CONCURRENCY AND COMPUTATION:PRACTICE AND EXPERIENCE
Concurrency Computat.:Pract.Exper.2008;20:347–360
Published online25July2007in Wiley InterScience(https://www.doczj.com/doc/382597531.html,).DOI:10.1002/cpe.1220
A taxonomy of grid work?ow
veri?cation and validation
Jinjun Chen?,?and Yun Yang
Centre for Information Technology Research,Faculty of Information and
Communication Technologies,Swinburne University of Technology,
P.O.Box218,Hawthorn,Melbourne,Vic.3122,Australia
SUMMARY
In a grid architecture,a grid work?ow management system is a type of high-level grid middleware which is supposed to support modelling,redesign and execution of large-scale sophisticated scienti?c and business processes in many complex e-science and e-business applications.To ensure the correctness of grid work?ow speci?cation and execution,grid work?ow veri?cation and validation must be conducted. However,current research on grid work?ow veri?cation and validation is at the infancy stage and very few projects focus on them.Therefore,a systematic investigation and an overall classi?cation of key issues in grid work?ow veri?cation and validation is helpful and should be presented so that we can keep on the right track and reduce unnecessary work as much as possible.As such,in this paper,we analyse the grid work?ow veri?cation and validation and present a taxonomy.Especially,we identify some important open points which are not discussed by the current research and hence need further investigation.The taxonomy is aimed at providing an overall picture of grid work?ow veri?cation and validation.Copyright ?2007John Wiley&Sons,Ltd.
Received23March2007;Accepted1April2007
KEY WORDS:grid work?ow management systems;grid work?ow veri?cation;grid work?ow validation
1.INTRODUCTION
In a grid architecture,a grid work?ow management system is a type of high-level grid middleware which is supposed to support modelling,redesign and execution of large-scale sophisticated scien-ti?c and business processes in many complex e-science and e-business applications such as climate modelling,astrophysics,medical surgery,disaster recovery,international?nance and insurance
?Correspondence to:Jinjun Chen,Centre for Information Technology Research,Faculty of Information and Communication Technologies,Swinburne University of Technology,P.O.Box218,Melbourne,Vic.3122,Australia.
?E-mail:jchen@https://www.doczj.com/doc/382597531.html,.au
Contract/grant sponsor:Australian Research Council Discovery Project;contract/grant number:DP0663841
Contract/grant sponsor:Linkage Project;contract/grant number:LP0669660
Copyright2007John Wiley&Sons,Ltd.
348J.CHEN AND Y.YANG
[1–6].A grid work?ow management system normally contains a number of grid work?ows in the form of their speci?cations and instances[2,7,8].Generally speaking,a grid work?ow management system works at three stages:build-time,run-time instantiation and run-time execution[9–12].At build-time stage,complex scienti?c or business processes are modelled or redesigned as grid work-?ow speci?cations[9,11,13,14].According to[9,10,15,16],conceptually,a grid work?ow contains a lot of computation or data-intensive activities as well as dependencies between them.These ac-tivities are implemented and executed by corresponding grid services[9,11,17].The dependencies de?ne activity execution orders[9,11,17].At run-time instantiation stage,grid work?ow instances are created,and especially grid services that are speci?ed in the build-time de?nition documents are discovered.This could include an instantiation service that is a high-level grid service[9,11]. At run-time execution stage,grid work?ow instances are executed.The execution is coordinated between grid services by the grid work?ow engine which itself is also a high-level grid service, hence automatically grid aware[9,11].
To ensure the well performing of large-scale sophisticated scienti?c and business process sup-port,we must ensure the correctness of grid work?ow speci?cation and execution.According to the IEEE Standard Glossary of Software Engineering Terminology[18],the correctness is de?ned as freedom from faults,meeting of speci?ed requirements,and meeting of user needs and expec-tations.Furthermore,based on the IEEE Software Veri?cation and Validation Standard[19],the correctness can be more commonly described as grid work?ow veri?cation and validation.Both grid work?ow veri?cation and validation are important.Veri?cation failure results in the grid work-?ow speci?cation and execution containing faults or?aws.Validation failure constitutes a breach of contract between the complex scienti?c and business process developer and the client.Clearly, neither is desirable.Figure1further depicts the relationship between correctness and veri?cation &validation.
Some research related to(grid)work?ow veri?cation and validation has been done.van der Aalst [20]and Adam et al.[21]discuss how to use Petri Net to model work?ow processes.van der Aalst et al.[22]present20work?ow patterns to represent the work?ow expressive power.van der Aalst[23]analyses the structure errors in work?ow speci?cations using Petri Net-based tech-niques.van der Aalst[24]uses the simulation method to compute the performance parameters.Li et al.[25]and Li and Yang[26]analyse resource constraints in the work?ow speci?cation and ex-ecution,and present some veri?cation algorithms.Bertino et al.[27]and Wu et al.[28]discuss the speci?cation,control,and enforcement of authorization constraints.Marjanovic and Orlowska[29] assign maximum and minimum durations to each activity and propose some veri?cation algorithms to check the consistency of relative and absolute temporal constraints.Chen and Yang[30]discuss temporal dependency and its impact on the temporal veri?cation effectiveness and ef?ciency in grid
Figure 1.Correctness of grid work?ow speci?cation and execution versus grid
work?ow veri?cation and validation.
Copyright2007John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2008;20:347–360
DOI:10.1002/cpe
A TAXONOMY OF GRID WORKFLOW VERIFICATION AND V ALIDATION349 work?ow systems.Chen et al.[31]and Chen and Yang[32–34]discuss checkpoint selection for temporal veri?cation at run-time execution stage.Chen and Yang[17]introduce four consistency states to a?xed-time constraint and develops corresponding veri?cation algorithms.Buyya et al.[3], Yu and Buyya[14]and Yu et al.[35]discuss Quality of Service(QoS)scheduling issues and presents a grid economy-based architecture with corresponding grid middleware.Sadiq and Orlowska[36] discuss how to analyse work?ow process models by using graph reduction techniques. However,as far as grid work?ow veri?cation and validation is concerned,current research is still in infancy.To the best of our knowledge,very little work has been done and very few projects focus on grid work?ow veri?cation and validation issues.Therefore,a systematic identi?cation and an overall classi?cation of key issues in the grid work?ow veri?cation and validation?eld are needed which can help us?nd out main points in the?eld and avoid some unnecessary work as much as possible.Hence,in this paper,we investigate the key issues in grid work?ow veri?cation and validation,and present a taxonomy for them.Especially,we identify some important points which are not stated by the current research and hence need further investigation.The taxonomy depicts an overall picture for grid work?ow veri?cation and validation.
The remainder of the paper is organized as follows.Section2discusses the taxonomy of grid work?ow veri?cation and validation.Section3conducts a survey of existing projects and presents
a further discussion.Section4concludes our contributions and points out future work.
2.TAXONOMY
We investigate grid work?ow veri?cation taxonomy in Section2.1and grid work?ow validation taxonomy in Section2.2.
2.1.Grid work?ow veri?cation
Speci?cally speaking,based on[19],grid work?ow veri?cation is mainly concerned with the spe-ci?c correctness of grid work?ow speci?cation and execution such as no deadlock,no livelock, no temporal violation or no resource con?ict.It aims at no faults in grid work?ow speci?cation and execution under the condition where complex scienti?c and business process requirements have already been correctly supported in grid work?ow speci?cation by the selected grid work?ow management system.Correspondingly,as shown in Figure2,grid work?ow veri?cation taxonomy consists of six elements:(a)structure veri?cation;(b)performance veri?cation;(c)resource ver-i?cation;(d)authorization veri?cation;(e)cost veri?cation;and(f)temporal veri?cation.In this section,we look at each element in detail.
Figure2.Elements of grid work?ow veri?cation.
Copyright2007John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2008;20:347–360
DOI:10.1002/cpe
350J.CHEN AND Y.YANG
2.1.1.Structure veri?cation
Structure veri?cation aims to verify structure consistency[23,24].It consists of syntactic structure veri?cation and semantic structure veri?cation[23,24,36].
In a grid work?ow speci?cation,main syntactic structure inconsistencies include deadlock,live-lock,lack of synchronization,misuse of modelling objects and constructs,active end,dead activity, etc.[23,24,36].Some methods such as Petri Net or directed graph with graph reduction technologies have been presented to model the structure and verify whether there are any structure inconsistencies [20,21,23].
