Conference Room 3
East Midlands Conference Centre
Nottingham

Wednesday 21 September, 17:00–19:00

Background information

The National Institute for Environmental eScience (NIEeS) and the Reading eScience Centre are organising a "Birds of a Feather" session at the 2005 All Hands meeting in Nottingham.

Although there is a strong focus on the role of data access and data management within the environmental sciences, there is also a high demand for computations based on model simulations or data analysis/assimilation. The first-round NERC environmental escience projects are now maturing, and new projects will be coming on stream soon. It is appropriate to take stock of how environmental escience projects are using computational grids and associated grid middleware. One particular aspect that may be unique to the environmental escience community is the essential integration of compute and data grids.

This “Birds of a Feather” session will follow successful sessions arranged at the previous two All Hands meeting, in which we bring together the NERC-funded environmental escience projects with similar projects funded by other research councils (such as BiodiversityWorld and Equator), with participation from an audience containing people with experience in other areas of escience. These environmental escience projects span a range of methodologies in grid computing, using a similarly wide range of metadata. Examples are

1. The eMinerals project, making use of core middleware products such as Globus, Condor (and Condor-G), portal technologies and the Storage Resource Broker, and developing the integration of these tools and the integration of web services with Condor.
2. The GENIE project, using condor for Earth system simulations, and using ICENE and the MyGrid workbench.
3. GODIVA and NERC DataGrid, using web services to access data for computations.
4. ClimatePrediction.net, using home computers and now based on Berkley’s BOINC (the software behind SETI@home).

Session content

Each presenter will be expected to show three powerpoint slides and talk for no more than five minutes, with slides following the sequence (but with titles of the presenter's chosing):

1. The context of the computational issues – what kinds of computations are required within the project
2. One or two key problems currently being faced by the project (ideally where is not yet a proven solution in sight)
3. One or two successful solutions to problems encountered by the project, which might act as useful hints to other projects

The slides will need to be given to the organising group in advance, and used to formulate a set of discussion questions to be given in the session. These will be used to guide the discussion following presentations (which will be ordered and grouped according to the issues to be raised by the presenters).

We anticipate that the discussions will encompass projects’ experience and development of a number of key technologies and methodologies, such as:

1. Implementation of grid technologies for performing calculations
2. Use of web and grid services to run and manage computations
3. Use of portals for running and managing computations
4. Visualisation of computational results
5. Integration of compute and data grids (e.g. with technologies such as the Storage Resource Broker)
6. Metadata challenges
7. Data interoperability between codes, e.g. using XML

A detailed programme will appear nearer the time.

Further details

Please contact Stuart Ballard tel 01223 765669) for more information, or if you would like to offer a contribution to this session.