 |
Petascale Data Storage
|
bof abstract
The Petascale Data Storage Institute is a DOE-funded collaboration of three universities and five national labs with the objective of anticipating the challenges of data storage for computing systems operating in the peta-operations per second to exa-operations per second and working toward the resolution of these challenges in the community as a whole. An important part of our agenda is outreach to other researchers and practitioners to share our resources and gather better understanding of the petascale issues ahead from all.
In this BOF we will:
1) Introduce the Petascale Data Storage Institute (PDSI),
2) Advertise PDSI gathered and released sources of useful data, including
- data sets of node and storage failures in large scale computing
- file access traces of non-trivial petascale computing applications
- collections of file systems statistics gathered from petascale computing systems
and other systems,
3) Discuss requirements for one or more petascale data storage systems and applications, and
4) Lead an open discussion of these and other issues for large scale data storage systems.
Organizer: Garth Gibson, Carnegie Mellon University and Panasas
Co-organizers: Peter Honeyman, U. Michigan/CITI; Darrell Long, U.C. Santa Cruz; Gary Grider, Los Alamos NL; Lee Ward, Sandia NL; Evan Felix, Pacific Northwestern NL; Phil Roth, Oak Ridge NL; Bill Kramer, Lawrence Berkeley NL
PRESENTATIONS
PDSI FAST 2008 BOF Introduction - Garth Gibson, CMU
PDF
The Computer Failure Data Repository (CFDR) -
Bianca Schroeder,
University of Toronto
PDF [141K]
File System Statistics -
Shobhit Dayal, CMU,
Garth Gibson, CMU,
Marc Unangst, Panasas
PDF [104K]
PNNL – Petascale Data
Storage Institute
Data release Update -
Evan Felix, PNNL
PDF [580K]
NERSC Reliability Data -
Bill Kramer,
Jason Hick,
Akbar Mokhtarani, NERSC
PDF [195K]
LANL SciDAC Petascale Data Storage Institute Operational Data Releases -
James Nunez,
Gary Grider, John Bent,
HB Chen, Meghan Quist, Alfred Torrez,
Los Alamos National Lab
PDF [213K]
Ceph: An Open-Source
Petabyte-Scale File System -
Ethan Miller, Storage Systems Research Center,
UCSanta Cruz
PDF [565K]
Special Presentation on HPC User Requirements:
I/O Requirements for HPC Applications:
A User Perspective
John Shalf, National Energy Research
Scientific Computing Center
(NERSC), LBNL
PDF [15.4M]
PDSI POSTER AT THE FAST '08 POSTER SESSION
PDSI Data Releases and Repositories
PDF [7.05M]