CN101038559A - Batch task scheduling engine and dispatching method - Google Patents
Batch task scheduling engine and dispatching method Download PDFInfo
- Publication number
- CN101038559A CN101038559A CN 200610126976 CN200610126976A CN101038559A CN 101038559 A CN101038559 A CN 101038559A CN 200610126976 CN200610126976 CN 200610126976 CN 200610126976 A CN200610126976 A CN 200610126976A CN 101038559 A CN101038559 A CN 101038559A
- Authority
- CN
- China
- Prior art keywords
- task
- batch
- weights
- scheduling
- subregion
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Disclosed is a batch task dispatching engine and a dispatching method, the dispatching method comprises the steps of: implementing a static characteristic analysis of the present batch task to obtain the independent relationship between the monomer characteristic of the tasks represented by parameters and the tasks; subgrouping the tasks in terms of the independent relationship between the tasks; computing the group authority of each group task in terms of the authority of each task, in which the group authority determines the main dispatching sequence; determining the secondary dispatching sequence of each subarea of each group; establishing the dispatching scheme of the present dispatching task in terms of the main dispatching sequence and the secondary dispatching sequence; executing the dispatching of the present dispatching task by using the scheme. The invention can realize the intensive utilization of the time resource and hardware resource at the utmost by using the batch processing, and the flexibility, maintenance and fault toleration of the batch processing can be improved in the maximum.
Description
Technical field
The present invention relates to database batch processing task scheduling technical field, particularly relate to the large database batch task scheduling technology of multitask, multiprocessor, batch processing The limited time.
Background technology
Along with each application to database technology be extensive use of and that company information is integrated is deep day by day, on the one hand, the data of enterprise information system are concentrated day by day, and increase rapidly, also emerge in an endless stream for batch machining, the statistical study demand of data; On the other hand, for guaranteeing normally carrying out of daily online business, will be limited in the limited time range in the processing time in batches again, and hardware resource has also carried out rigid restriction to resource occupation in batches.
Traditional batch task scheduling mode, be that different pieces of information processing, statistics task in the batch are carried out successively according to fixing order, new processing demands of every increase, just write a new task, add in the back, the a certain task handling program of perhaps clear-cut directly modification increases new processing code.Simultaneously, in order to improve the utilization rate of resource, for each task, as much as possible according to the concurrent execution of data partition condition (as: area, country, product classification etc.).
Fig. 9 is the structured flowchart that a kind of typical case of conventional batch task scheduling engine realizes.
As shown in Figure 9, batch task scheduling main control module 11 is software modules, be responsible for the logical process of scheduling, it is according to the batch play of user's appointment, each batch tasks module in this play of sequence call, and operation result recorded task instances log recording apparatus 13, but if certain task module concurrent scheduling wherein, then obtain the concurrent condition of data partition (such as a group of area, country or product classification etc.) of this play task by partitioned parameters pen recorder 12, press this batch tasks module of subregion concurrent invocation then, and the record result.
Partitioned parameters pen recorder 12 is one or one group of database table, and its attribute generally comprises: subregion sign (unique), partition name have ground area code, regional title or country code, name of the country etc. such as commonly used.
Task instances log recording apparatus 13 is one or one group of database table, and its attribute generally comprises: task identification, task names, successful failure flags, start time, deadline etc.
Figure 10 is a kind of typical workflow synoptic diagram of conventional batch task scheduling engine, but is that the concurrent scheduling module is an example with batch tasks module 1, has demonstrated the flow process that batch tasks module 1-n is scheduled, and each step is described below:
Step 200: bulk management person imports the concurrent process number.
Step 201: bulk management person confirms by this batch of this strategy execution.
Step 202: the concurrent partitioned parameters that from partitioned parameters pen recorder 12, obtains this batch.
Step 203: the concurrent process number according to step 200 input, start one group of treatment progress, each process according to a still untreated subregion, is called batch tasks module 1 respectively, and outcome record is taken office pragmatic routine log recording apparatus 13.
Step 204: whether determination step batch tasks module 1 runs succeeded, if success then carry out 205 steps, otherwise interrupt carrying out finishing scheduling in batches.
Step 205: because batch tasks module 2 can not concurrent scheduling, so process of startup is called batch tasks module 2, and outcome record is taken office pragmatic routine log recording apparatus 13.
Step 206: whether determination step batch tasks module 2 runs succeeded, if success then carry out 207 steps, otherwise interrupt carrying out finishing scheduling in batches.
Step 207,208: as 203,204 steps (but module of concurrent scheduling) or 205,206 steps the module of concurrent scheduling (can not), order be carried out all batch tasks modules, finishing scheduling.
The needs that this quasi-tradition and typical batch task scheduling mode all can not adaptive technique develop with application in all many-sides such as resource utilization, fault-tolerance, dirigibility, maintainabilitys.
Summary of the invention
The present invention has overcome the deficiency in the existing batch task scheduling engine, realized a kind of parametrization of concrete batch tasks, intelligent batch task scheduling engine of being totally independent of, the scheduling strategy of task has been decided by the analysis of the time of running to task static nature and behavioral characteristics fully.
In one aspect of the invention, provide a kind of batch task scheduling method, comprised step: the batch tasks to current play is carried out the static nature analysis, to obtain with the monomer feature of each task of parametric representation and the dependence between the task; Based on the dependence between each task task is divided into groups; According to the weights of each task, calculate the group weights of each group task, wherein organize weights and determine the master scheduling order; According to the subregion weights, determine the secondary dispatching sequence of the dispatching sequence of each subregion in the group; Based on described master scheduling order and described secondary dispatching sequence, form the scheduling strategy of the batch tasks of current play; And utilize described scheduling strategy, carry out the scheduling of the batch tasks of current play.
In addition, according to an embodiment of the invention, this batch task scheduling method also comprises step: according to the execution result of each task at each subregion, calculate the new task weights when dispatching next time; And the described new task weights that write down each task.
In addition, according to an embodiment of the invention, this batch task scheduling method also comprises step: according to the average execution result of all tasks of each subregion, and the new subregion weights when calculating the next time scheduling; And the described new subregion weights that write down each subregion.
In addition, according to an embodiment of the invention, this batch task scheduling method also comprises step: if certain task is in commission made mistakes, then cancel the execution that all and this task have the subtask of dependence, other tasks then continue to carry out.
In addition, according to an embodiment of the invention, described group of weights are to obtain by the weights addition that will organize interior task.