Current research of structure veri?cation mainly focuses on syntactic structure veri?cation.How-ever,semantic structure veri?cation is also very important.For example,suppose we have a grid work?ow speci?cation which is correct in terms of syntax.So,it is using the selective structure correctly.Suppose at a decision activity point A,it has two branches B and C.Syntactically,it can go to B or C.However,semantically,at run time,according to the execution results of A,it may need to go to D.Apparently,the grid work?ow execution will stop because no D can be selected.That is to say,semantically,the structure of the grid work?ow speci?cation is incorrect.Therefore,semantic structure veri?cation should also be investigated in addition to syntactic structure veri?cation.
2.1.2.Performance veri?cation
Performance veri?cation aims to verify whether user-de?ned performance parameters can be met [24].A grid work?ow speci?cation or execution is normally attached with some user-de?ned performance parameters such as average activity completion time,mean capacity utilization rate, mean activity queuing time,mean activity synchronization delay,mean activity set-up delay or mean resource allocation[24,37].Therefore,we need to verify whether these parameters can be met by the selected grid work?ow management system and current system run-time status.To conduct performance veri?cation,we need to compute system performance parameters,and then compare them with user-de?ned performance values.Currently,people try to use Markovian chain theory, queuing theory or simulation tools for system performance parameter computations[24,37].Based on Markovian chain theory,we try to incorporate some properties into a Markovian chain,and then use the analysis methods provided by the Markovian chain theory to calculate system performance parameters.However,not every aspect of a grid work?ow can be incorporated into the analysis and the Markovian chain analysis can also be very time consuming[24,37].The queuing theory can provide a good statistical analysis.However,many of the assumptions used in the queuing theory are not valid for grid work?ow processes[24].For example,in the presence of parallel routing,it is often impossible to apply the results obtained from the queuing theory because the distribution of activity completion time is normally not the one required for applying the queuing theory.Based on the simulation,we often convert a grid work?ow speci?cation into a simulation model,and then use some simulation tools to simulate grid work?ow execution[24].However,the challenge is that sometimes we cannot map a grid work?ow speci?cation into a simulation model without any semantic loss.
Further research might be conducted on how to extend the Markovian chain theory or queuing the-ory or simulation tools to allow for grid work?ow-speci?c characteristics,and on the extent to which a grid work?ow can be simpli?ed to meet the requirements of the use of these theories or tools.
Copyright2007John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2008;20:347–360
DOI:10.1002/cpe
A TAXONOMY OF GRID WORKFLOW VERIFICATION AND V ALIDATION351
2.1.
3.Resource veri?cation
Resource veri?cation aims to verify whether there are any resource con?icts between different activities or work?ows[25,26].In a single grid work?ow or among multiple ones,different ac-tivities may compete for the same resource such as a machine or a human being.Given that the grid computing infrastructure normally encompasses a number of heterogeneous and autonomous hosting environments,such resource competitions can often happen[7].Hence,we must verify them to make sure that resource sharing proceeds correctly.Currently,some fundamental research has been done[25,26].However,in grid work?ow management systems,run-time execution envi-ronments are very dynamic.One resource may serve more than one activities at the same time while sometimes more than one resource may serve one activity simultaneously.Therefore,we should incorporate the statistical property of grid work?ow execution into veri?cation mechanisms.For example,two activities may compete for the same machine during the same time interval.It may or may not lead to a resource con?ict.We need to conduct some statistical analysis to judge whether there are really any con?icts.Further research on resource veri?cation based on statistical analysis is a challenge because,as discussed in the above performance veri?cation section,some statistical analysis theories or tools cannot be equally applied to grid work?ow management systems.
2.1.4.Authorization veri?cation
A grid work?ow management system often needs to deploy heterogeneous and distributed hardware and software systems to execute a given grid work?ow[6,7,38].This gives rise to decentralized security policies and mechanisms that need to be managed.To ensure the security,authorization veri?cation must be conducted to check the consistency between resource access and participants or roles.Authorization means that no resources are accessed by unauthorized participants or roles at anytime[27].Therefore,some authorization constraints are set for different participants and roles,which show which kinds of resources can be accessed by them and how to access authorized resources[27,28,39].For the grid work?ow speci?cation,authorization veri?cation is conducted to judge whether access rules for corresponding participants or roles are correctly de?ned.For the grid work?ow execution,authorization veri?cation is conducted to show whether there is any unauthorized resource access so that grid work?ow management systems can take some protective measures immediately.One of the key challenges for authorization veri?cation is that we need to ?gure out what the grid work?ow security is since the underlying grid infrastructure has already supported the security to some extent[2,38].
2.1.5.Cost veri?cation
In the real world,scienti?c and business processes are normally budget constrained[3,14,35]. Therefore,we must verify the cost aspect of grid work?ow speci?cation and execution to check whether grid work?ow speci?cation and execution can meet corresponding budget requirements. Some grid work?ows have many local cost constraints,while others have only one end-to-end cost constraint.To the best of our literature review,no research has been done about cost veri?cation. Although some important and valuable work concerning cost scheduling has been done[14,35],cost scheduling cannot be used to judge whether grid work?ow speci?cation or execution is de?nitely Copyright2007John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2008;20:347–360
DOI:10.1002/cpe
352J.CHEN AND Y.YANG
going wrong when the scheduling assignment is not followed.This is normal as the judgement is one of the responsibilities of grid work?ow veri?cation.To conduct cost veri?cation,we need to represent activity cost information in a grid work?ow.We may need to enrich grid work?ow mod-elling languages so that they can accommodate activity cost information.We then need to compute the cost between two activities and compare it with cost constraints to check the consistency.In some cases where only one end-to-end cost constraint exists,to more ef?ciently control the activity cost for grid work?ow execution,we may need to investigate how to dynamically assign,verify and adjust some local cost constraints within the frame of the overall cost constraints.In fact,a grid work?ow normally consists of a number of activities and consequently lasts a long time.Therefore, only one end-to-end cost constraint is not suf?cient to control the budget aspect of grid work?ow execution.We may?nd that the end-to-end cost constraint is violated at the last activity.Then,it is too late to take any compensating or precaution measures.Therefore,if there is only one end-to-end cost constraint,we must investigate how to divide it into some smaller local cost constraints.Then, by these local cost constraints,we can better control the budget aspect of grid work?ow execution. With local cost constraints,we need to?gure out how to compute the cost between two activities for veri?cation purpose.The run-time uncertainty of activity cost must be taken into consideration so that corresponding statistical features can be considered.
2.1.6.Temporal veri?cation
Similar to the cost,in the real world,scienti?c and business processes normally stay in a temporal context and are often time constrained[3,33].Therefore,temporal veri?cation must be conducted to check whether there are any temporal constraint violations.We need to investigate four factors to conduct temporal veri?cation:temporal consistency states,assignment of?ne-grained temporal constraints,selection of checkpoints,and veri?cation of temporal constraints.
To verify temporal constraints,temporal consistency states must be de?ned.Some research on timed work?ow modelling has been done such as in[29,40–44].Two temporal consistency states have been de?ned by them.Chen and Yang[17]argue that in grid work?ow management systems, two states are too restricted as grid work?ow execution environments are very dynamic.Intermediate states exist and different exception handlings should be triggered to handle them[45].Accordingly, Chen and Yang[17]had introduced four temporal consistency states and developed corresponding veri?cation algorithms.However,it does not investigate how different exception handlings should be triggered to handle different states.
In many complex scienti?c and business processes such as a climate modelling process,users often only set a few coarse-grained temporal constraints rather than a large number of them[1,3]. However,only a few coarse-grained temporal constraints are not suf?cient to control grid work?ow execution in terms of time as we are not able to control grid work?ow execution locally at various activity points.Therefore,we need to investigate how to assign a number of?ne-grained temporal constraints based on a few user-de?ned coarse-grained ones.To do so,we need to solve two problems.One is where to assign some?ne-grained temporal constraints.The other is how to assign them.Chen and Yang[46]have solved the second problem.Further research is needed for solving the?rst problem.
