Support for Cooperative Experiments in e-Science: From Scientific Work flows to Knowledge Sharing - Identification of Ligand Binding Site and Protein-Protein Interaction Area - page 147

scheduled on geographical and grid-enabled computational resources. When data is
available for processing, it is passed/streamed directly to workflow components
enabling concurrent execution. The engine supports enacting workflow components
written in a number of programming and scripting languages (Java, C ++ , python,
SWIG), and also allows access to RPC-style web services and workflows (Korkhov
et al. 2007a, b ). The system has been used to prototype a number of application
(Inda et al. 2008 ; Zudilova-Seintra et al. 2002 ; Leguy et al. 2009 ) . Once an application
workflow is designed, it can be attached to PFT and become available for scientists
who can create multiple instances to meet their specific requirements such as setting
the different parameters and new input data sets. WS-VLAM offers not only an
intuitive way for the creation and execution of application workflows, but also
provides seamless access to the underlying complex grid-enabled infrastructure.
WS-VLAM has the ability to (1) interact and monitor the workflow at runtime,
(2) automatic redirection of the graphical output to the end-user default screen,
(3) easily adapt/change the application workflow to meet user-specific needs, and
(4) run workflows in batch mode.
The DC Analysis workflow for the MACS lab experiment was one of the first
scientific workflow to be ported to the earlier versions of WS-VLAM 1.5. Through
this workflow, scientists are able to query a database containing detailed informa-
tion about all data produced by the sample treatment process. The (raw) data sets
corresponding to query results are retrieved and piped into an apodisation routine
(also called a tapering function). The apodised data sets are subsequently submitted
to a fast Fourier transform and calibration procedure. The results are piped to a data
viewer for visualisation and a multivariate data analysis module for extraction of
principal components (Hendrikse et al. 2003 ) .

7.5.4 The Work fl ow Bus

From the state of art study presented in Sect. 7.4 it is clear that, at least in the near
future, a unique workflow management for e-Science is unlikely to emerge. All of
the presented systems have their advantages and disadvantages; moreover, most of
them are building small communities of users around themselves. It is evident that
at a certain point, in order to continue to promote sharing and reusability, there will
be no other way but to bridge these systems to allow scientists to re-assemble
workflows in different systems. To achieve this goal, a meta-execution framework is
needed for integrating different workflows, coordinating the execution of different
enactors and moving data around. This approach will become more feasible once
most of the workflow management systems have adopted a service-oriented
architecture where the engine and enactors are implemented as standalone services.
The basic idea of a workflow bus (Zhao et al. 2006 ) is to wrap a number of popular
and relatively mature legacy SWMSs as federated components, and to loosely
couple them as one meta-workflow system using a software bus. The workflow bus
is an interactive workflow environment, which provides an agent-based wrapper for