Supporting Enterprise Network Set Up
Combining ebXML, Semantic Tools and
Matteo BUSANELLI, Piero DE SABBATA
Nicola GESSA, Cristiano NOVELLI
ENEA, Via Martiri di Monte Sole 4, Bologna, 40127, Italy
Tel: +390516098671, Fax: +39 051 6098 084, Email: firstname.lastname@example.org,
email@example.com, firstname.lastname@example.org, email@example.com,
Abstract: The idea that undergoes the proposed paper is to present a collaborative
framework, mainly based on the ebXML standard, able to implement the extended
Smart Garment Organisation (xSGO) and Interoperability concepts in a useful set of
tools and reference specifications.
We assumed ebXML as a reference for the framework since ebXML represents, at
this moment, one of the most important initiatives for the standardisation of
collaborative eBusiness processes. The adoption of a standard framework, like
ebXML, should reduce the efforts required to set up an electronic collaboration.
Nevertheless ebXML lacks, for the moment, of practical implementations in real
cases of clusters of enterprises (whilst cases based on Public Administration are
known): with the proposed framework we would cover the gap between the ebXML
specifications and the needs for a real implementation of the extended Smart
Garment Organisation that is focused on a peculiar production chain like the
The scenario of global commerce relationships requires more and more new mechanisms
and tools that, adopting and implementing international standards like ebXML, can ease the
establishment and maintenance of new efficient business collaborations.
Starting from our previous experience supporting collaborations between companies,
from the Moda-ML FP5 project and the TexWeave standardisation initiative of
CEN/ISSS and from the concepts of eXtended Smart Garment Organization we create an
ebXML-based framework that is composed of a set of resources and a set of tools to
address specific problems that need to be solved before starting an industrial collaboration
between two or more industrial partners.
The eBusiness Watch report on B2B witnesses the difficulty of the T/C sector  in the
adoption of ICT in the sector to improve the collaboration between the partners: few
installations, regarding only large companies and the relationships with large retail
organisations rather than with suppliers and subcontractors.
To overcome these difficulties we adopt a standard-based approach: the main advantage
in using the developed tools is the reduction of the efforts required to create a new version
of standard documents for enterprise collaboration setup.
This paper describes, in the next section, the ebXML vision for the set-up of an e-
business collaboration, how we intend to support such vision and in which steps the
developed tools can be helpful. Section 3 will provide a general description of each
component that constitutes the framework. Finally, the open issues and the conclusions that
could by extracted from this experience will follow.
2. Setting-up an e-business collaboration: how to support the ebXML
Many efforts have been done to improve enterprise interaction . Our aim was to define a
framework that could, on one hand, support the enterprises to face several interoperability
issues and, on the other hand, that can rely upon a wide and complete standardisation
initiative, also in order to draw up the world of the standards with SME. On this
purpose ebXML represents one of the most complete standardisation initiatives .
The starting point in the ebXML context is the creation of a common understood and of
an shared XML document (called ebBP) that describes the whole business process
involving different partners, each of which with different roles in the production chain.
ebXML  does not specify explicitly the format of the exchanged documents: partners
are free to decide what is the format of the documents they want to manage and transfer.
But these documents must be explicitly indicated in the ebBP document with some
reference information. In some cases the necessity of creation of new documents arises
from the specific requests of the partners: in this case the people involved in the ebBP
creation or implementation should analyse the requirements of the parties and create the
new documents (or adapt the old ones).
It’s clear that modelling such type of business collaboration in a complex scenario
requires a great experience and a direct communication with the industries. It is also clear
that the statement of a standard, or of a public description of a business process, cannot be
made directly by one industry that, even if leader in a particular production sector, has not a
complete vision of the whole production process.
Our purpose is then mainly to develop enabling tools that can be used both by ICT
consultants and internal experts when setting up a business collaboration and that can be
easily reused when creating a new collaboration.
