CN104077188A - Method and device for scheduling tasks - Google Patents
Method and device for scheduling tasks Download PDFInfo
- Publication number
- CN104077188A CN104077188A CN201310109076.3A CN201310109076A CN104077188A CN 104077188 A CN104077188 A CN 104077188A CN 201310109076 A CN201310109076 A CN 201310109076A CN 104077188 A CN104077188 A CN 104077188A
- Authority
- CN
- China
- Prior art keywords
- task
- computing node
- processing time
- information
- minimum
- 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
Abstract
The invention relates to a method and a device for scheduling tasks. The device comprises a decision-making module and a distribution module. The decision-making module is used for determining tasks among at least one part of to-be-processed tasks according to stored information of processing time when information for requesting distribution of specified types of tasks is received, the information for requesting distribution of the specified types of tasks is transmitted by computing nodes in computing clusters, the processing time is spent on processing the various tasks by the computing nodes which transmit the information for requesting distribution of the specified types of tasks, the determined tasks belong to the specified types, and the computing nodes which transmit the information for requesting distribution of the specified types of tasks previously spend the shortest processing time on processing the determined tasks; the distribution module is used for distributing the determined tasks to the computing nodes which transmit the information for requesting distribution of the specified types of tasks, so that the determined tasks can be processed by the computing nodes. The method and the device have the advantage that the overall performance of distributed parallel computation can be improved by the aid of the method and the device.
Description
Technical field
The present invention relates to distributed parallel and calculate field, relate in particular to a kind of method and apparatus for task scheduling.
Background technology
MapReduce is the novel and effective distributed parallel computing architecture of one that Google (Google) proposes, and it becomes the most widely used framework in the cloud computing epoch.MapReduce framework is designed to carry out parallel computation in heterogeneous (heterogeneous) computer cluster, to improve the overall calculated performance of parallel computation.
In MapReduce framework, each work (Job) is divided into can be in multiple tasks (task) of calculating parallel running on multiple computing nodes of cluster, and these tasks can be divided into mapping (Map) task and reduction (Reduce) task according to its type.
For each task of each work, which task is processed and when is processed by task scheduling (task scheduling) and determine by calculating which computing node in cluster.Therefore, task scheduling is very important in parallel computation, and it can affect the overall performance of parallel computation.
Summary of the invention
Embodiments of the invention propose a kind of method and apparatus for task scheduling, and it can improve the overall performance that distributed parallel calculates.
According to a kind of method for task scheduling of the embodiment of the present invention, comprise: distribute while thering is the message of task of specified type when receiving request that the computing node that calculates in cluster sends, process the information in the processing time that each task spends according to the computing node of the stored described message of transmission, from least a portion task of pending task, determine belonging to described specified type and sending computing node its processing time minimum being spent with pre-treatment of described message of task; And, described definite task is distributed to the computing node of the described message of transmission and processed.
In a kind of specific implementation, described determining step comprises: the information in the processing time spending according to the described pending task of each computing node processing in stored described calculating cluster, add up the described pending task minimum treat time separately; From described pending task, find out the task that its minimum treat time is greater than the first assign thresholds; And, according to the stored information of processing the processing time that each task spends about the computing node that sends described message, from described task of searching, retrieve the task of computing node its processing time minimum being spent with pre-treatment that belongs to described specified type and send described message, as described definite task.
Wherein, described the first assign thresholds is to determine minimum treat time based on described pending task.
In another kind of specific implementation, described determining step comprises: the information in the processing time spending according to the described pending task of each computing node processing in stored described calculating cluster, add up the described pending task minimum treat time separately; From described pending task, find out it and belong to the task that described specified type and its minimum treat time are greater than the second assign thresholds; And, according to the stored information of processing the processing time that each task spends about the computing node that sends described message, from described task of searching, retrieve the task of computing node its processing time minimum being spent with pre-treatment that sends described message, as described definite task.
Wherein, described the second assign thresholds is to determine the minimum treat time of the particular task based in described pending task, and wherein said particular task is the task of belonging to described specified type.
In a kind of specific implementation, describedly also comprise: after the computing node from described calculating cluster receives the information of processing the processing time that a task spends about it, the information that storage receives.
