Information Technology Reference
In-Depth Information
Service is not only able to handle information for
which it has been designed rather it can handle
different information using different annotations.
Example:
For an application, there is the need to store
data about an enterprise's employees. A manager
is modeled in this way:
dissimilar to the way used in the closed systems
where information is being validated at different
systems layers. The first type of validation hap-
pens at the user/data interface. At the business
logic layer, there are probably some functions
that check the data from another perspective. At
the data layer, triggers and stored procedures are
used to validate the date from a third perspective.
All these validations perspectives are together
responsible for the validity of an object state.
At SOA, a component does not know which
validation is done at which layer (or component).
As a result, each component itself has to check
the validity of an object state. More precisely,
the object itself should be able to determine its
correct state. In semantic SOA-based model, this
functionality is realized in the validator compo-
nent. The reason to determine the object state at
this level is that it is the point where an object
reaches its persistent state. An object stored in a
database (represented by a Web Service) has to be
in a correct state, because all the other components
that are using this object expect its correctness. If
an object can be interchanged between different
applications, its state has to be validated too. The
validator acts as a test library to support these
requirements.
Each test represents an ontology or a part of
it. The relation between an ontology and a test is
documented using the Uniform Resource Identi-
fier (URI) (Berners-Lee, 2005). As a result, the
state of an object can be validated starting from
the semantic level, which is the abstract ontology,
down to the physical representation at the Web
Services level.
Figure 4 illustrated this idea clearly, it shows
that two different components A and B are using
the same component C. Component A probably
shares data via component C with component B.
The flow: {Semantic Interface → Component
A → Component C → Validator} can be affected
by the flow {Semantic Interface → Component
B → Component C → Validator}.
≔
class manager
{first_
name:string,last_
name:string,born:date,job_
description:string}
A developer is modeled as follows:
≔
class developer
{first_
name:string,last_
name:string,born:date,job_
description:string}
An object with the same attributes can represent
both employees. In object-oriented programming
(OOP) the solution would be the generalization
of the class employee.
class employee
{first_
name:string,last_
name:string,born:date,job_
description:string}
class manager
≔
employee
class developer
∶
∶
employee
The separation of the annotation from the Web
Service itself allows using similar techniques to
reduce the need of different services. Related to
the previous example, the annotation created by
the validator decides if an object is the representa-
tion of a manager or a developer.
information flow
In general, components used in SOA have to work
within different applications. Consequently, the
way that the validation of information performs is
Search WWH ::
Custom Search