CN100373343C - Service grid scheduling device and scheduling method - Google Patents

Service grid scheduling device and scheduling method Download PDF

Info

Publication number
CN100373343C
CN100373343C CNB2005101325393A CN200510132539A CN100373343C CN 100373343 C CN100373343 C CN 100373343C CN B2005101325393 A CNB2005101325393 A CN B2005101325393A CN 200510132539 A CN200510132539 A CN 200510132539A CN 100373343 C CN100373343 C CN 100373343C
Authority
CN
China
Prior art keywords
module
job
execution
scheduler
resource
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.)
Expired - Fee Related
Application number
CNB2005101325393A
Other languages
Chinese (zh)
Other versions
CN1790275A (en
Inventor
怀进鹏
胡春明
沃天宇
于雷
钟亮
雷磊
王贝贝
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beihang University
Beijing University of Aeronautics and Astronautics
Original Assignee
Beihang University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beihang University filed Critical Beihang University
Priority to CNB2005101325393A priority Critical patent/CN100373343C/en
Publication of CN1790275A publication Critical patent/CN1790275A/en
Application granted granted Critical
Publication of CN100373343C publication Critical patent/CN100373343C/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Abstract

The present invention provides a grid job scheduling device which comprises a job submission module of descriptive language analysis, a scheduling module and an information service acquiring module in order connection, and a resource matching module and a POSIX job scheduling module in mutual connection, wherein the scheduling module is also connected with the resource matching module, the job submission module of descriptive language analysis is connected with a job managing module in the scheduling module, the information service acquiring module is connected with a resource acquiring module in the scheduling module, the resource matching module is connected with the job scheduling module in the scheduling module, and the resource acquiring module and the job scheduling module are simultaneously connected with the job managing module. The present invention also comprises an executing module of assignment executing jobs, which is connected with the resource matching module. The present invention also provides a method scheduling the job. The executing performance of distributed jobs is effectively guaranteed, the resource balance is realized in the entire grid environment, and the speed and the quality for processing the jobs are increased by the scheduling device and the scheduling method of the present invention.