At build-time and run-time instantiation stages,temporal veri?cation is static because there are no any speci?c execution times.Each temporal constraint needs to be veri?ed only once with the Copyright2007John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2008;20:347–360
DOI:10.1002/cpe
A TAXONOMY OF GRID WORKFLOW VERIFICATION AND V ALIDATION353 consideration of all covered activities.Therefore,we need not decide at which activities we should conduct the veri?cation.At run-time execution stage,however,activity completion durations vary and consequently,we may need to verify each temporal constraint many times at different activities. However,conducting the veri?cation at every activity is not ef?cient as we may not need to do so at some activities such as those that can be completed within allowed time intervals.Hence,we need to?gure out where to conduct temporal veri?cation.The activities at which we conduct temporal veri?cation are called checkpoints[29,30,43].Correspondingly,a research topic comes into the picture which is checkpoint selection strategies(CSS).Currently,some CSS have been proposed
[29,31–34,43].Some of them often select some unnecessary checkpoints or ignore some necessary ones.Although Chen and Yang[34]develop a strategy which can select necessary yet suf?cient checkpoints,it does not operate in the whole context of temporal veri?cation.For example,it does not take into consideration temporal dependency between temporal constraints.With temporal dependency,we may be able to select fewer checkpoints.Hence,further research is needed. Once we know where to verify temporal constraints,we can start to do it.Some conventional temporal veri?cation work has been done such as in[29,44].Some other work is related to tem-poral veri?cation from other perspectives such as time QoS scheduling,project management,etc. [14,35,47–49].They try to apply the technologies from their?elds to grid work?ow temporal veri-?cation.However,it might not be consistent with grid work?ow management systems very well because different?elds have different characteristics.For example,a project management system normally has only one instance at run time,while a grid work?ow management system normally has a number of grid work?ow instances at run time.So,the project management temporal analysis is relatively static and hence cannot be equally applied to more dynamic grid work?ow management system environments.Another example is about time QoS scheduling.Time QoS scheduling is im-portant for ensuring that grid work?ow execution is going smoothly in terms of time[14,35,50,51]. However,time QoS scheduling cannot be used to judge whether a grid work?ow execution is de?-nitely going wrong when the scheduling assignment is not followed.Hence,grid work?ow-speci?c temporal veri?cation is needed.For example,as analysed above,a grid work?ow normally needs a series of temporal constraints.When a series of temporal constraints are veri?ed,they are often dependent on each other in terms of overall temporal veri?cation effectiveness and ef?ciency.This is because the later veri?cation may make the previous veri?cation ineffective and also the later veri?cation may utilize the previous veri?cation results to save current veri?cation computation for better ef?ciency.Therefore,temporal dependency between temporal constraints must be taken into consideration when temporal veri?cation is conducted so that we can further improve overall temporal veri?cation effectiveness and ef?ciency.Chen and Yang[30]have discussed temporal dependency based on conventional two consistency states.Further research is needed for temporal dependency based on multiple temporal consistency states which,according to Chen and Yang[17], allow for grid work?ow-speci?c characteristics.
2.2.Grid work?ow validation
Speci?cally speaking,based on[19],grid work?ow validation is mainly concerned with the consis-tency between complex scienti?c and business processes and grid work?ow speci?cations.When we model or redesign a complex scienti?c or business process as a grid work?ow speci?cation based on models and constructs provided by the selected grid work?ow management system,we must
Copyright2007John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2008;20:347–360
DOI:10.1002/cpe
354J.CHEN AND Y.YANG
Figure3.Elements of grid work?ow validation.
ensure that all complex scienti?c and business process requirements are modelled or redesigned in the grid work?ow speci?cation.Otherwise,the grid work?ow speci?cation is incomplete and incorrect from the perspective of user requirements and needs.For example,some grid work?ow management systems or architectures cannot support temporal constraint modelling[38].Then, for those time-critical scienti?c and business processes such as climate modelling processes for weather forecast,the corresponding grid work?ow speci?cations are incomplete or even incorrect as temporal information will be ignored.In fact,different grid work?ow management systems provide different basic models and constructs for complex scienti?c and business process modelling. And different models and constructs have different expressive power[38].A complex scienti?c or business process which can be modelled by one grid work?ow management system may not be completely modelled by another one.Therefore,a question is raised between complex scienti?c and business processes and grid work?ow modelling power.That is,whether current complex scienti?c and business process requirements can be supported by the selected grid work?ow management system.This is particularly important when e-scientists or e-business people develop or purchase corresponding grid work?ow management system products to support their complex scienti?c and business processes.To answer this question,grid work?ow validation must be conducted.We need to describe a complex scienti?c or business process as a proper representation.Then,we represent the expressive power of the selected grid work?ow management system.Finally,we need to develop some ef?cient approaches to validate the consistency between them.Hence,as shown in Figure3,the grid work?ow validation taxonomy consists of three elements:(a)representation of complex scienti?c and business processes;(b)representation of expressive power of grid work?ow management systems;and(c)validation approaches.
2.2.1.Representation of complex scienti?c and business processes
To check the consistency between complex scienti?c and business processes and corresponding grid work?ow speci?cations,?rstly,we need to capture and represent complex scienti?c and business processes.
Some methods such as Petri Net or process algebra(mainly pi-Calculus)might be used to represent complex scienti?c and business processes[20,21,52,53].However,a complex scienti?c or business process normally changes often[6,7,38].Changes in process requirements may lead to signi?cant state adjustments in the Petri Net-based representation or signi?cant equation adjustments in the process algebra-based representation[54].Some other models such as state machine or concurrent transaction logic also have similar problems[55].How to extend the expressive ability Copyright2007John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2008;20:347–360
DOI:10.1002/cpe
A TAXONOMY OF GRID WORKFLOW VERIFICATION AND V ALIDATION355 of Petri Net,process algebra or some other models to allow for the complexity and dynamics of sophisticated scienti?c and business processes needs further investigation.
2.2.2.Representation of expressive power of grid work?ow management systems
By representing the expressive power of a grid work?ow management system,we can get to know how complicated scienti?c and business process requirements can be supported by it[22].Cur-rently,most grid work?ow management systems support four basic control structures:parallel, selective,sequential and iterative[9,11,38].However,it is not enough for us to use only these four control structures to represent the grid work?ow expressive power.On the one hand,a grid work?ow management system may support other control structures.On the other hand,the diver-sity of complex scienti?c and business process requirements is calling for new control structures. Therefore,we need to tackle how to extend the control structures to allow for profound scienti?c and business process requirements and how to use them to evaluate the expressive power of the selected work?ow system.Although van der Aalst et al.[22]present20work?ow patterns,they are mainly for business processes.For scienti?c processes,the requirements are often different because scienti?c processes are normally computation or data intensive and lack intensive dependency be-tween different activities[7,9].It is still a problem whether we can equally apply the20patterns to complex scienti?c processes.And if not,how to extend or modify them to allow for complex scienti?c requirements would be open.
2.2.
3.Validation approaches
Based on the representation of complex scienti?c and business processes and based on the expressive power of grid work?ow management systems,we can fall to develop some validation approaches. These approaches can be used to check whether all scienti?c and business process requirements can be supported by the selected grid work?ow management system based on its control structures.Here, a key challenge is how to map the representation of a complex scienti?c or business process based on Petri Net or process algebra or some other models to the control structures.If the approaches can conduct the mapping,then if some part of the representation has no corresponding control structures provided by the selected grid work?ow management system,it means that we cannot correctly capture the scienti?c or business process as a grid work?ow speci?cation.In other words, the grid work?ow validation fails.Then,we may consider changing the grid work?ow management system to another one which has stronger expressive power.
3.SURVEY AND FURTHER DISCUSSION
Currently,very few projects focus on grid work?ow veri?cation and validation.This is not surprising as grid work?ow has been investigating for just a few years.Many efforts have been made on the research and development of functional issues such as modelling languages,grid work?ow engines and so on[9,11,13,38].Gradually,more and more attention has been paid to QoS scheduling and overall management[35,51,56].However,very little work has been done on grid work?ow veri?cation and validation which is an important non-functional issue.More or less,the very Copyright2007John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2008;20:347–360
DOI:10.1002/cpe
356J.CHEN AND Y.YANG
Table I.Grid work?ow veri?cation taxonomy mapping.
Project name Structure Performance Resource Authorization Cost Temporal
DILIGENT Supported Supported N/A N/A N/A N/A
CROWN Partly supported N/A N/A N/A N/A N/A
Discovery Net Supported N/A N/A N/A N/A N/A
SwinDeW-G N/A N/A N/A N/A N/A Supported
CAT Supported N/A N/A N/A N/A N/A
Table II.Grid work?ow validation taxonomy mapping.
