Towards Self-Configuration and Management of e-Service Provisioning in Dynamic Value Constellations [original]

Towards Self-Configuration and Management of e-Service

Provisioning in Dynamic Value Constellations

Jaap Gordijn

Department of Computer

Science

Free University Amsterdam

The Netherlands

gordijn@cs.vu.nl

Hans Weigand

Department of Information

Management

Tilburg University

The Netherlands

[email protected]

Manfred Reichert,

Roel Wieringa

Department of Electrical

Engineering, Mathematics and

Computer Science

University of Twente

The Netherlands

{m.u.reichert,roelw}@cs.utwente.nl

ABSTRACT

Networked value constellations are collections of enterprises

that jointly satify complex consumer needs. Increasingly,

such needs are satisfied by e-services, i.e. commercial ser-

vices that can be ordered and provisioned via the Internet.

Current research in dynamic web-service composition has

yielded run-time platforms to dynamically compose com-

plex web services, but there is still a considerable gap be-

tween web services and commercial e-services. To compose

e-services, an estimation of commercial profitability must be

made, which is absent from web service composition. In this

paper, we extend our earlier approach to e-service compo-

sition with a dynamic part, that ensures that a commercial

e-service can be dynamically composed from other commer-

cial e-services, and can be mapped on a web service compo-

sition process.composition of lower-level web services. We

propose a skeleton-oriented approach, that first composes

a network of enterprises, jointly satisfying need, based on

commercial considerations. Second, given a set of such can-

didate value constellations, the business processes providing

the services can be dynamically configured. We illustrate

this skeleton-driven composition of networked value constel-

lations by using a case study of clearing and repartitioning

of Intellectual Property Rights (IPR).

General Terms

Value web, value constellation, coordination process, e-service,

web service, dynamic composition

Keywords

Business networking, inter-organizational collaboration, con-

ceptual modeling

1. INTRODUCTION

Permission to make digital or hard copies of all or part of this work for

personal or classroom use is granted without fee provided that copies are

not made or distributed for profit or commercial advantage and that copies

bear this notice and the full citation on the first page. To copy otherwise, to

republish, to post on servers or to redistribute to lists, requires prior specific

permission and/or a fee.

SAC’08 March 16-20, 2008, Fortaleza, Cear´

a, Brazil

Enterprises increasingly participate in networked value

constellations [15]. By doing so, these enterprises can jointly

satisfy a more complex consumer need than they could ever

do on their own. Well known examples include Amazon.com,

Cisco and Dell; these enterprises have strong partner rela-

tionships with other enterprises to provide their products.

Many of these constellations actually provide services: ac-

tivities and deeds of a mostly intangible nature [9]. A key

problem in Service Sciences [14] is how to actually compose

and manage such service-provisioning value constellations.

In this paper, we concentrate on the following three main

issues.

Firstly, we focus on e-services specifically. These are com-

mercial services that can be ordered and provisioned online.

As an aside, we note that this is different from web services,

which are usually web API’s that can be accessed over a

network and are sometimes defined as software systems that

support interoperaibility over a network, usually using XML

technology. Web services are an implementation technique

for e-services.

Secondly, we specifically deal with dynamic value constel-

lations. In practice, enterprises enter and leave the con-

stellation over time, services evolve, and parties may play

different roles in different periods. Service provisioning con-

stellations therefore have to deal in an adequate way with

these dynamics, e.g. by reconfiguring themselves on-the-fly.

Thirdly, we consider self-configuring networked value con-

stellations. In classical work on coordination [10], a distinc-

tion is made between hierarchical and multi-lateral orga-

nized constellations. In a hierarchical organization, top-level

enterprises determine the selection of enterprises forming a

constellation (e.g. Cisco, Dell). In contrast, in multi-lateral

configurations, enterprises have no control over each other,

and the configuration emerges in a self-organizing way (e.g.

Skype). Both configurations can be empirically observed in

practice, and occur also in combination. We want to focus

on multi-lateral configurations. This means that our constel-

lations must be self-configuring without central coordinator.

The contribution of this paper is a skeleton-driven ap-

proach to the dynamic (re)composition of networked value

constellations for the provisioning e-services. This is in-

spired by the skeleton planning approach used by Ome-

layenko [11] for dynamic composition of semantic web ser-

vices. Our approach differs from Omelayenko’s because we

apply it to value-based e-service composition. To do so,

we propose two different, but related, skeletons, that each

can be instantiated, and together represent a generated net-

worked value constellation for service provisioning. These

skeletons are: (1) the business value skeleton (section 4),

(2) the business process skeleton (section 5).

