The DEISA project – Network operation and support – first experiences
Niederberger Ralph, Mextorf Olaf
Research Centre Jülich, D -52425 Jülich, Germany
e-mail: {R.Niederberger/O.Mextorf}@fz-juelich.de
DOI: 10.12921/cmst.2005.11.02.119-128
OAI: oai:lib.psnc.pl:587
Abstract:
The principal objective of DEISA1, 2, a consortium of major national supercomputing centres in Europe, is to advance computational science in leading scientific and industrial disciplines by deploying an innovative Grid-empowered infrastructure to enhance and reinforce High-performance Computing in Europe. The infrastructure, based on the tight coupling of homogeneous national supercomputers, provides a distributed supercomputing platform operating in multi-cluster mode. The production capability and capacity of the distributed platform will provide a substantial European added-value to the existing national infrastructures and will, in turn, be integrated into a larger heterogeneous Grid. The DEISA service activity “Network operation and support” defines, deploys and operates the network interconnect required by this infrastructure during the project lifetime and beyond. The paper will give an overview about the activities performed in this service activity, the experiences that have been gathered and the future tasks which have to be performed.
Key words:
European distributed multi-cluster supercomputer, GEANT IP Premium service, Grid infrastructure, virtually dedicated bandwidth network interconnect
References:
[1] The Japanese Earth Simulator, http://www.es.jamstec.go.jp/esc/eng/index.html
[2] TOP 500 supercomputer sites, http://www.top500.org
[3] Distributed European Infrastructure for Supercomputer Applications, http://www.deisa.org
[4] FP6 European Research Infrastructures, http://www.cordis.lu/infrastructures/home.html
[5] DEISA Extreme Computing Initiative, http://www.deisa.org/grid/initiative.php
[6] HPC-Europa – Pan-European Research Infrastructure on High Performance Computing, http://www.hpc-europa.org/
[7] The Enabling Grids for E-sciencE (EGEE) project, http://public.eu-egee.org/
[8] TeraGrid, homepage, http://www.teragrid.org
[9] VIOLA – Vertically Integrated Optical Testbed for Large Applications, http://www.viola-testbed.de
[10] Mupbed – Multi-Partner European Test Beds for Research Networking, http://www.ist-mupbed.org
[11] The Globus alliance, http://www.globus.org/
[12] D.Erwin Ed., Unicore Plus Final Report, Unicore Interface to Computing Resources, ISBN 300-011592-7, 2003
[13] GPFS: A Shared-Disk File System for Large Computing Clusters, F. Schmuck, R. Haskin, Proceedings of the Conference on File and Storage Technologies, 28–30 January 2002, Monterey, CA, pp. 231-244., http://www.almaden.ibm.com/StorageSystems/filesystems/GPFS/Fast02.pdf
[14] GÉANT/Dante description of the Premium IP service, http://www.dante.net/server/show/nav.00700a003
[15] DANTE, SEQUIN Project – SErvice QUality across Independently managed Networks,
http://archive.dante.net/sequin/
[16] N. Simar, Performance Monitoring – TF-NGN Meeting, Lissabon, Sept. 2004, http://www.dante.net/upload/pdf/NSTF-NGN-Sept_04.pdf
[17] M. Mathis, R. Reddy, J. Mahdavi, Advanced Network computing – Enabling High Performance Data Transfers, http://www.psc.edu/networking/projects/tcptune/, March 2005
[18] Iperf – Version 2.0.1, The National Laboratory for Applied Network Research (NLANR) Distributed Application support team, http://dast.nlanr.net/Projects/Iperf/
[19] About GÉANT2, http://www.dante.net/server/show/conWebDoc.1202
[20] D. Simeonidou, R. Nejabati (Editors), Optical Network Infrastructure for Grid, GFD-I.036,
http://forge.gridforum.org/projects/ghpn-rg/, August 2004
[21] V. Sander, Networking Issues for Grid Infrastructure, http://www.ggf.org/documents/GWD-I-E/GFD-I.037.pdf, Nov. 2004
[22] V. Sander, Design and Evaluation of a Bandwidth Broker that Provides Network Quality of Service for Grid Applications, NIC Series Volume 16, Forschungszentrum Jülich, ISBN 3-00-010002-4, 2003
[23] M. Fidler, W. Klimala, V. Sander, Path Allocation in Backbone Networks: Project Report,
http://webdoc.sub.gwdg.de/ebook/ah/dfn/PAB.pdf, June 2004
[24] The Large Hadron Collider Project, http://lhc.cern/lhc/general/gen_info.htm
The principal objective of DEISA1, 2, a consortium of major national supercomputing centres in Europe, is to advance computational science in leading scientific and industrial disciplines by deploying an innovative Grid-empowered infrastructure to enhance and reinforce High-performance Computing in Europe. The infrastructure, based on the tight coupling of homogeneous national supercomputers, provides a distributed supercomputing platform operating in multi-cluster mode. The production capability and capacity of the distributed platform will provide a substantial European added-value to the existing national infrastructures and will, in turn, be integrated into a larger heterogeneous Grid. The DEISA service activity “Network operation and support” defines, deploys and operates the network interconnect required by this infrastructure during the project lifetime and beyond. The paper will give an overview about the activities performed in this service activity, the experiences that have been gathered and the future tasks which have to be performed.