Representation of Representation of Validation
Project name process expressive power approaches
DILIGENT N/A N/A N/A
CROWN N/A Supported N/A
Discovery Net N/A N/A N/A
SwinDeW-G N/A N/A N/A
CAT N/A N/A N/A
few projects related to grid work?ow veri?cation and validation mainly include Digital Library Infrastructure on Grid Enabled Technology(DILIGENT)[57],China R&D Environment over Wide-area Network(CROWN)[58],Discovery Net[59],Swinburne Decentralised Work?ow for Grid (SwinDeW-G)[60]and Composition Analysis Tool(CAT)[61].We show the comparison of these projects based on grid work?ow veri?cation taxonomy in Table I,and based on grid work?ow validation taxonomy in Table II.
We now describe those projects brie?y.
DILIGENT[57]aims at supporting this new research operational mode by providing a knowl-edge infrastructure that manages a network of shared resources(e.g.archives,database and software tools)and enables the creation of on-demand digital libraries.It supports process design by com-posing existing services.Performance and structure veri?cation based on the process description are supported.
CROWN[58]is a grid toolkit developed by BeiHang University in China.It contains a grid work?ow modelling language called Grid Process Execution Language(GPEL).Based on GPEL, CROWN supports part of the structure veri?cation to identify deadlock and loss of synchro-nization.Besides,it also analyses the expressive power of GPEL.This is part of grid work?ow validation.
Discovery Net project[59]is an Engineering and Physical Sciences Research Council funded project to build the world’s?rst e-Science platform for scienti?c discovery from the data generated by a wide variety of high-throughput devices at Imperial College London.Its work?ow management supports structure veri?cation.
SwinDeW-G[60]is a peer-to-peer-based grid work?ow management system.For the moment, it can support only grid work?ow temporal veri?cation.It is on the way in developing approaches for other types of veri?cation and validation.
Copyright2007John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2008;20:347–360
DOI:10.1002/cpe
A TAXONOMY OF GRID WORKFLOW VERIFICATION AND V ALIDATION357 CAT[61]developed by University of Southern California is a tool that analyses grid work?ows and generates error messages and suggestions in order to help users compose complete and consistent work?ows.CAT supports structure veri?cation.
From Table I,we can see that each project can support at least one type of grid work?ow veri?cation,but not all types.From Table II,we can see that most of the projects do not support validation.Given that many grid work?ow management systems have been developed and can well support functional issues such as modelling languages or execution engines[38],more efforts should now be devoted to grid work?ow veri?cation and validation so that we can develop a set of complete veri?cation and validation approaches.Accordingly,we are able to check the correctness of grid work?ow speci?cation and execution.This correctness is important as it directly affects the applicability of a grid work?ow management system.
4.CONCLUSIONS AND FUTURE WORK
In this paper,we have systematically analysed grid work?ow veri?cation and validation,and clas-si?ed their key issues into a taxonomy.Some of them are not investigated by current research such as mapping validation approaches or cost veri?cation,while others may need further investigation. The taxonomy depicts a picture and provides some insights into further potential research points for grid work?ow veri?cation and validation.Our ongoing and future work is to investigate such points,develop and evaluate corresponding veri?cation and validation approaches. ACKNOWLEDGEMENTS
A preliminary version of this paper appeared in AusGrid2006[62].
REFERENCES
1.Abramson D,Kommineni J,McGregor JL,Katzfey J.An atmospheric sciences work?ow and its implementation with
web services.Proceedings of the4th International Conference on Computational Science,Part I(Lecture Notes in Computer Science,vol.3036).Springer:Berlin,Krakow,Poland,June2004;164–173.
2.Buyya R,Enugopal SV.The Gridbus toolkit for service oriented grid and utility computing:An overview and status report.
Technical Report,GRIDS-TR-2004-2,Grid Computing and Distributed Systems Laboratory,University of Melbourne, Australia,2004.Available at:https://www.doczj.com/doc/382597531.html,/papers/gridbus2004.pdf[1February2007].
3.Buyya R,Abramson D,Venugopal S.The grid economy.Proceedings of the IEEE2005;93(3):698–71
4.
4.Foster I,Kesselman C,Tuecke S.The anatomy of the grid:Enabling scalable virtual organizations.International Journal
of Supercomputing Applications2002;15(3):200–222.
5.Foster I,Kesselman C,Nick J,Tuecke S.The physiology of the grid:An open grid services architecture for distributed
systems integration.Globus Project,2002.Available at:https://www.doczj.com/doc/382597531.html,/alliance/publications/papers/ogsa.pdf [1February2007].
6.Simpson DR,Kelly N,Jithesh PV,Donachy P,Harmer TJ,Perrott RH,Johnston J,Kerr P,McCurley M,McKee S.
GeneGrid:A practical work?ow implementation for a grid based Virtual Bioinformatics Laboratory.Proceedings of the UK e-Science All Hands Meeting2004(AHM04),Nottingham,https://www.doczj.com/doc/382597531.html, Engineering and Physical Science Research Council:Swindon,U.K.,September2004;547–554.
7.Deelman E,Blythe J,Gil Y,Kesselman C,Mehta G,Vahi K.Mapping abstract complex work?ows onto grid environments.
Journal of Grid Computing2003;1(1):9–23.
Copyright2007John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2008;20:347–360
DOI:10.1002/cpe
358J.CHEN AND Y.YANG
8.Marinescu D.A grid work?ow management architecture.Global Grid Forum White Paper,2002.Available at:
https://www.doczj.com/doc/382597531.html,/mail archive/gce-wg/2002/Archive/pdf00003.pdf[1February2007].
9.Cybok D.A grid work?ow infrastructure.Concurrency and Computation:Practice and Experience,Special Issue on
Work?ow in Grid Systems2006;18(10):1243–1254.
10.Huang Y.JISGA:A JINI-BASED service-oriented grid architecture.The International Journal of High Performance
Computing Applications2003;17(3):317–327.
11.Krishnan S,Wagstrom P,Laszewski GV.GSFL:A work?ow framework for grid services.Technical Report,Argonne
National Laboratory,Argonne,U.S.A.,2002.Available at:https://www.doczj.com/doc/382597531.html,/cog/papers/gs?-paper.pdf [1February2007].
12.Work?ow Management Coalition:The Work?ow Reference Model.TC00-1003,1995.
13.Fahringer T,Pllana S,Villazon A.A-GWL:Abstract Grid Work?ow Language.Proceedings of the4th International
Conference on Computational Science,Part III,Krakow,Poland(Lecture Notes in Computer Science,vol.3038).Springer: Berlin,June2004;42–49.
14.Yu J,Buyya R.A novel architecture for realizing grid work?ow using tuple spaces.Proceedings of the5th IEEE/ACM
International Workshop on Grid Computing(GRID’04)—Volume00,Los Alamitos,U.S.A.IEEE Computer Society Press:Washington,DC,U.S.A.,November2004;119–128.
15.Amin K,Laszewski GV,Hategan M,Zaluzec NJ,Hampton S,Rossi A.GridAnt:A client-controllable grid work?ow.
Proceedings of the37th Annual Hawaii International Conference on System Sciences(HICSS’04),Hawaii.IEEE Computer Society Press:Washington,DC,U.S.A.,January2004;210–219.
16.Cao J,Jarvis SA,Saini S,Nudd GR.GridFlow:Work?ow management for grid computing.Proceedings of IEEE/ACM
International Symposium on Cluster Computing and the Grid(CCGrid2003),Tokyo.IEEE Computer Society Press: New York,U.S.A.,May2003;198–205.
17.Chen J.Yang Y.Multiple states based temporal consistency for dynamic veri?cation of?xed-time constraints in grid
work?ow systems.Concurrency and Computation:Practice and Experience2007;19(7):965–982.
18.IEEE.Standard Glossary of Software Engineering Terminology.In IEEE Software Engineering Standards Collection,
IEEE,Std610.12-190,1994.
19.IEEE-SA Standards Board.IEEE Standard for Software Veri?cation and Validation,IEEE,Std1012,1998
20.van der Aalst WMP.The application of Petri Nets to work?ow management.Journal of Circuits,Systems and Computers
1998;8(1):21–66.
21.Adam N,Aturi V,Huang W.Modeling and analysis of work?ows using Petri Nets.Journal of Intelligent Information
Systems,Special Issue on Work?ow and Process Management1998;10(2):131–158.
22.van der Aalst WMP,Hofstede AHMTer,Kiepuszewski B,Barros AP.Work?ow patterns.Distributed and Parallel
Databases2003;14(3):5–51.
23.van der Aalst WMP.Work?ow veri?cation:Finding control-?ow errors using Petri-net based techniques.Proceedings
of the International Conference on Business Process Management:Models,Techniques,and Empirical Studies(Lecture Notes in Computer Science,vol.1806).Springer:Berlin,Germany,2000;161–183.
