Call for participation to the Nordic Computational Grand Challenge Survey

The four Nordic countries Denmark, Finland, Norway and Sweden are carrying out a survey to determine scientific problems that are of high interest to the research community in the Nordic countries and whose solution requires the availability of large-scale e-Infrastructure. This survey is part of a larger study that also addresses the current status of the largest computational projects and the (need for) resource sharing policies in the Nordic countries.

The survey is commissioned by the Danish Agency for Science, Technology and Innovation, the Finnish IT center for Science, the Research Council of Norway, and the Swedish Research Council.

Introduction

Many of today's large scale computational facilities for academic research are shared resources. The facilities are expensive and are therefore usually being financed by large organizations (e.g., universities) and/or national funding agencies (e.g., Research Councils or Ministries). At any time, the facilities are used by multiple research projects. Several applications from different research groups are executed simultaneously on each facility. Even if the facilities are large in terms of the number of processors, memory, and storage, there are limitations to their capacities.

To distribute the available computational resources over the research groups that wish to use the facilities, national committees have been established that decide on the allocation of the resources (CPU-hours). Usually, the total amount of resources requested, by all research groups together, exceeds by far the total available capacity of the facilities. This means in practice that the research groups only get part of the resources that they applied for, even though they have a real need for a larger amount of resources. Ultimately, this may lead to a degradation of the research carried out, and the researchers may not be able to address challenging scientific questions that require computationally or data intensive computer simulations.

Goals

The main goal of the survey is to identify outstanding problems in science whose solution will greatly benefit from the use of large-scale e-Infrastructure (i.e., infrastructure for computational science that includes large computational resources, storage devices, and high-speed networks).

The aim of the survey is to get an overview of

• important scientific questions that are being studied today with the use of large computational resources but that can be addressed significantly better by having access to much larger computational resources
• important scientific questions that will have an emerging need for large computational resources (in the coming ten years)

With larger resources, we mean resources that are at least a factor of ten larger than the largest national facility available and which is likely to be more than the aggregate capacity of all available facilities on the national level.

Subgoals of the survey are:

• to identify groups in the Nordic countries that address similar (computational) challenges and that will benefit from (strengthened) cooperation and exploiting synergies
• to identify the need for e-Infrastructure that will significantly advance science (and significantly faster)
• to identify scientific problems and applications that justify a Nordic collaboration on e-Infrastructure
• to increase the visibility of Nordic research and Nordic computational science in a European and global context and enable participation in international Grand Challenges

Criteria

In the context of this survey, Grand Challenges have the following characteristics:

• Grand Challenges address problems in science that are widely recognized as important and challenging to solve, and have a wide international interest and relevance. Grand Challenges aim to find answers to fundamental scientific questions.
• Grand Challenges have a demonstrated need for large-scale e-Infrastructure. This can include capacity architectures (e.g., compute clusters), capability architectures (e.g., shared-memory or vector-processor architectures) or grid environments.
• The solution to Grand Challenge problems requires a long-term sustained effort with measurable intermediate goals.

The survey is particularly seeking Grand Challenges that involve novel (groundbreaking) research that will warrant publication in the most prestigious international scientific journals.

Industrial Grand Challenges are not included in the survey, unless they are of considerable interest to the international research community.

Proposal

Research groups are hereby invited to define and submit the Grand Challenge problem(s) that they consider the most important in their field and that should be included in this Nordic survey. The groups are also invited to describe their current scientific activity, how it is related to the proposed Grand Challenge and how the quality of their research will improve with access to large-scale e-Infrastructure.

A proposal consists of three parts:

• A description of the Grand Challenge. A preformatted form (doc) is available and must be used. The project description must not be longer than 2500 words (ca. five pages A4). If you experience problems with this form, you may try one of the following: form 1 (doc), form 2 (txt).
• The CVs of the main applicant and other key people (max four pages per CV).
• A list of key publications not older than five years.

Deadline for submission of the proposal is November 15, 2006 (extended deadline)

The proposal must be sent electronically to the contact person (see below) from the country of residence of the main applicant in the proposal.

Contact persons

Contact persons for the Nordic Computational Grand Challenge Survey within each country are:

• Denmark: Kurt V. Mikkelsen, University of Copenhagen, kmi@theory.ki.ku.dk
• Finland: Jura Tarus, CSC, Espoo, jtarus@csc.fi
• Norway: Jacko Koster, UNINETT Sigma AS, Trondheim, jacko.koster@uninett.no
• Sweden: Sinisa Krajnovic, Chalmers University of Technology, Göteborg, sinisa@chalmers.se

Contact any of these people for additional information.

Other information

The proposals that are received within the deadline will be bundled and presented to the national funding agencies and to the management of the national infrastructures for computational science. An evaluation process will be carried out to select the proposals that (best) satisfy the criteria for Grand Challenges. Incomplete proposals will not be included in the final material.

A seminar is planned in February 2007 to which a number of applicants will be invited to present their proposal.

The aim of the survey is to identify Grand Challenge that are of interest to the research community in the Nordic countries and establish the need for an appropriate (Nordic) e-Infrastructure. As of today, there is no approved funding for such infrastructure. The results of the survey will provide a basis to justify such funding.

The Nordic Data Grid Facility (NDGF) has decided to select a limited number of scientific applications from Nordic research groups for implementation in the NDGF GRID environment. The NDGF Steering Body wishes to use the Grand Challenge proposals as a basis for this selection.

More information on the survey will be provided on these pages in due time.

Last updated: 2006-10-24