Description

The method of service grid scheduling device and schedule job
Technical field
The present invention relates to the method for a kind of service grid scheduling device and schedule job, the grid job scheduler of carrying out the operation execution module is appointed in especially a kind of setting, and the method for utilizing this grid job scheduler that operation is dispatched.
Background technology
Along with improving constantly of community networkization, informationalized degree, requirement to computing power analyze every field such as scientific research from commerce in improves greatly, at present, traditional computing system based on unit/single node can not satisfy far away as the demand of types of applications such as processing, storage data to computing power.In order to satisfy demand to computing power, grid computing is arisen at the historic moment, grid computing is to utilize the computation schema that some idle processing poweies solve challenge in the network, promptly utilizes Internet technology, and the computing machine that is dispersed in diverse geographic location is formed a virtual supercomputer.The computing machine of each participation is exactly one of them " node ", and all computing machines have just been formed a node net-grid, and the computing power of grid is strong and expense is low.Substantially " grid computing " is a kind of Distributed Application, the fraction that in the grid each computing machine is just finished the work, though the arithmetic capability of single computer is limited, but thousands of computer set computing power altogether just can reach the computing power of supercomputer, therefore, utilizing gridding technique is a main path that improves computing power at present.In gridding technique, effectively carrying out task management, job scheduling and resource management is the problem that gridding technique must solve.
In the prior art, OMII (Open Middleware Infrastructure Institute, open middleware foundation structure research institute) adopted the operation based on Web service of a standard to submit to and monitoring GridSAM module in the grid job scheduler of developing, effectively realized operation is submitted to and monitoring, but there is following defective when utilizing this grid job scheduler schedule job: utilize this grid job scheduler simply to dispatch to the operation of submitting to, and when schedule job, do not consider the type of operation, can not be performed after causing the operation back of the body scheduling of some type, bring very big inconvenience to the user; Simultaneously, utilize this grid job scheduler not dispatch again to the operation of submitting to, promptly can not in whole grid scope, the load balancing according to the operation XM distribute operation, carry out and can only distribute to the operation execution module that is connected with this scheduler, like this, when the node ground load of a certain execution operation is overweight, speed and quality that operation is handled have just been influenced, do not make full use of simultaneously not busy surplus computational resource yet, fundamentally do not realize scheduling operation.
Summary of the invention
First purpose of the present invention is can not realize present situation to the scheduling of operation at grid job scheduler of the prior art, a kind of grid job scheduler is provided, in grid job scheduler, be provided with one and appoint execution operation execution module, appoint by this and carry out the effective scheduling of operation execution module realization operation.
Second purpose of the present invention is can not realize present situation to effective scheduling of operation at prior art, a kind of method of schedule job is provided, carries out the operation execution module by appointing of being provided with in the grid job scheduler analysis, the coupling of operation and resource realized the effective scheduling to operation.
For realizing above-mentioned first purpose, the invention provides a kind of grid job scheduler, comprise the operation submission descriptive language parsing module that connects successively, scheduler module and information service acquisition module, interconnective resource matched module and POSIX job scheduling module, described scheduler module also is connected with described resource matched module, described operation submits to the descriptive language parsing module to be connected with task management module in the described scheduler module, described information service acquisition module is connected with resource acquisition module in the described scheduler module, described resource matched module is connected with job scheduling module in the described scheduler module, and described resource acquisition module and described job scheduling module are connected with described task management module simultaneously; In addition, this grid job scheduler also comprises:
One appoints execution operation execution module, be connected with described resource matched module, the interface of submitting to the descriptive language parsing module to be connected with the operation of operation execution module and other grid job scheduler is provided, be used for according to homework type and with the described serviceability information of carrying out the operation execution module that the operation execution module is connected of appointing, operation appointed to other the grid job scheduler of appointing that execution operation execution module is connected with described dispatches.
For realizing above-mentioned second purpose, the invention provides a kind of method of schedule job, it carries out following steps:
Step 1, user import pending information;
Step 2, operation submit to the descriptive language parsing module that operation is analyzed and constructed to described pending information;
Step 3, described operation submit to the descriptive language parsing module that described operation is kept in the task management module in the scheduler module;
Resource acquisition module in step 4, the described scheduler module is obtained resource information and is described operation coupling resource by the information service acquisition module;
The schedule job for the treatment of that job scheduling module in step 5, the described scheduler module will have been mated resource sends to resource matched module, this resource matched module judges whether the operation execution module of this resource matched module place node supports the described schedule job for the treatment of, if support that then execution in step 6; Otherwise, redirect execution in step 8;
Step 6, described resource matched module judge describedly treat whether schedule job is the operation of POSIX type, are execution in step 7 then; Otherwise, redirect execution in step 8;
Step 7, described resource matched module treat that with described schedule job is transmitted to POSIX job scheduling module, by this POSIX job scheduling module this is treated that schedule job sends to an operation execution module to treat that to described schedule job handles redirect execution in step then 9;
Step 8, described resource matched module treat that schedule job is transmitted to appoint and carry out the operation execution module described, and this is appointed and carries out the operation execution module and treat that schedule job is appointed to a grid job scheduler and dispatch and carry out this is treated schedule job described according to the described type for the treatment of schedule job and with this serviceability information of appointing the node that execution operation execution module is connected;
Step 9, end.
By the technical scheme of above-mentioned grid job scheduler and dispatching method as can be known, the present invention has introduced to appoint in grid job scheduler and has carried out the operation execution module, this module can effectively guarantee to be assigned with the enforceability of operation according to the type for the treatment of schedule job to this job assignment operation; Simultaneously, also whether this operation is dispatched again, in whole grid environment, realized the equilibrium of resource, improved the speed and the quality of processing operation according to the serviceability information decision of other scheduler node that is connected with this grid job scheduler, effectively utilize resource, made things convenient for the user.