2. CASE STUDY: IPR CLEARANCE AND

REPARTITIONING

We explore our skeleton-approach by using a real case

study on Intellectual Property Rights (IPR). In The Nether-

lands, we have a long-standing relationship with IPR soci-

eties, who are operating in a large value web consisting of

other societies, IPR owners, and IPR users.

IPR consists of many elements, and in this paper we focus

on one right only, namely the right to make content public,

henceforth called publication right. Commercial entities (ra-

dio & television stations, bars, discotheques, etc.) have to

pay to IPR owners (artists, producers, text & song writers)

a fee for using intellectual property (e.g. a music track).

IPR societies act as a “man in the middle”; they collect

this money from the IPR users (called right clearance), and

they pay money to the IPR owners (called repartitioning).

As there exist many IPR owners and -users, and there is

one IPR society per type of [IPR owner,right] combination,

multiple societies are involved in clearing and repartition-

ing the money for one specific track used. For instance, in

The Netherlands, to clear and repartition neighboring rights,

there is the SENA for artists and producers, and BUMA for

song- and text writers. Even more societies will be involved

in case of music tracks with IPR owners of different nation-

alities, since each country has its own IPR societies.

The clearance and repartitioning activities are commer-

cial e-services provided by IPR societies, because they are

done automatically. Payment for the provisioning of these

e-services is also automated. Moreover, the constellation be-

comes increasingly dynamic. Currently, IPR handling is in

many countries a monopolistic activity as IPR societies are

appointed and controlled by a local government, but due to

EU-deregulation, private entities will also be allowed to de-

velop IPR activities. IPR users and owners may then select

their preferred IPR-society for clearance and repartitioning.

This will result in a situation that for each music track, a

different constellation has to be created, since each differ-

ent IPR owner of the track may select its own IPR society,

as well as the IPR user may do. Due to the large number

of music tracks played daily by different rights users, this

should be automatically done on-the-fly. Moreover, IPR

owners and users may change their preference over time.

This results in dynamic constellations, in terms of societies

and right-stakeholders leaving and entering a constellation.

In theory, building a IPR clearance and repartitioning net-

work can be done by a centralized authority. However, such

a solution is not acceptable to the existing IPR societies, as

a centralized authority is considered as too powerful. Conse-

quently, a self-organizing & managing solution is considered

as more promising by the industry. Here, a decentralized

mechanism forms and manages the IPR clearing and repar-

titioning constellation, thereby spreading power sufficiently.

3. SKELETONS AND INSTANTIATIONS

We assume that there is a given set of IPR users, soci-

eties, and owners. We call this set a value web. We do

allow value webs to grow and shrink, but we do not con-

sider here the mechanism by which this happens. We also

assume that the members of a value web have compatible

run-time platforms that can dynamically search, compose,

and execute web services. Several run-time platforms for

dynamic service composition are now available, that allow

dynamic discovery and matching of requested and offered

web services, including negotiation of service level agree-

ments (SLAs), monitoring of quality-of-service and dynamic

reconfiguration in case of SLA violation [1, 2, 5, 6]. Differ-

ent actors in the value web may have the same platform, or

they may have different platforms that satisfy the relevant

standards.

A value web exists to satisfy certain consumer needs, and

we consider consumers of this need to be part of the value

web too. Whenever a consumer need occurs, a particular

subset of the value web will provide services to each other,

and by doing so create a complex service that satisfies the

need of the consumer. We call a subset of the value web that

satisfies a particular consumer need a value constellation.

The purpose of a value constellation is to satisfy a particular

consumer need occurring at a particule time. Value constel-

lations are formed dynamically, so that different needs of the

same consumer may lead to different value constellations,

the same need occurring in different consumers may lead to

different value constellations, and even different occurrences

of the same need in the same consumer may lead to different

constellations at different times. For example, IPR societies

have a need to audit IPR users, but they may use differ-

ent auditors for this at different times. The determination

which value constellation must be formed for a given need

must be done on-the-fly in an automated manner. Our goal

is to develop techniques to do this in a value-driven way,

using currently available for dynamic service composition

allow dynamic binding of services, service level negotiation,

and quality of service (QoS) monitoring.

Our solution consists of the following elements.