According to a kind of device for task scheduling of the embodiment of the present invention, comprise: decision-making module, for calculating request distribution that the computing node of cluster sends while thering is the message of task of specified type when receiving, process the information in the processing time that each task spends according to the computing node of the stored described message of transmission, from least a portion task of pending task, determine belonging to described specified type and sending computing node its processing time minimum being spent with pre-treatment of described message of task; And, distribution module, the computing node that sends described message for described definite task is distributed to is processed.
In a kind of specific implementation, described decision-making module comprises: the first statistical module, be used for the information in the processing time spending according to the described pending task of each computing node processing of stored described calculating cluster, add up the described pending task minimum treat time separately; First searches module, for from described pending task, finds out the task that the minimum treat time is greater than the first assign thresholds; And, the first retrieval module, for according to the stored information of processing the processing time that each task spends about the computing node that sends described message, from described task of searching, retrieve the task that it belongs to described specified type and sends computing node its processing time minimum being spent with pre-treatment of described message, as described definite task.
Wherein, described the first assign thresholds is to determine minimum treat time based on described pending task.
In another kind of specific implementation, described decision-making module comprises: the second statistical module, be used for the information in the processing time spending according to the stored described pending task of each computing node processing about described calculating cluster, add up the described pending task minimum treat time separately; Second searches module, for from described pending task, finds out and belongs to the task that described specified type and its minimum treat time are greater than the second assign thresholds; And, the second retrieval module, for according to the stored information of processing the processing time that each task spends about the computing node that sends described message, from described task of searching, retrieve the task of computing node its processing time minimum being spent with pre-treatment that sends described message, as described definite task.
Wherein, described the second assign thresholds is to determine the minimum treat time of the particular task based in described pending task, and wherein said particular task is the task of belonging to described specified type.
In a kind of specific implementation, described device also comprises: collection module, and for receive the information of processing the processing time that certain task spends about it from the computing node of described calculating cluster after, the information that storage receives.
Can find out from description above, the scheme of the embodiment of the present invention is according to the stored information of processing the processing time that each task spends about each computing node, determine that its computing node using the pending processing time minimum that task was spent of pre-treatment is as the computing node for the treatment of this pending task, thereby make the processing of each pending task complete within a short period of time as much as possible, thereby improved the overall performance that distributed parallel calculates.
Brief description of the drawings
Further feature of the present invention, feature, advantage and benefit will become more apparent by the detailed description below in conjunction with accompanying drawing.
Fig. 1 shows according to the configuration diagram of the concurrent computational system of one embodiment of the invention.
Fig. 2 shows according to the process flow diagram of the method for task scheduling of one embodiment of the invention.
Fig. 3 A shows according to the schematic diagram of the device for task scheduling of one embodiment of the invention.
Fig. 3 B shows according to the schematic diagram of the decision-making module of one embodiment of the invention.
Fig. 3 C shows according to the schematic diagram of the decision-making module of another embodiment of the present invention.
Fig. 4 shows according to the schematic diagram of the equipment for task scheduling of one embodiment of the invention.
Embodiment
In practice, have some evaluation works normally repeatedly to carry out, analytic statistics work common one day, one month or a season etc. of such as system journal, Operation Log and call detail record will carry out once.Carry out the work that these carry out repeatedly under the framework of parallel computation time, each task that these work of repeatedly carrying out are divided is also to distribute to repeatedly each computing node processing of calculating in cluster conventionally.The scheme of the embodiment of the present invention is collected and storage is processed the information in the processing time that each task spends about calculating each computing node in cluster, then in the time receiving request that idle computing node sends and distribute the message of the task with specified type according to the collected information of processing the processing time that each task spends about this idle computing node, from pending task, choose this idle computing node with minimum task of its processing time being spent of pre-treatment and the task that this is chosen distributed to this idle computing node and carry out processing, thereby improve the overall performance that distributed parallel calculates.
Below, in connection with accompanying drawing each embodiment of the present invention in detail.
Referring now to Fig. 1,, it shows according to the configuration diagram of the concurrent computational system of one embodiment of the invention.As shown in Figure 1, concurrent computational system 10 is the systems that belong to Mapreduce framework, and it comprises multiple computing nodes 100 and main controlled node 200, and wherein each computing node 100 can communicate to connect main controlled node 200.The plurality of computing node 100 has formed the calculating cluster of concurrent computational system 10.
