Information Technology Reference
In-Depth Information
shift equation of Petri nets, we discover that the set of all candidate
configuration processes is identical to the set of all basis solutions of the
state-shift equation, which leads to the useful conclusion that the
functional constraint of a configuration can be replaced by the state-
shift equation. Then, after compiling the QoS attributes and aggregation
function to compute the QoS for the whole configuration, we formalize
the optimal Web service configuration problem as a linear programming
problem. Hence, more efficient algorithms can be applied and large-size
configuration problems can be solved. Finally, we implement the pro-
posed functions and algorithms as the service configuration manager and
incorporate the Web service registration, orchestration, and invocation
functions into the framework. The implementation framework is a
platform from component design to concrete deployment.
The proposed approach in this chapter lacks such features as
concurrency and synchronization. Actually, a service configuration
net models a set of components that are used to make a particular
business function. It models the way in which they are configured. It
does not model the time sequences involved in executing particular
service operations. The techniques proposed in this chapter need to be
extended to deal with business process sequencing and run-time-related
QoS. The approaches based on Petri nets and elementary siphons will be
explored [137].
5.7 BIBLIOGRAPHIC NOTES
QoS Specification and Description
. Menasce [138] presents an
overview of the current QoS-related research in Web services area.
Ran [139] summarizes a set of quantifiable QoS parameters and
measurements, for example, reliability, capacity, availability, and
cost, into multiple QoS categories. tModel in UDDI registries is
used to formalize representation models for each QoS parameter.
Perryea and Chung [140] group Web services into several commu-
nities according to nonfunctional requirements. Xiong and Fan
[133] propose a uniform QoS attribute definition framework in
order to coordinate QoS properties from both Web service
provider's and user's perspectives. Liu et al. [141] extend the
QoS model by including both generic and domain specific criteria.
The QoS information is usually collected from service providers,
