PHILharmonicFlows: Research and Design Methodology [original]

PHILharmonicFlows:

Research and Design Methodology

Vera K¨unzle and Manfred Reichert

Institute of Databases and Information Systems, Ulm University, Germany

vera.kuenzle,manfred.reichert@uni-ulm.de

Abstract. In comprehensive case studies we found out that many lim-

itations of existing Process Management Systems (PrMS) can be traced

back to the unsatisfactory integration of processes and data. In the PHIL-

harmonicFlows1project, we aim at a deep and extensive understanding

of the inherent relationships between processes and data, and thus want

to overcome some of the fundamental limitations known from activity-

centered PrMS. Overall, we target at a comprehensive framework pro-

viding integrated access to processes, data, and functions to its users.

1 Introduction

Enterprises spend a lot of time and money for introducing Information Technol-

ogy (IT) to improve their effectiveness and efficiency [1]. Effectiveness addresses

the distance between business goals and the capabilities offered by IT in order

to achieve these goals; i.e., the alignment of IT with different business perspec-

tives [2]. These perspectives, in turn, typically comprise business data, business

functions, and business processes.

Business data is typically represented by a number of business objects. These

cover domain-specific business entities like orders, customers, or products. Each

business object, in turn, is represented by a set of attributes. As example consider

the delivery date of an order or the name of a customer. In addition, business

objects are related to each other; e.g., an order may comprise several products.

Business functions, in turn, constitute a wide range of (various) activities. Typ-

ically, many of them are used to create and delete business objects or to change

their properties. As example consider user forms enabling humans to fill in de-

sired property values. Finally, business processes comprise a number of related

business functions to be executed in order to achieve a certain business goal.

To reach competitive advantages, it is further important to introduce new prod-

ucts and services as quickly as possible at the market. In addition, for dealing

with increasing competitive pressure and market dynamics, it should be possi-

ble to continuously adapt IT systems in a quick and effective way. Hence, rapid

development and improved maintenance are important success factors. In this

context, process management systems (PrMS) offer promising perspectives in

respect to comprehensive lifecycle support of business processes. In these PrMS,

1Process, Humans and Information Linkage for harmonic Business Flows

2 Vera K¨unzle and Manfred Reichert

business processes are modeled in terms of activities required for achieving a

particular business goal as well as their control flow defining the order and the

constraints for executing these activities [3]. Each activity is then linked with a

specific business function of an application service (e.g., business application).

In addition, most PrMS handle atomic data elements which are connected with

one or more activities reading or writing it. These data elements are also used

for process control (i.e., for evaluating routing conditions).

The remainder of this paper is structured as follows. We first motivate the prob-

lem addressed in our research in Section 2. In Section 3 a discussion on how

process support looks like in today’s companies follows. Section 4 then intro-

duces fundamental research questions and expected solutions, while Section 5

investigates related work along a well defined evaluation schema. The relevance

of our research is discussed in Section 6. Section 7 describes the research method-

ology we apply. Finally, Section 8 sketches the solution approach we target at

and Section 9 closes with a summary and outlook.

2 Problem Statement

Despite the widespread adoption of existing PrMS, there exist numerous pro-

cesses not adequately supported by these PrMS. In particular, traditional PrMS

have been primarily designed for supporting highly structured, repetitive busi-

ness processes [4]. For various other processes, in turn, there is a contradiction

between the way these processes can be defined and the preferred work prac-

tice [5, 6, 7]. These processes are often characterized as ”information-centric”

[8] or ”knowledge-intensive” [9]. Further, they are rather ”unstructured” [6] and

cannot be ”straight-jacked into activities” [9].

Moreover, existing PrMS focus on the business process perspective; i.e., activities

and their control flow. Business functions (which are linked with activities) and

business data, in turn, are usually out of the control of existing PrMS. For this

reason, generic support is only provided for process enactment in existing PrMS.

For realizing business functions, specific programming is required. Typically, this

consumes more time and efforts as the modeling of the corresponding processes.

Moreover, in existing PrMS business functions are treated as ”black-boxes”.

This means, what is done during activity execution is out of the control of

existing PrMS; i.e., business data is managed by the invoked business functions

themselves. Obviously, this missing link between business data and business

process prohibits integrated access to them. Consequently, users cannot access

and manage data at any point in time during process execution (assuming proper

authorization).

