Information Technology Reference
In-Depth Information
and forwarded to the EDF, which in turn passes the final outcomes to the PA. This is shown in Table 1
(Row 6 and Row 7).
The different steps involved in implementing the proposed model are shown in Table 2. Any FIPA-
compliant agent environment is used to support the agent architecture at the presentation, federation,
and foundation layers. At the presentation layer, personal agents with customized user interfaces are
deployed on a single or multiple hosts. The agent environment at the federation layer supports EMA,
EDF, and CC agents on a single system or across a distributed architecture. Schema managers are re-
sponsible for developing the OWL-S ontology documents at the foundation layer, and Web services are
configured at this layer to expose the OWL-S documents to the IC agents using reference URIs. The
agent architecture at this layer enables direct binding between the IC agents and the local schemas at
the individual host systems.
conclusion
Although FIS models operate efficiently in a tightly coupled federation, the evolution pattern of the Web
is forcing a switch toward using loosely coupled federations of information sources. Loosely coupled
FISs pose immediate formidable integration challenges, especially when flexibility to allow organiza-
tions to join or leave the federation is highly desired. This chapter presents an extensibility model with
sumptuous bargaining power over traditional models of loose coupling. By integrating the fundamental
concepts borrowed from the semantic Web model (Berners-Lee et al., 2001) with the FIS architecture,
we inherit a solution model that is natural and semantically eliminates reliance on permanent structures.
The use of the multiagent system and documents based on ontology Web languages helps to construe
an extensible FIS model that is not simply adaptable to changes, but can also encourage organizations
to create mutually beneficial federations that is void of physical or geographical rigidity.
The major contribution of this work is the demonstration that FIPA-compliant agents within a se-
mantic Web framework can offer support for extensibility in a loosely coupled FIS. This is increasingly
important as the semantic Web framework becomes widely adopted in the future. The research also
demonstrates how the use of ontology markup (OWL-S) allows semantic heterogeneity irrespective of
the structural heterogeneity in the data model at the participating organizations. The model provides
implementation details regarding the ability to publish and access the OWL-S documents as URI ref-
erences. This allows independence between the agent framework and the information sources while
allowing the agent architecture to reconfigure and adjust based on the individual data-source profile.
The chapter demonstrates this using a real-world example.
This research assumes that agents have direct access to ontology documents stored at the local data
sources. Although this may be viewed as a limitation of the model, it was necessary in order to clearly
present how an agent community can effortlessly adapt and operate in an FIS where new content sources
are added constantly or changes are made to existing content sources. For the same reason, we did not
address data quality explicitly, but recognize that it is an important issue in FISs. Both data-quality and
access-control issues are heavily researched areas, and there is no discernable reason why outcomes
from existing research cannot be included into the proposed model. Left for future research is the task
of performance testing and evaluation. Although the implementation and testing of the framework may
not be an easy task, it will be a worthwhile effort since simplistic forms of loosely coupled semantic
Web infrastructures are making inroads into personal and commercial computing environments.
Search WWH ::
Custom Search