Information Technology Reference
In-Depth Information
Thus, service requestors face a large number of choices of service
configurations that can provide a similar function. An environment
with changeable QoS of the component services makes the selection
problem even more challenging. Therefore, selecting an optimal
configuration remains difficult.
In this chapter, we address the optimal configuration issues by
concentrating on
(a)
the formal modeling and definition of Web service configura-
tion by using a formal methodology, i.e, Petri nets; and
(b)
an optimal QoS search algorithm under varying configuration
constraints.
For the first issue, we introduce a formally defined configuration net
named service configuration net (SC-net) based on Petri nets in order to
help select candidate configurations. The main advantages of adopting
Petri nets as a modeling language are twofold. First, as a graphical and
mathematical tool, Petri nets can provide a uniform environment for
dealing with large and complex Web service configurations. Second, by
firing the transitions in the SC-net, we can generate a set of candidate
configurations automatically and thus minimize the human efforts in
searching for a configuration. For the second issue, before we present our
optimal QoS search algorithm, we carry out structural analysis to
discover the essential properties of the SC-net we build. The analysis
results clarify that the set of basis solutions of a state-shift equation of an
SC-net net is identical to the set of the solutions that correspond to
realizable configuration processes. Based on this, we formulate Web
service configuration under single and multiple QoS attributes as a linear
programming problem. This greatly alleviates the solution complexity.
The rest of the chapter is organized as follows: Section 5.2 discusses
Web service quality attributes, their measurement, and aggregation
function to be used in optimal configuration. Section 5.3 introduces
assembly Petri nets and their properties. It also discusses how to model
the configuration problem formally. Section 5.4 specifies the optimal
Web service configuration problem and presents a real example to
certify the validity of the analysis results and solution methodology.
Section 5.5 introduces the implementation of the methodology. Finally,
Section 5.6 concludes the chapter.