3 Background

Many processes not adequately supported by existing PrMS are more or less

hard-coded within specific business applications. Consequently, long develop-

PHILharmonicFlows: Research and Design Methodology 3

ment cycles arise and even simple process changes require costly code adaptions

and high efforts for testing. However, the more specific a business application is,

the better it meets the requirements of the respective application domain; i.e.,

business-IT-alignment is improved. In particular, tailor-made business applica-

tions requiring a customer-specific programming of all functionalitites typically

fit business needs best.

Regarding the support of business processes, companies typically face Mortons

fork: Either they can use PrMS to enable rapid development as well as improved

maintenance and therefore achieve a high degree of efficiency, or they directly

apply tailor-made applications for providing the required effectiveness; i.e., ad-

equate business-IT-alignment (cf. Fig. 1).

Fig. 1. Dilemma between efficiency and effectiveness

Many companies have achieved considerable benefits from their investments in

domain-specific business applications (i.e., ERP packages). Regarding the latter,

some standard functionality is already pre-implemented. These applications can

be customized for the specific needs of an enterprise. Usually, customizability is

realized through configuration support (e.g., based on some settings one can con-

figure a particular process variant). Thus, domain-specific business applications

constitute a trade-off between effectiveness and efficiency. However, since con-

figurability depends on the range of preconfigured alternatives, domain-specific

enterprise applications are huge and complex [10]. An additional problem emerg-

ing in this context are the lack of transparency of the configurable processes.

4 Research Questions

Starting with the basic observation that there are business processes not ade-

quately supported by existing PrMS, we define the following research questions:

Research Question 1: What are the common properties of business processes

currently not adequately supported by PrMS?

Expected Solution: A collection of characteristic properties relating to different

4 Vera K¨unzle and Manfred Reichert

business perspectives; i.e., the inter-relationships between business processes,

business data, business functions, and users.

Research Question 2: Which requirements must be fulfilled by a PrMS to

adequately capture these properties?

Expected Solution: A set of elicited requirements for PrMS enabling the support

of the identified properties.

Research Question 3: How to support the requirements elicited in an inte-

grated process support framework?

Expected Solution: Concepts, methods and tools for realizing a PrMS enabling

process support in tight integration with data to overcome the aforementioned

limitations.

5 Related Work

Generally, we believe that the identified limitations of existing PrMS can be

traced back to the unsatisfactory integration of processes and data. In particular,

many processes necessitate object-awareness; i.e., they focus on the processing

of business data represented by business objects. The latter comprise a set of

object attributes and are related to each other. To understand the inherent re-

lationships between process and data, we investigate on processes currently not

adequately supported. This includes a systematic analysis of their properties. In

summary, our process analysis has revealed the following major characteristics

of object-aware processes:

1. Object behavior: The behavior of the involved business objects must be taken

into account during process execution.

2. Object interactions: Interactions between business objects must be ade-

quately considered; i.e., the behavior of individual objects must be coor-

dinated with the one of related business objects.

3. Data-driven execution: Since the progress of a process mainly depends on

available business objects and on their attribute values, process execution

has to be accomplished in a data-driven manner.

4. Integrated access: Authorized users must be able to access and manage

process-related objects at any point in time (assuming proper authoriza-

tion).

5. Flexible activity execution: Activities must be executable at different levels

of granularity. While one user may work on a particular object instance,

another one may process a number of related object instances in one go.

Though there exist several approaches targeting at a tighter integration of busi-

ness processes and business data [9, 11, 12, 13, 14, 7, 5], as illustrated in Fig. 2,

none of them supports all identified properties in an integrated and comprehen-

sive way. In addition, some approaches only deal with the modeling of processes,

PHILharmonicFlows: Research and Design Methodology 5

Fig. 2. Characteristics supported by related work

but exclude process execution; e.g., they do not provide a well-defined opera-

tional semantics for the automatic enactment of the defined processes. Conse-

quently, existing approaches provide generic support for only few of the identified

characteristics.

6 Relevance

Altogether we believe that a tighter integration of the different business per-

spectives (cf. Fig. 3) will provide an important contribution to overcome some of

the fundamental limitations known from contemporary PrMS. For this purpose,

we aim at a deep and extensive understanding of the inherent relationships that

exist between processes, data, functions, and users in order to enable generic

enactment for object-aware process management. Their support could impact

the realization of more flexible process management technology in which daily

work can be done in a more natural way.

Moreover, we assume that a tighter integration between process and data not

only fosters effectivity, but also efficiency. This means, not only generic pro-

cess support but also generic business functions become possible. In addition to

process-oriented views (i.e., work-lists) we aim at the automatic generation of

form-based activities and data-oriented views (e.g., overview table) at run-time.