Key words:
European distributed multi-cluster supercomputer, GEANT IP Premium service, Grid infrastructure, virtually dedicated bandwidth network interconnect
References:
[1] The Japanese Earth Simulator, http://www.es.jamstec.go.jp/esc/eng/index.html
[2] TOP 500 supercomputer sites, http://www.top500.org
[3] Distributed European Infrastructure for Supercomputer Applications, http://www.deisa.org
[4] FP6 European Research Infrastructures, http://www.cordis.lu/infrastructures/home.html
[5] DEISA Extreme Computing Initiative, http://www.deisa.org/grid/initiative.php
[6] HPC-Europa – Pan-European Research Infrastructure on High Performance Computing, http://www.hpc-europa.org/
[7] The Enabling Grids for E-sciencE (EGEE) project, http://public.eu-egee.org/
[8] TeraGrid, homepage, http://www.teragrid.org
[9] VIOLA – Vertically Integrated Optical Testbed for Large Applications, http://www.viola-testbed.de
[10] Mupbed – Multi-Partner European Test Beds for Research Networking, http://www.ist-mupbed.org
[11] The Globus alliance, http://www.globus.org/
[12] D.Erwin Ed., Unicore Plus Final Report, Unicore Interface to Computing Resources, ISBN 300-011592-7, 2003
[13] GPFS: A Shared-Disk File System for Large Computing Clusters, F. Schmuck, R. Haskin, Proceedings of the Conference on File and Storage Technologies, 28–30 January 2002, Monterey, CA, pp. 231-244., http://www.almaden.ibm.com/StorageSystems/filesystems/GPFS/Fast02.pdf
[14] GÉANT/Dante description of the Premium IP service, http://www.dante.net/server/show/nav.00700a003
[15] DANTE, SEQUIN Project – SErvice QUality across Independently managed Networks,
http://archive.dante.net/sequin/
[16] N. Simar, Performance Monitoring – TF-NGN Meeting, Lissabon, Sept. 2004, http://www.dante.net/upload/pdf/NSTF-NGN-Sept_04.pdf
[17] M. Mathis, R. Reddy, J. Mahdavi, Advanced Network computing – Enabling High Performance Data Transfers, http://www.psc.edu/networking/projects/tcptune/, March 2005
[18] Iperf – Version 2.0.1, The National Laboratory for Applied Network Research (NLANR) Distributed Application support team, http://dast.nlanr.net/Projects/Iperf/
[19] About GÉANT2, http://www.dante.net/server/show/conWebDoc.1202
[20] D. Simeonidou, R. Nejabati (Editors), Optical Network Infrastructure for Grid, GFD-I.036,
http://forge.gridforum.org/projects/ghpn-rg/, August 2004
[21] V. Sander, Networking Issues for Grid Infrastructure, http://www.ggf.org/documents/GWD-I-E/GFD-I.037.pdf, Nov. 2004
[22] V. Sander, Design and Evaluation of a Bandwidth Broker that Provides Network Quality of Service for Grid Applications, NIC Series Volume 16, Forschungszentrum Jülich, ISBN 3-00-010002-4, 2003
[23] M. Fidler, W. Klimala, V. Sander, Path Allocation in Backbone Networks: Project Report,
http://webdoc.sub.gwdg.de/ebook/ah/dfn/PAB.pdf, June 2004
[24] The Large Hadron Collider Project, http://lhc.cern/lhc/general/gen_info.htm