The ebXML standard also requires the parties involved in the process to create and
manage two different types of document. The first one, called Collaboration Protocol
Profile (CPP), defines the data about the party itself (like the role played in the
collaboration, the required/expected documents, the communication channels implemented
and the transport method available for each channel).
After the creation of the CPP, that is mainly based on the ebBP document, and that is in
some sense proprietary of each partner since it describes only the capabilities/requirements
of one party, the partners must find a common agreement to implement the collaboration,
by the comparison of the different CPPs.
This agreement is written in the Collaboration Protocol Agreement (CPA) and signed
by the partners that want to adopt it during the collaboration.
Fig. 1 depicts the “path” to define and establish e-business collaboration, the
relationships among all the components of the framework, and interaction between them.
The idea is to start from the definition of an ebBP document. This first step is
performed exploiting shared models of business scenarios and data structures. Both the
collaborations and the data models can be designed considering the semantic model
implemented in a set of OWL ontologies.
Fig. 1- A graphical representation of the interaction between different tools and resources
3. The basic components of the framework.
In the proposed framework, four components have been developed, under the LEAPFROG-
IP project  to cover the different aspects of the definition of an electronic collaboration:
1. the OntoMODA ontology and the Ontology Explorer, to define data models
2. the CPP Editor and the CPA MatchMaker, to define business processes.
All these tools, integrated together, support the set-up of a business collaboration.
3.1 OntoMODA Ontology and the Ontology Explorer
The OntoMODA ontology and the Ontology Explorer can be used together to build a part
of the ebBP document and, eventually, the electronic documents required during the
transaction (for example an Order Document that contains all information about the
provision of a fabric).
OntoMODA is a multilayered modular domain ontology oriented to the data modelling
and e-business data exchange. Its primary purpose is to model a part of the Textile and
Clothing sector knowledge through the semantic description of many aspects, like industrial
processes and treatments, product description (like fabrics, yarns and fibres) and their
characteristics and other information. It also is strictly related with the standardised
TexWeave vocabulary for which represents the semantic view.
In fact to support in a helpful manner the data modelling phase, our aim has been to
strongly interconnect the semantic model (the ontology) with a practical, formal and
standardised set of data structures, as that defined in the TexWeave initiative.
This interconnection is implemented adopting the W3C reccomendation for semantic
annotation  that allows adding the semantic information to XML Schema documents.
The figure shows the main parts of the architecture that we’ve implemented:
• OntoMODA, that is mainly composed of two sub-ontologies: Dynamic Ontology
(DO), Static Ontology (SO).
• Annotated XML Schemas and Type Libraries: this is a library of XS type and a set
of XS document annotated with the concepts defined in OntoMODA.
• ModaML Dictionary: this is a dictionary of business terms upon which it was based
the TexWeave standardisation specifications.
The Static Ontology models the Textile/Clothing domain knowledge, defining, for
example, concepts like “fabric” and specifying all its properties. It is connected with two
different types of connections to the Dynamic Ontology that, on the other hands, contains
all the semantic descriptions of the representation mechanisms adopted to exchange the
information modelled in the Static Ontology. Then, the Dynamic Ontology models the
XML components (types, elements and attributes) used as interchange data format in e-
business transactions. The Static Ontology itself is modular and therefore composed of
several sub-ontologies, each of which addresses different modelling and meta-modelling
aspects (i.e. ISO11179 standard, XML Schema meta modelling and the real sector
Fig.2 - OntoMODA overall architecture with annotated schemas and dictionary
The Dynamic Ontology is generated automatically from the ModaML Dictionary (that
can evolve in time - then it is dynamic) and it is split in three sub-ontologies concerning
Business Documents (like Order, Invoice, etc…), Business Processes (i.e. Fabric
production, Supplying etc…) and the XML Schema Components defined in the real XML
Schema files. Here the main connections connect respectively
1. the semantic representations of the business documents with their XML root
elements defined in XML Schema and
2. each semantic representation of the XML components with their real
representations in XML Schema files.
