Information Technology Reference
In-Depth Information
Chapter 3
Workflow Concept of WS-PGRADE/gUSE
Á
kos Balask
ó
Abstract This chapter introduces the data-driven workflow concept supported by
the WS-PGRADE/gUSE system. Workflow management systems were investigated
by Workflow Management Coalition, among others in aspects of implemented data
flow structures, and several workflow patterns are identi
ed as commonly used and
meaningful workflow structures. The workflow concept of gUSE is shown by
introducing the supported data patterns and illustrating their creation in the system.
Moreover, the possibilities of utilizing parallelization techniques are described, and
then the different views of a workflow design and management are described
covering the whole lifecycle of a workflow development. Finally, more complex
composition of patterns and their creation techniques are shown.
3.1 Introduction
Scienti
c workflow management systems are widely known and highly accepted
tools to connect stand-alone scienti
c applications and/or services together to
access, process,
c data in an automated way. In most
cases the composition of these steps is not a straight-forward process, moreover
usually there is huge amount of data to be processed and stored remotely. In
addition, the application can take a long time to deal with the data, which requires
parallel processing to achieve results in a reasonable time. Another case is when the
input data drives the interpretation of the workflow, and different applications
should be executed depending on the input content. Such concepts indicate that
formal investigations are needed to avoid adhoc
filter, and visualize scienti
hence generally not optimal
solutions.
Search WWH ::




Custom Search