Computing node 100 can be any equipment with computing power, such as, but be not limited to desk-top computer, server, notebook computer etc.Computing node 100 can have one or more the ability of processing in the mapping task of (Map) type and the task of reduction (Reduce) type.In the time that the available free time of computing node 100 can be processed the task of a certain type (mapping or conclusion) T, computing node 100 can send request the message of distributing the task with type T to main controlled node 200, to ask main controlled node 200 to distribute the task with type T to process.When receive the task of a type T who distributes to its processing from main controlled node 200, computing node 100 is processed the task of these distribution, and sends to main controlled node 200 information of processing the processing time that this task spends about it after the processing that completes this task.
Main controlled node 200 can be any equipment with computing power, such as, but be not limited to desk-top computer, server, notebook computer etc.The information in the processing time spending about each computing node 100 each tasks of processing is collected and stored to main controlled node 200, in the starting stage, main controlled node 200 for example can be distributed to same task each computing node 100 as far as possible and process, to can collect about the information of processing the processing time that each task spends before each computing node 100.When receiving while distributing the message of task of specified type from the request of calculating the arbitrary computing node 100 in cluster, main controlled node 200 can be according to the information in the stored processing time spending about these arbitrary computing node 100 each tasks of processing, from at least a portion task of pending task, determine that it belongs to the task of this specified type and this arbitrary computing node 100 its processing time minimum being spent with pre-treatment, and determined task is distributed to this arbitrary computing node 100 and process.
Referring now to Fig. 2,, it shows according to the process flow diagram of the method for task scheduling of one embodiment of the invention.As shown in Figure 2, at step S200, in the time that the available free resource of arbitrary computing node 100i in calculating cluster is processed the task of a certain type T, computing node 100i sends request the message of distributing the task with type T to main controlled node 200.Wherein, type T can be mapping task or conclusion task.
At step S204, after receiving and distributing from the request of computing node 100i and have the message of task of type T, main controlled node 200 is added up the pending task K minimum treat time separately.For example, the processing time that can spend from each stored computing node 100 each tasks of processing, search each computing node 100 and process in each processing time spending of pending task K the minimum processing time, as each minimum treat time of pending task K.
At step S208, main controlled node 200 calculates the weighted mean value of maximum minimum treat time in the pending task K minimum treat time separately and minimum minimum treat time, as assign thresholds YZ.
At step S212, main controlled node 200 finds out the task that its minimum treat time is greater than assign thresholds YZ from pending task K.
At step S216, main controlled node 200 searches out the task that type belongs to type T from the task of searching.
At step S220, main controlled node 200, according to the stored information of processing the processing time that each task spends about computing node 100i, retrieves the task Ki of computing node 100i with the processing time minimum of its cost of pre-treatment from searched for task.That is to say, in searched for task, the processing time that before computing node 100i, Processing tasks Ki spends is minimum.
At step S224, main controlled node 200 is distributed to computing node 100i retrieved task Ki.
At step S228, after main controlled node 200 receives distributed task Ki, computing node 100i Processing tasks Ki.
At step S232, after the processing of the Ki that finishes the work, computing node 100i sends the information in the processing time spending about computing node 100i Processing tasks Ki to main controlled node 200.
At step S236, main controlled node 200 is stored the information in the received processing time spending about computing node 100i Processing tasks Ki, to upgrade the information in the processing time being spent about computing node 100i Processing tasks Ki of being stored in the past.
Can find out from above description, the scheme of the present embodiment is stored each computing node and is processed the processing time that each task spends, then in the time there is idle computing node, process according to this stored idle computing node the processing time that each task spends, from pending task, choose this idle computing node with minimum task of its processing time being spent of pre-treatment and selected task distributed to this idle computing node and process, thereby make each task as much as possible within a short period of time processed, thereby improve the overall performance that distributed parallel calculates.