In another aspect of this invention, a kind of batch task scheduling engine is provided, comprise: the static nature pen recorder, carry out static analysis to the batch tasks of current play, to obtain with the monomer feature of each task of parametric representation and the dependence between the task; The behavioral characteristics pen recorder divides into groups to task based on the dependence between each task, according to the weights of each task, calculate the group weights of each group task, wherein organize weights and represent the master scheduling order,, determine the secondary dispatching sequence of the dispatching sequence of each subregion in the expression group according to the subregion weights; And main control unit, based on described master scheduling order and described secondary dispatching sequence, form the scheduling strategy of the batch tasks of current play, utilize described scheduling strategy, carry out the scheduling of the batch tasks of current play.
In addition, according to an embodiment of the invention, this batch task scheduling engine also comprises: monomer characteristic storage device is used to store the static nature of each task; The dependence memory storage is used to store the dependence between each task; Subregion weights memory storage is used to store the weights of each subregion; And task weights memory storage, be used to store the weights of each task.
In addition, according to an embodiment of the invention, this batch task scheduling engine also comprises: log recording apparatus, be used to store the actual schedule situation of all tasks, and after whole batch task scheduling finished, record in the task weights memory storage according to the new task weights of each task when dispatching the next time that the execution result of each subregion calculates, and the new subregion weights of storage when next time of calculating according to the average execution result of all tasks of each subregion dispatching record in the subregion weights pen recorder.
This engine improves the dirigibility and the maintainability of engine by the atomization of task and the parametrization of static call feature; By analysis to dependency tiee between task, limit the influence of local atomic task fault, improve fault-tolerance; Allow task level, 2 dimension parallel processings of data partition level (abbreviation partition level), promote the handling capacity and the resource utilization of scheduling to greatest extent; By record and the analysis to each scheduler task operation characteristic, the scheduling strategy of Automatic Optimal execution next time is realized the intelligent task scheduling of saturation type.
The present invention is by the realization of above characteristic, make the intensification utilization of the maximized realization of batch processing task energy to time resource and hardware resource, promote dirigibility, maintainability and the fault-tolerant ability of batch processing to greatest extent, thereby improve load capacity, running efficiency and the vitality of whole enterprise information system greatly.
Parametrization provided by the invention, intelligent batch task scheduling engine, scheduling in batches and concrete batch tasks are separated from each other, the batch tasks of each atom level all is kept at the form of its static nature with parameter in the database, static nature comprises two parts of dependence between job order body characteristics (task play, program name, parameter mode etc.) and task, the parametrization of static nature makes batch tasks to customize fast in the mode of parameter, adapts to the fast-developing needs that change in business field; Also guaranteed the stability of engine self code, reduced and safeguarded and the test risk.
The present invention can be limited in fault in the subrange when certain atomic task breaks down by the analysis to dependency tiee, and does not influence whole operation in batches, and the ability of fault-tolerant operation is provided for system.
The present invention has realized a kind of parametrization of concrete batch tasks, intelligent batch task scheduling engine of being totally independent of, scheduling strategy to task is decided by the analysis of the time of running to task static nature and behavioral characteristics fully, shows outstanding advantage in the following areas:
1. resource utilization
The present invention uses task level, 2 dimensions of data partition level to carry out concurrent scheduling, and according to subregion weights, task weights generation optimal scheduling strategy, the weights tuning mechanism of saturation type is provided simultaneously, thereby promoted the handling capacity and the resource utilization of task scheduling to greatest extent, shortened the bulk treatment time as far as possible.According to the test of typical case, its resource utilization has promoted 32%, and the whole batch processing time has shortened 43%.
2. fault-tolerance
By the analysis to dependency tiee, the present invention can be limited in fault in the subrange when certain task atom breaks down, and does not influence whole operation in batches.If certain task broke down in the batch in the past, bulk management person is often in a rush, Yi Bian will work out emergency plan, analyze other batch tasks one by one and can handle in advance, understand the method for the manual operation of these batch tasks, will worry that also above analytic process has delayed the deadline in batches; Another side is the while analyzing failure cause again, deals with problems.
The ability of fault-tolerant operation provided by the present invention, the coverage of batch fault is dwindled to greatest extent, and guarantee the normal process of other onrelevant tasks, make bulk management person can be absorbed in trouble spot itself, if fault is got rid of, bulk management person can give task again engine at any time and continue to handle.
3. dirigibility
The present invention is decided by the automatic analysis of the time of running to task static nature and behavioral characteristics fully to the scheduling strategy of task, so have extraordinary dirigibility.The one side user only need change all kinds of characteristic parameters and can adjust scheduling strategy voluntarily; System also utilizes the tuning mode of self feed back on the other hand, adapts to the variation of situation automatically.
With a simple case: certain batch processing scene has a group task to need scheduling, the processing time of 11 CPU of this group task use is 4 hours at ordinary times, average resource is 90.9% therebetween, task A can not press the concurrent task of subregion as one of them, because the processing time is shorter, the last execution that is placed all tasks automatically; Performance issue appearred in the A task after certain version was gone into operation, and needed to carry out 4 hours, and day this task of going into operation still places the end of task queue by original strategy, so the whole batch processing time is about 8 hours, the average resource rapid drawdown is 50%; Go into operation next day, engine places the top of task queue by the analysis of task weights with this task, and the processing time tuning of whole batch tasks is 4.4 hours.
4. maintainable
The present invention is separated from each other scheduling in batches and concrete batch tasks, each other only with the static nature of task and behavioral characteristics as the parametrization interface, make batch tasks to customize fast in the mode of parameter, adapt to the fast-developing needs that change in business field, guaranteed the stability of engine self code, reduced and safeguarded and the test risk.
5. ease for use
The present invention provides friendly man-machine interface simultaneously, by all kinds of task handling states of color differentiating, provides the execution feature of each each specific tasks of play by abundant form, helps the maintainer to carry out performance evaluation.
Description of drawings
From the detailed description below in conjunction with accompanying drawing, above-mentioned feature and advantage of the present invention will be more obvious, wherein:
Fig. 1 system architecture synoptic diagram of the present invention;
Fig. 2 logical organization synoptic diagram of the present invention;
Fig. 3 workflow synoptic diagram of the present invention;
Fig. 4 task scheduling interactive interface of the present invention;
Fig. 5 shows the how regional exemplary resource when concurrent of certain task and utilizes situation;
Fig. 6 shows according to the batch task scheduling engine of embodiment of the present invention in the how regional utilization of resources situation when concurrent of certain task;
Fig. 7 shows the batch task scheduling engine that utilizes prior art and carries out multitask batch task scheduling scene;
Fig. 8 shows utilization and carries out multitask batch task scheduling scene according to the batch task scheduling engine of embodiment of the present invention;
The logical organization synoptic diagram of Fig. 9 conventional batch task scheduling engine;
The workflow synoptic diagram of Figure 10 conventional batch task scheduling engine.
