Wide-ranging interoperability architectures and frameworks have been proposed to date. A comprehensive review of some of these has been documented by Chen et al (2008) and this section, therefore, concentrates on a discussion of the most pertinent interoperability architectures and frameworks relevant to this work. Early efforts fostered have resulted in
well-established reference architectures such as (1) the Computer Integrated Manufacturing Open System Architecture (CIMOSA) (AMICE, 1993), (2) the Purdue Enterprise Reference Architecture (PERA) (Williams, 1994), (3) the GRAI-GIM reference model (Chen and Doumeingts, 1996) and (4) the Reference Model of Open Distributed Processing (RM-ODP) (ISO/IEC 10746, 1996).
With the evolving view on interoperability at enterprise level, a majority of interoperability architectures and frameworks are being established according to the strategic principles related to the requirements for business interoperability, considerations for appropriate technological support and the chosen architecture perspective. The Zachman Framework (The Zachman Framework Website, 2009), IDEAS interoperability framework (Chen et al, 2004) and The Open Group Architecture Framework (TOGAF) (TOGAF Website, 2009), for example, all identify significant multi-level prerequisites for enabling enterprise interoperability.
In the IDEAS interoperability framework, which has been developed within the ATHENA project (Ruggaber, 2006), a specific dimension is acknowledged for the implications of semantics cutting across the “business”, “knowledge” and
“Information and Communication Technology” (ICT) levels within single enterprises and the need for integrating, unifying and federating across enterprise boundaries. This understanding is portrayed in the simplified IDEAS interoperability framework in Figure 2-12.
Business
Knowledge
ICT Systems Semantics Enterprise A
Business
Knowledge
ICT Systems
Semantics
Enterprise B
Integrated Unified Federated
Figure 2-12 IDEAS Interoperability Framework (Redrawn from Chen et al (2004))
In the context of international standards, a multi-dimensional framework has been proposed for enterprise interoperability (CEN/ISO 11354, 2008). The first elaborated part of the framework entails the requirements for enabling process interoperability across manufacturing enterprises. Figure 2-13, adapted from CEN/ISO 11354 (2008) illustrates the Framework for Enterprise Interoperability. There exist three dimensions to the framework notably (1) the barriers to interoperability such as conceptual and technological, (2) relevant concerns such as business and process and (3) the approaches to interoperability such as federated and unified. In Figure 2-13, PSL has been positioned according to the Framework for Enterprise Interoperability, and it can be seen that the “conceptual”, “process” and “unified” dimensions help position PSL in the right segment of the framework matrix.
Other architectures, such as the semantic-mediation architecture for business-to-business interoperability (Vujasinovic et al, 2007), have also been researched and industrially validated. In their work, Vujasinovic et al (2007) have implemented their architecture within the ATHENA research project (Ruggaber, 2006). Their implementation platform primarily harnesses Semantic Web tools with XML and RDF(S) capability. Vetere and Lenzerini (2005), on the other hand, have identified four different types of models for semantic interoperability in service-oriented architectures, by following an understanding of centralised and decentralised mappings between service
PSL approaches
barriers
concerns
conceptual technological organisational
business
process
service
data
integrated unified federated
Figure 2-13 Positioning PSL in the Framework for Enterprise Interoperability (Redrawn from CEN/ISO 11354 (2008))
schemas. However, this work has remained at a conceptual level since no test case implementation is proposed.
In current literature, very few contributions have coined the terms “semantic interoperability framework”. Amidst these contributions lies the extended COntext INterchange (eCOIN) framework (Firat et al, 2007), whose main purpose is to facilitate semantic reconciliation through the definition of reusable “conversion function networks” as mappings. The authors of eCOIN adopt a view that the achievement of semantic interoperability should take account of semantic heterogeneity as well as semantic reconciliation. It has been argued that the eCOIN uses a hybrid of ontology-based methods involving principles like ontology alignment through articulation axioms and ontology merging (Firat et al, 2007). However, the motivational scenarios that back up eCOIN remain broad in nature and have not been attuned to the world of product design and manufacture.
Specifically in the field of product design and manufacture, relatively few frameworks have been proposed in order to contribute to semantic interoperability. Patil et al (2005), for instance, have presented an approach to foster the semantic interoperability of product data utilising an ontology-based framework. This framework for semantic interoperability is identified in Figure 2-14.
Following the framework diagram proposed by Patil et al (2005), it is possible to identify two main reconciliation mechanisms present namely (1) the mapping of the semantics from a “System A” and “System B” into an intermediate product ontology (Product Semantic Representation Language (PSRL) which is DL-based) and (2) the translation of syntax and terminology
A‟s Semantics
A‟s Terminology
B‟s Semantics
B‟s Terminology PSRL Semantics
PSRL Syntax
System A System B
Figure 2-14 Framework for Semantic Interoperability by Patil et al (2005)
from “System A” to syntax in PSRL, which is then translated to the syntax of the target “System B”. It is to be noted that Patil et al (2005) have recognised that their approach does not support low levels of abstraction in product models, such as geometric entities, as a result of their preference for the domain of Description Logics.
Gupta and Gurumoorthy (2008) have argued a feature-based framework to support semantic interoperability of product models. The concept of “Domain Independent Form Feature” (DIFF) has been proposed, over which the framework is constructed. Figure 2-15 illustrates their schematic concept which enables semantic interoperability of product models. In the figure, the DIFF model supported by an ontology, provides a basis for the representation of features, and facilitates semantic interoperability between a source and a target system.
In their approach, Gupta and Gurumoorthy (2008) have focused on the definition of features in terms of their faces solely, and have looked exclusively at semantic interoperability problems occurring due to different labels that refer to the same shape and different representations for the same shape. This implies that other significant considerations for (1) feature function in design, (2) relationships between features and manufacturing processes and (3) other forms of semantic interoperability issues remain to be addressed. Furthermore, the authors have implemented their framework utilising the Protégé (Protégé Website, 2009) ontological environment. Since Protégé does not provide full support for First Order heavyweight semantics,
Ontology
Figure 2-15 Gupta and Gurumoorthy's (2008) Approach to Semantic Interoperability of Product Model
this implies that opportunities still exist for improving the expressiveness of semantics in product models.