24.van der Aalst WMP.Work?ow Management,Models,Methods,and Systems.The MIT Press:Cambridge,MA.
25.Li H,Yang Y,Chen TY.Resource constraints analysis of work?ow speci?cations.The Journal of Systems and Software
2004;73(2):271–285.
26.Li H,Yang Y.Dynamic checking of temporal constraints for concurrent work?ows.Electronic Commerce Research and
Applications2005;4(2):124–142.
27.Bertino E,Ferrari E,Atluri V.An authorisation model for supporting the speci?cation and enforcement of authorisation
constraints in work?ow management systems.ACM Transactions on Information System Security1999;2(1):65–104.
28.Wu S,Sheth A,Luo Z.Authorisation and access control of application data in work?ow systems.Journal of Intelligent
Information Systems2002;18(1):71–94.
29.Marjanovic O,Orlowska ME.1999.On modelling and veri?cation of temporal constraints in production work?ows.
Knowledge and Information Systems1999;1(2):157–192.
30.Chen J,Yang Y.Temporal dependency for dynamic veri?cation of?xed-date constraints in grid work?ow systems.
Proceedings of the7th Asia Paci?c Web Conference(Lecture Notes in Computer Science,vol.3399).Springer:Berlin, Germany,March2005;820–831.
31.Chen J,Yang Y,Chen TY.Dynamic veri?cation of temporal constraints on-the-?y for work?ow systems.Proceedings of
Asia-Paci?c Software Engineering Conference,Korea.IEEE CS Press:New York,U.S.A.,November/December2004;
30–37.
32.Chen J,Yang Y.Activity completion duration based checkpoint selection for dynamic veri?cation of temporal
constraints in grid work?ow systems.The International Journal of High Performance Computing Applications 2006,in press.Available at:https://www.doczj.com/doc/382597531.html,.au/personal/jchen/docs/Chen Yang IJHPCA.pdf[1February 2007].
33.Chen J,Yang Y.A minimum proportional time redundancy based checkpoint selection strategy for dynamic veri?cation
of?xed-time constraints in grid work?ow systems.Proceedings of the12th Asia Paci?c Software Engineering Conference (APSEC2005),Taiwan.IEEE Computer Society Press:Washington,DC,U.S.A.,December2005;299–306.
Copyright2007John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2008;20:347–360
DOI:10.1002/cpe
A TAXONOMY OF GRID WORKFLOW VERIFICATION AND V ALIDATION359
34.Chen J,Yang Y.Adaptive selection of necessary and suf?cient checkpoints for dynamic veri?cation of temporal constraints
in grid work?ow systems.ACM Transactions on Autonomous and Adaptive Systems(TAAS)2007,in press.Available at:https://www.doczj.com/doc/382597531.html,.au/personal/jchen/papers/ACMTAAScheckpoint.pdf[10May2007].
35.Yu J,Buyya R,Tham CK.QoS-based scheduling of work?ow applications on service grids.Proceedings of1st IEEE
International Conference on e-Science and Grid Computing(e-Science2005),Melbourne,Australia.IEEE Computer Society Press:Washington,DC,U.S.A.,December2005;140–147.
36.Sadiq W,Orlowska ME.Analysing process models using graph reduction https://www.doczj.com/doc/382597531.html,rmation Systems2000;
25(2):117–134.
37.Son JH,Kim MH.Improving the performance of time-constrained work?ow processing.The Journal of Systems and
Software2001;58(3):211–219.
38.Yu J,Buyya R.A taxonomy of scienti?c work?ow systems for grid computing.Special Issue on Scienti?c Work?ows,
SIGMOD Record2005;34(3):44–49.
39.Fabio C,Silvana C,Mariagrazia F.Managing work?ow authorisation constraints through active database technology.
Information Systems Frontiers2001;3(3):319–338.
40.Chinn S,Madey G.Temporal representation and reasoning for work?ow in engineering design change review.IEEE
Transactions on Engineering Management2000;47(4):485–492.
41.Li H,Fan Y.Work?ow model analysis based on time constraint Petri Nets.Journal of Software2004;15(1):17–26.
42.Li J,Fan Y,Zhou M.Timing constraint work?ow nets for work?ow analysis.IEEE Transactions on Systems,Man and
Cybernetics—Part A:Systems and Humans2003;33(2):179–193.
43.Zhuge H,Cheung T,Pung H.A timed work?ow process model.The Journal of Systems and Software2001;55(3):231–243.
44.Eder J,Panagos E,Rabinovich M.Time constraints in work?ow systems.Proceedings of the11th International Conference
on Advanced Information Systems Engineering(CAiSE’99),Heidelberg,Germany(Lecture Notes in Computer Science, vol.1626).Springer:Berlin,June1999;286–300.
45.Hagen C,Alonso G.Exception handling in work?ow management systems.IEEE Transactions on Software Engineering
2000;26(10):943–958.
46.Chen J,Yang Y.Dynamic setting,veri?cation and adjustment of upper bound constraints in grid work?ow systems.
Proceedings of the2nd International Conference on Semantics,Knowledge and Grid(SKG2006),Guilin,IEEE Digital Library,China.IEEE Computer Society Press:Washington,DC,U.S.A.,October/November2006.
47.Bussler C.Work?ow instance scheduling with project management tools.Proceedings of Workshop on Database and
Expert Systems Applications(DEXA’98)(Lecture Notes in Computer Science,vol.1460).Springer:Berlin,Germany, August1998;753–758.
48.Cardoso J,Sheth A,Miller J.Quality of service and semantic composition of work?ows.PhD Thesis,University of
Georgia,Athens,GA,2002.
49.Pozewaunig H,Eder J,Liebhart W.ePERT:Extending PERT for work?ow systems.Proceedings of EastEuropean
Symposium on Advances in Database and Information Systems(ADBIS’97),St Petersburg,Russia,September1997;
217–224.
50.Brandic I,Benkner S,Engelbrecht G,Schmidt R.Towards quality of service support for grid work?ows.Proceedings
of the European Grid Conference2005(EGC2005),Amsterdam,The Netherlands(Lecture Notes in Computer Science, vol.3470).Springer:Berlin,February2005;661–670.
51.Brandic I,Benkner S,Engelbrecht G,Schmidt R.QoS support for time-critical grid work?ow applications.Proceedings
of the1st IEEE International Conference on e-Science and Grid Computing(e-Science2005),Melbourne,Australia.
IEEE Computer Society Press:Washington,DC,U.S.A.,December2005;108–115.
52.Hennessy M.Algebraic Theory of Processes(Series in the Foundations of Computing).MIT Press:Cambridge,MA,
1998.
53.Sangiorgi D,Walker D.The Pi-Calculus:A Theory of Mobile Processes.Cambridge University Press:England,2004.
54.Amit KC,Munindar https://www.doczj.com/doc/382597531.html,mitments for?exible business processes.Proceedings of the3rd International Joint
Conference on Autonomous Agents and Multi Agent Systems(AAMAS2004),New York,U.S.A.IEEE Computer Society Press:New York,U.S.A.,August2004;1362–1363.
55.Davulcu H,Kifer M,Ramakrishnan CR,Ramakrishnan IV.Logic based modelling and analysis of work?ows.Proceedings
of ACM Symposium on Principles of Database Systems(PODS1998),Seattle,Washington,U.S.A.ACM Press:New York,June1998;25–33.
56.Al-Ali R,Amin K,Laszewski GV,Rana O,Walker D,Hategan M,Zaluzec N.Analysis and provision of QoS for
distributed grid applications.Journal of Grid Computing2004;2(2):163–182.
57.DELIGENT.Available at:https://www.doczj.com/doc/382597531.html,/.
58.CROWN Team.CROWN portal.Available at:https://www.doczj.com/doc/382597531.html,/en/.
59.Discovery Net project.Available at:https://www.doczj.com/doc/382597531.html,/.
60.SWINDEW-G Team.System Architecture of SwinDeW-G.Available at:https://www.doczj.com/doc/382597531.html,.au/personal/jchen/
SwinDeW-G/System Architecture.pdf.
Copyright2007John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2008;20:347–360
DOI:10.1002/cpe
360J.CHEN AND Y.YANG
61.Kim J,Spraragen M,Gil Y.An intelligent assistant for interactive work?ow composition.Proceedings of the9th
International Conference on Intelligent User Interface(IUI2004),Funchal,Madeira,Portugal.ACM Press:New York, January2004;125–131.