•Value perspective: In our approach each value constel-

lation is described by a value model, that shows which

actors are participating in this constellation, and who

is providing services to whom and against which re-

ciprocal service. We use the e3value notation [7] to

describe value models. Furthermore, for each kind of

consumer need we will define a value model skeleton,

or value skeleton for short, that will be instantiated to

a value model for each consumer need occurrence. We

will extend the e3value notation to be able to represent

these value skeletons.

•Process perspective: For each value skeleton we define

acoordination process skeleton, or coordination skele-

ton for short, that represents the same kinds of actors

as the value configuration skeleton, and tells us which

kinds of activities need to be performed by these kinds

of actors, and how these must be composed to deliver

a e-service to the consumer. A process skeleton can

be instantiated to a process model that describes for

a particular consumer need what coordination process

must be performed by which individual actors and how

this delivers the required e-service to a particular con-

sumer.

Instantiation of value and process skeletons is context

sensitive. In particular, the consumer need occurrence

e3value

legend

Value

Transfer

Value

port

Value

interface

Consumer

need

Connect.

element

Actor

Boundary

element

AND

element

Activity

Market

segment

Value

object

[...]

Explosion

element

e3value skeleton

[Right]

[MONEY]

[Right]

[MONEY]

[Audit]

Figure 1: A skeleton e3value model for the IPR case.

vervang “right” door “IPR”

should provide sufficient information to allow us to in-

stantiate the value model and process model skeletons,

respectively.

In the next two subsections we elaborate these two perspec-

tives.

4. BUSINESS VALUE PERSPECTIVE

4.1 Value Skeleton

The value skeleton shows, given a consumer need, how

that need is satisfied by providing valuable e-services by mul-

tiple enterprises. The focus is on what is exchanged in terms

of economic value, and not yet how this is accomplished (this

is considered by the business process perspective, see Sec. 5).

To specify a business value skeleton, we use the e3value

technique [7]. Elements in the template that need to be

instantiated are indicated by angled brackets ‘<’ and ‘>’.

This is an extension to the e3value notation.

For the case at hand, we have the following skeleton ac-

tors. First, the <IPR user>is a financially independent

actor, making money by playing a music tracks. Examples

include a radio station (usually earning money by advertise-

ments), a bar (music creates a certain atmosphere, thereby

attracting visitors), and a restaurant. An <IPR user>has a

need to <use a track>. Which track specifically is required,

is part of the instantiation process for the skeleton, and not

of the skeleton itself. To satisfy this need, the <IPR user>

usually has to obtain a set of rights (to use a track), and

in return pay for this. In Fig. 4.1, we have used an e3value

explosion element (annotated with ‘#1’) to denote that one

need for a track results in a set of (different) rights. So, a

series of rights can be transferred between the <IPR user>

and the <IPR society>(s).

[Right]

[MONEY]

Figure 2: An instantiated e3value model for the IPR

case.

Second, an <IPR society>is an independent organiza-

tion, who for a specific IPR owner clears a specific right. Be-

cause an IPR society clears only a subset of all existing rights

for a subset of all existing right owners, an instantiation of

this template typically requires more than one society. We

will demonstrate this in Sec. 4.2 further. Furthermore, for a

particular track, an <IPR society>usually clears rights for

more than one IPR owner. This is for example the case if an

orchestra played a track and some musicians in the orchestra

have the same nationality. Consequently, another explosion

element (annotated ‘#2’) has been added to model that an

issued usage right for a track to an <IPR user>requires

that multiple rights (from multiple <IPR owners)>) need

to be obtained.

Third, an <IPR owner>is an independent entity having

one or more rights on the used track. Examples include

artists, text and song writers, and record companies.

The value skeleton also shows an <auditor>that satisfies

a need of an <IPR society>. This is a different need and

it will be instantiated at different times, depending on the

goals of the IPR society, say once a month or once a year.

But in order to satisfy this need, an <IPR user>has to keep

logs of the music tracks played. These logs must be reported

to the <IPR society>, so that the clearing and repartition-

ing process can be carried out. As these logs are the source

for track clearing and therefore result in financial transac-

tions, it is attractive for the <IPR user>to commit a fraud

(e.g. to log not all the used tracks, but just a subset). To

accommodate this risk, the <IPR society>hires a auditing

service from an <auditor>, who occasionally checks the logs

of the <IPR user>for correctness. The operational details

of this must be specified in the process skeleton for auditing.

4.2 Value Model

To obtain a working value constellation that clears and

repartitions a particular music track at a particular time,

we have to instantiate the skeleton as introduced in Sec. 4.1.

The following description mentions some elements of an in-

stantation process that takes a consumer need and a value

skeleton as input and produces the instantiated value model

of figure 4.2 as output. The instantiation process will itself

be a coordination process implemented by means of web ser-

vices. The process will be decentralized: It can be started by

any partner in the value web and it can involve any partner

in the value web. We describe in the following the instanti-

ation process for our example value skeleton.

This instantiation process is triggered by the IPR user.

In Fig. 4.2, the parameter <IPR user>has been bound to

‘TopPop radio’, a radio station, who wants to clear the track

‘BiCycle’ by ‘The Cyclists’.

For a given track (such as ‘Bicycle’ by ‘The Cyclists’), the

instantiation process must obtain the list of IPR societies

required to clear this track. We assumed in Sec. 3 that par-

ticipating actors have registered themselves in the value web

beforehand, and, at that time have provided sufficient infor-

mation to guide the instantiation process. For this specific

case, the registration process includes - per society - the ad-

dress of a web-service that can be used as a directory service

to look up whether the society is capable of clearing (a sub-

set of) the rights for a specific track. The directory service

itself can be distributed and replicated. A distributed and

replicated solution is required to control power of individ-

ual enterprises. If the directory would be fully centralized,

a single enterprise (owning the database) would be able to

control the instantiation process of a value web into a value

constellation. IPR clearance societies have made it clear to

us that this would not be an acceptable solution for them.

The international ISRC code may be used to identify mu-

sic tracks. The web-services of the various societies are used

to instantiate the skeleton value model. As a result, in our

example, the ‘SENA’ and the ‘BUMA’ are found as the two

IPR societies that are required to clear the music track ‘Bi-

cycle’ by ‘TheCyclists’.

Note that two IPR societies are selected, whereas the

skeleton only shows one society. This is the result of the

explosion element (‘#1’) as shown in the skeleton model;

this element says that generally, multiple rights (from mul-

tiple IPR owners) have to be cleared. It corresponds to the

directory lookup in the instantiation process that we de-

scribed above. In Fig. 4.2, the explosion element in Fig. 4.1

is ‘expanded’ by using an e3value value interface (denoted

as a bounded rectangle with arrowheads inside).

A value interface models that an actor wants to exchange

something of value (here: money) for something else that

carries economic value (here: a series of rights on a mu-

sic track). As such, the value interface models the notion of

‘economic reciprocity’: stating that in order to obtain some-

thing of value, something else of value should be offered in

return. For the example, the value interface of ‘TopPop ra-

dio’ bundles four value object transfers: A right with ‘SENA’

(a IPR society) for money, and a right with ‘BUMA’ (an-

other society), also for money.

The IPR societies themselves have to look-up into their

own local directories the IPR owners for the track to be

cleared. Again, this is part of the value model instantiation

process and it precedes the execution of the coordination

process discussed later. In the example, ‘SENA’ finds ‘John

Doe’, ‘Roger Cha’ and ’CDgram’, and ‘BUMA’ finds ‘Chris

Cross’ and ‘M. Oldfield’ and they return these answers to

the instantiation process, which then includes these in the

value model being generated. The directories of the IPR so-

cieties can be centralized, meaning that they are operated by

‘SENA’a and ‘BUMA’. This reflects the hierarchical nature

of these societies; they operate on behalf of right-owners,

who registered on beforehand at ‘SENA’a and ‘BUMA’.

The value perspective is useful for the following reasons.

First, the instantiation process of a value skeleton into a

value model, describing a particular value constellation sat-

isfying a need, selects enterprises and the services they offer

based on economical motivations. From a consumer per-

spective, the instantiation process should therefore select a

bundle of services that reasonably satisfies the stated con-

sumer need. This means that services and needs should be

described such that an informed value-based service selec-

tion can be made. Current approaches, e.g. on web-service

composition, do not consider services as commercially valu-

able entities, and therefore can not select services on an

economic-value criterion. Rather, such web-services are se-

lected on grounds of interoperability concerns, or technical

quality concerns. From a supplier perspective, the instanti-

ation process should select those suppliers that reasonably

can make a profit by participating in the constellation. Ad-

ditionally, the instantiation process must balance the inter-

ests of both the consumer and suppliers, such that the pro-

vided bundle of services can be provisioned in a sustainable

way.

Second, by taking a business value perspective, reasoning

about formation of a constellation can be done more effi-

ciently. If such reasoning would start at the coordination

process level, many possibilities would have to be consid-

ered that from a commercial perspective would not be of

interest (e.g. because one participant can not make a profit

by participating in the constellation at hand). Such alterna-

tives are discarded beforehand, by pre-selecting only com-

mercially viable constellations.

5. DISCUSSION AND FURTHER RESEARCH

5.1 Coordination process instantiation

A value model is not a process model because a value

model states the objects, services, rights, experiences and

other things of economic value are exchanged among eco-

nomic actors, but does not tell us what they should do to

realize these exchanges [8]. Fig. 3 shows a process skeleton

that is corresponds to the value skeleton by means of an

activity diagram. It says that when a track is scheduled,

certain payments have to be performed; the explosion ele-

ment in the value skeleton has been translated into a multi-

instance activity in the process skeleton.

The instantion process to create the process instance takes

the generated instance (value model) from the value skele-

ton as a starting point. First, the value model gives the

enterprises involved in the constellation. who, from a eco-

nomic value perspective, have agreed to participate in the

constellation at hand. Second, the value model lists the se-

lected commercial services, which need to be provisioned by

the coordination process.

The instantiation process that creates the process instance

from the process skeleton is analogous to the one for value

models. Figure 4 shows an instantiated process that corre-

sponds to the value model of figure 3.

5.2 Profitability estimations

This small example illustrates several issues to be resolved

to make this approach work. Our e3value models come with

techniques to make estimations of the number of value trans-

schedule

track

send

payment

(to right society)

concurrent

right societies

receive

payment

(from right user)

send

payment

(to right owner)

receive

payment

(from right society)

right owners

concurrent

play track

at scheduled time

Activity

Multi Instance

Activity

AND-Split / -Join

Process

Start

Process

Termination

Branch

Termination

Legend

Deadline!

Figure 3: Process skeleton for the IPR case.

schedule

track

send

payment

(to SENA)

TopPop radio SENA BUMA

send

payment

(to John Doe)

play track

at scheduled time

receive

payment

(from TopPop)

send

payment

(to BUMA)

receive

payment

(from TopPop)

send

payment

(to CDGram)

send

payment

(to Roger Cha)

send

payment

(to Chris Cross)

send

payment

(to M. Oldfield)

Figure 4: Instatiated process model for IPR case.

actions over a time period, and tools to make net present

value compuations of the profitability for a business based on

these estimations. This allows a business to decide whether

or not to join a value web on business-economic grounds.

However, these techniques are geared to the static case.

In order to decided whether or not to join a value web in

which value constellations are formed dynamically, a busi-

ness needs to estimate how often it will be part of a value

constellation within that web over a period of time. This re-

quires simulation techniques that are currently not yet part

of e3value .

5.3 Process skeleton design

We currently have methods to design a process model from

a value model. Weigand et al. [16] use a language-action

approach to suggest patterns for inter-business communica-

tion that support the economic value exchanges represented

in a value model (rights and money, in our case). This is

typically done by means of a request/accept communica-

tive transaction. Pijpers and Gordijn [12] propose a step-

wise transformation approach that first considers ownership

rights transfers, then possession right transfers and finally

the time ordering of these transfers. These approaches work

for the design of coordination processes from value models,

but we here need to design a process skeleton from a value

skeleton. Guidelines for this design process are closely inter-

twined with the directives for instantiating a process model

(or value model) from a process skeleton (value skeleton)

and we need to look into this in more detail.

5.4 Consistency of views

Second, we do have criteria for checking the consistency

of a coordination process with a value model in monetary

terms [3, 4]. Basically, these criteria say that the values

estimated to be exchanged in a time period by the value

model, are indeed exchanged in the coordination process.

But these criteria work at the value model level and we need

to lift them to the skeleton level.

5.5 Quality and risk

Third, different ways of realizing a value model in a pro-

cess model differ in efficiency (do we perform some actions

in a batched way?) and risk (do we perform audits once a

week or once a year?). These issues now come up due to

the possibility of automatic self-reconfiguration. The gen-

eral form of the problem is that there are objectives to re-

alize and choice of means to realize them. Since we want

to work in a decentralized manner, the best we can get is

local optimization and mutual adjustment on the basis of

negotiations. Risk estimations and trust assumptions play

a central role here: If we trust that violations by IPR users

hardly occur, then we can audit once a year; if we do not

rust IPR users, we may want to audit once a day but then

the costs for IPR societies will be correspondingly larger.

Current work on quality of service negotiations in dynamic

web service composition provides techniques to compare and

negotiate QoS levels of different services [17, 13], but it does

not take the risk assumptions of the service delivery process

into account. This needs to be dealt with top make dynamic

service composition an economic reality.

Acknowledgements

The authors thank Bert Hazelaar of SENA for offering the

IPR case study and the fruitful discussions about the IPR

case study.

6. REFERENCES

[1] L. Baresi, E. D. Nitto, C. Ghezzi, and S. Guinea. A

framework for the deployment of adaptable web

service compositions. Service Oriented Computing and

Applications, 1(1):75–91, April 2007.

[2] B. Benatallah, M. Dumas, and Q. Sheng. Facilitating

the rapid development and scalable orchestration of

composite web services. Distributed and Parallel

Databases, 17:5–37, 2005.

[3] L. Bodenstaff, P. Ceravolo, C. Fugazza, and

A. Wombacher. Toward semantic-aware representation

of digital business processes. In accepted for the

Second Knowledge Management in Organization

Conference (KMO), 2007.

[4] L. Bodenstaff, A. Wombacher, M. U. Reichert, and

R. J. Wieringa. Monitoring collaboration from a value

perspective. In E. Chang and F. K. Hussain, editors,

2007 Inaugural IEEE International Conference on

Digital Ecosystems and Technologies, Cairns,

Australia, volume 1 of IEEE-DEST 2007, pages

134–140, Los Alamitos, February 2007. IEEE

Computer Society Press.

[5] M. Colombo, E. D. Nitto, and M. mauri. SCENE: a

service composition execution environment supporting

dynamic changes disciplined through rules. In A. Dan

and W. Lamersdorf, editors, International Conference

of Service-Oriented Computing (ICSOC 2006), pages

191–202. Springer, 2006. LNCS 4294.

[6] A. Dan, D. Davis, R. Kearney, A. keller, R. King,

D. Kuebler, H. Ludwig, M. Polan, M. Spreizter, and

A. Youssef. Web services on demand: WSLA-driven

automated management. IBM Systems Journal,

43(1):136–158, 2004.

[7] J. Gordijn and H. Akkermans. Value based

requirements engineering: Exploring innovative

e-commerce idea. Requirements Engineering Journal,

8(2):114–134, 2003.

[8] J. Gordijn, H. Akkermans, and H. Van Vliet. Business

modelling is not process modelling. In Conceptual

Modeling for E-Business and the Web, ECOMO 2000,

volume 1921 of LNCS, pages 40–51. Springer, 2000.

[9] C. Gr¨onroos. Service Management and Marketing: A

Customer Relationship Management Approach, 2nd

edition. John Wiley & Sons, Chichester, UK, 2000.

[10] T. W. Malone, J. Yates, and R. I. Benjamin.

Electronic markets and electronic hierarchies.

Communications of the ACM, 30(6):484–497, 1987.

[11] B. Omelayenko. Web Service Configuration on the

Semantic Web: Exploring How Semantics Meets

Pragmatics,. PhD thesis, Vrije Universiteit,

Amsterdam, NL, 2005.

[12] V. Pijpers and J. Gordijn. Bridging business value

models and business process models in aviation value

webs via possession rights. In Proceedings of the 20th

Annual Hawaii International Conference on System

Sciences, page cdrom. Computer Society Press, 2007.

[13] S. Shen, G. Yu, T. Nie, R. Li, and X. Yang. Modeling

QoS for semantic equivalent web services. In Advances

in Web-Age Information Management (WAIM04),

pages 478–488. Springer, 2004. LNCS 3129.

[14] J. Spohrer and D. R. (eds). Special issue on services

science. Communications of the ACM, 49(7):30–87,

2006.

[15] D. Tapscott, D. Ticoll, and A. Lowy. Digital Capital -

Harnessing the Power of Business Webs. Nicholas

Brealy Publishing, London, UK, 2000.

[16] H. Weigand, P. Johannesson, B. Andersson,

M. Bergholtz, A. Edisuriya, and T. Ilayperuma. Value

object analysis and the transformation from value

model to process model. In G. Doumeingts, J. Mueller,

G. Morel, and B. Vallespir, editors, Enterprise

Interoperability: new challenges and approaches, pages

55–65. Springer-Verlag, London, 2007.

[17] L. Z. Yutu Liu, Anne Ngu. QoS computation and

policing in dynamic web service selection. In Proc

WWW2004, ACM, pp.66-73, 2004.