As said before the content of OntoMODA is split into its static and dynamic part. The
first one mainly describes products, their properties and treatments. It contains a
classification of the products and in particular describes the fabrics and their properties.
Each fabric property is classified for a particular kind of application. Moreover, a set of
relationships has been defined in order to interconnect the classes of the ontology and to
model the properties of the instances.
OntoMODA is also a great knowledge source that could be used for documentation
purpose. Thanks to the textual description of many concepts it can offer many interesting
information useful for who needs to know product definition, industrial treatments,
processes and fabric properties.
In order to search and read information through the OWL ontology we developed a web
application named Ontology Explorer. The tool lets the user surf the entire OntoMODA,
Annotated XML Schema
Dynamic Ontology (DO)
(Processes, Documents and XML
Static Ontology (SO)
(Fabric properties, Sector Knowledge
and Meta Models for data modeling)
starting from the taxonomy and picking up from it the desired concepts to see more detailed
information through apposite panels.
We integrate the ontology with the CPP Editor and the CPA Match Maker to allow an
easy and rapid access to the description of the processes defined into OntoMODA. This was
done adding simple links that open in a separate window the OE with the ontology loaded
on a particular business process.
There are many tools to edit and browse ontology. Protégé  is one of the most used
one, but there are many others. On the other hand, our aim is to simplify the operation of
The Ontology Explorer allows the user to navigate, in a simple way, ontologies (it is not
strictly related to the OntoMODA ontology and it can show all the online ontologies written
in OWL language) and to find concepts and information. Actually all the tools that manage
ontologies are really hard to use and to understand: sectorial experts could not be so skilled
in computer science or in ontology development to use these tools. Nevertheless, semantic
annotation and description ease the comprehension of the information for modelling data
and process defining a business collaboration. Then, a relevant problem in developing an
ontology for a classical industrial sectors, like the Textile/Garment one, is to create tools to
use it easily: the Ontology Explorer is a configurable web tool that is mainly oriented to the
Domain Expert rather then to the Ontology Expert or developer.
Usually, domain experts have great knowledge about concepts that concern their
expertise area, but their knowledge about ontology implementation is quite absent.
An example of a typical user of the Ontology Explorer (OE) could be a textile expert
who consults a sectorial ontology (like OntoMODA) to understand which properties can be
used to describe or to characterize a generic fabric.
The Ontology Explorer provides more and better functionalities than other tools
dedicated to the same purpose. To enable these functionalities, the Ontology Explorer is
designed to be intuitively to use (also for the inexpert user) and many visualization and
navigation configuration alternatives are available to the user. It also implements dynamic
components that respond to user input, thus enhancing interactivity.
3.2 The CP-NET tool set
CP-NET (Collaboration Protocol – Networking Enterprises Technology) is a software
application set to enable the enterprises, cooperating through a collaborative framework
ebXML-based, to establish and to perform Business Collaborations.
To achieve a Business Collaboration it is necessary to provide, for each couple of
enterprises, a common base upon which to start doing business. This base is basically a
Business Agreement and it is built, following the ebXML standard model, by comparison
and by match of two company Business Profiles. ebXML provides a XML standard to
describe both Profiles and Agreements: ebXML Collaboration Protocol Profile and
CP-NET provides two web applications to handle the ebXML CPPA specification: the
CPP editor and the CPA Match Maker.
The CPP editor allows the enterprises to create and modify their own CPPs
(Collaboration Protocol Profile), required to set up the collaboration with other partners and
reducing the number of errors, using a simple interface with the aid of a simple inline help.
In fact actually the CPPs are created by hand, directly writing the XML, because no tool
exists that allows to create it using a human friendly interface. The idea of the CPP editor is
to cover this gap, allowing a non XML expert to write a correct CPP.