62.Chen J,Yang Y.Key Research issues in grid work?ow veri?cation and validation.Proceedings of4th Australasian
Symposium on Grid Computing and e-Research(AusGrid2006),vol.28,no.7,ACSW Frontiers2006.Australian Computer Science Communications:Hobart,Australia,January2006;97–104.
Copyright2007John Wiley&Sons,Ltd.Concurrency Computat.:Pract.Exper.2008;20:347–360
DOI:10.1002/cpe
二次开发培训文档 一、ECOLOGY系统框架结构 1、主要的程序结构 Ecology Classbean 存放编译后的CLASS文件 js 系统中使用的JA V ASCRIPT和VBSCRIPT脚本 Css 系统中JSP页面使用的样式 Images Images_face Images_frame 系统中使用的图片的存放目录 Crm Workflow 该功能分文件夹存放每个功能的文件 WEB-INF Prop 系统配置文件存放 Service 系统的接口配置文件的存放 二、说明一个JSP页面，一个JA V A程序的基本组成，如何阅读JSP页面 1、一个jsp页面通常需要包含什么内容 2、如何阅读一个JSP页面 由于ECOLOGY系统支持多语言，因此在JSP页面上一般不出现中文，全部使用 标签的形式来显示中文： 比如：在IE上显示“姓名”那么在JSP页面中将通过 <%=SystemEnv.getHtmlLabelName(413,user.getLanguage())%>这样的形式来表示，其中的数字413就是表示姓名，同时可以通过“select labelname from htmllabelinfo where indexid=413 and languageid=7”来获取到“姓名”这个显示名称，其中 languageid=7表示中文显示名称,languageid=8表示英文显示名称. delete from HtmlLabelIndex where id=81249 delete from HtmlLabelInfo where indexid=81249 INSERT INTO HtmlLabelIndex values(81249,'选择范围') INSERT INTO HtmlLabelInfo VALUES(81249,'选择范围',7) INSERT INTO HtmlLabelInfo VALUES(81249,'Range of choice',8) INSERT INTO HtmlLabelInfo VALUES(81249,'選擇範圍',9) 3、JA V A程序的基本组成 在ECOLOGY中开发JA V A程序建议继承weaver.general. BaseBean，在BaseBean 中主要封装了两个方法：写日志文件，获取配置文件中的参数值。 public String getPropValue(String fname , String key) public void writeLog(Object obj)
1WorkflowRuntime 名称空间:System.Workflow.Runtime.WorkflowRuntime 文件:system.workflow.runtime.dll 在工作流的运行环境（宿主）中,为工作流提供运行的引擎 1．WorkflowRuntime在宿主中以自已独立的线程运行例 2．WorkflowRuntime可以加载多个工作流实例，每个工作流实例在WorkflowRuntime有独立的线程 3.同一宿主可可以实例化多个引擎，他们可以同时工作 2引擎的状态管理 3使用引擎创建实例 3.1使用工作流类创建实例
3.2使用XmlReader创建实例 3.3用自定义的GUID创建实例
3.4创建实例时指定参数 可以在用引擎创建一个工作流实例时，用System.Collections.Generic.Dictionary
4从引擎中得到工作流实例4.1得到内存中所有工作流实例的集合 4.2得到指定GUID的工作流实例
5引擎事件5.1与引擎状态相关的事件 事件类型:EventHandler
5.3与实例运行相关的事件 事件类型:EventHandler
Ora E-B O RA Wor Author: MSN: Creation Last Up Docume Version Approv
Workflow File Ref: 深入浅出Oracle EBS之Workflow实例详解.docx (v. DRAFT 1A ) Document Control ii Document Control Change Record Date Author Version Change Reference 17-Apr-07 Jianhua.Huang Draft 1a No Previous Document Reviewers Name Position Distribution Copy No. Name Location 1Library Master Project Library 2Project Manager 3 4 Note To Holders: If you receive an electronic copy of this document and print it out, please write your name on the equivalent of the cover page, for document control purposes. If you receive a hard copy of this document, please write your name on the front cover, for document control purposes.
INFORMATICA关于WORKFLOW Manager系统的元数据 解析 INFORMATICA关于WORKFLOW Manager系统的元数据解析关键词：INFORMATICA,WOR Manager,元数据 informaica是一个很强大的ETL工具。其WORKFLOW MANAGER负责对ETL调度流程进行设计与管理和执行！informatica在在资料库中提供以下表来存储调动流程的相关信息。 以便WORKFLOW MANAGER对用户所设计的调动流程进行管理和执行。 opb_wflow_dep:描述workflow执行步骤相关信息和每个步骤执行的条件信息 opb_wflow_dep_run：描述workflow执行步骤运行时相关信息 opb_wflow_expr ：描述workflow中相关的表达式或条件的相关信息 opb_wflow_perval：描述workflow可持续性变量相关信息opb_wflow_run:描述workflow运行日志相关信息 opb_wflow_var:描述workflow变量相关信息
opb_task:描述任务对象的基本信息 opb_task_attr:描述任务对象相关的属性的信息 opb_task_inst:描述任务对象实例的基本信息 opb_task_inst_run:描述任务对象实例运行日志相关信息opb_task_val_list:描述任务对象实例中command信息WORKFLOW MANAGER系统中常用的有这几个模块，Command模块， Session模块， Waiting_Event模块， Raising_Event模块， Assignment模块， Worklet模块 WORKFLOW MANAGER系统中上述的这些模块统称为任务(Task).如果你对一个模块进行了 复制后新的模块就称作该任务的任务实例（Task_Inst）. WORKFLOW MANAGER系统中Worklet模块可以有其他非Worklet模块组成。 在WORKFLOW MANAGER系统中一个工资流被称作Workflow,Workflow由各种任务模块组合而成。 同时一个Workflow也是一个任务。以下是WORKFLOW 元数据表的详细说明， -----------------------------------------------------------------------
导航 ?首页 ?日志 ?相册 ?音乐 ?收藏 ?博友 ?关于我 日志 小翼 程序员就像男人,语言就像女人,每个男人都想要很多女人,却很少有男人能真正了解一个女人,因为男人总是朝三暮四,而女人每天都在变,甚至有些是经过变性和整容的。 加博友关注他 最新日志 ?飘荡 ?游走... ?过客.. ?...记录 ?拯救... ?迷失? 该作者的其他文章
博主推荐 相关日志 随机阅读 首页推荐 ?俄监狱举办选美大赛 ?街拍“大宋”美女游郑州 ?韩女星遗书曝被迫卖身 ?我在美国野外露营装备 ?为毛泽东看病的六大御医 ?3种易被小三勾引的男人 更多>> 对“推广广告”提建议 应届毕业生必看（转） （转载）和我一起学Windows Workflow Foundation(2) 让WF通过参数接收数据 （转载）和我一起学Windows Workflow Foundation(1)-----创建和调 试一个WF实例） 工作流研究2009-03-13 12:09:13 阅读196 评论0 字号：大中小订阅 今天开始，我打算开始学习WWF，从网上搜索到了部分相关资料，也找到了一些文档和实验。但是，资料以英文的占多数，所以，在学习起来似乎比较吃力，不过相信我能坚持下来，顺便提高点英语阅读能力，不过本人英文水平实在是差，解释的不到位或错误的地方请大家谅解（千万别笑话我，哈哈）。毕竟我也是从空白开始学习WWF，让我们一起进步。 首先，我们需要安装WinFX（下载）和Visual Studio 2005 extensions for .NET Framework 3.0 (Windows Workflow Foundation)。这是我们必备的开发组件。 这些例子和教程是从微软的labs上下载到的，分为10个部分。先来学最基础的第一部分先：） 第一部分的目的 这个试验的目的是介绍Windows Workflow Foundation的工作流概念的关键点 完成了这个试验以后，我们可以学习到: ·使用Visual Studio 2005 为WWF设计顺序工作流 ·配置和使用Visual Studio 2005调试器调试你的工作流