This way, integrated access to business process, business functions, and business

data shall be provided to users.

6 Vera K¨unzle and Manfred Reichert

Fig. 3. Business perspectives in object-aware PrMS

7 Research Design and Methods

Regarding research in the field of information technology (IT), there are two

kinds of sciences: design science and natural science [15, 1]. Natural science re-

search is a knowledge-producing activity comprising the two steps discovery and

justification [15]. Design science, in turn, is a knowledge-using activity [15]. It

aims at developing IT systems. Here, building and evaluation as the two major

activities [1].

Generally, doing research means applying natural science. Regarding IT, how-

ever, design research is considered as being more successful and important. Nev-

ertheless, technology and behavior cannot be separated from each other [1]. Thus,

in accordance with [15, 1], it is an opportunity for IT research to make significant

contributions by engaging in both. As illustrated in Fig. 4, IT research calls for

synergistic efforts between natural and design science research [1].

Our main research activities are as follows (cf. Fig. 5): We start with natural

research to identify the characteristic properties of object-aware processes (cf.

Research Question 1). To deal with Research Question 2, we evaluate existing ap-

proaches (using already available and applicable knowledge) to elicit the require-

ments for a PrMS supporting the identified properties. Finally, we address Re-

search Question 3 and develop a comprehensive framework and proof-of-concept

prototype for object-aware process management based on design research.

7.1 Doing Natural Research: Property Investigation and Justification

To discover the properties of those business processes not adequately supported

by current PrMS, we perform a detailed property investigation by analyzing

these business processes. We then justify our findings with an extensive litera-

ture study.

PHILharmonicFlows: Research and Design Methodology 7

Fig. 4. IT Research [1]

Fig. 5. Research Methodology

8 Vera K¨unzle and Manfred Reichert

Process analysis Data Source: Due to the limitations of contemporary PrMS

there exist numerous business applications (e.g., ERP or CRM systems) which

are process-aware, but do not rely on PrMS. Instead they contain hard-coded

process logic; i.e., process logic interwoven with application code. To ensure that

the processes we analyze are not ”self-made” examples, but constitute real-world

processes of high practical relevance, we select processes as implemented in exist-

ing business applications. Amongst others, we analyze the processes implemented

in the human resource management system Persis and the conference reviewing

system Easychair [16, 17]. In particular, our evaluation is not restricted to the

inspection of user interfaces solely. In addition, we rely on extensive practical

experiences gathered during the development of contemporary business applica-

tions; i.e., we have deep insights into their application code and process logic.

Finally, we underpin our results by interviewing system users as well as business

consultants being familiar with the respective business applications.

Selection Criteria: We evaluate the processes (and additional features) based on

the main business perspectives. These comprise processes, data, functions, and

users. In particular, we focus on their interdependencies.

Literature study Ensuring importance: We complement our process analyses by

an extensive literature study. This way we want show that other researchers

consider some of the properties we identified as being relevant as well.

Ensuring completeness: To not exclude important properties already identified

by other researchers, we compare our analysis results with existing literature.

However, to set a focus we exclude properties in respect to process change and

process evolution. Instead, our focus is on process modeling, execution and mon-

itoring.

Ensuring generalisation: Interestingly, some authors refer to similar application

examples as we do, while addressing different properties. Based on these insights

we contrast the different application examples with the total set of identified

properties. This way, we are able to demonstrate two things: first, the properties

are related to each other. Second, broad support for them is required by a variety

of processes from different application domains.

7.2 Using applicable knowledge: Requirements Engineering

Concerning Research Question 2, we first discuss to what degree existing PrMS

cover the identified properties. More precisely, we evaluate which properties can-

not be directly supported when applying traditional imperative and declarative

process support paradigms [18]. Based on this evaluation we then elicit the basic

requirements as inquired by Research Question 2.

PHILharmonicFlows: Research and Design Methodology 9

7.3 Doing Design Research: Framework Design and Proof-of-Concept

Hevner et al [1] consider solution design as search process being inherently it-

erative. This has been confirmed by other authors [19, 20]. Simon [19], in turn,

describes the nature of the design process as a ”Generate/Test Cycle”. The spi-

ral model [20], in turn, defines an approach in which one and the same step is

repeated several times, each time improving the results of the previous outcome.

For this purpose, we perform iterative walkthroughs. In particular, we revise

our solution and improve it step by step. This leads to different development

versions. Additionally, we investigate in user interface design [21, 22]. This way,

shortcomings concerning the usability of the framework design are identified at

early project stages and can be considered in subsequent iterative revisions.

