<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>Cross-Organizational Workflows: A Classification of Design Decisions</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <string-name>Pascal van Eck</string-name>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Rieko Yamamoto</string-name>
          <email>r.yamamoto@jp.fujitsu.com</email>
          <xref ref-type="aff" rid="aff2">2</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Jaap Gordijn</string-name>
          <email>gordijn@cs.vu.nl</email>
          <xref ref-type="aff" rid="aff1">1</xref>
        </contrib>
        <contrib contrib-type="author">
          <string-name>Roel Wieringa</string-name>
          <email>roelw@cs.utwente.nl</email>
          <xref ref-type="aff" rid="aff0">0</xref>
        </contrib>
        <aff id="aff0">
          <label>0</label>
          <institution>Department of Computer Science, University of Twente P.</institution>
          <addr-line>O. Box 217, 7500 AE Enschede</addr-line>
          ,
          <country country="NL">The Netherlands</country>
        </aff>
        <aff id="aff1">
          <label>1</label>
          <institution>Department of Computer Science, Vrije Universiteit Amsterdam De Boelelaan 1081</institution>
          ,
          <addr-line>1081 HV Amsterdam</addr-line>
          ,
          <country country="NL">The Netherlands</country>
        </aff>
        <aff id="aff2">
          <label>2</label>
          <institution>Fujitsu Laboratories Ltd., IT Core Laboratory 4-1-1 Kamikodanaka, Nakahara-ku, Kawasaki Kanagawa</institution>
          ,
          <addr-line>211-8588</addr-line>
          ,
          <country country="JP">Japan</country>
        </aff>
      </contrib-group>
      <abstract>
        <p>Web service technology enables organizations to open up their business processes and engage in tightly coupled business networks to jointly offer goods and services. This paper systematically investigates all decisions that have to be made in the design of such networks and the processes carried out by its participants. Three areas of different kinds of design decisions are identified: the value modeling area, which addresses economic viability of the network, the collaboration modeling area, which addresses how business partners interact to produce the goods or services identified in the value modeling area, and the workflow modeling area, which addresses the design of internal processes needed for the interactions identified in the collaboration modeling area. We show, by reporting on a real-world case study, that there are significant differences between these areas: design decisions are unique for each area, IT support for collaboration processes is orthogonal to IT support for workflows, and the role of web choreography standards such as BPEL4WS differs for both of them.</p>
      </abstract>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>Introduction</title>
      <p>Recently, a number of standards have been proposed for machine-readable specification
of inter-organizational business processes, such as ebXML BPSS, BPEL, and WSCI.
Although these formalisms differ in many respects, they are all based on the same idea:
they provide (and only provide) XML-based syntactic constructs to specify valid
sequences of web service invocations of some business partners. Thus, these business
partners are assumed to work together by invoking each other’s web services. Although
a considerable amount of research has been published on these standards, for instance
on their fundamental and relative expressiveness, not much is currently known about
guidelines for actually designing cross-organizational processes. In this paper, we show
that three groups of design decisions must be made before the designer even considers
using web services (instead of something else, such as EDI) and before the designer
considers using one of the available web service choreography standards. Briefly, these
This work is part of the Freeband A-MUSE and FRUX projects. Freeband (http://www.
freeband.nl) is sponsored by the Dutch government under contract BSIK 03025.
decisions concern the commercial viability of the business activities being coordinated,
the coordination mechanisms required to realize these commercial activities, and the
workflows for each participating organization to support this coordination. We will
argue that a workflow internal to a business, which is private but yet must support an
interorganizational coordination process, involves decisions that are absent from purely
business-internal workflow design.
2</p>
    </sec>
    <sec id="sec-2">
      <title>Design decisions</title>
      <p>Before committing considerable resources to creating a collaboration with another
organization, an organization usually seeks answers to the following two questions: is the
collaboration economically viable (e.g., generating a positive cash flow), and is the
collaboration feasible (e.g., does it not overextend organizational and IT support change
capabilities). We approach these questions by dividing them in three areas of concern:
value modeling (economic viability), coordination modeling (feasibility/impact with
respect to relations of organizations with each other), and workflow modeling
(feasibility/impact with respect to internal structure and processes). For each area of concern,
we identify a number of design decisions, which are presented in Table 1. In the next
sections, for each area of concern we discuss these issues and illustrate them using a
case study.