Below by drawings and Examples, technical scheme of the present invention is described in further detail.
Description of drawings
Fig. 1 is a structural representation of grid job scheduler of the present invention;
Fig. 2 is another structural representation of grid job scheduler of the present invention;
Fig. 3 is the process flow diagram of network job scheduling method of the present invention.
Embodiment
Referring to Fig. 1, Figure 1 shows that a concrete structure synoptic diagram of grid job scheduler of the present invention, as shown in Figure 1, this grid job scheduler comprises operation submission descriptive language (the JobSubmission Description Language that connects successively, JSDL) parsing module 1, scheduler module 2 and information service acquisition module 3, interconnective resource matched module 4 and POSIX (Portable Operating SystemInterface Standard) job scheduling module 5, and appointing of being connected with resource matched module 4 carried out operation execution module 6.Wherein, scheduler module 2 comprises resource acquisition module 21, task management module 22 and the job scheduling module 23 that connects successively, it specifically is to be connected with task management module 22 that JSDL parsing module 1 is connected with scheduler module 2, it specifically is to be connected with resource acquisition module 21 that information service acquisition module 3 is connected with scheduler module 2, and resource matched module 4 is connected with job scheduling module 23.Appoint execution operation execution module 6 that the interface that is connected with the JSDL parsing module and the operation execution module of job scheduler is provided, can appoint serviceability such as the cpu load of carrying out other scheduler node that operation execution module 6 is connected according to the type for the treatment of schedule job and with this, free memory, information such as hard disk free space appoint the suitable operation execution module of carrying out operation execution module 6 place nodes to be carried out job assignment to this, maybe this operation is appointed to other grid job scheduler to dispatch again, this job assignment carries out for the lighter node of load the most at last, in whole grid environment, realized the equilibrium of resource, the speed and the quality of processing operation have been improved, effectively utilize resource, made things convenient for the user.
Referring to Fig. 2, Figure 2 shows that another concrete structure synoptic diagram of grid job scheduler of the present invention, as shown in Figure 2, this grid job scheduler is in grid job scheduler shown in Figure 1, also set up a Web job scheduling module 7, this Web job scheduling module 7 is connected with resource matched module 4, the Web service interface is provided, be used for the called Web service execution module and carry out the operation of Web service type, make to the processing of the operation of Web service type efficient and conveniently, further improved the speed of schedule job.
Referring to Fig. 3, Figure 3 shows that an idiographic flow of network job scheduling method of the present invention, it specifically carries out following steps:
Step 31, user are by pending information of client input;
Step 32, client send the pending information of user's input to JSDL.Parsing module, this JSDL, parsing module treat process information and analyze and construct operation, and the concrete operations step is as follows:
Step 321, client send the pending information of user's input to the JSDL parsing module;
It is JSDL file or Job object or JSDL document object that step 322, JSDL parsing module are analyzed pending information, if the Job object, execution in step 324; If JSDL document object 323, execution in step; If the JSDL file is a JSDL document object with this JSDL file document analysis then;
Step 323, JSDL parsing module resolve to a Job object with the JSDL document object;
Step 324, JSDL parsing module are according to the operation of Job object structure;
Step 33, JSDL parsing module send to scheduler module with operation, and scheduler module is kept at this operation in the task management module;
Step 34, resource acquisition module scan the task management module, when existing by the new operation of coupling resource in finding this task management module, obtain resource information by the information service acquisition module, and are new operation coupling resource according to resource information;
Job scheduling module in step 35, the scheduler module scans the task management module, in finding this task management module, exist mated resource but be not scheduled as yet treat schedule job the time, this is treated that schedule job sends to resource matched module, this resource matched module judges whether each operation execution module of its place node supports this to treat schedule job, if support that then execution in step 36; Otherwise, redirect execution in step 38;
Step 36, resource matched module judge this treats whether schedule job is the operation of POSIX type, are execution in step 37 then; Otherwise, redirect execution in step 38;
Step 37, resource matched module will treat that schedule job is transmitted to POSIX job scheduling module, this POSIX job scheduling module will be treated performance such as the load of schedule job according to the operation execution module that connects, Information Selection such as CPU one operation execution module, this is treated that schedule job sends to selected operation execution module to handle redirect execution in step then 39;
Step 38, resource matched module will be treated that schedule job is transmitted to appoint and carry out the operation execution module, this is appointed and carries out the operation execution module and at first judge the type for the treatment of schedule job, carries out the higher operation execution module of selection one load minimum in operation execution module that the operation execution module is connected and performance and carries out this operation appointing with this according to this homework type; Carry out the operation execution module that the operation execution module is connected and homework type does not all match or its performance is lower or load when higher when appointing with this, this grid job scheduler of appointing execution operation execution module just this operation to be entrusted to other is dispatched;
Carry out the operation execution module and realized effective coupling by appointing according to the performance of type and operation execution module to operation, avoided because of homework type do not match or the performance of operation execution module lowly to the influence of the speed and the quality of processing operation, guaranteed to be assigned with the enforceability of operation; Simultaneously,, in whole grid scope, realized the equilibrium of resource, improved the speed and the quality of processing operation, effectively utilized resource, made things convenient for the user by scheduling again to operation;
Step 39, end.
In above-mentioned network job scheduling method, in the step 35, when resource matched module was judged this type for the treatment of schedule job, if treat that schedule job is the operation of POSIX type, then execution in step 37; If the operation of Web service type, execution in step 351; If neither the operation of POSIX type neither the Web service type operation, then redirect execution in step 38;
Step 351, resource matched module will treat that schedule job is transmitted to Web job scheduling module, and this Web job scheduling module is selected a concrete Web service execution module to carry out this and treated schedule job; Execution in step 39 then.
It should be noted last that, above embodiment is only unrestricted in order to technical scheme of the present invention to be described, although the present invention is had been described in detail with reference to preferred embodiment, those of ordinary skill in the art is to be understood that, can make amendment or be equal to replacement technical scheme of the present invention, and not break away from the spirit and scope of technical solution of the present invention.

