Information Technology Reference
In-Depth Information
FIS is limiting and inadequate. There is a compelling demand for flexibility that allows data sources to
join and leave an FIS. The work presented in this chapter rides on a stern belief that the need for flex-
ibility can be accomplished by extending the loosely coupled FIS model to include concepts from the
semantic Web model. The beliefs are validated by a detailed discussion of the proposed model and how
the model adds value in a real-world loosely coupled FIS setting. Whereas current FIS models function
efficiently in a tightly coupled federation, a loosely coupled federation poses formidable integration
challenges. Interweaving the fundamental FIS model with the semantic Web model (Berners-Lee et
al., 2001) consisting of documents based on ontology description languages and a multiagent system
could be a viable solution to liberate the loosely coupled federation from the clutches of prior concerns
that limited their practical adoption.
A MoTIVATIng exAMPLe
The role of federating schemas in an FIS can be best explained using a running example. Consider the
case of a home buyer requesting prequalifying for a loan prior to the loan application process. To the
potential home owner, this is where the rubber meets the road. The prequalifying process essentially
determines the amount that the borrower will qualify for. To the loan officer, it is a critical process that
can identify red flags before the loan approval and sanction. In reviewing the preloan request applica-
tion, the loan officer interacts with numerous sources of information in an attempt to paint the most
accurate picture of the borrower. For the sake of contextualization, assume that the loan officer oper-
ates in a federated environment where he or she constantly accesses specific data sources pertaining
to the following four categories of information: employment history, financial stability, credit check,
and property valuation. The information from these sources will help the loan officer to determine
the correct loan amount based on an applicant's unique credit and employment history, income and
debt, and property investment goals. The employment history provides pertinent information regard-
ing employment dates, addresses, and salary data, while financial stability is shown through multiple
other data sources to validate information regarding investment accounts, stocks and bonds, checking
and savings account balances, retirement plans, and life insurance policies. In addition, the final loan
amount is also dependent upon the applicant's residential history (real estate rented, owned, or sold)
and the credit-rating scores from one or many credit-reporting agencies. The loan officer interacts with
the finite set of data sources in the foundation layer via the federation layer using the presentation layer.
This is represented in Figure 2.
As organizations evolve and grow over time, new information becomes available, and changes to the
federation become inevitable. New data sources may be added to the federation while some data sources
are no longer retained or undergo structural and semantic changes. If the federation layer is incapable
of adjusting to changes in the foundation layer, the federated schema is no longer in correspondence
with the local data sources, thereby severely falling short in meeting the levels of expectations of the
presentation layer. For example, the FIS for an online loan-approval system may decide to include ad-
ditional insurance companies in their listing or may terminate partnership with a specific credit-report-
ing agency. Irrespective of whether the federated schema is tightly coupled or loosely coupled, changes
in the foundation layer have to be reflected in the overlying architectural layers of the FIS to reduce
inconveniences to the loan officer and to reduce the time lost during the approval process. Deficiencies
of this nature commonly experienced when working with multiple heterogeneous data sources can be
Search WWH ::
Custom Search