CN101034362A - Method for implementing network job scheduling using mobile proxy - Google Patents
Method for implementing network job scheduling using mobile proxy Download PDFInfo
- Publication number
- CN101034362A CN101034362A CN 200710019979 CN200710019979A CN101034362A CN 101034362 A CN101034362 A CN 101034362A CN 200710019979 CN200710019979 CN 200710019979 CN 200710019979 A CN200710019979 A CN 200710019979A CN 101034362 A CN101034362 A CN 101034362A
- Authority
- CN
- China
- Prior art keywords
- agency
- resource
- job
- control agent
- subjob
- 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
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention is a method for implementing operation scheduling in network environment by mobile agent, for a to-be-processed operation in the network, using resolving agent to resolve it, depending on management system-provided information in the network and using the mobile agent to migrate it to a proper resource for executing. And the solution can implement self-adpating to resources and operation management in the network, reduce network communication quantity and raise network utilization ratio, and concurrently find the solution of operation, so as to achieve the target of raising network resource utilization ratio and network calculation executing efficiency, and speed up task executing and thus raising the processing efficiency of a distributed system.
Description
Technical field
The present invention is a kind of being used under grid computing environment, uses mobile agent to realize the scheme of job scheduling, belongs to the interleaving techniques application of grid computing and Distributed Calculation.
Background technology
Grid computing is accompanied by Internet technology and develops rapidly, specially the novel computation schema that calculates at complicated science.Grid computing is conceived to large-scale application item, in dynamic change, has in the complicated Virtual Organization of a plurality of departments or group, flexibly, collaborative resource sharing and problem solving safely.So-called Virtual Organization is exactly the dynamic combined of some individuals, tissue or resource.In grid computing environment, the Virtual Organization that is formed by different autonomous territories as shown in Figure 1.
In grid computing environment, the user normally uses resource in the mode of submit job.Operation is the set of personal code work, data, task and related resource descriptor.Job scheduling makes each operation of shared resource obtain maximum performance.Job scheduling has two different targets, i.e. high-performance calculation and high-throughput calculate.The former is conceived to minimizing of each application program execution time, adopts the mode of parallel processing usually.And the latter was conceived in a long time cycle, increased the processing power of system by dispatching a series of independently tasks.
Basic job scheduling algorithm has the user directly to assign UDA (User Direct Assigning) algorithm and the shortest deadline MCT algorithm at present.
The main thought of the direct assignment algorithm of user is that each operation is all directly assigned by the user and carried out on which gridding resource.Appointment person does not need to understand how be assigned state of resources.The job manager of this situation just replaces the user to be transferred to the destination, when can carry out as for operation, will see resource state at that time.If resource is free time or underloading by chance, operation can move soon; If resource is heavy duty just in time, when operation moves, and depends on the task management strategy of resource this locality.
The advantage of this algorithm is simply, realize easily, but its shortcoming is obvious and fatal, the response time of operation is not guaranteed, being submitted to the last complete needed time from operation is a underrange fully, and is not within the compass of competency of hoping a management organization.
This algorithm provides the chance of specifying gridding resource to the user, so the ability of user's allocated resource can influence the time limit that end is carried out in operation.A professional who knows quite well network and principle of work can be by the gridding information service under the situation of having inquired about gridding information, operation is assigned on the adequate resources carries out, and can accomplish like this to execute operation in the short as far as possible time.But it can not guarantee that operation finishes fast under the situation of resource underloading, carries out because other users also can be assigned to same resource to the operation of oneself at the same time.
The shortest deadline, (Minimum Completion Time, MCT) algorithm was to assign each task to the node with best expected performance time from order arbitrarily, and does not consider whether this node has the shortest execution time to task.This algorithm is also referred to as quick greed (Fast Greedy) algorithm.What this algorithm was paid attention to is to carry out the concluding time the earliest, rather than the shortest execution time.
The shortest deadline job scheduling algorithm, care be to fulfil assignment as early as possible, suitable scheduling requires the operation of response fast, this may cause job request person to pay bigger cost, because the time that operation moves on resource may be longer.
