Member of the International Virtual Observatory Alliance

Current German E-Science Projects: Status 2009

by Iliya Nickelt (GAVO/AIP), Harry Enke (AstroGrid-D/AIP), 25.06.2009

Over the last years numerous smaller and larger e-Science projects were carried out in Germany. Products and services are now used in daily scientific work, some of them with impressive quantity. This short overview is intended to give examples for the outcome so far and list the research in the field.

GAVO

Attach:Millenium-merger.jpg Δ The German participation in the international Virtual Observatory alliance started in 2003 and the BMBF will in its current form continue funding until 2011. Research institutes that participate in the 3rd phase of GAVO are ZAH, AIfA, MPE, AIP, TUM and IAAT. A total of seven students and postdocs are working on GAVO projects full or part time.

As the oldest German e-Science project GAVO produced a number of services. From the scientific viewpoint the most successful example is the Millenium Database where over a hundred publications made use of this GAVO service that presents a novel way to open up the vast resources of a cosmological simulation to the scientific community.

In general the results of theoretical models have become a focus of current GAVO research, with general model description ans well as services like TheoSSA to calculate model spectra on the fly. Of course, GAVO also publishes many sets of observational data, most of it via the GAVO data center. This service offers an fast and assisted way for every German astronomer to release results, and we hope to see an increased demand in the near future. GAVO also supports the community with tutorials for efficient VO use for colleagues.

Attach:VO-tools.png Δ

For further details about GAVO, see the Jenam 2008 overview or the GAVO website at http://www.g-vo.org.


AstroGrid-D

The German Astrophysical Community Grid project ran from mid-2005 to beginning of 2009. The BMBF funded the development as part of the D-Grid initiative with the goal to set up a nationwide Grid infrastructure for sciences. Consequently, large parts of the German astrophysical community participated in the project (AEI, AIP, MPA, MPE, MPIA and ZAH as well as LRZ, TUM and ZIB) with over a dozen researchers. AstroGrid-D successfully implemented a middleware infrastructure for compute or data intensive research, with several specifically developed elements.

Attach:Geo600.png Δ In this first phase of grid implementation scientific results served mostly as example use cases to test the development with research software. Different small and mid-sized applications have successfully been implemented, such as a connector for robotic telescopes or a grid interface to the Planck Satellite workflow engine. Due to substantial BMBF investment in hardware, the D-Grid also made massive amounts of compute resources available for scientific calculations. Especially the Geo600 use case puts these ressources to extensive use. The application analyses the signals of gravitational wave detectors. Using Grid techniques Geo600 can allocate about 100,000 CPU hours per day.

After the end of the project funding the AstroGrid-D infrastructure is currently maintained by the participating partners. Efforts are now focused on applying the existing developments to new scientific projects. Also the sustainability of the infrastructure needs further development. For more details about AstroGrid-D, see the website http://www.astrogrid-d.org.


WissGrid

"Grid für die Wissenschaft - WissGrid" has only just started, in May 2009, and will be funded until 2012. WissGrid is run jointly by the institutions from the former scientific grid-projects: AstroGrid-D, C3-Grid (Climate-Comunity), HEP Grid (High Energy Physics), TextGrid (Humanities) and MediGrid. The goal is to further improve the use of the Grid in the academic realm. WissGrid will achieve this by forming expert teams to support new academic communities that make use of D-Grid infrastructure. It will address obstacles for Grid usage with funding agencies or with resource providers. And it will address issues of persistent access to and long time archival of scientific data.


Future projects

Several initiatives in Germany are currently ongoing for other e-Science projects. They range from small scale research (e.g. there are several DFG-proposals that explicitly include e-Science infrastructure) to the European scale. The E-Science section is also becoming a part of the research community, if sometimes as part of IT. With the continuous increase of data and the rising acceptance of new the tools, e-Science is likely to become a common component of mid-level research and large scale projects.


What is e-Science?

With the experience of two large-scale research activities in the field, this question is still not conclusively answered. Development is still on-going. Some of the visions of e-Science could not (yet) be realised, e.g. the idea of a vastly simplified general access to national compute resources, which still is not achieved due to security constraints, legal restrictions and the complexity of the middleware.

It is part of the scientific process to openly discuss the coming steps and re-define targets whenever necessary. The feedback of the Astronomical community must be constantly heard and included in the process, and to encourage further active participation is a perpetual task.

We want to contribute to these goals with the upcoming Splinter meeting during the conference of the German Astronomical Society in September 2009 in Potsdam. We would like to invite all interested colleagues to participate with comments, criticism, and recommendations.