DCSC logo  
 
ABOUT-DCSC
DCSC/SDU
DCSC/AU
DCSC/AAU
DCSC/DTU
DCSC/KU
 
+Open all         -Close all
 
 

DCSC Organisation Specifics

The Danish organisational and funding model for providing Scientific Computing (SC), is as of December 2007 as follows:

  • SC for public research is provided by a partnership agreement between five of Denmark’s biggest universities and the Ministry of Science, Technology and Innovation within the setting of the Danish Centre for Scientific Computing (DCSC).

  • DCSC has a total budget of approximately EUR 2.1 million provided by the Ministry, which also provides the statutes.

  • DCSC is governed by a board of ten appointed by the Ministry; five of them are representing the stakeholder universities.

  • DCSC is administrated by a secretariat consisting of a director only, with a total budget of some EUR 200.000.

  • The five universities host and pay for regional operation centres (i.e. buildings, cooling, electricity, staff etc.), estimated at approximately the same as the DCSC budget (I.e. the universities contribute an estimated 50% co-funding).

  • From the remaining part of the DCSC budget (90% = EUR 1.91 million) a 20% overhead charge is deducted by the universities hosting the regional operation centres. The rest (72% = EUR 1.53 million) is spent primarily on SC hardware, determined by the researchers.

  • The DCSC board receives applications based on an annual open call for SC based research projects. The board decides which research teams get SC resources, based on research quality, competition and international peer review recommendations.

  • The hardware requested by researchers must be purchased by, located at, and run by one of five DCSC regional operating centres. The hardware is often added to an existing installation, but resources are dedicated proportionally to research teams holding DCSC grants.

  • 10% of the new hardware’s SC capacity (cycles) will be reserved by DCSC for the so called pool of free resources, used for young scientists wanting to try out SC; for research projects that do not require substantial SC resources, or only for a limited time span; as emergency resources; or for the sudden possibility for "breakthrough SC runs" etc.

  • Researchers can alternatively apply for access to existing hardware belonging to the pool of free resources.

One of the immediate advantages of this model is that it is the users of the research infrastructure, the researchers themselves, who decide which SC resources are best suited for their research aims, and hence what to invest in. Furthermore, the model guarantees that precious funding resources are not invested on various add-ons, deemed of marginal use by the specific researchers, e.g., such as excessive technical support staff; redundant technical utensils; code optimisers for already vender optimised code, etc. The close-to-lab utilisation guarantees that system administrators "know their customers". University support (by means of co-financing DCSC’s grant decisions) guarantees legitimacy both for the investments in SC infrastructure in general, as well as for DCSC specifically.

On the other hand, the apparent weakness of the Danish organisational model is that it presumes a user community that has substantial experience with SC, knows what it wants, and why. While efforts have been made to give guidance and support to newcomers, young researchers and fields of research new to SC, DCSC sees a need to do more to provide SC to such new users. And this task is ever more important as the need for SC permeates virtually all areas of science. These weaknesses are to be dealt with in the period of 2008 - 2010 via the two phase expansion plan (Scientific Computing and Grid Infrastructure for Scientific Calculation, Simulation, and Modelling) submitted to and subsequently financed by the Danish Ministry of Science, Technology and Innovation.