Summary of the invention
Technical matters: the purpose of this invention is to provide a kind of method of using mobile agent to realize job scheduling in the grid environment.The method that the application of the invention proposes can realize the self-adaptation to resource in the grid and task management, reduce mesh traffic, improve the utilization factor of network, form the parallel of operation and find the solution, thereby reach the target of the execution efficient of the utilization ratio that improves gridding resource and grid computing.
Technical scheme: mobile proxy technology is a kind of emerging technology that occurs along with the development of Internet (the Internet), is the product of distributed computing technology and artificial intelligence combination.Mobile agent be an energy in heterogeneous network, according to certain rules independently from a host migration to another main frame, and can be mutual with other mobile agents or resource, representative of consumer is finished the program of specific task.In fact it is the combination of agent skill group and distributed computing technology.
The final purpose of grid is exactly for a kind of convenient environment that carries out high-performance calculation is provided to the user.For as close as possible data source is carried out in the operation that makes us, reduce cost on network communication, save bandwidth, balanced load is accelerated task executions, thereby improves the treatment effeciency of distributed system, and we have proposed a kind of job scheduling scheme based on mobile agent.For the operation that needs in the grid to handle, use and decompose the agency its decomposition, rely on the information that management system provided in the grid, and utilize mobile agent (Mobile agency) that it is migrated on the adequate resources and carry out.
Job scheduling in the grid computing environment comprise operation decomposition, resource discovering and choose, Task Distribution, task run, task supervision and recovery, task coordinate and six aspects such as integrated.The utilization mobile agent realizes that the concrete steps of job scheduling are as follows:
1) user in the grid node submits to the task management agency with operation;
2) the task management agency is the job creation operation control agent of user's submission, and this operation control agent is responsible for the management and the control of operation;
3) the operation control agent is created operation and is decomposed the agency, and operation is decomposed the agency and according to character such as the size of operation, authority and decomposition strategy user job is decomposed into several portions;
4) operation control agent establishing resource is chosen the agency, and resource is chosen the agency and carried out alternately with the grid resource subsystem; Resource is chosen the authority that the agency authorizes according to the node of submitting this operation to resource is carried out choosing the first time, according to the operation control agent description of operation resource requirement is carried out choosing the second time again, after the process of choosing for twice took place, some resource nodes were chosen by this operation;
5) operation is decomposed agency and resource and is chosen the agency and choose information with job assignment by the grid management system mutual resource, and reports that to the operation control agent its work finishes;
6) the operation control agent is created the job assignment agency, and job assignment act on behalf of according to certain job assignment strategy, selected resource node allocation step 3 in step 4)) in the subjob of decomposition;
7) the job assignment agency finishes the work, to the situation of operation control agent report job assignment;
8) the operation control agent is created job scheduling agency and operation monitoring agent;
9) the job scheduling agency acts on behalf of according to acting on behalf of the information creating experimental process operation that obtains alternately with job assignment, and the subjob agency who carries subjob is sent on the resource node; Job scheduling agency needs to consider subjob acts on behalf of how mutual transmission security, resource reservation, subjob agency be with the local scheduling strategy, a series of problems such as safety in operation of subjob agency when creating these subjobs and act on behalf of;
10) the job scheduling agency finishes to operation control agent report subjob agency's establishment;
11) the operation control agent is created operation and is coordinated the agency, task coordinate agency's main task be and the subjob agency between coordination and integrated;
12) operation monitoring agent and subjob agency, operation are coordinated to carry out information interaction between the agency, and the operation implementation status is fed back to the operation control agent.If there is certain subjob agency to break down, then the operation agency will give the operation control agent this information feedback, from the 3rd) go on foot again and carry out;
13) experimental process operation agency is on the node of transferring to separately, the operation that the authority of giving according to the place node utilizes the resource parallel processing to distribute, and and the job scheduling agency is mutual coordinates, the integrated information that operation is finished feeds back to operation and coordinates the agency;
14) operation of carrying as each subjob agency is all finished, and operation is coordinated the agency and given the operation control agent information feedback after comprehensive;
15) the operation control agent is to the operation result of user's submit job.
