 | The processor resource optimization method of calculate node, calculate node and server cluster |
| CN107766156A (en) * | 2017-10-20 | 2018-03-06 | 北京小度信息科技有限公司 | Task processing method and device |
| CN107783840A (en) * | 2017-10-27 | 2018-03-09 | 福州瑞芯微电子股份有限公司 | A kind of Distributed-tier deep learning resource allocation methods and device |
| WO2018176998A1 (en) * | 2017-03-29 | 2018-10-04 | 华为技术有限公司 | Data storage method and device |
| CN108958942A (en) * | 2018-07-18 | 2018-12-07 | 郑州云海信息技术有限公司 | A kind of distributed system distribution multitask method, scheduler and computer equipment |
| CN110837421A (en) * | 2019-11-13 | 2020-02-25 | 北京知道创宇信息技术股份有限公司 | Task allocation method and device |
| US11032858B2 (en) | 2017-03-28 | 2021-06-08 | Huawei Technologies Co., Ltd. | Network access method, terminal, access network, and core network |
| CN115114034A (en) * | 2022-08-29 | 2022-09-27 | 岚图汽车科技有限公司 | Distributed computing method and device |
| CN116225669A (en) * | 2023-05-08 | 2023-06-06 | 之江实验室 | Task execution method and device, storage medium and electronic equipment |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101604264A (en) * | 2009-07-08 | 2009-12-16 | 深圳先进技术研究院 | The method for scheduling task of supercomputer and system |
| CN102012891A (en) * | 2010-11-18 | 2011-04-13 | 浪潮(北京)电子信息产业有限公司 | Computer cluster management method, device and system |
| CN102073546A (en) * | 2010-12-13 | 2011-05-25 | 北京航空航天大学 | Task-dynamic dispatching method under distributed computation mode in cloud computing environment |
| CN102541640A (en) * | 2011-12-28 | 2012-07-04 | 厦门市美亚柏科信息股份有限公司 | Cluster GPU (graphic processing unit) resource scheduling system and method |
| CN103092698A (en) * | 2012-12-24 | 2013-05-08 | 中国科学院深圳先进技术研究院 | System and method of cloud computing application automatic deployment |
| CN104050042A (en) * | 2014-05-30 | 2014-09-17 | 北京先进数通信息技术股份公司 | Resource allocation method and resource allocation device for ETL (Extraction-Transformation-Loading) jobs |
| CN104598318A (en) * | 2014-12-30 | 2015-05-06 | 北京奇艺世纪科技有限公司 | Node calculating capability reporting method and calculating node |
-
2016
- 2016-06-23 CN CN201610474608.7A patent/CN106095586A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101604264A (en) * | 2009-07-08 | 2009-12-16 | 深圳先进技术研究院 | The method for scheduling task of supercomputer and system |
| CN102012891A (en) * | 2010-11-18 | 2011-04-13 | 浪潮(北京)电子信息产业有限公司 | Computer cluster management method, device and system |
| CN102073546A (en) * | 2010-12-13 | 2011-05-25 | 北京航空航天大学 | Task-dynamic dispatching method under distributed computation mode in cloud computing environment |
| CN102541640A (en) * | 2011-12-28 | 2012-07-04 | 厦门市美亚柏科信息股份有限公司 | Cluster GPU (graphic processing unit) resource scheduling system and method |
| CN103092698A (en) * | 2012-12-24 | 2013-05-08 | 中国科学院深圳先进技术研究院 | System and method of cloud computing application automatic deployment |
| CN104050042A (en) * | 2014-05-30 | 2014-09-17 | 北京先进数通信息技术股份公司 | Resource allocation method and resource allocation device for ETL (Extraction-Transformation-Loading) jobs |
| CN104598318A (en) * | 2014-12-30 | 2015-05-06 | 北京奇艺世纪科技有限公司 | Node calculating capability reporting method and calculating node |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106874120A (en) * | 2017-02-28 | 2017-06-20 | 联想(北京)有限公司 | The processor resource optimization method of calculate node, calculate node and server cluster |
| US11144366B2 (en) | 2017-02-28 | 2021-10-12 | Lenovo (Beijing) Limited | Computing node processor resource optimization method, computing node and server cluster |
| US11032858B2 (en) | 2017-03-28 | 2021-06-08 | Huawei Technologies Co., Ltd. | Network access method, terminal, access network, and core network |
| WO2018176998A1 (en) * | 2017-03-29 | 2018-10-04 | 华为技术有限公司 | Data storage method and device |
