 |
WP3 Goals Summary
Triana is written in Java, and this GAT will be also designed to assist the
interfacing of other Java programs to Grids.
While Cactus and Triana were originally developed to support complementary
types of applications, there is also some overlap which will be exploited
through this project. As part of this WP we will extend Triana to include
heterogeneous modules executing on remote machines: different languages, automatic
compilation, use of control scripts, and a set of APIs that form an extensible interface
with Cactus. The TGAT will demonstrate greatly enhanced capabilities over
the current Triana application model, such as
automatic transfer of execution to remote machines, remote program steering
using metadata, two-way interfacing of data-flow and control using Cactus.
These capabilities will be tested for data analysis applications
on user testbeds, where they will deliver larger capacity, faster time to solution,
and collaborative interaction with analysis.
WP3 Tasks
| T3.1 |
Scheduler Development (month: 3-15) |
|
Develop schedulers for Triana workflow that jointly schedule
data transfers and work on remote machines. These schedulers will
execute Triana networks using a common XML format which can be
used to schedule other non-triana resources (interacting with task 3.3).
|
| T3.2 |
TGAT and Native Code Extensions (month: 1-36) |
|
Prototype the Triana Grid Application Toolkit (TGAT)to interface with
Cactus and other applications. Extend Triana execution model to include
heterogeneous modules executing on remote machines in different languages
with automatic compilation. Integrate or extend an existing wrapper generator
(such as SWIG, JCI or the XML based wrapper developed at the Dept. Computer
Science, Cardiff) to interface with native codes. Implement a drag and drop
graphical user interface to connect the wrapped code to other resources and
to provide a graphical data-mediation layer between these components.
|
| T3.3 |
XML Metadata Extensions (month: 1-21) |
|
Develop metadata associated with main data flows e.g. history of processing,
duration and cost of execution. Standardize metadata into a common XML
format and interface to databases of data, programs, and scripts that use
metadata as indices.
|
| T3.4 |
Optimisation of Workflow (month : 24-36) |
|
Incorporate appropriate metadata into schedulers that make
intelligent decisions about the flow of work using tools developed
in the GAT workpakage.
|
| T3.5 |
Mobile Communications and Stearing (month 30-36) |
|
Develop mobile capabilities, signalling events to mobile users (SMS, email),
remote steering of execution and workflow from small platforms (eg handhelds).
|
| T3.6 |
TGAT Testbed (month 30-36) |
|
Verify tools using GEO and LIGO gravitational wave signal data.
Binary black hole inspiral signals will be artificially injected into
GEO and LIGO data and TGAT will be used to make coincidence searches for
the presence of these signals. Such simulations will help in optimising
the search algorithms. Once the search algorithms are optimised they will
be applied to real data without the synthetic signals. After the completion
of the tests the methodology will be disseminated to the entire gravitational
wave community.
|
|