Beneficial effect:
(1) overcome that the response time is not guaranteed and shortcoming that working time on resource may be long etc.
(2) can independently calculation task be moved to another node from a node under the isomery lattice computing environment that on the region, distributes; And can with other agency or resource alternately to realize the management and the self-adaptation of operation and resource.
(3) mobile agent can be moved on the client servers at different levels or central server of grid computing environment, carries out local high-speed communication with it, and it no longer takies Internet resources, thereby greatly reduces the traffic of grid, and has improved utilization efficiency of network resources.
(4) mobile agent is by the two-way information such as transmitting corresponding resource information, load information, traffic and task execution sequence that moves between LAN server.These information are as the reference frame of resource management, load balance, communication adjustment, task scheduling etc., and mobile agent is according to the situation of the judgement management domain of these data intelligences and make respective handling.This will improve the performance and the intelligent level of system greatly, improve the reliability of grid computing and carry out efficient.
(5) mobile agent is carried out to other server end by server requests being acted on behalf of dynamic migration, make this act on behalf of less dependency network transmission link and direct server resource in the face of visiting, thereby avoided the network between mass data to transmit, reduced the dependence of system the network bandwidth.
(6) in grid computing, mobile agent does not need unified scheduling.Can be asynchronous by the agency that the user creates in the operation of various computing node, finish again and send the result to user etc. task.Same user or same computing node can be created multiple agency, in one or more node operations, form the parallel ability of finding the solution simultaneously.
(7) mobile agent is owing to have collaborative and mobility, and has OO feature and considering to realize that we just have more elasticity in the safety practice.
Description of drawings
Fig. 1 is the Virtual Organization's structural representation under the grid computing environment.
Fig. 2 is based on the structural drawing of the network job scheduling system of mobile agent.
Fig. 3 is the state transition graph of grid work.
Fig. 4 is the entity pie graph of grid computing resource.
Fig. 5 is the process flow diagram that the utilization mobile agent is realized network job scheduling.
Embodiment
One. architecture
Main grid assembly based on mobile agent:
The mobile agent back-up environment: as the middleware of mobile agent operation, provide that mobile agent moves, safe and intelligent basic-level support, can be integrated with other grid assembly.
Node: be the supplier of grid computing resource, general reference various device, instrument etc.
Grid management system: be responsible for unified command and Coordination Treatment that different grid users use resource; The information service of grid computing is provided, can adopts information inquiry, collection and dissemination method based on mobile agent.
Operation agency: be to be used for the collaborative grid task of finishing a complexity according to the mobile agent (or sub agent) that certain job description standard generates.
The structure of network job scheduling system:
Job scheduling in the grid computing environment comprise operation decomposition, resource discovering and choose, Task Distribution, task run, task supervision and recovery, task coordinate and six aspects such as integrated.Fig. 2 has provided the structure based on the network job scheduling system of mobile agent.
Operation is decomposed: major function is that will submitting to of task resolves into the subjob of a plurality of high as far as possible degree of parallelisms, and determines to have which agency to carry out them when.Operation in the grid decompose can with decompose set by step, by Function Decomposition, undertaken by three kinds of modes of data decomposition.
Resource searching and choosing: be the process of a both sides coupling, should comprise following four steps:
(1) resource that should issue of resource owner and access strategy are given the resource media.
(2) these resources of resource media storage releases news.
(3) its resource requirement information of resource requestor issue is given the resource media.
(4) the resource media is chosen adequate resources according to the demand information of resource requestor and is gathered to resource requestor.
Task Distribution: we suppose that the task of submission has resolved into a pack module and made the traffic of intermodule as far as possible little.Suppose that also allocation model is: an operation is broken down into m task T={T1, T2 ..., Tm} has n available resource R={R1 in the system, R2 ..., Rn}.The purpose of Task Distribution be exactly with this m module assignment in n resource, make the performance objective functional value minimum of expection.