Claims (5)

1. grid job scheduler, comprise the operation submission descriptive language parsing module that connects successively, scheduler module and information service acquisition module, interconnective resource matched module and POSIX job scheduling module, described scheduler module also is connected with described resource matched module, described operation submits to the descriptive language parsing module to be connected with task management module in the described scheduler module, described information service acquisition module is connected with resource acquisition module in the described scheduler module, described resource matched module is connected with job scheduling module in the described scheduler module, and described resource acquisition module and described job scheduling module are connected with described task management module simultaneously; It is characterized in that also comprising:
One appoints execution operation execution module, be connected with described resource matched module, the interface of submitting to the descriptive language parsing module to be connected with the operation of operation execution module and other grid job scheduler is provided, be used for according to homework type and with the described serviceability information of carrying out the operation execution module that the operation execution module is connected of appointing, operation appointed to other the grid job scheduler of appointing that execution operation execution module is connected with described dispatches.
2. grid job scheduler according to claim 1 is characterized in that described resource matched module also is connected with Web job scheduling module, is used to provide the Web service interface, and the called Web service execution module is carried out the operation of Web service type.
3. the method for a schedule job, wherein carry out following steps:
Step 1, user import pending information;
Step 2, operation submit to the descriptive language parsing module that operation is analyzed and constructed to described pending information;
Step 3, described operation submit to the descriptive language parsing module that described operation is kept in the task management module in the scheduler module;
Resource acquisition module in step 4, the described scheduler module is obtained resource information and is described operation coupling resource by the information service acquisition module;
The schedule job for the treatment of that job scheduling module in step 5, the described scheduler module will have been mated resource sends to resource matched module, this resource matched module judges whether the operation execution module of this resource matched module place node supports the described schedule job for the treatment of, if support that then execution in step 6; Otherwise, redirect execution in step 8;
Step 6, described resource matched module judge describedly treat whether schedule job is the operation of POSIX type, are execution in step 7 then; Otherwise, redirect execution in step 8;
Step 7, described resource matched module treat that with described schedule job is transmitted to POSIX job scheduling module, by this POSIX job scheduling module this is treated that schedule job sends to an operation execution module to treat that to described schedule job handles redirect execution in step then 9;
Step 8, described resource matched module treat that schedule job is transmitted to appoint and carry out the operation execution module described, and this is appointed and carries out the operation execution module and treat that schedule job is appointed to a grid job scheduler and dispatch and carry out this is treated schedule job described according to the described type for the treatment of schedule job and with this serviceability information of appointing the node that execution operation execution module is connected;
Step 9, end.
4. the method for schedule job according to claim 3, wherein said step 2 is specially:
It is that operation submits to descriptive language file or manipulating object or operation to submit the descriptive language document object to that step 21, described operation submit to the descriptive language parsing module to analyze pending information, if manipulating object, execution in step 23; If descriptive language document object, execution in step 22 are submitted in operation to; If the descriptive language file is submitted in operation to, then this operation being submitted to the descriptive language document analysis is that the descriptive language document object is submitted in an operation to;
Step 22, described operation submit to the descriptive language parsing module to submit to the descriptive language document object to resolve to a manipulating object operation;
Step 23, described operation submit to the descriptive language parsing module according to the operation of described manipulating object structure.
5. according to the method for claim 3 or 4 described schedule jobs, wherein described step 8 is replaced with:
Described resource matched module is judged and is describedly treated whether the type of schedule job is the operation of Web service type, if the operation of Web service type, then treat that with described schedule job is transmitted to Web job scheduling module, this Web job scheduling module selects a Web service execution module to carry out the described schedule job for the treatment of, execution in step 9 then; If not the operation of Web service type, treat that schedule job is transmitted to appoint and carry out the operation execution module then that this is appointed and carries out the operation execution module and treat that schedule job is appointed to a grid job scheduler and dispatch and carry out this is treated schedule job described according to the described type for the treatment of schedule job and with this serviceability information of appointing the node that execution operation execution module is connected with described.
CNB2005101325393A 2005-12-26 2005-12-26 Service grid scheduling device and scheduling method Expired - Fee Related CN100373343C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CNB2005101325393A CN100373343C (en) 2005-12-26 2005-12-26 Service grid scheduling device and scheduling method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CNB2005101325393A CN100373343C (en) 2005-12-26 2005-12-26 Service grid scheduling device and scheduling method

