Six years after its final adoption as ITU-T
Recommendation and ISO/IEC International Standard, the Reference Model of
Open Distributed Processing (RM-ODP) is increasingly relevant, mainly
because the size and complexity of current IT systems is challenging most
of the current software engineering methods and tools. These methods and
tools were not conceived for use with large, open and distributed systems,
which are precisely the systems that the RM-ODP addresses. As a result, we
are now witnessing many major companies and organizations investigating RM-ODP as a promising alternative for specifying their IT systems, and for
structuring their large-scale distributed software designs.
There are however some issues that may hamper the widespread
adoption of ODP by these companies. One of them is the fact that the Reference
Model does not prescribe any specific notation for representing its concepts and
viewpoint languages, which really hinders the development of commercial tools
for writing and analysing ODP system specifications.
The wide adoption of UML by industry, the number of available
UML tools, and the increasing interest for model-driven development and the MDA
initiative, motivated ISO/IEC and ITU-T to launch a joint project in 2004, which
aims to define the use of UML for ODP system specifications (ITU-T Rec. X.906 |
ISO/IEC 19793). Thus, ODP modellers could use the UML notation for expressing
their ODP specifications in a standard graphical way, and UML modellers could
use the RM-ODP concepts and mechanisms to structure their UML system
specifications. In addition, modelling tool suppliers will be able to develop
UML-based tools capable of expressing RM-ODP viewpoint specifications.
However, the use of UML as the language and notation of
choice for ODP system modelling is not straightforward. For instance, the object
models followed by UML and ODP do not match completely (e.g., UML is
class-based, whilst ODP is object-based; their behavioural models are different;
etc.). In addition, the loose semantics of UML may represent an impediment to
achieving the precise specification and analysis of ODP systems and the
development of tools, and some ODP concepts are difficult to map into UML (e.g.,
policy and accountability concepts). On the contrary, the simplicity and
elegance of RM-ODP contrasts with the complexity and semantic vagueness of UML.
Thus, clarification and simplification of UML semantics may be based on the
RM-ODP in this context. Furthermore, using ODP for business modelling may
provide an excellent basis for establishing technology- and tool-independent
communication mechanisms between business and IT stakeholders.
A second issue is the fact that some UML modellers are used to other design
methodologies and architectural practices (e.g., RUP, IEEE Std. 1471, the “4+1”
views model, the Zachman’s framework, etc.) that define different sets of
viewpoints on a system. The continuing relevance of RM-ODP in the context of
these software architectural practices, and their relationship to RM-ODP needs
to be determined in order to accomplish the degree of interoperability required.
Finally, the precise role played in this picture by model
driven architectures (MDA), service oriented architectures (SOA), component-based
architectures (CBA), event driven architectures (EDA) and other architectural
approaches is still unclear—although the RM-ODP might well serve as a
coordinating framework for them all.
This Workshop will provide a discussion forum where researchers,
practitioners, system modelers, tool developers and representatives of
standardization bodies can meet and exchange experiences, problems and
ideas related to the use of UML for ODP system specifications, and its
relationship with other architectural practices and approaches (e.g., MDA,
SOA, CBA, EDA) in the realm of Enterprise Distributed Computing.
The duration of the workshop is one day and will be held on September 19,
Topics of interest include, but are not limited to:
Ö UML and RM-ODP
Ö UML Profiles for RM-ODP
- specification of systems using approaches based on UML and the RM-ODP;
- current issues, limitations and problems of using UML to represent ODP
- Representation of specific concepts of ODP viewpoint languages in UML,
- policy concepts (obligations, permissions, authorizations,
- accountability concepts (commitments, prescriptions, evaluations,
declarations and delegations);
- multiparty synchronous interactions.
- impact of UML 2.0 on the specification of RM-ODP concepts;
- use of UML and ODP for specific application domains (finance, telecomms,
health care, aerospace, etc.).
Ö Architectural practices/approaches and RM-ODP
- formal semantics for subsets of UML or UML Profiles to support the
specification of RM-ODP concepts;
- foundations of RM-ODP based on exact philosophy of science (such as that
discussed by Bunge, Hayek and others);
- definition of UML Profiles for ODP viewpoints and for extensions and
refinements of the RM-ODP for particular application domains.
- use of MDA concepts and mechanisms for ODP system specifications;
- use of other modelling languages for ODP system specification and
- relationship and integration of RM-ODP with other architectural
practices (e.g., IEEE Std. 1471, RUP) and architectural approaches (e.g.,
SOA, CBA, or EDA).
Ö Case studies and experiences
- requirements for supporting tools or tool chains (currently in use or
- use of existing tools for ODP system specifications.
- knowledge gained from experiences with graphical notations for large
distributed system specifications;
- application of the RM-ODP for large-scale software projects;
- identification of potential problems for, and benefits from, the
industrial adoption of the UML for ODP system specification.
To enable lively and productive discussions, attendance will be limited
to 25 participants and submission of a paper or a position statement is
required. All submissions will be formally peer reviewed.
Submissions should be 4 to 8 pages long in
IEEE Computer Society format and
include the author's name, affiliation and contact details. They should be
submitted by e-mail as postscript or PDF files before June 13, 2005, to
the Workshop Chairs (
Authors will be notified of acceptance by July 11, 2005. At least one
author of accepted papers should participate in the Workshop. Workshop
proceedings will be published on the conference CD-ROM, and all accepted papers
will appear in the IEEE Digital Library.
- Workshop papers due: 13 June 2005
Author notification: 11 July 2005
Final papers due: 15 August 2005
Workshop date: 19 September 2005
Peter F. Linington, University of Kent (UK)
Akira Tanaka, Hitachi (Japan)
Sandy Tyndale-Biscoe Open-IT (UK)
Antonio Vallecillo University of
- Dave Akehurst
University of Kent (UK)
- Joao P. Almeida
University of Twente (The Netherlands)
- Jean Bérubé
- Jonathan Billington
University of South Wales (Australia)
- Celso González
- Haim Kilov
Stevens Institute of Technology (US)
- Lea Kutvonen
University of Helsinki (Finland)
- Peter F. Linington
University of Kent (UK)
- Arve Meisingset
- Joaquin Miller
X-Change Technologies (US)
- Tom Rutt
Coast Enterprises, INCITS T3 IR (US)
- Akira Tanaka
- Sandy Tyndale-Biscoe Open-IT (UK)
- Antonio Vallecillo
University of Málaga (Spain)
- Bryan Wood