In general, m>n.For the ease of problem analysis, we can set up following hexa-atomic group: W=(T, R,<, Q, C, X), and wherein, T={T1, T2 ..., Tm} is the set of task; R={R1, R2 ..., Rn} is the set of resource; "<" is the priority of task relation on the T, and Ti<Tj represents that task Ti must finish before task Tj carries out; Q is a m * n matrix, and its element Qij represents the execution time (working time of supposing each task in advance know) of task Ti on resource Rj; C is a m * n matrix, and Cij represents the communication overhead between task Ti and the Tj; X is the Task Distribution matrix of a m * n, and wherein Xij=1 represents to be assigned to execution, otherwise Xij=0.
Task run: comprise that mainly resource reservation, submission task are to resource, eligible task, the task content aspect four of operation tasks etc. under the management of local scheduling strategy.
Task monitors and recovers: various resources are scattered in various places, need carry out global coordination and management, and these work can be finished by Agent.Except finishing specific function, Agent must be supported monitoring of overall importance.Each has the incident of potential significance under their real time record, for the real-time monitoring of the overall situation provides support.Be in the Agent of system performance hub site, also need to note some critical performance parameters, support analysis and adjustment system performance.
Task monitors that two purposes are arranged: 1. be convenient to mutual between user and the operation.2. be the job control program feedback information in time, be convenient to job control program and make decisions fast.
Task coordinate and integrated: that carries out between we can finish the work by a coordinator is synchronous.After all tasks were finished, we must integrate their execution result, become the result of whole task.After all subjob scheduling, being finished, the execution result of integrated subjob is as one group of target knowledge that had both got, and it is the result of whole task.Collect the knowledge that they executable operations obtained by agency to all Task agencies, just can obtain the task executions result as the coordinator.
Two. method flow
Grid work is carried out flow process:
Generally speaking, the execution of grid work is all carried out on remote node, and the entire flow of a grid work is made up of following several steps:
◆ write the operation code, fill in job description, realize grid work.The general tool software of using always that uses is in subscriber's local or editor's realization on grid;
◆ by the grid clients submit job.The difference of the submission equipment that uses according to the user and submit to environmental differences can have order line to submit to and graphic user interface submission dual mode.Bring into operation after operation is submitted to or etc. pending;
◆ the inquiry job state.After user's submit job, in the time of need knowing job state, the state of the interface inquiry job that provides with the grid job management module;
◆ the operation interaction data is provided.Some operation needs the user to import some information in the process of implementation, with next step execution flow process of decision operation;
◆ the end of job, check the operation result of operation.After some operation is carried out and finished the operation execution result is returned on the equipment of submission person's use, after some operation is carried out and finished the operation execution result is kept at certain position of grid, at this moment will tell the concrete position of operation submission person and check way.
Operation is given after the grid, and grid is placed on operation in the suitable formation according to the administrative mechanism of oneself and the requirement description of operation, waits for scheduled for executing.As shown in Figure 3, the operation that is under the mesh operation supervisor can have following several state:
1. submit to.It is the request of user's submit job to mesh-managing mechanism and operation is put into the process of processing queue.Operation this moment has obtained the one-time job number of oneself;
2. ready.Operation forwards under the control of destination node under the control of mesh-managing mechanism, but does not also begin at local runtime, waits for the scheduling of local job scheduling mechanism;
3. operation.Operation begins to move under the control of local management mechanism, and operation has obtained local job number or process number.This state comprises all normal conditions under the local organization management;
4. wait for.Be in the operation under the local management mechanism controls, because certain incident takes place, can not continue operation in this locality and go down, local management mechanism tells mesh-managing mechanism this situation, waits for the processing of grid job management mechanism;
5. finish.Operation on the remote node is finished, but operation is not also cancelled from mesh-managing mechanism.The operation that is in this state has produced result of calculation.
6. cancellation.Operation is cancelled from mesh-managing mechanism, and the job number of distribution lost efficacy.After being cancelled, an operation illustrates that this operation has not existed.