Publications (2)

Publication Number Publication Date
CN1790275A CN1790275A (en) 2006-06-21
CN100373343C true CN100373343C (en) 2008-03-05

Family

ID=36788164

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB2005101325393A Expired - Fee Related CN100373343C (en) 2005-12-26 2005-12-26 Service grid scheduling device and scheduling method

Country Status (1)

Country Link
CN (1) CN100373343C (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102103518A (en) * 2011-02-23 2011-06-22 运软网络科技(上海)有限公司 System for managing resources in virtual environment and implementation method thereof

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101196832B (en) * 2007-12-29 2012-12-19 北京航空航天大学 Resource proxy machine facing illumination variant equipment
CN103049326B (en) * 2013-01-16 2015-04-15 浪潮(北京)电子信息产业有限公司 Method and system for managing job program of job management and scheduling system
CN103701894A (en) * 2013-12-25 2014-04-02 浙江省公众信息产业有限公司 Method and system for dispatching dynamic resource
CN107656741B (en) * 2016-07-25 2020-07-28 中国科学院计算机网络信息中心 Method and system for using application package in high-performance computing environment

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1581806A (en) * 2003-08-05 2005-02-16 国际商业机器公司 Method and system for balancing working load in network computing environment
CN1671135A (en) * 2004-03-19 2005-09-21 联想(北京)有限公司 A network-oriented machine group working management system and realizing method thereof
US20050283782A1 (en) * 2004-06-17 2005-12-22 Platform Computing Corporation Job-centric scheduling in a grid environment
US20050283534A1 (en) * 2004-06-17 2005-12-22 Platform Computing Corporation Goal-oriented predictive scheduling in a grid environment

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1581806A (en) * 2003-08-05 2005-02-16 国际商业机器公司 Method and system for balancing working load in network computing environment
CN1671135A (en) * 2004-03-19 2005-09-21 联想(北京)有限公司 A network-oriented machine group working management system and realizing method thereof
US20050283782A1 (en) * 2004-06-17 2005-12-22 Platform Computing Corporation Job-centric scheduling in a grid environment
US20050283534A1 (en) * 2004-06-17 2005-12-22 Platform Computing Corporation Goal-oriented predictive scheduling in a grid environment

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102103518A (en) * 2011-02-23 2011-06-22 运软网络科技(上海)有限公司 System for managing resources in virtual environment and implementation method thereof

Also Published As

Publication number Publication date
CN1790275A (en) 2006-06-21

Similar Documents

Publication Publication Date Title
Schopf A general architecture for scheduling on the grid
Elmroth et al. A grid resource broker supporting advance reservations and benchmark-based resource selection
US20030028645A1 (en) Management system for a cluster
CN100373343C (en) Service grid scheduling device and scheduling method
Prakash et al. A novel scheduling model for computational grid using quantum genetic algorithm
Shih et al. Performance study of parallel programming on cloud computing environments using mapreduce
CN110806928A (en) Job submitting method and system
Zhang et al. Meteor: Optimizing spark-on-yarn for short applications
In et al. Sphinx: A scheduling middleware for data intensive applications on a grid
Mao et al. GreenPipe: a Hadoop based workflow system on energy-efficient clouds
Yu et al. Resource management for elastic cloud workflows
Cao et al. Performance prediction technology for agent-based resource management in grid environments
Mishra et al. A memory-aware dynamic job scheduling model in Grid computing
Yang et al. A workflow-based computational resource broker with information monitoring in grids
Pandey et al. Data-intensive workflow execution using distributed compute resources
Lin et al. Tree-based task scheduling model and dynamic load-balancing algorithm for P2P computing
Ding et al. Customized plug-in modules in metascheduler csf4 for life sciences applications
Afgan et al. Dynamic Task Distribution in the Grid for BLAST.
Verdicchio et al. Introduction to High-Performance Computing
Maghawry et al. Enhancing query optimization technique by conditional merging over cloud computing
Panhwar et al. Workflow-based Approach in Cloud Computing Environment
Kousalya et al. K-Level with Buddy Memory Allocation (BMA) Approach for Parallel Workload Scheduling
Sun et al. An Energy-Efficient Scheduling Method for Real-Time Multi-workflow in Container Cloud
Afgan et al. Grid resource broker using application benchmarking
Du et al. The triangular pyramid scheduling model and algorithm for pdes in grid

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
C17 Cessation of patent right
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20080305

Termination date: 20121226