Embodiment
Below, describe preferred implementation of the present invention with reference to the accompanying drawings in detail.In the accompanying drawings, though be shown in the different accompanying drawings, identical Reference numeral is used to represent identical or similar assembly.For clarity and conciseness, be included in here known function and the detailed description of structure will be omitted, otherwise they will make theme of the present invention unclear.
Fig. 1 is a system chart of the present invention, and it has represented the general structure of a typical enterprise information system.
As shown in Figure 1, user terminal apparatus 1 is the client of system, can be an individual PC, enterprise information system for B/S (Browser/Server) structure, browser software should be installed, it also can be the device that other can operate browser software, as NC, Windows graphic terminal etc.; For the enterprise information system of C/S (Client/Server) structure, the client software of this system should be installed.In addition, it also needs to possess display device and input media, and input media can be keyboard and mouse.
Transaction Processing engine 2 is one or one group of software module, is used for the request of sending according to user terminal apparatus 1, is distributed to corresponding Transaction Processing module 3.For the enterprise information system of B/S structure, it is installed on the WEB/ application server (can be PC server or minicomputer etc.); For the enterprise information system of C/S structure, it generally is installed on the application server (can be PC server or minicomputer etc.).
Transaction Processing module (1...n) 3 is one group of software module, is used to respond the Transaction Processing request that the Transaction Processing engine sent in 2 minutes, and access service data storage device 10 is finished such as issued transaction such as newly-increased, renewal, deletion, inquiries.Its general and Transaction Processing engine 2 is installed on the same physical equipment.
External data source apparatus 4 can be the data storage device (for example database, exchanges data tape, CD etc.) of another infosystem or be the data file bag that another infosystem is come by Network Transmission, it also can be a plant information monitoring collecting device, be used to data capture engine 5 that external data information is provided, it also is the external data interface of whole enterprise information system.
Data capture engine 5 is one or one group of software module, is used to gather external data source, and calls service data acquisition module (1...n) 6, finishes collection, conversion, filtration, the refinement of external data and handles.Its general and Transaction Processing engine 2 is installed on the same physical equipment.
Service data acquisition module (1...n) 6 is one group of software module, is used for the service data acquisition request that the response data acquisition engine sent in 5 minutes, and access service data storage device 10 is finished such as processing such as data acquisition, conversion, filtration, refinements.Its general and data capture engine 5 is installed on the same physical equipment.
Administrator terminal device 7 is clients of system's batch processing, can be an individual PC, and the batch client software of this enterprise information system should be installed.It also needs to possess display device and input media, and input media can be keyboard and mouse.
Batch task scheduling engine 8 is subsystem module, it is the present invention residing position in whole enterprise information system, be used for the request of sending according to administrator terminal device 8, scheduling is finished the batch processing work of system with relatively-stationary one group of batch tasks module (1...n) 9.It generally is positioned on the database server.
Batch tasks module (1...n) 9 is one group of software module, be used to respond the batch tasks processing request that batch task scheduling engine sent in 8 minutes, access service data storage device 10 is finished such as data statistics, sampling, analysis, processing, generation form, large data work of treatment such as data externally is provided.Its general and batch task scheduling engine 8 is installed on the same physical equipment.
Business datum memory storage 10 can be a PC server, minicomputer or main frame, the service data base management system, be responsible for storage and business datum is provided, it generally obtains various external datas from professional data acquisition module (1...n) 6, and obtain all kinds of online business datums from Transaction Processing module (1...n) 3, and finish the statistics of business datum by batch tasks module (1...n) 9, adopt, analyze, processing, generate works of treatment such as form, and by forms such as report file or online displayings, with the raw data before handling or the derivation information after handling is interactively offers the user.
More than corresponding physical equipment room between each system module, local area network communication by enterprise, can be Ethernet (Ethernet), it also can be other LAN (Local Area Network), as Fiber Distributed Data Interface (FDDI), token ring (Token-Ring) etc., can also the LAN (Local Area Network) of its each branch offices be connected to form bigger intranet (Intranet) by renting modes such as special line in addition.
More than each system module may reduce according to the actual needs of enterprise information system, for example non-industrial information system, external data source is not necessarily arranged, thereby external data source apparatus 4, data capture engine 5, service data acquisition module (1...n) 6 can be reduced.
Fig. 2 is the parametrization according to embodiment of the present invention, the structured flowchart of intelligent batch task scheduling engine.
As shown in Figure 2, batch task scheduling main control module 81 is a software module, be responsible for the logical process of scheduling, it is according to the batch play of user's appointment, obtain the static nature that this play relates to task by task static nature pen recorder 82, by the behavioral characteristics of task behavioral characteristics pen recorder 83 analysis task, determine optimum scheduling strategy again.Finish the work after the scheduling by strategy, batch task scheduling module 81 records this task executions feature in the task behavioral characteristics pen recorder 83 again, prepares for the policy optimization of next time carries out data.For each being finished of task and subregion, batch task scheduling module 81 records task instances log recording apparatus 88 with execution result.
Task static nature pen recorder 82 is responsible for obtaining the static nature data from job order body characteristics pen recorder 84 and task dependence pen recorder 85.
Task behavioral characteristics pen recorder 83 is responsible for obtaining the behavioral characteristics data from subregion weights pen recorder 86 and task weights pen recorder 87.
Job order body characteristics pen recorder 84 is one or one group of database table, is responsible for preserving and providing the monomer feature of task, and its main attribute comprises:
The batch tasks numbering: but unique identification marking (character string type) that batch task scheduling engine distributes for each batch tasks;
Task names: the Chinese of this batch tasks is called for short;
The batch play: identify this batch tasks and belong to what play in batches, play is generally according to divisions such as execution time (morning, noon, evening etc.) in batches, significance levels in batches, and the play that can adjust each task at any time belongs to;
Batch storing process: the batch tasks module entry program name of this batch tasks correspondence;
The parameter transfer mode: this batch tasks is imported the enterprise schema of parameter into, such as, first parameter is the batch date, second parameter is ground area code etc., generally can stipulate some kinds of parameter mode standards by batch task scheduling engine, requires each batch tasks processing module to select to use, batch task scheduling engine can be according to " storing process in batches " and this parameter, generate complete batch tasks module entry program call instruction string, such as: call program name (batch date, ground area code).
Whether permission is parallel: whether identify this batch tasks by the parallel scheduling of data partition condition.