7. make mistakes.In the whole implementation of operation, abnormal conditions occur, made operation can not enter next normal condition, waited for error handling processing.Through after the error handling processing, just can enter other states.
Coupling scheduling and order scheduling problem: when we use grid in reality, always to consider specifically that a user once can only submit an operation to or can once can successively submit the plurality of grids operation to actually, and whether be in the management of same mesh operation supervisor for the operation that different grid users is submitted to, whether can use same resource between them, remove this, the mutual relationship between operation and operation, operation and resource, resource and the resource finally also can influence the dispatching sequence of operation and the coupling of operation and resource.These problems are that we are the primary key issue that must consider in the planning grid job scheduling.
Static load problem under the grid environment: because the composition structure of computational resource is very complicated in the grid computing environment, but it can be made of the LAN (Local Area Network) of up to ten thousand single PCs, a plurality of cluster even several tissues.Fig. 4 has provided grid computing resource volume entity and has constituted.Owing to the difference of computational load, the difference of processor architecture, the reasons such as difference of high-speed cache service efficiency, all can cause the unbalanced of computational load between each resource node.The computational resource node idle waiting that has, the excessive phenomenon of computational resource node load that has.
We are by being quantitatively described the computing power of computational resource and the computation requirement of parallel task, make that the distribution of task each time all requires the computing power of resource node to satisfy the computation requirement of task node, therefore, can avoid the bigger task of calculated amount to be assigned on the resource of computing power difference, perhaps the less task of calculated amount is assigned on the strong resource of computing power, to realize static load balance.
Can see, if the computation requirement amount of the computing power parameter of computational resource and parallel task can reflect real situation more exactly, the resource that computing power is strong in the system can obtain more task so, and this meets the demand of the load balance of grid environment.
The network service load: why grid has powerful distributed computation ability, has benefited from it and can make full use of gridding resource.Therefore but this has also brought our problem that need pay close attention to of another one: the network service load.At present, grid is to be based upon on the basis of network, and it migrates to remote resource node, interprocess communication etc. none does not need the support of network the processing of grid work such as job entity.These will inevitably produce a large amount of network service loads, and how reducing these loads as much as possible also is the problem that our designing institute will be considered.The utilization mobile agent is realized the flow process of network job scheduling:
Fig. 5 has provided the process flow diagram of utilization mobile agent realization network job scheduling.For convenience of description, we have following application example:
1. certain user among the node A in the grid submits to the task management agency with operation Job1.
2. the task management agency creates the operation control agent for Job1, and this operation control agent is responsible for management and the control to Job1.
3. the operation control agent is created operation and is decomposed the agency, and operation is decomposed the agency and Job1 is decomposed into Task1, Task2, Task3 according to character such as the size of operation, authority and certain decomposition strategy.
4. operation control agent establishing resource is chosen the agency, and resource is chosen the agency and carried out alternately with the grid resource subsystem.Resource is chosen the agency and according to the authority that node A is authorized resource is carried out choosing the first time; According to the operation control agent description of operation resource requirement is carried out choosing the second time again.Suppose that the resource place node of being chosen by this operation is Node B and node C.
5. operation is decomposed the agency and is chosen the information of acting on behalf of mutual resource and job assignment with resource, reports that to the operation control agent its work finishes.
6. the operation control agent is created the job assignment agency, and the job assignment agency chooses node and distributes subjob.Suppose operation assignment agency according to certain job assignment strategy, running job Task1 and Task2 on Node B, running job Task3 on node C.
7. the job assignment agency finishes the work, to the situation of operation control agent report job assignment.
8. the operation control agent is created job scheduling agency and operation monitoring agent.
9. the job scheduling agency acts on behalf of according to acting on behalf of three subjobs of information creating that obtain alternately with job assignment, and the subjob agency who carries Task1 and Task2 is sent on the Node B, and the subjob agency who carries Task3 is sent on the node C.Job scheduling agency needs to consider subjob acts on behalf of how mutual transmission security, resource reservation, subjob agency be with the local scheduling strategy, subjob acts on behalf of a series of problems such as safety in operation when creating these subjobs and act on behalf of.