Other modification
Those skilled in the art are to be understood that, although in the above embodiments, assign thresholds YZ is the weighted mean value of maximum minimum treat time in the pending task K minimum treat time separately and minimum minimum treat time, but the present invention is not limited thereto.In some other embodiment of the present invention, also can utilize the minimum treat time of the task K of other any suitable mode based on pending to calculate assign thresholds YZ.The weighted mean value that for example, can calculate the pending task K minimum treat time is separately as assign thresholds YZ.Again for example, can use the minimum treat time of the task that in pending task K, its minimum treat time mediates as assign thresholds YZ.
Those skilled in the art are to be understood that, although in the above embodiments, at step S208-S220, first main controlled node 200 calculates the weighted mean value of maximum minimum treat time in the pending task K minimum treat time separately and minimum minimum treat time as assign thresholds YZ, then from pending task K, find out the task that its minimum treat time is greater than assign thresholds YZ, then from the task of searching, search out the task that its type belongs to type T, the last task Ki of computing node 100i with the processing time minimum of its cost of pre-treatment that retrieve from searched for task, but, the present invention is not limited thereto.
In some other embodiment of the present invention, also can carry out alternative steps S208-S220 by following steps: first main controlled node 200 searches out the task KP that its type belongs to type T from pending task K, then the weighted mean value of the maximum minimum treat time in the minimum treat time of the task KP that calculating is searched for and minimum minimum treat time is as assign thresholds TH, then from searched for task KP, find out the task KPP that its minimum treat time is greater than assign thresholds TH, the last task Ki of computing node 100i with the processing time minimum of its cost of pre-treatment that retrieve from searched task KPP, as the task of distributing to computing node 100i processing.
Here can also utilize the minimum treat time of the task KP of other any suitable mode based on searched for to calculate assign thresholds TH.The weighted mean value of minimum treat time that for example, can calculate searched for task KP is as assign thresholds TH.Again for example, can use the minimum treat time of the task that in searched for task KP, its minimum treat time mediates as assign thresholds TH.
Those skilled in the art are to be understood that, although in the above embodiments, its minimum treat time from pending task K is greater than to be chosen suitable task in the task of assign thresholds and distributes to idle computing node 100i, but the present invention is not limited thereto.In some other embodiment of the present invention, also can from all tasks of pending task K, choose suitable task and distribute to idle computing node 100i.
Although it will be appreciated by those skilled in the art that in the above embodiments, task is divided into map type and reduction type, but the present invention is not limited thereto.In some other embodiment of the present invention, the type of task can be divided arbitrarily as required.
Although it will be appreciated by those skilled in the art that in the above embodiments, concurrent computational system 10 is the systems that belong to Mapreduce framework, but the present invention is not limited thereto.In some other embodiment of the present invention, concurrent computational system 10 can be any concurrent computational system that task division is become to multiple subtasks and distribute to each computing node processing in calculating cluster.
Referring now to Fig. 3 A,, it shows according to the schematic diagram of the device for task scheduling of one embodiment of the invention.Device shown in Fig. 3 A can be arranged in main controlled node 200, and can utilize software, hardware (such as integrated circuit or FPGA etc.) or the mode of software and hardware combining to realize.
As shown in Figure 3A, can comprise decision-making module 310 and distribution module 320 for the device 300 of task scheduling.Wherein, decision-making module 310 is for for calculating the request that the computing node of cluster sends and distribute while having the message of task of specified type when receiving, according to the stored information of processing the processing time that each task spends about the computing node that sends described message, from least a portion task of pending task, determine its belonging to described specified type and sending computing node its processing time minimum being spent with pre-treatment of described message of task.Distribution module 320 is processed for the computing node that sends described message for described definite task is distributed to.
Wherein, in a kind of specific implementation, as shown in Figure 3 B, decision-making module 310 can comprise that the first statistical module 311, first searches module 312 and the first retrieval module 313.Wherein, the first estimation module 311, for the information in processing time of processing described pending task according to stored each computing node about described calculating cluster and spending, is added up the described pending task minimum treat time separately.First searches module 312 for from described pending task, finds out the task that its minimum treat time is greater than the first assign thresholds.The first retrieval module 313 is for according to the stored information of processing the processing time that each task spends about the computing node that sends described message, from described task of searching, retrieve the task that it belongs to described specified type and sends computing node its processing time minimum being spent with pre-treatment of described message, as described definite task.