Whether allow to reform: identify this batch tasks and run into external environment condition reason (full, hardware power down etc.) and have no progeny in causing, can continue to continue down to do without artificial treatment from breakpoint such as disk space.
Task dependence pen recorder 85 is one or one group of database table, be responsible for to preserve and provide dependency tiee between task, can divide into groups to task in view of the above, delivers to task behavioral characteristics pen recorder 83 calculating group weights; When certain atomic task breaks down in the batch task scheduling main control module 81, can utilize the dependence data, fault is limited in the subrange, its main attribute comprises:
Sub-batch tasks numbering: certain bar on the sign dependency tiee concern that the child node in the line numbers, and constitutes main external key incidence relation with the batch tasks number field of job order body characteristics pen recorder 84.
Father's batch tasks numbering: certain bar on the sign dependency tiee concern that the father node in the line numbers, and constitutes main external key incidence relation with the batch tasks number field of job order body characteristics pen recorder 84.
Subregion weights pen recorder 86 is one or one group of database table, is responsible for preserving and subregion weights information being provided, and can determine the subregion dispatching sequence of each task inside in view of the above; After every batch is finished, the batch task scheduling main control module is according to the average execution time of each all tasks of subregion, the subregion weights when adjusting the next time scheduling automatically, for long subregion of execution time when dispatching next time, strive for more priority scheduling power, its main attribute comprises:
The batch play: the present invention allows for each different subregion weight of play definition in batches, so can specify the corresponding play in batches of this subregion weights at this;
The subregion sign: the sign of subregion will pass to as parameter and allow concurrent batch tasks processing module;
Partition name: the abbreviation of subregion;
The subregion weight: the weighted value of subregion (abbreviation weights), weights will have more priority scheduling power than bigdos when next batch task scheduling.This parameter can be after every play batch tasks be all over, and is arranged to average deadline of all tasks of this subregion automatically by batch task scheduling main control module 81, but also manual setting, to change the subregion preference strategy of scheduling next time.This parameter is initially set to maximal value 999999, promptly for the subregion of carrying out batch tasks first, under the situation of unknown its time cost, supposes its cost for infinitely great, the override scheduling.
Task weights pen recorder 87 is one or one group of database table, be responsible for preserving and providing task weights information, (complementary task constitutes a task grouping in conjunction with the determined task grouping of the dependence between task, the task of no dependence has one group of one's own), can be with the weights addition of each task in the group, obtain the task groups weights, then according to the dispatching priority order between task groups weights decision task groups; By the dispatching priority order of each task in the weight decision task groups of the dependence in the group and each task, finally determine the dispatching sequence of this each task of batch.After every batch task scheduling is finished,, adjust the task weights when dispatching next time automatically according to the average implementation status of each task at each subregion.Its main attribute comprises:
Batch tasks numbering: constitute main external key incidence relation with the batch tasks number field of job order body characteristics pen recorder 84.
Task weight: the weighted value of task (abbreviation weights), the big task of weights will have more priority scheduling power when next lot size scheduling.This parameter can be after every play batch tasks be all over, and is arranged to average deadline of all tasks of this subregion automatically by batch task scheduling main control module 81, but also manual setting, to change the task priority strategy of scheduling next time.This parameter is initially set to maximal value 999999, promptly for the batch tasks of carrying out first, under the situation of unknown its time cost, supposes its cost for infinitely great, the override scheduling.
Task instances log recording apparatus 88 is one or one group of database table, is responsible for preserving and providing the actual implementation status of each play batch task scheduling processing, for inquire about, analyze use afterwards.The attribute of record mainly comprises in the task instances log recording apparatus 88:
Task identification, task names, successful failure flags, start time, deadline etc.
The batch processing date: the date of the actual execution of batch processing task
Subregion sign: the subregion sign that the batch processing task is actual when being called
Batch tasks numbering: the unique identification of batch processing task constitutes main external key incidence relation with the batch tasks number field of job order body characteristics pen recorder 84.
Batch play: corresponding parameter value in the job order body characteristics pen recorder 84 of this play batch task scheduling strategy institute foundation (influence of all previous parameter variation to the batch task scheduling strategy is used to audit).
Batch storing process: corresponding parameter value in the job order body characteristics pen recorder 84 of this play batch task scheduling strategy institute foundation (influence of all previous parameter variation to the batch task scheduling strategy is used to audit).
Parameter transfer mode: corresponding parameter value in the job order body characteristics pen recorder 84 of this play batch task scheduling strategy institute foundation (influence of all previous parameter variation to the batch task scheduling strategy is used to audit).
The task packet number: batch task scheduling main control module 81 is according to the determined task group character of task dependence (unique in batches at every play, the dependence between all previous task of being used to audit changes the influence to the lot size scheduling strategy).
The task groups weight: batch task scheduling main control module 81 is according to the determined task groups of task dependence, and the weights of interior each task of task groups, the task groups weights of determining, the size of weights have identified the priority of this task groups in task scheduling strategy at that time (dependence between all previous task of being used to audit changes the influence to the lot size scheduling strategy).
Serial number in the group: batch task scheduling main control module 81 is according to the determined task groups of task dependence, and the weights of interior each task of task groups, dispatching sequence number in the group of determining (at every play in batches, unique in the grouping of each task, the dependence between all previous task of being used to audit changes the influence to the lot size scheduling strategy).
Task weight: respective weights value in the task weights pen recorder 87 of this play lot size scheduling strategy institute foundation (influence of all previous weights variation to the lot size scheduling strategy is used to audit).
Subregion weight: respective weights value in the subregion weights pen recorder 86 of this play lot size scheduling strategy institute foundation (influence of all previous weights variation to the lot size scheduling strategy is used to audit).
The operation treatment state: this batch processing task executions result, the present invention have stipulated unified operation treatment state code for it, comprise be untreated, handle, success, cancel, make mistakes (can reform), make mistakes (can not reform)
Database session number: the database session of correspondence number when this batch processing task is carried out.
Operating system process number: relative operation system process number when this batch processing task is carried out.
Zero-time: the time that this batch tasks begins to carry out.
Deadline: this batch tasks finishes the time of execution.
Batch operation person: this batch tasks is carried out constantly, corresponding batch operation person number.
Batch tasks module return code: the return code of batch tasks module self (0 identifies successfully, all the other can by the layout voluntarily of batch tasks module)
Batch tasks module return message: the return message of batch tasks module, mainly return error reason etc. when makeing mistakes.
Fig. 3 shows the process flow diagram according to the batch task scheduling method of embodiment of the present invention.As shown in Figure 3:
Step 100: batch task scheduling main control module 81 access task static nature pen recorders 82, carry out the static nature analysis.