10. the job scheduling agency finishes to operation control agent report subjob agency's establishment.
11. the operation control agent is created operation and is coordinated the agency.Task coordinate agency's main task be and three subjobs agencies between coordination and integrated.
12. operation monitoring agent and three subjob agencies, operations are coordinated to carry out information interaction between the agency, and the operation implementation status is fed back to the operation control agent.If there is certain subjob agency to break down, then the operation agency will give the operation control agent this information feedback, go on foot again from the 3rd and carry out.
13. the operation that three subjobs agency on the node of transferring to separately, utilizes the parallel processing of resource of Node B and C to distribute, and and the job scheduling agency is mutual coordinates, the integrated information that operation is finished feeds back to operation and coordinates the agency.
14. all finish when the operation that three subjob agencies carry, operation is coordinated the agency and is given the operation control agent information feedback after comprehensive.
15. the operation control agent is to the operation result of user's submit job.
Claims (1)
1. a method of using mobile agent to realize job scheduling in the grid environment is characterized in that the job scheduling in this method has used mobile proxy technology to realize, concrete steps are as follows:
1) user in the grid node submits to the task management agency with operation;
2) the task management agency is the job creation operation control agent of user's submission, and this operation control agent is responsible for the management and the control of operation;
3) the operation control agent is created operation and is decomposed the agency, and operation is decomposed the agency and according to character such as the size of operation, authority and decomposition strategy user job is decomposed into several portions;
4) operation control agent establishing resource is chosen the agency, and resource is chosen the agency and carried out alternately with the grid resource subsystem; Resource is chosen the authority that the agency authorizes according to the node of submitting this operation to resource is carried out choosing the first time, according to the operation control agent description of operation resource requirement is carried out choosing the second time again, after the process of choosing for twice took place, some resource nodes were chosen by this operation;
5) operation is decomposed agency and resource and is chosen the agency and choose information with job assignment by the grid management system mutual resource, and reports that to the operation control agent its work finishes;
6) the operation control agent is created the job assignment agency, and job assignment act on behalf of according to certain job assignment strategy, selected resource node allocation step 3 in step 4)) in the subjob of decomposition;
7) the job assignment agency finishes the work, to the situation of operation control agent report job assignment;
8) the operation control agent is created job scheduling agency and operation monitoring agent;
9) the job scheduling agency acts on behalf of according to acting on behalf of the information creating experimental process operation that obtains alternately with job assignment, and the subjob agency who carries subjob is sent on the resource node; Job scheduling agency needs to consider subjob acts on behalf of how mutual transmission security, resource reservation, subjob agency be with the local scheduling strategy, a series of problems such as safety in operation of subjob agency when creating these subjobs and act on behalf of;
10) the job scheduling agency finishes to operation control agent report subjob agency's establishment;
11) the operation control agent is created operation and is coordinated the agency, task coordinate agency's main task be and the subjob agency between coordination and integrated;
12) operation monitoring agent and subjob agency, operation are coordinated to carry out information interaction between the agency, and the operation implementation status is fed back to the operation control agent.If there is certain subjob agency to break down, then the operation agency will give the operation control agent this information feedback, from the 3rd) go on foot again and carry out;
