Biology Reference
In-Depth Information
7.4.3
Results Dissemination
Although most of the presented systems already provide visualization tools to view
and analyze experiment results, we found limited support for sharing analysis results
with other (future) users. Virtual Data systems used in Pegasus as well as provenance
support in Taverna allow for sharing and reuse of experiment results, while in other
systems collaborative sharing of experimental results has to be performed manually.
Kepler and ICENI increase reusability and dissemination. In Kepler a new com-
ponent is being developed and which allow adding provenance information for each
workflow. This new component is attached to the workflow and can be used later as
a logbook or to search for experiments. Similar developments are taking place in
both Triana and Taverna. In ICENI a backend component called application mapper
is used to find the specific appropriate workflow.
While some support for design and dissemination exists, this support is incom-
plete especially when it comes to sharing resources. While there is no system that
can even match all the computational requirements of the limited number of
applications, trying to create one SWMS for all e-Science applications seems
unfeasible. In the next section we will present our own attempts at workflow support
and explain how our approach may help create a framework where sharing between
all e-Science applications becomes practical.
7.5
An e-Science Virtual Laboratory
In this section we describe the main components composing the architecture of the
e-Science virtual laboratory, which helps scientists to develop CPU- and data-
intensive complex applications and allow them to use a distributed and complex
computing infrastructure. The architecture follows a service-oriented approach; the
main components are either simple web services or WSRF compliant services.
6
This approach has a major benefit that is the virtualization of the complex and
distributed computing and storage infrastructure (Fig.
7.3
). It also allows building
a loosely coupled system, something that is highly required in a dynamic and non-
reliable environment such as the grid.
7.5.1
Process Flow Template (PFT)
This initial prototype of the e-Science framework introduces a two-level abstraction
approach where the Process Flow Template helps enforce best practices and
increases the reusability within a given domain by (Kaletas
2004
; Afsarmanesh
Search WWH ::
Custom Search