3</p>
    </sec>
    <sec id="sec-3">
      <title>Value modeling</title>
      <p>
        In the value modeling area, we address the enterprises and final customers that
participate in a collaboration. As such, the value model presents who is offering what of
economic value to whom and expects what in return. The latter refers to the notion of
economic reciprocity; an important notion in commercial trade. In addition, the value
model shows whether valuable objects are offered as a bundle (potentially by different
suppliers) or not. Bundling [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ] is an important notion in business to increase total sales
and is in e-business settings of specific interest because information integration enable
multi-supplier bundles. Finally, a value model shows the assignment of value activities
(activities that yield profit) to performing actors. In the recent past, we have seen in the
context of e-business many shifts of such activities from the one enterprise to another
enterprise.
      </p>
      <p>
        The value modeling design decisions can be represented by an e3-value value model
[
        <xref ref-type="bibr" rid="ref2 ref3">2,3</xref>
        ] such as depicted in Fig. 1. For self containment of this paper, we introduce the
main e3-value constructs: e3-value represents actors (enterprises and final customers)
that exchange value objects (goods and services) with each other through value
interfaces. These interfaces consist of ports offering or requesting value objects. Final
customers have a customer need. To satisfy such a need, a series of value exchanges
need to be executed by all enterprises collaborating in satisfying that need. We
represent these exchanges by a dependency path. A dependency path consists of the need,
the interfaces exchanging objects contributing to need satisfaction, and internal actor
dependencies between interfaces. If an actor has a need, he will exchange objects of
value through one of his interfaces to satisfy the need. Additionally, exchanges via an
Value modeling – Which consumer needs do exist?
– How are these consumer needs satisfied by items of economic
value that can be produced or consumed by enterprises and
endcustomers, and are by definition of economic value?
– Who is offering/requesting value objects to/from the environment?
– What are the reciprocal value object exchanged between
      </p>
      <p>enterprise/end-customers?
– What bundles of value objects exist?
– What partnerships do exist?
Coordination modeling Coordination process design decisions:
Workflow modeling
– Which information is exchanged between business partners, and in</p>
      <p>which order?
– What are the trust relations between the actors?
– Are additional actors needed to resolve trust issues (e.g., trusted</p>
      <p>third parties?)
– Who is responsible for the coordination activities at each business</p>
      <p>partner?
IT support design decisions:
– What technology to use (e.g., HTML forms, web services)?
– Synchronous or asynchronous information exchange?
– What is the format of the message data exchanged?
Workflow design decisions are mainly concerned with issues in
operations management and organization theory. As an example, we mention
the choice between make-to-client-order (goods are manufactured after
a client order has been received) or make-to-stock (goods are
manufactured in advance, orders are satisfied from stock) processes that always
have to be made in the manufacturing industry.</p>
      <p>IT support design decisions:
– What information systems are needed?
– What functions do these information systems need to offer?
– Distribution decisions, e.g. central IT facilities or facilities per
location
interface may cause exchanges via another interface of such an actor (e.g. to buy raw
materials to produce the object requested).</p>
      <p>
        We illustrate the cross-organizational design decisions using a case study on portals
for music fans, based on earlier work of one of the authors [
        <xref ref-type="bibr" rid="ref4">4</xref>
        ]. Fans of a particular artist
are interested in information about the artists, merchandise, and song scores, but also
want to chat about their favorite artist. In sum, fans want to have a portal that organizes
information and services related to their artist. The value modeling design decisions
presented in the first row of Table 1 can be represented by an e3-value value model,
such as the one depicted in Fig. 1.