13) experimental process operation agency is on the node of transferring to separately, the operation that the authority of giving according to the place node utilizes the resource parallel processing to distribute, and and the job scheduling agency is mutual coordinates, the integrated information that operation is finished feeds back to operation and coordinates the agency;
14) operation of carrying as each subjob agency is all finished, and operation is coordinated the agency and given the operation control agent information feedback after comprehensive;
15) the operation control agent is to the operation result of user's submit job.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200710019979 CN101034362A (en) | 2007-02-05 | 2007-02-05 | Method for implementing network job scheduling using mobile proxy |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200710019979 CN101034362A (en) | 2007-02-05 | 2007-02-05 | Method for implementing network job scheduling using mobile proxy |
Publications (1)
Publication Number | Publication Date |
---|---|
CN101034362A true CN101034362A (en) | 2007-09-12 |
Family
ID=38730932
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 200710019979 Pending CN101034362A (en) | 2007-02-05 | 2007-02-05 | Method for implementing network job scheduling using mobile proxy |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN101034362A (en) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101907989A (en) * | 2010-06-01 | 2010-12-08 | 南京大学 | Mobile agent-based application seamless migration method |
CN102103521A (en) * | 2011-01-31 | 2011-06-22 | 中国科学院计算技术研究所 | HPC system and method for dynamically dispatching task based on HPC system |
CN101222522B (en) * | 2008-01-17 | 2011-08-10 | 武汉理工大学 | Gridding task scheduling method considering gridding task importance and time urgency |
CN102360314A (en) * | 2011-10-28 | 2012-02-22 | 中国科学院计算技术研究所 | System and method for managing resources of data center |
CN101388844B (en) * | 2008-11-07 | 2012-03-14 | 东软集团股份有限公司 | Data flow processing method and system |
CN102486720A (en) * | 2010-12-02 | 2012-06-06 | 北大方正集团有限公司 | Method for sharing data in pre-press operations and pre-press operating system |
CN102567085A (en) * | 2010-12-24 | 2012-07-11 | 北大方正集团有限公司 | Method and device for multi-user cross-task coordination work of printing flow software |
CN103001798A (en) * | 2012-11-22 | 2013-03-27 | 华为技术有限公司 | Application service management method, device and system |
CN103294483A (en) * | 2013-06-27 | 2013-09-11 | 曙光信息产业(北京)有限公司 | Processing method used for GROMACS computing program |
CN103294481A (en) * | 2013-06-27 | 2013-09-11 | 曙光信息产业(北京)有限公司 | Processing method used for CST (computer simulation technology) computing program |
CN105320660A (en) * | 2014-06-05 | 2016-02-10 | 中国石油化工股份有限公司 | Method and device for automatic task submitting in numerical simulation parallel computing |
CN107590589A (en) * | 2017-08-25 | 2018-01-16 | 北京科技大学 | The calculating accelerated method of city common buildings group's seismic demage analysis based on GPU cluster |
CN109062695A (en) * | 2018-08-02 | 2018-12-21 | 中国水利水电科学研究院 | A kind of VIC land-surface model grid data calculation method |
CN110991041A (en) * | 2019-12-03 | 2020-04-10 | 北京仿真中心 | Automatic multi-professional collaborative simulation method based on high-performance computing system |
-
2007
- 2007-02-05 CN CN 200710019979 patent/CN101034362A/en active Pending
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101222522B (en) * | 2008-01-17 | 2011-08-10 | 武汉理工大学 | Gridding task scheduling method considering gridding task importance and time urgency |
CN101388844B (en) * | 2008-11-07 | 2012-03-14 | 东软集团股份有限公司 | Data flow processing method and system |
CN101907989A (en) * | 2010-06-01 | 2010-12-08 | 南京大学 | Mobile agent-based application seamless migration method |
CN102486720A (en) * | 2010-12-02 | 2012-06-06 | 北大方正集团有限公司 | Method for sharing data in pre-press operations and pre-press operating system |
CN102486720B (en) * | 2010-12-02 | 2015-04-22 | 北大方正集团有限公司 | Method for sharing data in pre-press operations and pre-press operating system |
CN102567085B (en) * | 2010-12-24 | 2014-06-18 | 北大方正集团有限公司 | Method and device for multi-user cross-task coordination work of printing flow software |