Step 101: in task static nature pen recorder 82, obtain the job order body characteristics that this play need be carried out in batches, determine that this needs to carry out which batch tasks module in batches, these batch tasks modules are used information such as what entry program call instruction string execution again.
Step 102: in task static nature pen recorder 82, according to the dependence between task to carrying out of task is divided into groups (complementary task constitutes a task grouping, and the task of no dependence has one group of one's own).
Step 103: batch task scheduling main control module 81 access task behavioral characteristics pen recorders 83, carry out dynamic Feature Analysis.
Step 104: in task behavioral characteristics pen recorder 83, obtain the weights information of each atomic task, the weights addition with each task in each task groups obtains the task groups weights, then according to the dispatching priority order between task groups weights decision task groups; By the dispatching priority order of each task in the weight decision task groups of the dependence in the group and each task, finally determine the dispatching sequence of this each task of batch.
Step 105: in the dynamic special representative record device of task, according to the subregion weights, the dispatching sequence of inner each subregion that sets the tasks.
Step 106: batch task scheduling main control module 81 generates these plays full schedule strategy in batches, but the i.e. equal concurrent scheduling of each task groups, by task groups weights decision priority; The necessary sequential scheduling of task in each task groups is by what task weights decision priority of dependence; But the subregion concurrent scheduling in each task is by subregion weights decision priority.
Step 107: bulk management person imports concurrent process number (being generally CPU number-2).
Step 108: bulk management person confirms by this batch of this strategy execution.
Step 109: the concurrent process number according to input, start one group of treatment progress, each process is selected next executable task and subregion according to scheduling strategy respectively.
Step 110: whether determination step 109 exists task to need to carry out, if exist then carry out 111 steps, otherwise forwards 114 steps to;
Step 111: actually execute the task and write down execution result and take office in the pragmatic routine log recording apparatus 88.
Step 112: whether the judgement task runs succeeded, if successfully turn to step 109, otherwise turns to step 113.
Step 113: cancel this task of makeing mistakes relevant subtask, and turn to step 109.
Step 114: after this play batch tasks is all finished,, calculate the task weights when dispatching next time according to the average implementation status of each task at all subregions.
Step 115:, calculate the subregion weights when dispatching next time according to the average implementation status of each all task of subregion.
Step 116: if having cancellation or makeing mistakes of task, bulk management person can select part or all of task to carry out once more, has this generic task to need to carry out again, then forwards step 117 to, otherwise finishes.
Step 117: specify the cancellation and the task of makeing mistakes that need execution, forward step 109 to.
Fig. 4 shows task scheduling interactive interface of the present invention.As Fig. 4, wherein with different color shows one group task by the sight of task, two dimension concurrent schedulings of subregion, wherein black is represented pending task or subregion, orange representative is handled, and (the red mistake of representing common fault, brown representative to need manual intervention, grey representative cancellation are arranged in addition) handled successfully in green representative.
Fig. 5 illustrates the how regional exemplary resource when concurrent of certain task of knowing clearly and utilizes situation (Y-axis is a utilization of resources number percent), as can be seen because the difference of data volume, the each department task processing time there are differences, if task according to random fashion scheduling each department, some big area has been come when carrying out after the little area, other areas can occur finishes, the situation in only etc. several big areas, so resource utilization is low more more to the end, know that easily the level of resources utilization only has 60% after calculating by Fig. 5.
As mentioned above, according to the embodiment of the present invention, if set up weights (area to be called regional weights the scale in area, subregion is called the subregion weights), carry out sequential scheduling according to the weights size during scheduling, then can produce the effect that big or small task pairing is carried out, thereby improve concurrency.Utilization of resources situation when Fig. 6 shows and executes the task according to the batch task scheduling engine of embodiment of the present invention.Be illustrated in fig. 6 shown below, its resource utilization has brought up to 85%, and the task deadline has shortened 25%.
In addition, traditional task scheduling engine only provides by data partition condition (as: country, area, product classification etc., below still be example with the area) and carries out the one dimension concurrent scheduling, and does not support concurrent between each task, be to remain serial between the task, the utilization factor of resource is not high like this.
Fig. 7 shows the batch task scheduling engine that utilizes prior art and carries out multitask batch task scheduling scene.Because the B task must wait until that last area of A task is complete, even at this moment have only a CPU to be used, also can only wait for; And the D task be one can not be by the concurrent task in area, its resource utilization will be lower, the resource utilization of whole batch task scheduling only has 56.7%, even introduce subregion weights mechanism, estimates also only to rise to about 80% according to the empirical value of front case.
But, according to the embodiment of the present invention,, introduce the concurrent ability of task dimension under the prerequisite of dependence, make the level of resources utilization of engine obtain significantly to promote.Fig. 8 shows utilization and carries out multitask batch task scheduling scene according to the batch task scheduling engine of embodiment of the present invention.As shown in Figure 8, under the situation of introducing subregion weights and task concurrent mechanism, its resource utilization can be near 90%.
Top description only is used to realize embodiments of the present invention; it should be appreciated by those skilled in the art; the any modification or partial replacement that is not departing from the scope of the present invention; all should belong to claim of the present invention and come restricted portion; therefore, protection scope of the present invention should be as the criterion with the protection domain of claims.
Claims (9)
1. batch task scheduling method comprises step:
Batch tasks to current play is carried out the static nature analysis, to obtain with the monomer feature of each task of parametric representation and the dependence between the task;
Based on the dependence between each task task is divided into groups;
According to the weights of each task, calculate the group weights of each group task, wherein organize weights and determine the master scheduling order;
According to the subregion weights, determine the secondary dispatching sequence of the dispatching sequence of each subregion in the group;
Based on described master scheduling order and described secondary dispatching sequence, form the scheduling strategy of the batch tasks of current play; And
Utilize described scheduling strategy, carry out the scheduling of the batch tasks of current play.
2. batch task scheduling method as claimed in claim 1 is characterized in that, also comprises step:
According to the execution result of each task, calculate the new task weights when dispatching next time at each subregion; And
Write down the described new task weights of each task.
3. batch task scheduling method as claimed in claim 2 is characterized in that, also comprises step:
According to all task executions results of each subregion, calculate the new subregion weights when dispatching next time; And
Write down the described new subregion weights of each subregion.
4. batch task scheduling method as claimed in claim 1 is characterized in that, also comprises step:
If certain task is in commission made mistakes, then cancel the execution that all and this task have the subtask of dependence, other tasks then continue to carry out.