The CPA MatchMaker wants to simplify the agreement process required to start up the
collaboration: it allows to create and to modify, from two CPP Profiles, the Collaboration
Protocol Agreements (CPA) for a couple of enterprises. Currently the two CPPs are
compared by hand, identifying both the possible problems and the agreements: the
problems are solved in a direct contact, using the phone or the fax, by the partners. At the
end of the process nowadays one of the partners must write down all the defined
agreements in a XML structured document. This document is the final CPA.
This process is very long in time, because the agreement process is, normally, not in
real time: when a possible conflict arises during the CPP comparison, the CPA writer must
contact the other party and negotiate about the modifications.
The CPA tool simplifies this agreement process reducing the comparison time and
highlighting directly the conflicts between the two CPPs. At the end of the agreement
process it writes down directly the CPA in the XML format. This reduces the time required
by the whole agreement process.
CP-NET framework supports the ebXML Business Process Specifications (ebBP
standard), therefore, into the CPP Profiles and into the CPA Agreements the enterprises can
describe their characteristics related to one or more Business Processes.
3.3 CP-NET Requirements
The CP-NET tools, CPP-Editor and CPA-MatchMaker, provide:
- a support to upload and store the ebXML CPP and CPA documents, checking and
validating them against the proper XML schema;
- a set of Data Access Object to read/write from/to generic DBMS (particularly
MySQL and Microsoft Access), remote ebXML ebBP documents, local ebXML
CPP and CPA documents;
- web interfaces, both web applications and web services, that guide users through a
logic step sequence to view/change information;
- a software architecture under MVC (Model View Controller) paradigm;
A general vision of the initial objective is showed in the diagram depicted in fig. 3. CP-
NET provides an infrastructure to access/edit XML files (particularly ebXML CPP and
CPA files) under the MVC (Model View Controller) paradigm. The application is
developed in Java language and includes a library set to implement further characteristics
not expected by the framework.
To make the application accessible from the web, the tools run on Apache Tomcat Web
Server with the support of the Apache Struts framework to develop web applications and
Apache Axis to publish web services.
The software architecture is structured in different layers. We can separate them in two
- Client Layer: users can access to CP-NET tools through a web browser or their own
web service client implementation.
- Web Layer: two ways are provided to access to CP-NET interface, through web
applications or web services calls. Both ways are linked with the same Business Delegate
layer and the information is arranged in java bean structures.
2) Core Application
- Business Delegate Layer: all the external accesses must go cross this border that is
the general interface to CP-NET core application. This layer provides a set of
methods that are called from Web Layer. Each method of Business Delegate is
called by Servlets of web applications and published as a single web service too;
- Business Logic Layer: this layer contains all the methods to perform the main tasks
that characterize CP-NET tools: to receive get/set commands from web layer, to
retrieve/insert information from/to XML files and database, to prepare java bean
structures, to reply to method calls, to handle errors and exceptions. To access to
resources (XML files and database) the business logic use Data Access Business
Objects (next layer);
- Data Access Business Object (DAO) Layer: CP-NET tools need to access to two
types of resources: XML files (ebXML CPP, CPA and ebBP documents) and
database (MySQL or Microsoft Access). The created DAO classes provide all the
methods to get/set information from/to resources.
- X-Lab libraries Layer: the DAO classes of previous layer are specific for CP-NET
tools to access to ebXML standards and to CP-NET database. To access to CPP,
CPA and ebBP files, the DAO classes extend the XLabDOM class that provides
constructors, methods and functionalities for generic XML files. To access to CP-
NET database, the DAO classes extend the Xdatabase class that provides
constructors, method and functionalities for generic databases. XLabDOM and
XDatabase are classes of org.xlab package, developed to reuse and sharing
commons procedures into ENEA XML-Laboratory.
- Java libraries Layer: other standard java libraries are included to implement
previous layer (for example: Xerces, Xalan, JDOM, …) to access to resources.