4
      </p>
    </sec>
    <sec id="sec-4">
      <title>Coordination modeling</title>
      <p>Coordination is the interaction between a number of actors, in this case business
partners, needed to produce a result. Coordination is needed because of dependence
between the activities of the actors. Therefore, the actors exchange information to keep
each other informed on the current state of affairs in their joint effort to produce a result.</p>
      <p>We distinguish between on the one hand, coordination process by which business
coordinate their behavior in a collaboration (coordination modeling area), and on the
other hand, internal processes of each of the businesses participating in a collaboration
(workflow modeling area). A coordination process consists only of interactions between
two or more parties in the collaboration. These interactions involve externally visible
behavior of each of the coordinating businesses. The set of all interactions of one
business is called its abstract business process. In general, there will be one or more internal
business processes that jointly realize the abstract process of a business. Most of these
internal processes will be confidential, as it contains confidential business rules and uses
confidential data.</p>
      <p>In the value model (Fig. 1), a dependency path shows which value exchanges are
needed to satisfy a customer need. A dependency path is not itself a coordination
process or workflow. Instead, one or more coordination processes have to be designed to
coordinate all activities and information exchange needed to realize the value exchanges
connected by a dependency path.</p>
      <p>IT support for coordination processes involves supporting information exchange
as specified in the coordination processes. This information exchange can take many
forms, e.g. web services but also conventional forms of information exchange.</p>
      <p>In our case study, information exchange between fanclub members and the
portal provider is based on HTML forms, as fanclub members (who are human end
consumers) want to interact with the portal directly. The portal software designers have
complete freedom in designing these forms. They may also choose to not design these
forms themselves, but use a ready-made shopfront offered by a third party. As this is
usually not a free service, in this case the value model (Section 3) needs to be adapted
and the business case reconsidered. Information exchange with the settlement service is
most probably governed by the standard procedures published by the settlement service
provider.</p>
      <p>While some parts of a coordination process may be supported by technology other
other than web services, for other parts web services may be chosen. For these parts,
web service choreography specification formalisms such as BPEL4WS can be used
to describe them. In fact, a coordination process such as the one depicted in Fig. 2
actually requires two BPEL4WS specifications, one for each business partner. (It is not
possible to specify sequences of both partners in one BPEL4WS specification as in
BPEL4WS, only a choreography of one partner can be described.) Thus, a complex
business collaboration as described by the dependency path that we have focused on in
our case results in a number of relatively small web service choreography specifications.
5</p>
    </sec>
    <sec id="sec-5">
      <title>Workflow modeling</title>
      <p>
        Workflows are internal business processes that realize the atomic activities that are
visible in coordination processes. For a large part, design choices in workflow design fall
outside Computer Science as they are in the domains of Organization Theory [
        <xref ref-type="bibr" rid="ref6">6</xref>
        ],
process design [
        <xref ref-type="bibr" rid="ref7">7</xref>
        ], or Operations Management [
        <xref ref-type="bibr" rid="ref8">8</xref>
        ]. Moreover, a number of design decisions
are dictated by outside stakeholders, e.g. stemming from legal requirements and
generally accepted accounting principles. In workflow modeling, there are two roles for a
web service choreography standard such as BPEL4WS. First, there is an
implementation relation between collaboration processes (which may be described in BPEL4WS)
and internal workflows. So, a BPEL4WS description can be considered a partial
specification of the workflows that have to be designed. Second, in a number of cases,
internal workflows have to be formally specified and supported by workflow management
systems. In these cases, the workflows can be described as BPEL4WS executable
processes and executed by the BPEL4WS execution engines that are currently emerging in
the market.
6
      </p>
    </sec>
    <sec id="sec-6">
      <title>Conclusion</title>
      <p>In this paper, we have systematically identified all design decisions that need to be made
when designing multi-party collaborations. This revealed a clear distinction between
value modeling, coordination modeling and modeling internal workflows. IT support
is different for each of the latter two, as is the role of web service technology and
choreography standards. For each type of modeling, we have shown examples (using
our case study) of modeling techniques such as e3-value and BPMN. These modeling
techniques are relatively lightweight while still providing enough insight to support
design decisions. Moreover, these techniques are simple enough to be understood by
non-technical stakeholders, which is important as design decisions made by software
engineers influence design decisions that have to be made by business stakeholders and
the other way around.</p>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          1.