5. batch task scheduling method as claimed in claim 1 is characterized in that, described group of weights are to obtain by the weights addition that will organize interior task.
6. batch task scheduling engine comprises:
The static nature pen recorder carries out static analysis to each batch tasks of current play, to obtain with the monomer feature of each task of parametric representation and the dependence between the task;
The behavioral characteristics pen recorder divides into groups to task based on the dependence between each task, according to the weights of each task, calculate the group weights of each group task, wherein organize weights and represent the master scheduling order,, determine the secondary dispatching sequence of the dispatching sequence of each subregion in the expression group according to the subregion weights; And
Main control unit based on described master scheduling order and described secondary dispatching sequence, forms the scheduling strategy of the batch tasks of current play, utilize described scheduling strategy, under the prerequisite of not violating dependence, concurrent with task, subregion two dimensions, carry out the scheduling of the batch tasks of current play.
7. batch task scheduling engine according to claim 6 is characterized in that, also comprises:
Monomer characteristic storage device is used to store the static nature of each task;
The dependence memory storage is used to store the dependence between each task;
Subregion weights memory storage is used to store the weights of each subregion; And
Task weights memory storage is used to store the weights of each task.
8. batch task scheduling engine as claimed in claim 6 is characterized in that, also comprises:
Log recording apparatus, be used to store the actual schedule situation of all tasks, and after whole batch task scheduling finished, record in the task weights memory storage according to the new task weights of each task when dispatching the next time that the execution result of each subregion calculates, and the new subregion weights when dispatching according to the next time that the average execution result of all tasks of each subregion calculates record in the subregion weights pen recorder.
9. batch task scheduling engine as claimed in claim 6 is characterized in that, described group of weights are to obtain by the weights addition that will organize interior task.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2006101269769A CN100444121C (en) | 2006-09-11 | 2006-09-11 | Batch task scheduling engine and dispatching method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2006101269769A CN100444121C (en) | 2006-09-11 | 2006-09-11 | Batch task scheduling engine and dispatching method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101038559A true CN101038559A (en) | 2007-09-19 |
| CN100444121C CN100444121C (en) | 2008-12-17 |
Family
ID=38889476
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2006101269769A Active CN100444121C (en) | 2006-09-11 | 2006-09-11 | Batch task scheduling engine and dispatching method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100444121C (en) |
Cited By (35)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101882161A (en) * | 2010-06-23 | 2010-11-10 | 中国工商银行股份有限公司 | Application level asynchronous task scheduling system and method |
| CN102004666A (en) * | 2010-11-25 | 2011-04-06 | 中国工商银行股份有限公司 | Large-scale computer operation scheduling equipment and system |
| CN102214236A (en) * | 2011-06-30 | 2011-10-12 | 北京新媒传信科技有限公司 | Method and system for processing mass data |
| CN102236580A (en) * | 2010-04-26 | 2011-11-09 | 阿里巴巴集团控股有限公司 | Method for distributing node to ETL (Extraction-Transformation-Loading) task and dispatching system |
| CN102609875A (en) * | 2012-01-20 | 2012-07-25 | 中国工商银行股份有限公司 | Method and system for concurrently and dynamically processing bank data |
| CN102929695A (en) * | 2012-09-18 | 2013-02-13 | 曙光信息产业(北京)有限公司 | Batch job scheduling method based on subscription |
| CN103365708A (en) * | 2012-04-06 | 2013-10-23 | 阿里巴巴集团控股有限公司 | Method and device for scheduling tasks |
| CN104598234A (en) * | 2015-01-13 | 2015-05-06 | 北京京东尚科信息技术有限公司 | Method and device for processing service |
| CN105573866A (en) * | 2009-07-14 | 2016-05-11 | 起元技术有限责任公司 | Method and system for fault tolerant batch processing |
| CN106055315A (en) * | 2016-05-19 | 2016-10-26 | 乐视控股（北京）有限公司 | Method and system for processing batched tasks |
| CN106293920A (en) * | 2016-08-15 | 2017-01-04 | 北京票之家科技有限公司 | Method for scheduling task and device |
| CN102722381B (en) * | 2011-03-25 | 2017-03-01 | 微软技术许可有限责任公司 | The technology of optimization and upgrading task |
| CN106648842A (en) * | 2016-12-27 | 2017-05-10 | 中国建设银行股份有限公司 | Method and device for continuous running of batch jobs |
| WO2017088189A1 (en) * | 2015-11-27 | 2017-06-01 | 华为技术有限公司 | User data processing method, related device and system |
| CN107370808A (en) * | 2017-07-13 | 2017-11-21 | 盐城工学院 | A kind of method for being used to carry out big data task distributed treatment |
| CN107506906A (en) * | 2017-08-02 | 2017-12-22 | 贵州工程应用技术学院 | A kind of task intelligent scheduling engine design method based on data processing |
| CN107665144A (en) * | 2016-07-29 | 2018-02-06 | 北京京东尚科信息技术有限公司 | The balance dispatching center of distributed task scheduling, mthods, systems and devices |
| CN107885600A (en) * | 2017-10-19 | 2018-04-06 | 用友网络科技股份有限公司 | Dispatching method, device and the computer equipment of batch timed task |
| CN108228329A (en) * | 2018-02-02 | 2018-06-29 | 观相科技（上海）有限公司 | A kind of packet scheduling system |
| CN108537543A (en) * | 2018-03-30 | 2018-09-14 | 百度在线网络技术（北京）有限公司 | Method for parallel processing, device, equipment and the storage medium of block chain data |
| CN108920261A (en) * | 2018-05-23 | 2018-11-30 | 中国航天***科学与工程研究院 | A kind of two-stage self-adapting dispatching method suitable for large-scale parallel data processing task |
| CN108958923A (en) * | 2017-05-25 | 2018-12-07 | 上海华为技术有限公司 | A kind of method and controlling equipment of scheduling hardware accelerator |