Wherein, described the first assign thresholds can be to determine minimum treat time based on described pending task.
In another specific implementation, as shown in Figure 3 C, decision-making module 310 can comprise that the second statistical module 315, second searches module 316 and the second retrieval module 317.The second statistical module 315, for the information in processing time of processing described pending task according to stored each computing node about described calculating cluster and spending, is added up the described pending task minimum treat time separately.Second searches module 316 for from described pending task, finds out it and belongs to the task that described specified type and its minimum treat time are greater than the second assign thresholds.The second retrieval module 317 is for according to the stored information of processing the processing time that each task spends about the computing node that sends described message, from described task of searching, retrieve the task of computing node its processing time minimum being spent with pre-treatment that sends described message, as described definite task.
Wherein, described the second assign thresholds can be to determine the minimum treat time of the particular task based in described pending task, and wherein said particular task is the task of belonging to described specified type.
Wherein, device 300 can also comprise collection module 330, for receive the information of processing the processing time that certain task spends about it from the computing node of described calculating cluster after, and the information that storage receives.
Referring now to Fig. 4,, it shows according to the schematic diagram of the equipment for task scheduling of one embodiment of the invention.As shown in Figure 4, equipment 400 can comprise storer 410 and the processor 420 for stores executable instructions.Wherein, the executable instruction that processor 420 is stored according to storer 410, the performed operation of modules of actuating unit 300.
The embodiment of the present invention also provides a kind of computing node computer-readable recording medium, and stores executable instructions on it, in the time that this executable instruction is performed, makes machine carry out the performed operation of processor 420.
It will be appreciated by those skilled in the art that disclosed each embodiment can make various changes and modifications in the situation that not departing from invention essence above.Therefore, protection scope of the present invention should be limited by appending claims.
Claims (12)
1. for a method for task scheduling, comprising:
Distribute while thering is the message of task of specified type when receiving request that the computing node that calculates in cluster sends, process the information in the processing time that each task spends according to the computing node of the stored described message of transmission, from least a portion task of pending task, determine belonging to described specified type and sending computing node its processing time minimum being spent with pre-treatment of described message of task; And
Described definite task being distributed to the computing node of the described message of transmission processes.
2. the method for claim 1, wherein described determining step comprises:
The information in the processing time spending according to the described pending task of each computing node processing in stored described calculating cluster, adds up the described pending task minimum treat time separately;
From described pending task, find out the task that its minimum treat time is greater than the first assign thresholds; And
Process the information in the processing time that each task spends according to the computing node of the stored described message of transmission, from described task of searching, retrieve the task that it belongs to described specified type and sends computing node its processing time minimum being spent with pre-treatment of described message, as described definite task.
3. method as claimed in claim 2, wherein,
Described the first assign thresholds is to determine the minimum treat time based on described pending task.
4. the method for claim 1, wherein described determining step comprises:
The information in the processing time spending according to the described pending task of each computing node processing in stored described calculating cluster, adds up the described pending task minimum treat time separately;
From described pending task, find out and belong to the task that described specified type and its minimum treat time are greater than the second assign thresholds; And
Process the information in the processing time that each task spends according to the computing node of the stored described message of transmission, from described task of searching, retrieve the task of computing node its processing time minimum being spent with pre-treatment that sends described message, as described definite task.
5. method as claimed in claim 4, wherein,
Described the second assign thresholds is to determine the minimum treat time of the particular task based in described pending task, and wherein said particular task belongs to the task of described specified type.
6. the method as described in any one in claim 1-5, wherein, also comprises:
After the computing node from described calculating cluster receives the information in the processing time spending about task of its processing, the information that storage receives.
7. for a device for task scheduling, comprising:
Decision-making module, for calculating request distribution that the computing node of cluster sends while thering is the message of task of specified type when receiving, process the information in the processing time that each task spends according to the computing node of the stored described message of transmission, from least a portion task of pending task, determine belonging to described specified type and sending computing node its processing time minimum being spent with pre-treatment of described message of task; And
Distribution module, the computing node that sends described message for described definite task is distributed to is processed.