To evaluate our framework we develop a proof-of-concept prototype for the mod-

eling as well as the run-time environment. In addition, we apply the prototype to

real-world cases. In particular, we use scenarios from the medical domain, order

processing, and house building which are different from the ones we consider

in the context of our process analyses. Finally, we elaborate the benefits of our

approach when applying it to these processes as well as lessons learned.

8 Contribution

In the course of our PHILharmonicFlows project we are developing a compre-

hensive framework supporting the modeling,execution and monitoring of object-

aware processes. Opposed to traditional process support paradigms, we provide

a uniform methodology for modeling processes on well-defined levels of gran-

ularity. The proper execution as well as termination of processes at run-time

is further ensured by a set of correctness rules. In addition, a well-defined op-

erational semantics not only enables generic support for process execution, but

also generic realization of business functions. In particular, PHILharmonicFlows

enables integrated access to business processes, business functions, and business

data. For this purpose, we automatically create end-user components like work-

lists, form-based activities, and overview tables during run-time based on the

corresponding models.

One fundamental pillar of our framework enables the processing of individual

business objects to enforce object behavior. Like existing work considering object

behavior during process execution [12, 23, 11, 24, 8, 13, 25, 14], our approach

applies the well established concept of modeling object behavior in terms of

states and state transitions. Opposed to existing approaches, however, PHIL-

harmonicFlows enables a mapping between attribute values and objects states

and therefore ensures compliance between them [26]. Moreover, this mapping

allows us to combine object behavior with data-driven execution.

Another fundamental pillar of our framework enables the modeling and execu-

tion of processes involving multiple business objects (i.e., objects of the same and

of different type). Thereby, we consider their individual behavior as well as their

inter-relationships. More precisely, we enable the coordination of object-specific

10 Vera K¨unzle and Manfred Reichert

processes that represent object behavior taking the relationships between the in-

volved business objects into account. Regarding the latter, we obtain a complex

process structure comprising inter-related, object-specific processes. Opposed to

existing work [12, 7] we hide this complex structure form modelers as well as

from end users. For this purpose, our approach provides sophisticated concepts

for defining aggregations abstracting from individual business objects. This way,

different cardinality constraints can be taken into account. In addition, coordi-

nation is not only possible along direct object relations (e.g., a review directly

refers to a paper). In fact, in PHILharmonicFlows the processing of business ob-

jects can be coordinated based on complex inter-relationships taking transitive

as well as transverse relationships between business objects into account.

The presented execution paradigm applies data-driven process execution in a

comprehensive way; i.e., in respect to object behavior and object interactions.

Since in some situations explicit user decisions and commitments are required,

PHILharmonicFlows combines its data-driven execution paradigm with activity-

oriented aspects. Finally, such tight integration of process and data necessitates

advanced concepts for user integration; i.e., process authorization must be com-

pliant with data authorization and vice versa. For this purpose, we introduce an

advanced concept for actor assignment and authorization.

9 Summary and Outlook

Our basic mission in the PHILharmonicFlows project is to develop generic con-

cepts, methods and tools for realizing object- and process-aware information sys-

tems. In particular, we aim at the flexible integration of business data, business

processes, and business functions to overcome limitations known from purely

activity-centered PrMS. In this paper, we report on our overall vision and on

the research methodology we apply.

We have already conducted extended process analysis in the areas of human

resource management and paper reviewing. In [16, 17] we reported on the basic

challenges for integrating processes, data, functions and users, and we described

the properties of object-aware process management in detail [27]. Based on a

detailed comparison of traditional process support paradigms [18] we elicit the

major requirements for object-aware process support [18, 27]. Currently, we are

developing a comprehensive framework for object-aware process management [27,

26] as well as a proof-of-concept prototype. In [27] we give an overview about

our framework, whereas [26] reports on our modeling paradigm for integrating

processes and data at the micro level (i.e., for realizing object behavior). In

future work we elaborate more detailed issues in the context of our framework

and apply it to other real world processes for evaluation purpose (i.e., health

care, order processing, house building).