CN102567085A (en) * | 2010-12-24 | 2012-07-11 | 北大方正集团有限公司 | Method and device for multi-user cross-task coordination work of printing flow software |
CN102103521A (en) * | 2011-01-31 | 2011-06-22 | 中国科学院计算技术研究所 | HPC system and method for dynamically dispatching task based on HPC system |
CN102360314A (en) * | 2011-10-28 | 2012-02-22 | 中国科学院计算技术研究所 | System and method for managing resources of data center |
CN103001798B (en) * | 2012-11-22 | 2016-03-16 | 华为技术有限公司 | The methods, devices and systems of management application service |
CN103001798A (en) * | 2012-11-22 | 2013-03-27 | 华为技术有限公司 | Application service management method, device and system |
CN103294483A (en) * | 2013-06-27 | 2013-09-11 | 曙光信息产业(北京)有限公司 | Processing method used for GROMACS computing program |
CN103294481A (en) * | 2013-06-27 | 2013-09-11 | 曙光信息产业(北京)有限公司 | Processing method used for CST (computer simulation technology) computing program |
CN105320660A (en) * | 2014-06-05 | 2016-02-10 | 中国石油化工股份有限公司 | Method and device for automatic task submitting in numerical simulation parallel computing |
CN105320660B (en) * | 2014-06-05 | 2019-01-18 | 中国石油化工股份有限公司 | Method and device is submitted in the operation of numerical simulation parallel computation automatically |
CN107590589A (en) * | 2017-08-25 | 2018-01-16 | 北京科技大学 | The calculating accelerated method of city common buildings group's seismic demage analysis based on GPU cluster |
CN109062695A (en) * | 2018-08-02 | 2018-12-21 | 中国水利水电科学研究院 | A kind of VIC land-surface model grid data calculation method |
CN109062695B (en) * | 2018-08-02 | 2019-04-26 | 中国水利水电科学研究院 | A kind of VIC land-surface model grid data calculation method |
CN110991041A (en) * | 2019-12-03 | 2020-04-10 | 北京仿真中心 | Automatic multi-professional collaborative simulation method based on high-performance computing system |
CN110991041B (en) * | 2019-12-03 | 2024-01-26 | 北京仿真中心 | Automatic multi-specialty co-simulation method based on high-performance computing system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN101034362A (en) | Method for implementing network job scheduling using mobile proxy | |
Kansal et al. | Energy-aware virtual machine migration for cloud computing-a firefly optimization approach | |
Cao et al. | Grid load balancing using intelligent agents | |
Wang et al. | Optimizing load balancing and data-locality with data-aware scheduling | |
Wang et al. | A three-phases scheduling in a hierarchical cloud computing network | |
Singh et al. | Task scheduling in cloud computing | |
CN100570569C (en) | Operation cross-domain control method under the grid computing environment | |
Sharma et al. | A survey of job scheduling and resource management in grid computing | |
CN1516419A (en) | Syste mand method for processing customer computer request on host computer network | |
CN1845075A (en) | Service oriented high-performance grid computing job scheduling method | |
Bessis et al. | Modelling requirements for enabling meta-scheduling in inter-clouds and inter-enterprises | |
Selvi et al. | Resource allocation issues and challenges in cloud computing | |
Shruthi et al. | Deep learning-based resource prediction and mutated leader algorithm enabled load balancing in fog computing | |
He et al. | Extending the EGEE grid with XtremWeb-HEP desktop grids | |
Cao et al. | Performance prediction technology for agent-based resource management in grid environments | |
Goswami et al. | Deadline stringency based job scheduling in computational grid environment | |
Wang et al. | A hard real-time scheduler for Spark on YARN | |
Pallickara et al. | Swarm: Scheduling large-scale jobs over the loosely-coupled hpc clusters | |
Patel et al. | Improved PSO based job scheduling algorithm for resource management in grid computing | |
Liang et al. | A grid-enabled software distributed shared memory system on a wide area network | |
Ding et al. | Data locality-aware and QoS-aware dynamic cloud workflow scheduling in Hadoop for heterogeneous environment | |
Afgan et al. | Dynamic BLAST–a grid enabled BLAST | |
Pop et al. | The Art of Scheduling for Big Data Science. | |
Singhal et al. | An economic allocation of resources in grid environment | |
Touzene et al. | Analytical Model for Performance Evaluation of Load Balancing Algorithm for Grid Computing |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C12 | Rejection of a patent application after its publication | ||
RJ01 | Rejection of invention patent application after publication |