Biology Reference
In-Depth Information
The ProxyWS approach has helped optimize a number of applications where the
adoption of a service-oriented architecture has led to a decrease in overall perfor-
mance, like the
Service-Oriented Visualisation applied to medical data analysis
(Zudilova-Seintra et al.
2002
) , and the
indexing and Name Entry Recognition
services
developed in Adaptive Information Disclosure project.
11
ProxyWS enables
exposure of data-intensive applications as web services without loss of performance.
Experiments such as MACS Lab and, more specifically, the
DC analysis
work fl ow
can thus be exposed as a set of web services which can be easily used by a wider
scienti fi c community.
7.6
Summary
In this chapter, we discussed some of the workflow-related issues in e-Science.
First, we analyzed the requirements for supporting different application domains,
and then discussed the necessity and importance of developing a unified framework
for e-Science. We illustrated this vision by describing the approach followed in the
context of the VL-e Project to develop such a framework.
The emergence of grid environments gives scientists new ambitions to tackle
more complex and large-scale problems, which can lead to new methodologies
in scientific research and problem solving. The core idea of e-Science is to allow
scientists representing different domains to share resources and knowledge and to
collaborate in their research. In this chapter, we highlighted two issues pertinent to
developing a collaboration environment for e-science: Modeling workflows on the
level of the entire scientific experiment lifecycle enables knowledge transfer between
scientists where successful experiment results and templates can be applied for new
problems as reusable resources.
Our state of the art analysis showed that the development of many SWMSs are
highly application-driven. The specifics of different application domains result in dif-
ferent workflow models and different styles of user support, which limits the opportu-
nities for sharing and interoperability. The other lesson to be learned from the survey
is that while practical support exists in some form for all the stages of the e-Science
experiment workflow, methodological support is lacking. A simplistic approach to
address this problem will likely fail; more advanced solutions have to be investigated
with the aid of recent achievements in semantics, Web 2.0, and ontologies.
One of the conclusion of the report from the NSF/Mellon Workshop on Scientific
and Scholarly Workflow organized 2007 (Klingenstein et al.
2007
) is that there will
be no single SWMS which is usable for all e-Science experiments. The absence of
standards in the field of SWMS seriously limits the sharing of resources between
diverse e-Science applications across scientific fields and various SWMSs.
A potential approach to lower the interoperability problem is through the use of
service-oriented architectures. Standards such as WSDL for publishing services
Search WWH ::
Custom Search