Back to Table of contents

Primeur weekly 2013-11-12

Special

e-Infrastructure Commons should connect the supercomputing, Big Data, networking and HPC islands in Europe ...

Big European research centres define road map to e-Infrastructure Commons ...

Bert's Bike measures Dutch cities' microclimates and other stories from the Dutch eScience conference ...

The Cloud

New publication: Developing Cloud Software - Algorithms, Applications, and Tools ...

Amazon Web Services deploys NVIDIA GRID GPUs ...

New computing model could lead to quicker advancements in medical research ...

Red Hat brings OpenShift Online Platform-as-a-Service to 14 new countries and expands gear size offerings for developers ...

Desktop Grids

BOINC 7.2.28 released to the public ...

EuroFlash

EUDAT and RDA/RDA Europe go to China ...

ESI releases a new version of ProCAST, the leading software solution for casting process simulation ...

SysFera to showcase strength in HPC management and remote visualization at Supercomputing Conference 13 ...

DLR takes new supercomputer into operation ...

New project "Execution Models for Energy-Efficient Computing Systems" has launched ...

Fortissimo project to issue open call for proposal submissions for HPC-Cloud-based application experiments ...

PRACE Summer of HPC Awards Excellence ...

Allinea DDT leaps ahead with support for NVIDIA CUDA 5.5 and CUDA on ARM ...

USFlash

Blue Waters and XSEDE sign collaborative agreement ...

Cray expands data management portfolio with new Tiered Adaptive Storage solution ...

Lustre file system version 2.5 released ...

Los Alamos National Laboratory selects DataDirect Networks' high-performance storage to support petascale collaboration among scientists & researchers ...

Micron's Hybrid Memory Cube earns high praise in next-generation supercomputer ...

Lawrence Livermore, Intel, and Cray produce Big Data machine to serve as catalyst for next-generation HPC clusters ...

SDSC's high-performance computing systems benefit San Diego-area companies ...

Kemal Ebcioglu named recipient of 2013 IEEE Computer Society B. Ramakrishna Rau Award ...

University of New Hampshire celebrates first-in-the-state supercomputing capabilities ...

Innodisk to debut 1M+ IOPS FlexiRemap technology at Supercomputing Conference 2013 ...

Suzuki accelerates the speed of vehicle design and innovation with DataDirect Networks high performance scale-out storage ...

DataCore Software's latest SANsymphony-V release magnifies the scope and sets the new standard for software-defined storage platforms ...

Big European research centres define road map to e-Infrastructure Commons


6 Nov 2013 Vilnius - At the e-IRG workshop in Vilnius, Bob Jones presented the vision of the EIROForum on the e-IRG proposed e-Infrastructure Commons. EIROForum is a collaboration between a number of large European Research Centres, including CERN, ESA, EFDA, and EMBL. EIROForum proposes to form "Research Accelerator Hubs" (ReAcH) that should be networked together. a ReAcH should be, in the vision of EIROForum, a sustainable, inclusive, flexible, integrated, innovative user driven e-Infrastructure. ReAcH can be owned and operated by a mixture of commercial companies and public organisations offering a portfolio of services.

According to Bob Jones, Research Accelerator Hub examples are EMBL-EBI that integrates genome datasets and services and integrates with, for instance, Helix Nebula HPC services.

The ReAcH at Bob Jones' CERN will focus on a data-centric service platform on which more sophisticated services can be developed around physics data. It will use the resources installed by CERN at the Wigner Research Centre for Physics in Budapest, Hungary.

Most European distributed computing infrastructures depend heavily on short term project funding. Bob Jones identifies some challenges here. In addition he notes there is a lack of control. Bodies such as EGI.eu and NGIs do not have ownership of the computing sites they co-ordinate. There are also high operational costs as there are a large number of small-scale sites offering identical services. Project funding also means a lack of interoperability and flexibility, as the project structure means users cannot easily combine and swap services.

Long lived ReAcHes that have a mix of long term funding and pay-per-usage services, that own large scale resources are better positioned, said Bob Jones. He proposes to consolidate distributed computing infrastructure sites into a reduced number of ReAcH with sufficient capacity to host a larger and more diverse portfolio of services. Volunteer computing infrastructures should be integrated into the e- infrastructure Commons. Publicly funded computer sites should be integrated with commercial Cloud services providers in a hybrid platform.

Not all distributed computing on a European scale can be done in the new ReAcHes. What should distributed computing infrastructure sites do that do not become a ReAcH?

Volunteer computing structures offer an avenue by which they can continue to contribute but with reduced operational costs, said Bob Jones. The DEGISCO project and International Desktop Grid Federation are instrumental in organising these communities. Volunteer computing should technically be integrated into the overall e-infrastructure Commons. Here, according to Bob Jones, EDGI developed a bridge between volunteer computing and Grids and Clouds. This route also offers a channel for engaging the general public and citizen scientists.

Bob Jones encourages PRACE supercomputer centres to become Research Accelerator Hubs.

The prototype ReAcH and the business models will be tested in 2014.

More information on EIROForum is available at http://www.eiroforum.org/

The e-IRG website can be visited at http://e-irg.eu

The Desktop Grid Federation portal can be accessed at http://desktopgridfederation.org
Ad Emmen

Back to Table of contents

Primeur weekly 2013-11-12

Special

e-Infrastructure Commons should connect the supercomputing, Big Data, networking and HPC islands in Europe ...

Big European research centres define road map to e-Infrastructure Commons ...

Bert's Bike measures Dutch cities' microclimates and other stories from the Dutch eScience conference ...

The Cloud

New publication: Developing Cloud Software - Algorithms, Applications, and Tools ...

Amazon Web Services deploys NVIDIA GRID GPUs ...

New computing model could lead to quicker advancements in medical research ...

Red Hat brings OpenShift Online Platform-as-a-Service to 14 new countries and expands gear size offerings for developers ...

Desktop Grids

BOINC 7.2.28 released to the public ...

EuroFlash

EUDAT and RDA/RDA Europe go to China ...

ESI releases a new version of ProCAST, the leading software solution for casting process simulation ...

SysFera to showcase strength in HPC management and remote visualization at Supercomputing Conference 13 ...

DLR takes new supercomputer into operation ...

New project "Execution Models for Energy-Efficient Computing Systems" has launched ...

Fortissimo project to issue open call for proposal submissions for HPC-Cloud-based application experiments ...

PRACE Summer of HPC Awards Excellence ...

Allinea DDT leaps ahead with support for NVIDIA CUDA 5.5 and CUDA on ARM ...

USFlash

Blue Waters and XSEDE sign collaborative agreement ...

Cray expands data management portfolio with new Tiered Adaptive Storage solution ...

Lustre file system version 2.5 released ...

Los Alamos National Laboratory selects DataDirect Networks' high-performance storage to support petascale collaboration among scientists & researchers ...

Micron's Hybrid Memory Cube earns high praise in next-generation supercomputer ...

Lawrence Livermore, Intel, and Cray produce Big Data machine to serve as catalyst for next-generation HPC clusters ...

SDSC's high-performance computing systems benefit San Diego-area companies ...

Kemal Ebcioglu named recipient of 2013 IEEE Computer Society B. Ramakrishna Rau Award ...

University of New Hampshire celebrates first-in-the-state supercomputing capabilities ...

Innodisk to debut 1M+ IOPS FlexiRemap technology at Supercomputing Conference 2013 ...

Suzuki accelerates the speed of vehicle design and innovation with DataDirect Networks high performance scale-out storage ...

DataCore Software's latest SANsymphony-V release magnifies the scope and sets the new standard for software-defined storage platforms ...