| CN109389286A (en) * | 2018-09-03 | 2019-02-26 | 阿里巴巴集团控股有限公司 | A kind of business processing, scheduling, batch processes, device, equipment and medium |
| CN109639653A (en) * | 2018-11-29 | 2019-04-16 | 中国人民银行清算总中心 | Message transmitting method and system based on distributed internet banking system |
| CN110083442A (en) * | 2019-03-18 | 2019-08-02 | 咪咕文化科技有限公司 | A kind of operating method of user task, server and can storage medium |
| CN110287018A (en) * | 2019-07-04 | 2019-09-27 | 中国工商银行股份有限公司 | Batch tasks method of combination and device |
| CN104991821B (en) * | 2015-06-29 | 2019-12-06 | 北京奇虎科技有限公司 | method and device for processing monitoring tasks in batches |
| CN110673939A (en) * | 2019-09-23 | 2020-01-10 | 汉纳森（厦门）数据股份有限公司 | Task scheduling method, device and medium based on airflow and yarn |
| CN110704170A (en) * | 2019-09-09 | 2020-01-17 | 深圳壹账通智能科技有限公司 | Batch task processing method and device, computer equipment and storage medium |
| CN110795219A (en) * | 2019-10-24 | 2020-02-14 | 华东计算技术研究所（中国电子科技集团公司第三十二研究所） | Resource scheduling method and system suitable for multiple computing frameworks |
| CN112035230A (en) * | 2020-09-01 | 2020-12-04 | 中国银行股份有限公司 | Method and device for generating task scheduling file and storage medium |
| CN112148491A (en) * | 2020-09-27 | 2020-12-29 | 京东数字科技控股股份有限公司 | Data processing method and device |
| CN112562114A (en) * | 2020-12-18 | 2021-03-26 | 北京百度网讯科技有限公司 | Data acquisition method and device |
| WO2022067531A1 (en) * | 2020-09-29 | 2022-04-07 | 深圳大学 | Computing resource aware task scheduling method |
| CN115373804A (en) * | 2022-10-27 | 2022-11-22 | 合肥综合性国家科学中心人工智能研究院(安徽省人工智能实验室) | Virtual machine scheduling method facing network test in cloud infrastructure |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104252496A (en) * | 2013-06-28 | 2014-12-31 | 国际商业机器公司 | Method and system for providing resource access |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7093250B1 (en) * | 2001-10-11 | 2006-08-15 | Ncr Corporation | Priority scheduler for database access |
| US20040244005A1 (en) * | 2003-05-30 | 2004-12-02 | Ancier Leland J. | Automatic urgency calculator and task scheduler |
| CN100345132C (en) * | 2003-07-28 | 2007-10-24 | 华为技术有限公司 | Parallel processing method and system |
| CN100377091C (en) * | 2006-03-16 | 2008-03-26 | 浙江大学 | Grouped hard realtime task dispatching method of built-in operation system |
-
2006
- 2006-09-11 CN CNB2006101269769A patent/CN100444121C/en active Active
Cited By (53)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105573866B (en) * | 2009-07-14 | 2018-11-13 | 起元技术有限责任公司 | The method and system of batch input data is handled with fault-tolerant way |
| CN105573866A (en) * | 2009-07-14 | 2016-05-11 | 起元技术有限责任公司 | Method and system for fault tolerant batch processing |
| CN102236580B (en) * | 2010-04-26 | 2013-03-20 | 阿里巴巴集团控股有限公司 | Method for distributing node to ETL (Extraction-Transformation-Loading) task and dispatching system |
| CN102236580A (en) * | 2010-04-26 | 2011-11-09 | 阿里巴巴集团控股有限公司 | Method for distributing node to ETL (Extraction-Transformation-Loading) task and dispatching system |
| CN101882161A (en) * | 2010-06-23 | 2010-11-10 | 中国工商银行股份有限公司 | Application level asynchronous task scheduling system and method |
| CN101882161B (en) * | 2010-06-23 | 2012-07-04 | 中国工商银行股份有限公司 | Application level asynchronous task scheduling system and method |
| CN102004666A (en) * | 2010-11-25 | 2011-04-06 | 中国工商银行股份有限公司 | Large-scale computer operation scheduling equipment and system |
| CN102722381B (en) * | 2011-03-25 | 2017-03-01 | 微软技术许可有限责任公司 | The technology of optimization and upgrading task |
| CN102214236A (en) * | 2011-06-30 | 2011-10-12 | 北京新媒传信科技有限公司 | Method and system for processing mass data |
| CN102214236B (en) * | 2011-06-30 | 2013-10-23 | 北京新媒传信科技有限公司 | Method and system for processing mass data |
| CN102609875A (en) * | 2012-01-20 | 2012-07-25 | 中国工商银行股份有限公司 | Method and system for concurrently and dynamically processing bank data |
| CN103365708A (en) * | 2012-04-06 | 2013-10-23 | 阿里巴巴集团控股有限公司 | Method and device for scheduling tasks |
| CN103365708B (en) * | 2012-04-06 | 2017-03-22 | 阿里巴巴集团控股有限公司 | Method and device for scheduling tasks |
| CN102929695B (en) * | 2012-09-18 | 2016-12-21 | 曙光信息产业(北京)有限公司 | A kind of batch jobs dispatching method based on reservation |
| CN102929695A (en) * | 2012-09-18 | 2013-02-13 | 曙光信息产业(北京)有限公司 | Batch job scheduling method based on subscription |
| CN104598234B (en) * | 2015-01-13 | 2018-10-30 | 北京京东尚科信息技术有限公司 | A kind of method and device of processing business |
| CN104598234A (en) * | 2015-01-13 | 2015-05-06 | 北京京东尚科信息技术有限公司 | Method and device for processing service |
| CN104991821B (en) * | 2015-06-29 | 2019-12-06 | 北京奇虎科技有限公司 | method and device for processing monitoring tasks in batches |
| WO2017088189A1 (en) * | 2015-11-27 | 2017-06-01 | 华为技术有限公司 | User data processing method, related device and system |
| CN106055315A (en) * | 2016-05-19 | 2016-10-26 | 乐视控股（北京）有限公司 | Method and system for processing batched tasks |
| CN107665144A (en) * | 2016-07-29 | 2018-02-06 | 北京京东尚科信息技术有限公司 | The balance dispatching center of distributed task scheduling, mthods, systems and devices |
| CN107665144B (en) * | 2016-07-29 | 2021-02-26 | 北京京东尚科信息技术有限公司 | Balanced scheduling center, method, system and device for distributed tasks |
| CN106293920A (en) * | 2016-08-15 | 2017-01-04 | 北京票之家科技有限公司 | Method for scheduling task and device |
| CN106648842A (en) * | 2016-12-27 | 2017-05-10 | 中国建设银行股份有限公司 | Method and device for continuous running of batch jobs |
| CN108958923A (en) * | 2017-05-25 | 2018-12-07 | 上海华为技术有限公司 | A kind of method and controlling equipment of scheduling hardware accelerator |
| CN107370808A (en) * | 2017-07-13 | 2017-11-21 | 盐城工学院 | A kind of method for being used to carry out big data task distributed treatment |
| CN107506906A (en) * | 2017-08-02 | 2017-12-22 | 贵州工程应用技术学院 | A kind of task intelligent scheduling engine design method based on data processing |