Fig. 3 - General vision
The proposed framework has been developed under the LEAPFROG-IP project, in order to
define new tools to improve e-business interoperability between the enterprises of the
Textile/Clothing sector. The framework will improve the ability of the enterprises to set-up
business collaboration, thanks to a complete set of tools that allows the modelling of some
relevant aspects related with the definition of a business agreement (from the definition of
the data format to exchange information to the definition of the business processes). The
benefits of this architecture could be perceived in the next year, when the adoption of ICT
tools for data modelling and agreement building will allow the enterprises to formalize the
e-business collaborations and then to automize the exchange of business documents. The
adoption of the developed framework will also bring to the definition of enterprise profiles
that can be used by the enterprises to find and better understand possible collaborations
among new and heterogeneous partners. The definition of such profiles represents the
premises to populate shared registry of enterprise profiles; CP-NET tools can moreover
ease the adoption of standardised documents, like UBL.
One of the next steps in our work will regard the strong integration of such types of
tools with applications specific for the exchange of business data (in the ebXML language,
a Message Service Handler – MSH), that we are now developing.
The introduction of the semantic vision of the different concepts will ease the usability
of the framework itself, allowing a simple access and comprehension also for non ICT
skilled users. A learnt lesson during these activities regards the complexity of existing
business documents: this complexity makes the adoption of e-business data formats really
hard (“customisation” of standard documents is one of the main issues for the enterprises
). The testing phase we are now starting includes the exploiting of the ontology to
build new, interoperable data format, and the creation of shared business models (ebBP)
that will be provided to the enterprise to allow then to design their own profile, following
the CPPA specification. These tests will involve both enterprises and domain experts (in the
context of the LEAPFROG-IP project) to evaluate the business collaboration design process
and to evaluate how to make this operation easy enough for SME.
It is worth to note that our effort does not want to create a brand new interoperability
framework, but aim to reinforce and support the adoption of shared standardisation
specifications with which it is strongly interconnected. These standards, ebXML and
TEXWeave, lack of practical implementations, especially in domains like the
Textile/Clothing one, and are consequently not easy to be adopted by the enterprises.
 N. Gessa, P. De Sabbata, M. Fraulini, T. Imolesi, G. Cucchiara, M. Marzocchi, F. Vitali, "Moda-ML, an
interoperability framework for the textile-clothing sector", IADIS International Conference WWWInternet
2003, p. 61-68, ISBN: 972-98947-1-X , 11/2003, Carvoeiro, Portogallo.
 [Moda-ML]: http://www.Moda-ML.org
 [TEXWeave]: http://www.texweave.org
 [ebXML]: http://www.ebXML.org/
 [ebBP 201]: “ebXML Business Process Specification Schema Technical Specification v2.0.1“, July 2005
 [CPPA21]: “Collaboration-Protocol Profile and Agreement Specification Version 2.1”, July 2005
 [LEAPFROG IP] (www.leapfrog-eu.org)
 e-Business Watch, sector report n. 01-11, August 2004, electronic Business in the Textile, Clothing and
 B. Medjahed, B. Benatallah, A. Bouguettaya, A. H. H. Ngu, and A. K. Elmagarmid, “Business-to-business
interactions: issues and enabling technologies”, VLDB J., 12(1):59–85, 2003.
 C. Bussler, “B2B Integration, Concepts and Architecture”, Springer - Verlag, 2003, ISBN 3-540-43487-
 K. Jakobs, “Standardisation and SME Users Mutually Exclusive?”, Proc. Multi-Conference on Business
Information Systems, Cuvillier Verlag, 2004
 [SAWSDL] http://www.w3.org/TR/sawsdl/
 [Protégé] Protégé home page: http://protege.stanford.edu/
 P. De Sabbata, N. Gessa, C. Novelli, A.Frascella, F. Vitali, "B2B: Standardisation or Customistation?", in
"Innovation and the Knowledge Economy Issues, Application, Case Studies", e-Challenges 2005 conference,
Ljubljiana, October 19-21 2005, pp 1556-1566, Dublin, Ireland, IOS PRESS, ISBN 1-58603-563-0.