Section: Partnerships and Cooperations

International Initiatives

• Title: Ulta-scalable I/O and storage for Exascale systems

• Inria principal investigator: Gabriel Antoniu

• International Partners (Institution - Laboratory - Researcher):

  • Argonne National Laboratory (United States) - Mathematics and Computer Science Division - Rob Ross

  • University of Illinois at Urbana Champaign (United States) - Marc Snir

• Duration: 2013 - 2015

• See also: http://www.irisa.fr/kerdata/data-at-exascale/

• Description: as the computational power used by large-scale scientific applications increases, the amount of data manipulated for subsequent analysis increases as well. Rapidly storing this data, protecting it from loss and analyzing it to understand the results are significant challenges, made more difficult by decades of improvements in computation capabilities that have been unmatched in storage. For many applications, the overall performance and scalability becomes clearly driven by the performance of the I/O subsystem. As we anticipate Exascale systems in 2020, there is a growing consensus in the scientific community that revolutionary new approaches are needed in computational science storage. These challenges are at the center of the activities of the Joint Inria-UIUC Lab for Petascale Computing, recently extended to Argonne National Lab. This project gathers researchers from Inria, Argonne National Lab and the University of Illinois at Urbana Champaign to address 3 goals: 1) investigate new storage architectures for Exascale systems; 2) investigate new approaches to the design of I/O middleware for Exascale systems to optimize data processing and visualization, leveraging dedicated I/O cores and I/O forwarding techniques; 3) explore techniques enabling adaptive cloud data services for HPC.

Politehnica University of Bucharest

(since 1 January 2013, just after the end of our former DataCloud@work Associate Team).

Joint Inria-UIUC Lab for Petascale Computing (JLPC),

since 2009. Collaboration on concurrency-optimized I/O for post-Petascale platforms (see details in Section  4.1 ). A joint project proposal with the team of Rob Ross (Argonne National Lab) has been completed in 2012. It served to prepare the creation of the Data@Exascale Associate Team with ANL and UIUC (2013-2015).

FP3C ANR-JST project (2010–2014).

This project co-funded by ANR and by JST (Japan Science and Technology Agency) started in October 2010 for 42 months. It focuses on programming issues for Post-Petascale architectures. In this framework, KerData collaborates with the University of Tsukuba on data management issues.