| CN107885600A (en) * | 2017-10-19 | 2018-04-06 | 用友网络科技股份有限公司 | Dispatching method, device and the computer equipment of batch timed task |
| CN108228329A (en) * | 2018-02-02 | 2018-06-29 | 观相科技（上海）有限公司 | A kind of packet scheduling system |
| CN108537543B (en) * | 2018-03-30 | 2023-07-28 | 百度在线网络技术（北京）有限公司 | Parallel processing method, device, equipment and storage medium for blockchain data |
| CN108537543A (en) * | 2018-03-30 | 2018-09-14 | 百度在线网络技术（北京）有限公司 | Method for parallel processing, device, equipment and the storage medium of block chain data |
| CN108920261A (en) * | 2018-05-23 | 2018-11-30 | 中国航天***科学与工程研究院 | A kind of two-stage self-adapting dispatching method suitable for large-scale parallel data processing task |
| CN108920261B (en) * | 2018-05-23 | 2020-03-24 | 中国航天***科学与工程研究院 | Two-stage adaptive scheduling method suitable for massive parallel data processing tasks |
| CN109389286A (en) * | 2018-09-03 | 2019-02-26 | 阿里巴巴集团控股有限公司 | A kind of business processing, scheduling, batch processes, device, equipment and medium |
| CN109389286B (en) * | 2018-09-03 | 2022-11-18 | 阿里巴巴集团控股有限公司 | Business processing, scheduling and batching method, device, equipment and medium |
| CN109639653A (en) * | 2018-11-29 | 2019-04-16 | 中国人民银行清算总中心 | Message transmitting method and system based on distributed internet banking system |
| CN110083442A (en) * | 2019-03-18 | 2019-08-02 | 咪咕文化科技有限公司 | A kind of operating method of user task, server and can storage medium |
| CN110083442B (en) * | 2019-03-18 | 2021-12-07 | 咪咕文化科技有限公司 | User task operation method, server and storage medium |
| CN110287018A (en) * | 2019-07-04 | 2019-09-27 | 中国工商银行股份有限公司 | Batch tasks method of combination and device |
| CN110704170A (en) * | 2019-09-09 | 2020-01-17 | 深圳壹账通智能科技有限公司 | Batch task processing method and device, computer equipment and storage medium |
| CN110673939A (en) * | 2019-09-23 | 2020-01-10 | 汉纳森（厦门）数据股份有限公司 | Task scheduling method, device and medium based on airflow and yarn |
| CN110673939B (en) * | 2019-09-23 | 2021-12-28 | 汉纳森（厦门）数据股份有限公司 | Task scheduling method, device and medium based on airflow and yarn |
| CN110795219A (en) * | 2019-10-24 | 2020-02-14 | 华东计算技术研究所（中国电子科技集团公司第三十二研究所） | Resource scheduling method and system suitable for multiple computing frameworks |
| CN110795219B (en) * | 2019-10-24 | 2022-03-18 | 华东计算技术研究所（中国电子科技集团公司第三十二研究所） | Resource scheduling method and system suitable for multiple computing frameworks |
| CN112035230A (en) * | 2020-09-01 | 2020-12-04 | 中国银行股份有限公司 | Method and device for generating task scheduling file and storage medium |
| CN112035230B (en) * | 2020-09-01 | 2023-08-18 | 中国银行股份有限公司 | Task scheduling file generation method, device and storage medium |
| CN112148491A (en) * | 2020-09-27 | 2020-12-29 | 京东数字科技控股股份有限公司 | Data processing method and device |
| CN112148491B (en) * | 2020-09-27 | 2023-12-05 | 京东科技控股股份有限公司 | Data processing method and device |
| WO2022067531A1 (en) * | 2020-09-29 | 2022-04-07 | 深圳大学 | Computing resource aware task scheduling method |
| CN112562114A (en) * | 2020-12-18 | 2021-03-26 | 北京百度网讯科技有限公司 | Data acquisition method and device |
| US11733693B2 (en) | 2020-12-18 | 2023-08-22 | Beijing Baidu Netcom Science Technology Co., Ltd. | Data acquisition method and apparatus |
| CN115373804A (en) * | 2022-10-27 | 2022-11-22 | 合肥综合性国家科学中心人工智能研究院(安徽省人工智能实验室) | Virtual machine scheduling method facing network test in cloud infrastructure |
| CN115373804B (en) * | 2022-10-27 | 2023-03-07 | 合肥综合性国家科学中心人工智能研究院(安徽省人工智能实验室) | Virtual machine scheduling method facing network test in cloud infrastructure |
Also Published As
| Publication number | Publication date |
|---|---|
| CN100444121C (en) | 2008-12-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101038559A (en) | Batch task scheduling engine and dispatching method | |
| CN1300693C (en) | Device for adjusting used of system resource and its method | |
| CN1959717A (en) | System and method for preprocessing mass remote sensing data collection driven by order form | |
| CN1956457A (en) | Method and apparatus for arranging mesh work in mesh computing system | |
| CN100341281C (en) | Method for realizing task management for network system | |
| CN1956456A (en) | Method and apparatus for presenting resource demand in mesh computing system | |
| CN1913525A (en) | Method and system for automatic responses to job requests from client | |
| CN101047549A (en) | Data synchronous system and its method | |
| CN1223025A (en) | Real-time pricing control system and method regarding same | |
| CN1267427A (en) | Telecommunications performance management system | |
| CN1828541A (en) | Implementation method for timing task in Java operating system | |
| CN1975720A (en) | Data tapping system based on Wcb and control method thereof | |
| CN1906600A (en) | Hierarchical resource management for a computing utility | |
| CN1450469A (en) | Data process device, method and program | |
| CN101031886A (en) | Network system, management computer, cluster management method, and computer program | |
| CN1755661A (en) | Information processing apparatus and method and program | |
| CN1229728C (en) | WEB application system having management/distributed management function and mechanism for operating the same | |
| CN1889045A (en) | Cocurrent event processing device and method in multi-task software system | |
| CN1842078A (en) | Gridding service container | |
| CN1315045C (en) | A method for implementing centralized concurrent management to cluster | |
| CN1259623C (en) | Structure method of five-hierarchical system structure base on J2EE | |
| CN1306407C (en) | Method for autonomic self-learning in selecting resources for dynamic provisioning | |
| CN1835450A (en) | Mesh operation supervisor | |
| CN1949223A (en) | Multidimensional data reading and writing method and apparatus in on-line analysing processing system | |
| CN101056210A (en) | An event processing system and method of network central management platform |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |