SEMANTIC INTEROPERABILITY FOR E-BUSINESS IN THE ISP
SERVICE DOMAIN
Jane Hall
Fraunhofer FOKUS, Kaiserin-Augusta-Allee 31, 10859 Berlin, Germany
Stefanos Koukoulas
Eworx S.A., 66 Jean Moreas Str, 152 31 Halandri, Athens, Greece
Keywords: Ontology, semantic interoperability, eBusiness, virtual organisation, ISPs.
Abstract: Enterprise interoperability is a challenging goal that has many facets, one of which being the need for
cooperating enterprises to have a precise understanding of the information that they exchange with each
other. Internet Service Providers (ISPs) collaborating in a virtual cluster to market, customize and provision
services to customers require the same understanding of these services and their features. This paper
describes the work being undertaken in the European IST project VISP (IST-FP6-027178) to meet this
requirement. ISP services are being classified and formally specified and an ontology is being developed for
the ISP service domain.
1 INTRODUCTION
Enterprises wishing to undertake eBusiness by
cooperating in a virtual organization must agree on
the definitions and meanings of concepts used in
their cooperation, and the software infrastructure
supporting their eBusiness transactions must be
aligned with this understanding so that services can
be offered and provisioned for customers
dynamically, efficiently and effectively. This paper
is concerned with the work being developed in the
European IST project VISP (IST-FP6-027178) to
support semantic interoperability for a virtual cluster
of small ISPs collaborating to market, customize and
provision services for business customers.
The structure of the paper is as follows. First the
eBusiness context of small ISPs and their
collaboration in a virtual cluster is briefly outlined.
Reference is then made to the significance of
knowledge and semantics for enterprise
interoperability in the European research area and to
research projects here. Work that has been
undertaken in the VISP project on the specification
and classification of ISP services is introduced. Then
the ontology that has been developed for the services
in the ISP domain is presented, including an
application of its use in composing services for
eBusiness. A brief summary of the work undertaken
concludes the paper.
2 ISPS AND VISP
Since the advent of the Internet, the traditional ISP
market has been in constant evolution due to the
gradual globalization and commoditization of ISP
services and to deregulation initiatives aimed at
fostering competition. Small ISPs are best at
targeting niche markets; they can respond rapidly
and provide excellent customer service. However,
small ISPs cannot offer the wide range of services
now required by increasingly demanding business
customers nor offer the geographical coverage that is
necessary in a globalized marketplace.
Collaborating in a virtual cluster to undertake
eBusiness enables small ISPs to offer a wider range
of services, particularly when customizing tailored
services that are composed from individual services.
Tailoring composite services from the services of
other partners in the cluster in order to offer
innovative and higher value solutions to customers is
indeed one of the principle advantages and reasons
390
Hall J. and Koukoulas S. (2008).
SEMANTIC INTEROPERABILITY FOR E-BUSINESS IN THE ISP SERVICE DOMAIN.
In Proceedings of the International Conference on e-Business, pages 390-396
DOI: 10.5220/0001910103900396
Copyright
c
SciTePress
behind the idea of collaborating in a virtual VISP
cluster with other small SMEs.
However, when composing services it is
necessary to have a precise understanding of the
features of each service element so that the
composition of individual services into a service
bundle is consistent and the composite service can
be validated. In order to meet these requirements,
work was undertaken within the VISP project to
develop a service decomposition and
characterization methodology and to establish an
ontology for semantic interoperability within the
cluster.
3 KNOWLEDGE MANAGEMENT
AND ONTOLOGY FOR
ENTERPRISE
INTEROPERABILITY
The relevance of knowledge management and
semantic interoperability for cooperating enterprises
is a topic of current significance. One of the four
Grand Challenges in the Enterprise Interoperability
Research Roadmap (Li et al., 2006) is Knowledge-
Oriented Collaboration, now renamed Knowledge-
Oriented Collaboration and Semantic
Interoperability to emphasize the significance of
semantics in enterprise interoperability (Charalabidis
et al., 2008). Knowledge-based collaboration allows
the sharing of knowledge within virtual
organisations to the mutual benefit of all partners.
The advances being undertaken in the area of
knowledge management and ontology development
are intended to be of direct benefit to enterprise
collaboration. Those sharing the knowledge need
have the same understanding of it, hence the
requirement for ontology definitions in this area.
Despite a rapidly changing research area, work
already undertaken was studied in the VISP project
to establish the state of the art in ontology
specification methods and ontology management
techniques. Existing approaches were examined and
their strengths and weaknesses evaluated in
conjunction with the specific requirements of the
VISP project for the use of an ontology.
The On-to-Knowledge IST project
(www.ontoknowledge.org) was investigated and
although the standards used in the project have
evolved, the experiences gained were found to be
useful input (Davies et al., 2002). The OTS
methodology and architecture in particular were
used as input to the VISP ontology work.
The DIP IST project (http://dip.semanticweb.org)
objective was to develop and extend Semantic Web
and Web service technologies, and it undertook an
exhaustive state of the art analysis in many of the
fields relevant to VISP (DIP, 2004). It developed the
DIP Ontology Management Suite, which is an
integrated set of tools for efficiently and effectively
managing ontologies. VISP was able to input this
work into its own state of the art analysis.
The Semantically-Enabled Knowledge
Technologies (SEKT) IST project (www.sekt-
project.com) objective was to develop and exploit
the knowledge techniques underlying Next
Generation Knowledge Management. It also
performed a state of the art analysis on ontology
evolution which was used as input to the work
undertaken by VISP (SEKT, 2004).
Other projects undertaking work in this area
include INTEROP, where ontology was one of its
three key thematic components and research was
undertaken to investigate ontology in conjunction
with interoperability (http://interop-vlab.eu);
FUSION, which is developing technologies for the
semantic integration of a heterogeneous set of
business applications within SMEs
(www.fusionweb.org/fusion); SUPER, which is
aiming to integrate Semantic Web Services and
business processes into one consolidated technology
(www.ip-super.org); and SYNERGY, one of the
aims of which is to provide semantic ontology-based
modelling of knowledge structures on collaborative
working (Popplewell et al., 2008).
The work carried out in these and other projects
provided generic results that could be evaluated and
used in VISP as a basis for developing a specific
ontology for services in the ISP domain.
4 SPECIFYING ISP SERVICES IN
VISP
A VISP cluster can only work efficiently if all
partners have the same view not only of what a
service is and what it offers but also of relationships
and constraints between services, service
characteristics and characteristic values. In a
federation of independent autonomous entities, there
is no centralized control but there must be a common
understanding between the partners of services and
their properties in order to collaborate dynamically
in real time. Properly specifying the properties of the
elementary service components is “a key aspect of
supporting the proper specification of (composite)
SEMANTIC INTEROPERABILITY FOR E-BUSINESS IN THE ISP SERVICE DOMAIN
391
services’ semantics” (Leymann, 2007). The
composite service must behave correctly when
provisioned for the customer.
A primary objective of the VISP project is to
build an ISP service knowledge base. A service has
been defined in VISP as representing an atomic set
of capabilities or activities that can be provisioned
and used alone, or that can be grouped with other
services to form a composite service. A set of
elementary services for the virtual cluster has been
investigated, classified and specified. The result is a
knowledge base of ISP services formally represented
using standard representations.
The VISP Service Knowledge Base (SKB) is a
service catalogue that contains the technical
information relating to the individual services that
can be used to offer services to customers either
standalone or as part of a composite service. The
SKB thus provides a common technical
understanding of services in the cluster, i.e. a
common “language” shared by all partners that
ensures consistency within the cluster.
Services described in the SKB are organized in
categories and sub-categories. The nine categories
comprise access services, bandwidth services,
hosting services, application services, network
services, security services, support services, and also
software supplies and material supplies. Each of the
nine categories defined is divided into sub-
categories. The SKB can be centralized, distributed
or replicated at each partner, depending on the
functional architecture adopted.
A basic service modelling approach was adopted.
A service is specified by a list of characteristics and
each characteristic comprises a list of values. The
service, each characteristic and each value has an
associated description plus other parameters. This
approach is very simple, generic and reusable. It
does not depend on any particular technical or
business environment and is applicable to many
fields apart from the ISP/telecom domain.
Each service is thus characterized by a general
description, a set of mandatory and optional
parameters with their possible values, usage
limitations and exclusions together with information
on how they can be composed. All services offered
by partners in the cluster are described using this
specification. This information constitutes the VISP
service knowledge base that, as the reference for ISP
services in a VISP cluster, is being used to specify
the VISP ontology.
5 ONTOLOGY FOR THE ISP
DOMAIN
An ontology is being developed in the VISP project
for VISP purposes, i.e. to enable partners in the
cluster to collaborate in providing tailored services
to customers as dynamically and in as automated a
manner as possible. An ontology can be regarded as
a “set of shared conceptualisations of entities within
an application domain” (Wilson et al., 2006). In
order to carry out eBusiness dynamically and in
close cooperation with other partners in a virtual
enterprise, all partners in the cluster need to adopt a
common semantic model for their interactions, a
shared understanding that can be automated by using
a software infrastructure supporting the eBusiness of
the ISPs. The VISP ontology thus constitutes one of
the innovative results of the project.
Although problems are associated with the
development of ontologies (Hepp, 2007), it was felt
that the requirement for automating what can be a
complex task meant that an ontology was essential
for a VISP cluster. Work undertaken on developing
an ontology for service components is often of a
quite generic nature, for example (Sheth et al., 2006)
and (Wang and Xu, 2008). However, the ontology
here is being developed for VISP cluster purposes,
to support the validation of ISP services composed
automatically and dynamically from other services
in an eBusiness context. It therefore has a restricted
purpose and a limited circle of users and it should
not be the bottleneck mentioned by Hepp when
ontologies that are created by a small community are
intended for much wider use.
The main use of an ontology in VISP is in the
area of service composition and tailoring of
innovative and often complex customized services.
When new services are composed of individual
component services, it must be ensured that the
features of each service comprising the composite
service do not conflict or result in inconsistency in
the tailored service. The use of ontology is intended
to assist the representation of the complex semantic
and the relationships between the services offered by
the partners in a VISP cluster. Validation of a
service bundle requires a validation of the rules and
restrictions among all services that form the service
bundle. Semantic information such as restrictions,
relationships between services, rules for the values
of characteristics, etc. is static. It was therefore
decided to model the content of the SKB using an
ontology whose instance data are the service
descriptions of all VISP services.
ICE-B 2008 - International Conference on e-Business
392
An analysis was performed within the VISP
project to compare ontology tools and languages and
to decide on the approach to take given the VISP
requirements. Languages such as Topic Maps
(www.isotopicmaps.org/sam/sam-model), RDF
Schema specification (www.w3.org/TR/rdf-schema),
OWL Web Ontology Language,
(www.w3.org/TR/owl-ref/), OWL-S
(www.w3.org/Submission/OWL-S) and F-Logic
(Kifer et al., 1995) were investigated. Ontology
editors were evaluated based on (Gomez-Perez et al.,
2002) and work performed in SEKT (SEKT, 2004).
Repositories for storing and retrieving ontologies
and ontology change management systems were also
evaluated.
Based on the evaluation results and the
requirements of the VISP cluster in doing eBusiness,
Protégé (http://protege.stanford.edu) was selected as
the ontology editing tool, providing not only
ontology editing capabilities but also constituting
ontology library systems and offering various
functions for managing, adapting and standardizing
groups of ontologies while enabling ontology reuse.
No tool fulfilled all the requirements for a reliable,
secure, interoperable ontology management
environment although most seemed to provide
adequate solutions for most of the main issues in the
VISP eBusiness environment. Protégé was one of
the first editors available, it is open source and
supports a variety of plugins and import formats,
such as RDF Schema and OWL. It is also widely
used; 68.2% of respondents in a survey were using
Protégé as their ontology editor (Cardoso, 2007).
The principle requirement for an ontology
specification language in VISP is expressive power
and OWL was selected here. It supports the RDF
data model, is used in the Semantic Web world and
was felt to be the most appropriate choice for the
eBusiness domain to be modelled in VISP. It too is
widely used; 75.9% of respondents in the survey
mentioned above are using OWL to develop their
ontologies (Cardoso, 2007).
A two-phase methodology was developed for
designing, developing and using the ontology-based
knowledge management system for the VISP
software environment. In the first phase, the
application area and use of the ontology were
determined. In the second phase, the ontology was
built based on the information sources identified in
the first phase and the concepts used there. Although
the area of tools and technologies enabling the
application of ontologies on a business level is not
considered mature and much work is still in
progress, the use of an ontology in VISP was
recommended. Given the increasingly complex
requirements of the VISP eBusiness environment
and the need for rich, consistent and reusable
semantics, ontologies represent an optimal answer to
the demand for an intelligent system that operates
more closely to the human conceptual level.
5.1 Description of the SKB Ontology
The ontology has been developed in VISP to
represent the semantic information included in the
VISP domain, i.e. ISP services. This ontology is
intended to model the SKB but would also describe
the classification of service descriptions, the
attributes of service descriptions as well as the rules,
constraints and relationships between them.
The service descriptions were classified as
presented in section 4 above. The goal of the SKB
ontology definition is to support this classification of
service descriptions, to include the attributes of
service descriptions and the relationships between
them as well as to demonstrate the rules and the
constraints for these attributes in a formal way that
can be used to validate a composite service, or
service set as it is termed in VISP.
Figure 1: Class hierarchy for the SKB ontology in Protégé.
The Protégé OWL editor provides three views
for the definition of an ontology, namely description
of classes (see Figure 1), description of properties
(see Figure 2) and description of individuals. For the
description of classes the SKB ontology is primarily
based on the schema of the SKB repository.
However, the schema includes several details that
SEMANTIC INTEROPERABILITY FOR E-BUSINESS IN THE ISP SERVICE DOMAIN
393
are not required in the definition of an ontology, the
primary purpose of which is to enable the validation
of a service set from within the VISP software
infrastructure. Such information includes mainly
fields that are used to describe static attributes inside
the SKB and which are of no special interest inside
the SKB ontology, for example, Description,
GenericDeploymentInformation, References.
The second view is the description of properties.
OWL properties model all relationships between
individuals of different classes. Apart from this pure
listing of the ontology properties, this model also
includes a description of the restrictions that each
property of each class should satisfy. This aspect is
further discussed in the following section.
Figure 2: List of OWL properties in the SKB ontology.
The third view is the description of individuals.
Individuals are instances of classes such as in object-
oriented programming languages. Thus the
descriptions of classes mentioned above also provide
general descriptions for the individuals of these
classes.
5.2 Restrictions to be Represented
The use of an ontology cannot only be justified by
the need to model the domain of interest in a fairly
simple way by using OWL classes and properties
that resemble capabilities of UML diagrams and
objected-oriented design. The need for an ontology
has evolved due to the need to define complicated
rules and constraints governing the domain of
interest (in this case the ISP service domain). The
definition of these restrictions is probably the most
important part of the SKB ontology in the sense that
it differentiates it from a flat database schema such
as the SKB itself. Such rules and constraints can be
defined in OWL provided that it is possible to
express them as logical expressions.
Therefore, the definition of the SKB ontology is
a two-step process:
1. First, the “skeleton” ontology model has to be
specified. The class hierarchy, the listing of
properties and the definition of individuals form
the necessary schema elements of this model and
will be utilized to describe the domain
restrictions in the next step.
2. Then the rules and constraints governing the ISP
services domain have to be transformed into
logical expressions that can then be expressed
using the expressive syntax of OWL.
Although the work on the classification of ISP
services and their specification using the service
description template (as described in section 4
above) has been almost completed and fully
documented, the work on the documentation of the
exact rules that each service will have to comply
with has not started yet. Since it is expected that this
task will involve a great amount of work which will
be continued even after the end of the project, a
workaround has been used to ensure that OWL is
adequate to express all possible rules and restrictions
in the ISP world. A demo service was described and
the following possible types of restrictions that the
modeller could be asked to represent were identified:
1. The set of selectable enumerated values of a
characteristic depends on the value of another
characteristic.
2. The existence of a characteristic depends on the
value of another characteristic.
3. The range of values of a characteristic depends
on the values of another characteristic (very
similar to case 1 but more complex to represent
because deals with ranges).
4. Characteristics are mutually exclusive, but at
least one must be present although they are
indicated as optional.
The first type of restriction essentially constitutes
an “if” statement: if the value of characteristic A is
greater than 5, characteristic B can only take values
1, 2 and 3. The representation of an “if” statement
can be easily achieved with the help of the logical
operators AND, OR and NOT. If X then Y is
equivalent to NOT X OR Y. Of course, OWL allows
the definition of such a logical restriction.
As the same analysis has been performed for all
the above types of restrictions, it is clear that OWL
and the definition of the SKB ontology model satisfy
ICE-B 2008 - International Conference on e-Business
394
the requirements of the ISP domain as well as of the
VISP use case called “validation of a service set”.
A technical restriction imposed by the current
OWL specification is the lack of support for numeric
ranges. While OWL has cardinality restrictions,
datatype values cannot be further restricted. One of
the solutions evaluated was to extend OWL files
with URI references into an (external) XML Schema
file. This means that a separate XML Schema file
together with the OWL file will have to be
maintained. Another solution is to define an
extension ontology to be imported by the SKB
ontology, which defines RDF properties that can be
used to represent XML Schema facets. Other
solutions are also being evaluated and hopefully the
next OWL specification will be enriched with the
capability of expressing restricted datatypes.
5.3 Use of the SKB Ontology
The use of the ontology comprises two concepts:
how to manipulate the ontology and how to exploit it
to enable the validation of a service set.
When manipulating the SKP ontology, Protégé
allows for a series of actions regarding a definition
of an ontology, such as to:
create the ontology definition
browse the ontology
edit/update the ontology definition
query the ontology
store the ontology persistently
apply various ontology checks
compute the inferred taxonomy
The last two actions are performed with the
assistance of an OWL reasoner, which can be
combined with Protégé if it supports the DIG
interface. An OWL reasoner provides the following
standard inference services:
Consistency checking: determines whether the
ontology contains any contradictory facts.
Concept satisfiability: determines whether it is
possible for all classes to have any instances.
Classification: computes the inferred class
hierarchy.
In general, Protégé supports all features expected
of an OWL editor. The SKB ontology will be
manipulated through the use of Protégé, at least until
the end of release 3 of the VISP software
infrastructure. In the future it may be desirable to be
able to perform certain actions on the ontology from
the VISP platform itself.
The primary reason for the definition of the SKB
ontology was to enable the implementation of the
validation of a service set. When a customer requests
one or more services from VISP, a sales
representative creates a service set that comprises
various services offered by partners in the cluster.
After a service set has been created, it needs to be
validated to ensure it can be instantiated. The sales
representative uses the VISP GUI to click on the
“validate” button to validate this service set.
The use of the SKB ontology allows for a more
sophisticated validation strategy than the rather
simple and incomplete validation undertaken in the
first two releases of the VISP infrastructure
software. The ontology is being used to enhance the
first implementation, e.g. checking intra-service
constraints between characteristics and values,
checking inter-service constraints between services,
characteristics and values, etc.
Compared with the current implementation, the
additional necessary architectural element will be a
framework enabling the manipulation and the usage
of the ontology. Such a framework can be Jena
(http://jena.sourceforge.net/index.html). Jena is a
Java framework for building applications related to
the Semantic Web in general. It includes:
RDF API
OWL API
Rule-based inference engine
In-memory and persistent storage
SPARQL query engine (www.w3.org/TR/rdf-
sparql-query/)
Its most useful feature for VISP is that it allows an
ontology or a data set to be represented with an
object model, as well as incorporating reasoning
capabilities.
A main prerequisite before the validation can
start is the existence of a consistent SKB ontology.
A service description is added to the SKB ontology
when it is approved in the VISP cluster with the
appropriate restrictions. The next steps are then
followed inside the implementation code of the
validation of a service set:
1. The system reads the service set that is to be
validated either through the repository
containing information about the service sets and
their component services or through an XML
file.
2. It creates a model representing the data set of
this service set.
3. It loads the ontology model of the SKB
ontology. This can be read and created each time
from the respective OWL file or it can be stored
persistently inside a database.
4. It validates the data model against the existing
ontology model, i.e. it performs a global check
across the schema and instance data looking for
inconsistencies.
SEMANTIC INTEROPERABILITY FOR E-BUSINESS IN THE ISP SERVICE DOMAIN
395
5. The result will report whether the validation
check has passed and detail any detected
inconsistencies.
This approach is quite simple and can be seamlessly
integrated with the current implementation of the
service set validation since Jena is a Java framework
that can used by the existing classes.
6 CONCLUSIONS
The VISP project has developed both a methodology
for categorizing and describing ISP services and an
ontology for the ISP service domain. The need to
add semantic content to ISP service descriptions is
fundamental for ISPs carrying out eBusiness for
innovative, complex and composite ISP services.
The VISP ontology has been produced for a
specific context and is being applied in this context.
This is indeed a recommended approach when the
aim is a lightweight ontology that can be cost
effective to design, build and maintain (Alani et al.,
2008). Ontologies do not need to be large and
complex and the intention of the VISP project was
innovative in that it has designed an ontology that
could be part of the VISP platform in a running
environment. Decisions taken in developing the
ontology tended therefore to be pragmatic with the
context, stakeholders and potential users in mind.
The intention is to further develop the results of
the VISP project for commercial use in a virtual
cluster of small ISPs. The ontology is an essential
part of this aim and has been designed for such use.
The work depicted here has built on existing
research work in the area and has applied it to a
specific context. This has required a lightweight
approach avoiding any undue complexity and the
development of a small and highly focussed
ontology that fits in with the specific requirements
of the VISP cluster and its partners. First results
have shown that the ontology has met these
requirements and that ISP composite services can be
automatically validated with the ontology during
service provisioning.
ACKNOWLEDGEMENTS
This work was carried out with partial funding from
the EU through the IST project VISP (IST-FP6-
027178). Further information on VISP can be
obtained from www.visp-project.org.
REFERENCES
Alani, H., Hall, W., O’Hara, K, Shadbolt, N., Szomszor,
M, Chandler, P., 2008. “Building a Pragmatic
Semantic Web”, IEEE Intelligent Systems, 23(3),
pp.61-68.
Cardoso, J., 2007. “The Semantic Web Vision: Where Are
We?”, IEEE Intelligent Systems, 22(5), pp.84-88.
Charalabidis, Y., Gionis, G., Hermann, K.M., Martinez,
C., 2008. Enterprise Interoperability Research
Roadmap, Update (Version 5.0), European
Commission.
Davies, J., Fensel, D., van Harmelen, F., 2002. Towards
the Semantic Web: Ontology-driven Knowledge
Management, Wiley, Chichester.
DIP, 2004. D2.1 Report on Requirements Analysis and
State of the Art (WP2 – Ontology Management), DIP
project IST-FP6-507483, Deliverable D2.1.
Gomez-Perez, A., Angele, J., Fernandez-Lopez, M.,
Christophides, V., Stutt, A., Sure, Y., 2002. A survey
on ontology tools, OntoWeb deliverable 1.3,
Universidad Politecnia de Madrid.
Hepp, M., 2007. “Possible Ontologies. How Reality
Constrains the Development of Relevant Ontologies”,
IEEE Intelligent Systems, 22(1), pp.90-96.
Kifer, M., Lausen, G., Wu, J., 1995. “Logical Foundations
of Object-Oriented and Frame-Based Languages”,
Journal of the ACM, 42(4), pp.741-843.
Leymann, F., 2007. “Semantics for the Cloud”, IEEE
Intelligent Systems, 22(5), pp.15-17.
Li, M-S., Cabral, R., Doumeingts, G., Popplewell, K.,
2006. Enterprise Interoperability Research Roadmap,
Final Version (Version 4.0), European Commission.
Popplewell, K., Stojanovic, N., Abecker, A., Apostolou,
D., Mentzas, G., Harding, J., 2008. “Supporting
Adaptive Enterprise Collaboration though Semantic
Knowledge Services”. In Enterprise Interoperability
III. New Challenges and Industrial Approaches, ed. K.
Mertins, R. Ruggaber, K. Popplewell, X. Xu,
Springer-Verlag, London, pp.381-393.
SEKT, 2004. D3.1.1 Ontology Management and Evolution
– Survey, Methods and Prototypes, SEKT project IST-
2003-506826, Deliverable D3.1.1.
Sheth, A., Verma, K., Gomadom, K., 2006. “Semantics to
energize the full services spectrum”, Communications
of the ACM, 49(7), pp.55-61.
Wang, Z., Xu, X., 2008. “Ontology-based Service
Component Model for Interoperability of Service
Systems”. In Enterprise Interoperability III. New
Challenges and Industrial Approaches, ed. K. Mertins,
R. Ruggaber, K. Popplewell, X. Xu, Springer-Verlag,
London, pp.367-380.
Wilson, M., Li, M-S., Jaekel, F-W., Ruggaber, R., Chen,
D., Mehandjiev, N., Zelm, M, 2006. Enterprise
Interoperability Roadmap. Annex I – Indicative
Research Challenges, Final Version (Version 4.0),
European Commission.
ICE-B 2008 - International Conference on e-Business
396