          <string-name>
            <surname>Choi</surname>
            ,
            <given-names>S.Y.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Stahl</surname>
            ,
            <given-names>D.O.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Whinston</surname>
            ,
            <given-names>A.B.</given-names>
          </string-name>
          :
          <article-title>The Economics of Doing Business in the Electronic Marketplace</article-title>
          .
          <source>MACMillan Technical Publishing</source>
          , Indianapolis, IN (
          <year>1997</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          2.
          <string-name>
            <surname>Gordijn</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Akkermans</surname>
          </string-name>
          , J.:
          <article-title>Value-based requirements engineering: Exploring innovative ecommerce ideas</article-title>
          .
          <source>Requirements Engineering Journal</source>
          <volume>8</volume>
          (
          <year>2003</year>
          )
          <fpage>114</fpage>
          -
          <lpage>134</lpage>
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          3.
          <string-name>
            <surname>Gordijn</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Akkermans</surname>
            ,
            <given-names>J.M.</given-names>
          </string-name>
          :
          <article-title>Designing and evaluating e-Business models</article-title>
          .
          <source>IEEE Intelligent Systems - Intelligent e-Business</source>
          <volume>16</volume>
          (
          <year>2001</year>
          )
          <fpage>11</fpage>
          -
          <lpage>17</lpage>
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          4.
          <string-name>
            <surname>Yamamoto</surname>
            ,
            <given-names>R.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Ohashi</surname>
            ,
            <given-names>K.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Yamamoto</surname>
            ,
            <given-names>K.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Inomata</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Matsuda</surname>
            ,
            <given-names>T.</given-names>
          </string-name>
          :
          <article-title>Lessons learned from requirements analysis for implementing integrated services</article-title>
          .
          <source>In: Proceedings of the International Workshop on Service-Oriented Requirements Engineering - SORE</source>
          <year>2004</year>
          .
          <article-title>(</article-title>
          <year>2004</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref5">
        <mixed-citation>
          5.
          <string-name>
            <surname>White</surname>
            ,
            <given-names>S.</given-names>
          </string-name>
          :
          <article-title>Business process modeling notation (BPMN) (</article-title>
          <year>2004</year>
          ) http://www.bpmn.org/ Documents/BPMN%20V1-
          <fpage>0</fpage>
          %20May%
          <fpage>203</fpage>
          %
          <fpage>202004</fpage>
          .pdf, visited
          <volume>20041123</volume>
          .
        </mixed-citation>
      </ref>
      <ref id="ref6">
        <mixed-citation>
          6.
          <string-name>
            <surname>Daft</surname>
            ,
            <given-names>R.L.</given-names>
          </string-name>
          :
          <article-title>Organization Theory and Design. Sixth edn</article-title>
          .
          <source>Thomson Publishing</source>
          (
          <year>1998</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref7">
        <mixed-citation>
          7.
          <string-name>
            <surname>Ould</surname>
            ,
            <given-names>M.A.</given-names>
          </string-name>
          :
          <article-title>Business Processes-Modelling and Analysis for Re-engineering and Improvement</article-title>
          . Wiley (
          <year>1995</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref8">
        <mixed-citation>
          8.
          <string-name>
            <surname>Slack</surname>
            ,
            <given-names>N.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Chambers</surname>
            ,
            <given-names>S.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Harland</surname>
            ,
            <given-names>C.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Harrison</surname>
            ,
            <given-names>A.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Johnston</surname>
          </string-name>
          , R.: Operations Management. Second edn. Pitman Publishing (
          <year>1998</year>
          ) ISBN:
          <fpage>0</fpage>
          -
          <lpage>273</lpage>
          -62688-4.
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>