| US10972542B2 (en) | 2017-03-29 | 2021-04-06 | Huawei Technologies Co., Ltd. | Data storage method and apparatus |
| US11575748B2 (en) | 2017-03-29 | 2023-02-07 | Huawei Technologies Co., Ltd. | Data storage method and apparatus for combining different data distribution policies |
| CN107766156B (en) * | 2017-10-20 | 2020-07-10 | 北京星选科技有限公司 | Task processing method and device |
| CN107766156A (en) * | 2017-10-20 | 2018-03-06 | 北京小度信息科技有限公司 | Task processing method and device |
| CN107783840B (en) * | 2017-10-27 | 2020-08-21 | 瑞芯微电子股份有限公司 | Distributed multi-layer deep learning resource allocation method and device |
| CN107783840A (en) * | 2017-10-27 | 2018-03-09 | 福州瑞芯微电子股份有限公司 | A kind of Distributed-tier deep learning resource allocation methods and device |
| CN108958942A (en) * | 2018-07-18 | 2018-12-07 | 郑州云海信息技术有限公司 | A kind of distributed system distribution multitask method, scheduler and computer equipment |
| CN110837421A (en) * | 2019-11-13 | 2020-02-25 | 北京知道创宇信息技术股份有限公司 | Task allocation method and device |
| CN115114034A (en) * | 2022-08-29 | 2022-09-27 | 岚图汽车科技有限公司 | Distributed computing method and device |
| CN116225669A (en) * | 2023-05-08 | 2023-06-06 | 之江实验室 | Task execution method and device, storage medium and electronic equipment |
| CN116225669B (en) * | 2023-05-08 | 2024-01-09 | 之江实验室 | Task execution method and device, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106095586A (en) | A kind of method for allocating tasks, Apparatus and system | |
| CN107613025B (en) | Message queue sequence reply-based implementation method and device | |
| CN105242957A (en) | Method and system for cloud computing system to allocate GPU resources to virtual machine | |
| MX350298B (en) | Methods and systems for data context and management via dynamic spectrum controller and dynamic spectrum policy controller. | |
| US9686330B2 (en) | Topic-based messaging using consumer address and pool | |
| CN103078965B (en) | The IP address management method of virtual machine | |
| CN112671830A (en) | Resource scheduling method, system, device, computer equipment and storage medium | |
| CN108924221A (en) | The method and apparatus for distributing resource | |
| CN110866167B (en) | Task allocation method, device, server and storage medium | |
| US20110004500A1 (en) | Allocating a resource based on quality-of-service considerations | |
| CN106603711B (en) | Communication method, system and message server | |
| CN102495759A (en) | Method for scheduling job in cloud computing environment | |
| CN105227489A (en) | A kind of bandwidth management method and electronic equipment | |
| CN107241421A (en) | A kind of cloud host resource method for obligating and device | |
| CN102970381A (en) | Multi-source load balance method and system for proportional polling based on content distribution network | |
| CN105763485A (en) | Resource distribution method, device and server | |
| CN106293945A (en) | A kind of resource perception method and system across virtual machine | |
| EP3279795B1 (en) | Method and apparatus for deleting cloud host in cloud computing environment, server and storage medium | |
| CN113535398B (en) | Resource allocation adjustment method, device, electronic equipment and readable storage medium | |
| CN109831474B (en) | Keep-alive system, method, server and readable storage medium for http long connection | |
| JP6721800B2 (en) | Cooperative distributed system, cooperative distributed management device, cooperative distributed method, and program | |
| CN105162560A (en) | Data processing method and system | |
| CN109729182B (en) | Network address allocation method and system of virtual network function and VNFM | |
| JP2015197892A (en) | Data center power saving system | |
| CN108090029B (en) | Computation scheduling device in matrix inversion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20161109 |