8. device as claimed in claim 7, wherein, described decision-making module comprises:
The first statistical module, for process the information in the processing time that described pending task spends according to each computing node of stored described calculating cluster, adds up the described pending task minimum treat time separately;
First searches module, for from described pending task, finds out the task that the minimum treat time is greater than the first assign thresholds; And
The first retrieval module, for process the information in the processing time that each task spends according to the computing node of the stored described message of transmission, from described task of searching, retrieve the task of computing node its processing time minimum being spent with pre-treatment that belongs to described specified type and send described message, as described definite task.
9. device as claimed in claim 8, wherein,
Described the first assign thresholds is to determine the minimum treat time based on described pending task.
10. device as claimed in claim 7, wherein, described decision-making module comprises:
The second statistical module, for process the information in the processing time that described pending task spends according to each computing node of stored described calculating cluster, adds up the described pending task minimum treat time separately;
Second searches module, for from described pending task, finds out and belongs to the task that described specified type and its minimum treat time are greater than the second assign thresholds; And
The second retrieval module, for process the information in the processing time that each task spends according to the computing node of the stored described message of transmission, from described task of searching, retrieve the task of computing node its processing time minimum being spent with pre-treatment that sends described message, as described definite task.
11. devices as claimed in claim 10, wherein,
Described the second assign thresholds is to determine the minimum treat time of the particular task based in described pending task, and wherein said particular task belongs to the task of described specified type.
12. devices as described in any one in claim 7-11, wherein, also comprise:
Collection module, for receive the information in the processing time spending about task of its processing from the computing node of described calculating cluster after, the information that storage receives.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310109076.3A CN104077188A (en) | 2013-03-29 | 2013-03-29 | Method and device for scheduling tasks |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310109076.3A CN104077188A (en) | 2013-03-29 | 2013-03-29 | Method and device for scheduling tasks |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104077188A true CN104077188A (en) | 2014-10-01 |
Family
ID=51598458
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310109076.3A Pending CN104077188A (en) | 2013-03-29 | 2013-03-29 | Method and device for scheduling tasks |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104077188A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106600220A (en) * | 2016-11-29 | 2017-04-26 | 叶飞 | Distributed calculation method |
| CN107357639A (en) * | 2016-05-09 | 2017-11-17 | 腾讯科技（深圳）有限公司 | A kind of distributed processing system(DPS), the method and apparatus of data processing |
| CN107613025A (en) * | 2017-10-31 | 2018-01-19 | 武汉光迅科技股份有限公司 | A kind of implementation method replied based on message queue order and device |
| CN108804378A (en) * | 2018-05-29 | 2018-11-13 | 郑州易通众联电子科技有限公司 | A kind of And Methods of Computer Date Processing and system |
| CN109144697A (en) * | 2018-08-30 | 2019-01-04 | 百度在线网络技术（北京）有限公司 | A kind of method for scheduling task, device, electronic equipment and storage medium |
| CN110197314A (en) * | 2018-02-27 | 2019-09-03 | 北京京东尚科信息技术有限公司 | A kind of dispatching method and device |
| WO2020119117A1 (en) * | 2018-12-14 | 2020-06-18 | 平安医疗健康管理股份有限公司 | Distributed computing method, apparatus and system, device and readable storage medium |
| CN114697072A (en) * | 2022-02-18 | 2022-07-01 | 广州理工学院 | Cloud desktop unified operation and maintenance control system and control method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
| CN102096602A (en) * | 2009-12-15 | 2011-06-15 | 中国移动通信集团公司 | Task scheduling method, and system and equipment thereof |
| CN102393839A (en) * | 2011-11-30 | 2012-03-28 | 中国工商银行股份有限公司 | Parallel data processing system and method |
-
2013
- 2013-03-29 CN CN201310109076.3A patent/CN104077188A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102096602A (en) * | 2009-12-15 | 2011-06-15 | 中国移动通信集团公司 | Task scheduling method, and system and equipment thereof |