4 YAWL：Yet Another Workflow Language 定义3：是一个扩展工作流网。 ， 。此外，辅助函数， 是用来定义每个节点各自分配的预设和后置。对任何节点，均满足：和 。 注意：预设和后置函数和依赖于上下文，例如：扩展工作流网中的函数的使用性。如果一个节点在多个网中使用，就会不清楚这个节点的预设和后置函数是和哪个网相关的。因此，我们当有歧异产生的时候我们增加对这个网中的节点的预设和后置的注释：表示网N中节点x的预设，表示在网N中节点x的后置。 定义 4. 每当我们介绍一个工作流规范，我们假设 按照下面的方式定义： 表示一系列的原子任务。 表示一系列的符合任务。 表示一系列的单实例任务。 表示一系列（潜在的）多实例任务，和 是所有条件的扩展集。 如果表示不明确，我们将会使用下标，如：。如果在上下文中只有一个工作流规范，我们将省略下标。此外，域函数 在不同的中没有重叠的部分，我们也可以省
略下标。 一个工作流规范定义了一个树形结构，为了有效的使用这种结构，我们定义了函数，给出一系列的节点（如任务和条件），函数将会返回这些节点和所有的孩子节点。 定义5：是一个工作流规范，定义函数： ，如下，对： 注意：返回在X中的每个节点和被这些节点包含的节点。对于原子任务和条件，不需要展开。对于复合任务，被这些任务包含的所有节点都包含在结果中，如递归的横穿在X中的所有复合任务。 图六给出了一个工作流规范的例子来解释注意。
是一系列的原子任务，是复合任务。 和。如D是唯一的含有多实例的任务（潜在的）。除了输入和输出条件，只有两个明确的条件。在集中所有的其他的条件都是含蓄的，如是一个含蓄的条件，它和连接A和B之间的弧相对应。如，如果D是展开的，那么D所包含的所有的任务和条件（包括含蓄的）都会被添加进来。 4.2 Semantics 定义2从数学的角度定义了工作流规范的语法。在这个定义的基础上，可以直接的给出一个工作流语言，如和XML相关的语言。然而在定义2中没有给出任何的语义，因此到目前为止我们只是给出了一个对工作流规范S的动态行为的直观的描述。在本节的其余部分我们将给出正式的语义。主要分3步进行，首先我们定义一些预备知识（包和标识符）。然后我们定义和工作流规范S相对应的状态空间。最后，我们给出状态转换的可能性。 Preliminaries(预备知识) 和工作流规范相对应的状态空间的定义是受有色Petri 网中的标记的概念的启发才这样定义的。状态空间包括了一些带有值的标记。既然我们从这篇文章中提取数据，因此要满足每个标记都有一个名字。我们需要依靠包来解决多个标记在同一个库所中有相同的名字的问题。此外，我们需要构造一系列的案例标识符来关联子实例和父实例，这是解决多实例问题所必须的。 在这篇文章中，包被定义为字母表A中的有限多集的元素.在字母表A中上的包可以被认为是从A到自然数N的一个函数，这样就只有A中的有限数目的元素可以分配一个非0的函数值。A上的包X，且，则表示a在X中出现的次数，通常称为X中a的基数。A上的所有包集用表示。对于一个包的明确显示，一个注释类
工作流引擎（Workflow Engine ） 所谓工作流引擎是指workflow作为应用系统的一部分，并为之提供对各应用系统有决定作用的根据角色、分工和条件的不同决定信息传递路由、内容等级等核心解决方案。 工作流引擎（Workflow Engine ） 什么是工作流引擎（Workflow Engine ） 例如开发一个系统最关键的部分不是系统的界面，也不是和数据库之间的信息交换，而是如何根据业务逻辑开发出符合实际需要的程序逻辑并确保其稳定性、易维护性（模块化和结构化）和弹性（容易根据实际业务逻辑的变化作出程序上的变动，例如决策权的改变、组织结构的变动和由于业务方向的变化产生的全新业务逻辑等等）。 Workflow 引擎解决的就是这个问题：如果应用程序缺乏强大的逻辑层，势必变得容易出错（信息的路由错误、死循环等等）。 就好比一辆汽车，外表做得再漂亮，如果发动机有问题就只是一个摆设。应用系统的弹性就好比引擎转速方面的性能，加速到100 公里需要1 个小时（业务流程发生变动需要进行半年的程序修改）还能叫好车吗？引擎动不动就熄火（程序因为逻辑的问题陷入死循环）的车还敢开吗？ 工作流解决方案与传统管理软件的关系 传统的管理软件注重解决企业应用层现存的问题（例如提高企业的资源配置率或提高单一员工的生产效率）。例如：EXCEL 可以提高员工画表格的效率、财务软件可以规范财务人员的工作并提高账目查询的效率、CRM 可以规范客户管理从而使客户资源掌握在公司手中而不是被一部分业务人员把持并提高客户响应时间、ERP 解决的是如何配置企业资源：使企业的人力资源、财力资源和物资资源能够根据业务的需求实现最大化配置。workflow 关注的是如何缩短流程闲置时间，从而提高企业的业务处理能力并使企业能够关注于真正对企业有意义的增值业务上。从建立企业神经系统的角度也许更能理解两者的区别。传统软件不能解决工作流的问题，例如ERP 关注的是企业的资源配置，但不可能解决资源传输过程中的损耗和降低传输（流程）的成本；同样workflow也不能完全解决传统管理软件所能解决的问题，例如对生产管理的MRP 系统所能解决的生产过程控制通过workflow很难实现。但一个好的传统软件如果希望能自动化地在整个企业
工作流控制方案 1流程方案 流程控制整体思路为利用Wordflow自身机制对文档进行路由，路由完毕后对活动的可能所有者、活动的读者按照定制域设定及所选人员进行修正。各类路由机制详细说明如下： 1.1 一般路由 从活动A路由到活动B，不关心活动B是否已完成，取新选择的人员作为活动可能所有者，这是流程中最常用的路由类型，路由图示说明如下： 活动A可以多人也可以单人，活动A为多人处理时，若非决策路由（设置DispChoice定制域），多人操作相同，都不选择人员，活动B的活动所有者根据公式自动计算，如发文会签完毕到拟稿人确定的路由、甘肃督办拟稿人确定到会签的路由等；若有多个决策路由，可能仅第一个处理的人选择决策完成路由，其他人不再处理（甘肃机制、西北网未设置MultiChoice时机制），目前无此实例；也可能多人操作相同，都不选择人员，活动B的活动所有者根据公式自动计算（西北网设置MultiChoice时机制，同时需设置NoSelChoice），如西北网部主任批示到处长批办的路由；活动A为单人处理时，可以由公式自动计算也可以手动选择人员，如起草到部主任审核的路由、秘书核稿到领导签发的路由、西北网部主任批示到处长批办的路由等。若活动B的活动所有者为自动计算，计算结果为空时则会引起路由出错。 此路由之后对流程的各种特殊操作情况有： 1)、收回。活动A最后一个处理人可以收回，收回后工作路由到活动A，活动所有者为活动A的最后一个处理人，收回后不能再返上一人。
2)、返回上一人。活动B的第一个处理人返回上一人，工作路由到活动A，活动所有者为活动A的最后一个处理人。返回上一人后活动B的第一个处理人可以再收回，收回后工作路由到活动B，活动所有者为活动B的第一个处理人；活动A的最后一个处理人也可以返回上一人，返回后工作路由到活动B，活动所有者为活动B的第一个处理人，以后可以在A、B活动间循环。 3)、特送。活动B的任一处理人可以进行特送到另一活动C，特送按照流程配置选择活动C的活动所有者，特送后特送人可以收回，收回后工作路由到活动B，活动所有者为特送人；活动C的第一个处理人可以返上一人，返回后工作路由到活动B，活动所有者为特送人，若特送人在返回上一人，则会在特送人和活动C的第一个处理人间循环；活动C的部分人处理后，特送人可以上一节点收回，收回后工作路由到活动B，活动所有者为特送人。 4)、上一节点收回。活动B的部分人处理后，活动A最后一个处理人可以上一节点收回，上一节点收回后工作路由到活动A，活动所有者为活动A的最后一个处理人。 5)、管理特送。应用管理员可以管理特送到活动C，特送时可在组织机构中选择任一人员作为活动C的活动所有者，管理特送同时更新了流程版本（西北网），但必须保证管理特送时按照正常流程到活动C也可以正常流转才可特送成功，管理特送后不能再收回、返回上一人、上一节点收回。 1.2 返回路由 从活动B路由到活动A，活动A在之前已经完成并记录了处理人（通过设置destitem定制域），取活动A的最后一个处理人（西北网机制）或活动A的全部处理人（甘肃机制），路由图示说明如下：
报销流程开发概述
徐栋 微软金牌讲师 北京中达金桥技术服务有限公司
议程
? ? ? ? Windows i d Workflow kfl Foundation简介 d i 简介 报销流程概述 WF开发生态系统 总结
Windows Workflow Foundation
为微软产品及相关应用程序提供通用的工作流 设计平台和开发工具
? 统一的工作流技术 可用于Windows之上的所有应用 可用于跨应用场景 ? 重新定义工作流 构建以工作流为中心的可扩展框架及API 可用于Human和System的工作场景 ? 最核心的工作流框架 集成的开发环境/语言 Office12工作流引擎的基础部分 Microsoft Confidential 强大的 Partner支持
Windows Workflow Foundation
? 核心概念
A Workflow An Activity App Activity Library Workflow Foundation Base Activity Library ? workflow constructs Runtime Engine ? intrinsic behavior Runtime Services ? hosting flexibility Host Process ? my app or server
? Workflow是一组Activities ? Workflow在一个宿主应用 程序中运行：任意应用程序 或服务 ? 开发人员可以创建自己的 Activity库
