Applications of WSMO - Enabling Semantic Web Services: The Web Service Modeling Ontology - page 157

able to resolve a request as in the scenario outlined by applying semantically
enabled technologies for Web service usage.
For the purpose of this example, we shall omit the provision of the WSMO
element definitions but shall discuss how the VISP platform achieves its func-
tionality by applying the techniques for discovery, composition, mediation,
and execution described in detail in Chapter 9. For discovery, the ISP aims
to detect the cheapest offer among all suppliers that satisfies the customer's
request. Therefore, it creates a WSMO goal containing an axiom stating that
the product to be delivered must have the lowest price among the applicable
products. This allows to run a discovery engine that finds the Web service
of the supplier that offers the requested product at the cheapest rate. More-
over, the choreography interface descriptions, along with a grounding to an
executable Web service communication technology, allow the ISP to automat-
ically invoke and execute the suppliers' Web services.
If no supplier can provide the requested product as a whole but it can
be provided as an aggregate by combining the offers from several suppliers,
the VISP platform can compose the respective Web services and provide the
result as a new, virtually created product bundle to the customer. Imagine
that providing such a composed product bundle requires the Web services
of provider X and provider Y. However, several types of heterogeneity that
hamper the necessary interaction of the Web services might occur: X might
use UBL 2 for describing its Web services, while Y might use WSML; X might
define the customer name as a string, while Y might define it as consisting
of a given name, middle initial, and surname. Furthermore, it might occur
that the composition of X and Y defines an interleaved process, meaning that
X requires some input from the ISP that can only be given after the ISP
has received a certain output from Y. Therefore, a WW mediator

can be
utilized that provides the mediation facilities for handling these mismatches.
In this case,

M

would incorporate two OO mediators - one for translating
between UBL and WSML, and one that resolves terminology mismatches
between the ontologies - and would utilize a process mediator that allows to
resolve the communication mismatches between X and Y.
In conclusion, we observe that Semantic Web services allow one to handle
the EAI problem by lifting information and process descriptions to a seman-
tic level. A specific characteristic of WSMO is that its four top-level concepts
correspond to the core elements of a Semantic Web service based on an EAI
platform and define them in a concise and unambiguous manner. Ontologies
provide the semantically formalized terminology, WSMO Web services se-
mantically annotate and execute computational functionalities, WSMO goals
precisely specify client objectives in a machine-processable way, and WSMO
mediators provide an infrastructure and techniques for handling and resolving
potential heterogeneities that would prevent successful interaction of entities.

M

2 OASIS Universal Business Language, see http://docs.oasis-open.org/ubl/
prd-UBL-2.0 .