PHILharmonicFlows: Research and Design Methodology 11

References

1. Hevner, A.R., March, S.T., Park, J., Ram, S.: Design Science in Information

Systems Research. MIS Quarterly 28(1) (2004) 75–105

2. Mutschler, B., Reichert, M., Bumiller, J.: Unleashing the Effectiveness of Process-

Oriented Information Systems: Problem Analysis, Critical Success Factors, and

Implications. IEEE Transactions on Systems, Man, and Cybernetics 38(3) (2008)

280–291

3. van der Aalst, W.M.P., van Hee, K.: Workflow-Management - Models, Methods

and Systems. MIT Press, Cambridge, MA, USA (2004)

4. Leymann, F., Roller, D.: Production Workflow. Prentice-Hall (2000)

5. Sadiq, S., Orlowska, M., Sadiq, W., Schulz, K.: When workflows will not deliver:

The case of contradicting work practice. In: Proc. BIS’05. (2005)

6. Silver, B.: Case Management: Addressing unique BPM Requirements. BPMS

Watch (2009) 1–12

7. van der Aalst, W.M.P., Barthelmess, P., Ellis, C.A., Wainer, J.: Workflow Modeling

using Proclets. In: Proc. CoopIS’00. (2000) 198–209

8. Redding, G., Dumas, M., ter Hofstede, A.H.M., Iordachescu, A.: Transforming

Object-oriented Models to Process-oriented Models. In: Proc. BPM’07 Workshops.

LNCS 4928 (2007) 132–143

9. van der Aalst, W.M.P., Weske, M., Gr¨unbauer, D.: Case Handling: A new Paradigm

for Business Process Support. DKE 53(2) (2005) 129–162

10. Klaus, H., Rosemann, M., Gable, G.G.: What is ERP? Information Systems

Frontiers 2(2) (2000) 141–162

11. Bhattacharya, K., Hull, R., Su, J. In: A Data-Centric Design Methodology for

Business Processes. IGI Global (2009) 503–531

12. M¨uller, D., Reichert, M., Herbst, J.: Data-Driven Modeling and Coordination of

Large Process Structures. In: Proc. CoopIS’07. LNCS 4803 (2007) 131–149

13. Redding, G.M., Dumas, M., ter Hofstede, A.H.M., Iordachescu, A.: A flexible,

object-centric approach for business process modelling. SOCA (2009) 1–11

14. Vanderfeesten, I., Reijers, H.A., van der Aalst, W.M.P.: Product-Based Workflow

Support: Dynamic Workflow Execution. In: Proc. CAiSE’08. LNCS 5074 (2008)

571–574

15. March, S.T., Smith, G.: Design and natural science research on information tech-

nology. Decision Support Systems 15 (1995) 251–266

16. K¨unzle, V., Reichert, M.: Towards Object-aware Process Management Systems:

Issues, Challenges, Benefits. In: Proc. BPMDS’09. (2009) 197–210

17. K¨unzle, V., Reichert, M.: Integrating Users in Object-aware Process Management

Systems: Issues and Challenges. In: Proc. BPD’09. (2009) 29–41

18. K¨unzle, V., Weber, B., Reichert, M.: Object-aware Business Processes: Fundamen-

tal Requirements and their Support in Existing Approaches. International Journal

of Information System Modeling and Design 2(2) (2010) 19–46

19. Simon, H.A.: The Sciences of the Artificial (3rd ed.). MIT Press Cambridge USA

(1996)

20. Boehm, B.: A spiral model of software development and enhancement. SIGSOFT

Softw. Eng. Notes 11(4) (1986) 14–24

21. Scheb, C.: Entwicklung eines Usability-Konzepts f¨ur die Laufzeitumgebung eines

datenorientierten Prozess-Management-Systems. Diploma thesis, Ulm University

(2010)

12 Vera K¨unzle and Manfred Reichert

22. Wagner, N.: Entwicklung eines Usability-Konzepts f¨ur die Modellierungsumge-

bung eines datenorientierten Prozess-Management-Systems. Diploma thesis, Ulm

University (2010)

23. Gerede, C.E., Su, J.: Specification and Verification of Artifact Behaviors in Busi-

ness Process Models. In: Proc. ICSOC’07. (2007) 181–192

24. K¨uster, J., Ryndina, K., Gall, H.: Generation of Business Process Models for

Object Life Cycle Compliance. In: Proc. BPM’07. LNCS 4714 (2007) 165–181

25. Reijers, H.A., Liman, S., van der Aalst, W.M.P.: Product-Based Workflow Design.

Management Information Systems 20(1) (2003) 229–262

26. K¨unzle, V., Reichert, M.: A Modeling Paradigm for Integrating Processes and

Data at the Micro Level. In: Proc. BPMDS’11, Springer (2011)

27. K¨unzle, V., Reichert, M.: PHILharmonicFlows: Towards a Framework for Object-

aware Process Management. Journal of Software Maintenance and Evolution: Re-

search and Practice 23(4) (June 2011) 205–244