? Base Activity Library：内 内 置的基本的Activity ? Runtime Engine：运行 Workflow与状态管理 ? Runtime Services：宿主 Workflow与通讯 ? Visual Designer：控件， 可以在应用程序中调用设计 Microsoft Confidential 器
? 组件
工作流包装 1.运行时包装 using System; using System.Collections.Generic; using System.Linq; using System.Text; using System.Threading; using System.Workflow.Runtime; namespace Nocturne.Workflow.Hosting { public class WorkflowRuntimeManager:IDisposable { private WorkflowRuntime _workflowRuntime; private Dictionary
手把手教你使用Gocad（实例） 1、首先通过DXF文件载入等高线，curve—9110、9120，由pointset mode生成等高点，如上图白色的点 2、通过wizards生成面 在surface creation中选择From data (without internal border) 3、点击creat outline curve from points和optiamize outline curve tobetter fit points,同时可以多次点击右侧的螺旋形按钮，多次优化；达到你满意的效果后，如果你需要保存生成outline就要在objects中进行重命名和保存。 4、点击creat surface来生成面，hide constraints用来隐藏不需要看到的contraints 5、上面就是生成的带有网格的等高面 6、将网格去掉（可以在attributes中去） 6、选择我们要载入其他的点（点必须是有坐标xyz），在生成我们向要的地层。在载入点后，curve mode—new—curvehull—of object来生成outline curve（这是另外一种生成outline curve的方法） 7、通过生成的outline curve和载入的点来得到一个面 8、这是生成的去掉网格以后的面 9、载入其他点，生成其他我们需要的面（古滑坡堆积体层） 9、宣威组滑带 10、现在我们来生成一个侧面，主要用来作为以后编辑切割其他地层的一个界面。在camera中仅显示等高面的outline courve，surface mode—new—build informs—tube 11、在expansion中键入你需要的值，如需要一个垂直的侧面则只修改Z值就可，同时选择twoway 从正负两个方向在延伸，这种方法还可以生成封闭的面（closed surface），选择seal ends即可 12、生成的侧面如图所示，白色的面是等高面的边界线 13、这时将刚才生成的洪积层、冰川堆积层、古滑坡面、宣威组滑带面和玄武岩层分别延伸surfacemode—border—expand--one
仅供个人参考 二次开发培训文档 一、ECOLOGY系统框架结构 1、主要的程序结构 Ecology For personal use only in study and research; not for commercial use Classbean 存放编译后的CLASS文件 js 系统中使用的JA VASCRIPT和VBSCRIPT脚本 Css 系统中JSP页面使用的样式 Images Images_face Images_frame 系统中使用的图片的存放目录 Crm Workflow 该功能分文件夹存放每个功能的文件 WEB-INF Prop 系统配置文件存放 Service 系统的接口配置文件的存放 二、说明一个JSP页面，一个JA VA程序的基本组成，如何阅读JSP页面 1、一个jsp页面通常需要包含什么内容 2、如何阅读一个JSP页面 由于ECOLOGY系统支持多语言，因此在JSP页面上一般不出现中文，全部使用标签的形式来显示中文： 比如：在IE上显示“姓名”那么在JSP页面中将通过 <%=SystemEnv.getHtmlLabelName(413,user.getLanguage())%>这样的形式来表示，其中的数字413就是表示姓名，同时可以通过“select labelname from htmllabelinfo where indexid=413 and languageid=7”来获取到“姓名”这个显示名称，其中languageid=7表示中文显示名称,languageid=8表示英文显示名称. delete from HtmlLabelIndex where id=81249 delete from HtmlLabelInfo where indexid=81249 INSERT INTO HtmlLabelIndex values(81249,'选择范围') INSERT INTO HtmlLabelInfo VALUES(81249,'选择范围',7) INSERT INTO HtmlLabelInfo VALUES(81249,'Range of choice',8) INSERT INTO HtmlLabelInfo VALUES(81249,'選擇範圍',9)
泛微O A e c o l o g y二次开发实例开发完整 说明
二次开发培训文档 一、ECOLOGY系统框架结构 1、主要的程序结构 Ecology Classbean 存放编译后的CLASS文件 js 系统中使用的JAVASCRIPT和VBSCRIPT脚本 Css 系统中JSP页面使用的样式 Images Images_face Images_frame 系统中使用的图片的存放目录 Crm Workflow 该功能分文件夹存放每个功能的文件 WEB-INF Prop 系统配置文件存放 Service 系统的接口配置文件的存放 二、说明一个JSP页面，一个JAVA程序的基本组成，如何阅读JSP页面 1、一个jsp页面通常需要包含什么内容 2、如何阅读一个JSP页面 由于ECOLOGY系统支持多语言，因此在JSP页面上一般不出现中 文，全部使用标签的形式来显示中文：
比如：在IE上显示“姓名”那么在JSP页面中将通过 <%=SystemEnv.getHtmlLabelName(413,user.getLanguage())%>这样的形 式来表示，其中的数字413就是表示姓名，同时可以通过“select labelname from htmllabelinfo where indexid=413 and languageid=7”来获 取到“姓名”这个显示名称，其中languageid=7表示中文显示名 称,languageid=8表示英文显示名称. delete from HtmlLabelIndex where id=81249 delete from HtmlLabelInfo where indexid=81249 INSERT INTO HtmlLabelIndex values(81249,'选择范围') INSERT INTO HtmlLabelInfo VALUES(81249,'选择范围',7) INSERT INTO HtmlLabelInfo VALUES(81249,'Range of choice',8) INSERT INTO HtmlLabelInfo VALUES(81249,'選擇範圍',9) 3、JAVA程序的基本组成 在ECOLOGY中开发JAVA程序建议继承weaver.general. BaseBean，在 BaseBean中主要封装了两个方法：写日志文件，获取配置文件中的参 数值。 public String getPropValue(String fname , String key) public void writeLog(Object obj) 三、页面权限控制的说明,怎样在页面中引用权限，怎么样新增一个权限，如 何在新开发的模块上引入权限控制 在这一部分将描述：新增的页面如何保持和ECOLOGY的风格保持一 致；新增的页面上引用ECOLOGY中的权限；新增的页面上引用新的 ECOLOGY中还没有的权限； 1、可以根据<泛微协同商务系统(Ecology)_JSP式样编写指南>保证新开 发的页面在风格上和原有系统保持一致
Workflow简明教程 注：我们将通过作一个workflow的实例来演示一个workflow工作流的建立。 首先建立new的项目 在Workflows的file选择菜单中的选择向导选项，用向导作比较直观一些！ 选择后结果如下： 上面两个栏位是新建立的项目的名称internal name 是程序需要的名字，display是显示的名字。new process是项目中工作流过程的名称，同理，interal name是程序的名字，display name 是显示的名字 需要注意的是，internal name是能用中文的，而且最好用大写，display name没有要求 其余两个选项不作要求 输入你需要的名字然后点ok，这时你的workflow整体框架就出来了，如图： 左边是导航区，右边是工作区，
跟form的风格和相似，注意右边他默认了两个图标，这是workflow流的两个端点，start和end 。其中间的流过程是设计者来完成 下面我将作一个关于审核工作票的工作流，其流程图如下
第一步：我们要建立attribute ，即你在工作流中用到的所有的属性， 这个例子中要用到4个属性，分别是send (审核人)，view（修改人），gzpbg（工作票编号），p_url（打开的url地址）建立审核人，点右键，如图： 建立新的attrib属性 填写属性内容，注意如果是人员角色的属性，则要选择相对应的角色选项，如下 填写完毕，确定，然后同理创建其余3个属性， 这个时候属性创建完毕 第二步！创建流程的节点，也就是关键点 如图，我要创建一个审批的节点（p_check），还要创建一个审批不合格，需要重新审批的节点（p_ok） 在导航区的notification上点右键，新增节点