| 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 |
| CN102393839A (en) * | 2011-11-30 | 2012-03-28 | 中国工商银行股份有限公司 | Parallel data processing system and method |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107357639A (en) * | 2016-05-09 | 2017-11-17 | 腾讯科技（深圳）有限公司 | A kind of distributed processing system(DPS), the method and apparatus of data processing |
| CN107357639B (en) * | 2016-05-09 | 2019-09-17 | 腾讯科技（深圳）有限公司 | A kind of distributed processing system(DPS), data processing method and apparatus |
| US10432455B2 (en) | 2016-05-09 | 2019-10-01 | Tencent Technology (Shenzhen) Company Limited | Distributed processing system, data processing method, and control node device |
| CN106600220A (en) * | 2016-11-29 | 2017-04-26 | 叶飞 | Distributed calculation method |
| CN107613025A (en) * | 2017-10-31 | 2018-01-19 | 武汉光迅科技股份有限公司 | A kind of implementation method replied based on message queue order and device |
| CN110197314A (en) * | 2018-02-27 | 2019-09-03 | 北京京东尚科信息技术有限公司 | A kind of dispatching method and device |
| CN108804378A (en) * | 2018-05-29 | 2018-11-13 | 郑州易通众联电子科技有限公司 | A kind of And Methods of Computer Date Processing and system |
| CN109144697A (en) * | 2018-08-30 | 2019-01-04 | 百度在线网络技术（北京）有限公司 | A kind of method for scheduling task, device, electronic equipment and storage medium |
| CN109144697B (en) * | 2018-08-30 | 2021-03-09 | 百度在线网络技术（北京）有限公司 | Task scheduling method and device, electronic equipment and storage medium |
| WO2020119117A1 (en) * | 2018-12-14 | 2020-06-18 | 平安医疗健康管理股份有限公司 | Distributed computing method, apparatus and system, device and readable storage medium |
| CN114697072A (en) * | 2022-02-18 | 2022-07-01 | 广州理工学院 | Cloud desktop unified operation and maintenance control system and control method |
| CN114697072B (en) * | 2022-02-18 | 2023-10-31 | 广州理工学院 | Cloud desktop unified operation and maintenance control system and control method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104077188A (en) | Method and device for scheduling tasks | |
| CA2897338C (en) | Data stream splitting for low-latency data access | |
| CN108924250B (en) | Service request processing method and device based on block chain and computer equipment | |
| US9378053B2 (en) | Generating map task output with version information during map task execution and executing reduce tasks using the output including version information | |
| US8903981B2 (en) | Method and system for achieving better efficiency in a client grid using node resource usage and tracking | |
| US20160188376A1 (en) | Push/Pull Parallelization for Elasticity and Load Balance in Distributed Stream Processing Engines | |
| US20140201753A1 (en) | Scheduling mapreduce jobs in a cluster of dynamically available servers | |
| CN104731595A (en) | Big-data-analysis-oriented mixing computing system | |
| JP2012079242A (en) | Composite event distribution device, composite event distribution method and composite event distribution program | |
| CN111258978B (en) | Data storage method | |
| CN103164253A (en) | Virtual machine deployment system and virtual machine deployment method | |
| KR20150062596A (en) | System and method for distributing big data | |
| US11221890B2 (en) | Systems and methods for dynamic partitioning in distributed environments | |
| CN106874067B (en) | Parallel computing method, device and system based on lightweight virtual machine | |
| US20170371892A1 (en) | Systems and methods for dynamic partitioning in distributed environments | |
| CN111831713A (en) | Data processing method, device and equipment | |
| CN104468710A (en) | Mixed big data processing system and method | |
| CN104281636A (en) | Concurrent distributed processing method for mass report data | |
| KR101029416B1 (en) | Ranking data system, ranking query system and ranking computation method for computing large scale ranking in real time | |
| CN111061557B (en) | Method and device for balancing distributed memory database load | |
| US20230300086A1 (en) | On-demand resource capacity in a serverless function-as-a-service infrastructure | |
| EP2765517B1 (en) | Data stream splitting for low-latency data access | |
| Mirtaheri et al. | Optimized load balancing in high‐performance computing for big data analytics | |
| Pechenkin et al. | Architecture of a scalable system of fuzzing network protocols on a multiprocessor cluster | |
| CN112541038A (en) | Time series data management method, system, computing device and storage medium |
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 |
Application publication date: 20141001 |
|
| RJ01 | Rejection of invention patent application after publication |