Application Projects/Toolkits that will be deployed over EMERGE
Prototyping an Earth System Grid
Particle Physics Grid
Prototyping a Combustion Corridor
Visualization Environment for ASCI/SSI
ASCI Data and Visualization Corridors
|Prototyping an Earth System Grid
Peter Beckman1, Ian Foster2, Steve Hammond3, Carl Kesselman4, Bob Lucas5, Bob Malone1, John Michalakes2, Don Middleton3, Jerry Potter6, Brian Tierney5, Dean Williams6
1Los Alamos National Laboratory, 2Argonne National Laboratory, 3National Center for Atmospheric Research, 4USC Information Sciences Institute, 5Lawrence Berkeley National Laboratory, 6Lawrence Livermore National Laboratory
The DoE's Accelerated Climate Prediction Initiative (ACPI) envisions the creation of a national climate modeling system in which the needs of researchers of global-change-impacts are served by a network of regional centers that provide reliable and rapid access to a range of climate data products derived from the results of very high-resolution, long-duration simulations performed at one or more climate modeling centers. Central to this vision is the creation of the Global Systems Grid that will allow end-user requests for derived data products to be translated efficiently and effectively into some combination of accesses to data caches at regional center(s) (ANL, LBNL, NCAR), the execution of downscaling simulations, requests to central data archives (LANL in the first prototype), and/or new large-scale simulations.
Particle Physics Data Grid
Harvey B. Newman1, Richard P. Mount8, Julian J. Bunn1, Les Cottrell8, Ian Foster4, Bruce Gibbard6, Andrew B. Hanushevsky8, Matthias Kasemann7, Miron Livny3, Stewart C. Loken5, David Millsom8, Reagan Moore2, James C.T. Pool1, Lawrence Price4, Arie Shoshani5, Margaret Simmons2, Victoria White7, Roy Williams1
1California Institute of Technology, 2San Diego Supercomputer Center, 3University of Wisconsin, 4Argonne National Laboratory, 5Berkeley Laboratory, 6Brookhaven National Laboratory, 7Fermi National Accelerator Laboratory, 8Stanford Linear Accelerator Center
The DoE NGI proposal "The Particle Physics Data Grid" proposes to create a large-scale, distributed prototype data analysis grid toolkit for the next generation of particle physics experiments, such as the Large Hadron Collider (LHC) at CERN. This data grid is intended to support both bulk data transfers between a central site and several satellite sites and on-demand data requests from end user applications used by a client base of several thousand physicists. Client applications will generate requests for physics event data; the data grid will then access relevant datasets either in local caches or at remote sites. Data movement will be handled by sophisticated middleware with end-to-end scheduling used to provide performance guarantees and caching used to reduce overall network load. DiffServ is to be exploited in the prototype and will be vital for a production system.
The Particle Physics Data Grid project will rely on EMERGE as a testbed environment for end-to-end scheduling techniques. MREN connects a number of Particle Physics Data Grid partners (ANL, FNAL, UWisconsin) and provides a realistic testbed for future particle physics data grids which must necessarily extend beyond DoE networks to reach university collaborators. The opportunity provided by EMERGE for large-scale testing in a heterogeneous multi-network environment will be essential to the networking goals of this Data Grid project.
Prototyping a Combustion Corridor
Robert F. Lucas1, John Bell1, Ian Foster2, Joe Grcar3, Nancy Johnston1, Arie Shoshani1, Chris Rutland4
1Lawrence Berkeley National Laboratory, 2Argonne National Laboratory, 3Sandia National Laboratory, 4University of Wisconsin
LBNL, ANL, SNL-CA, and U Wisconsin propose to prototype a Corridor for the DoE SSI Combustion research community (CMSI). Envisioned is a Combustion Corridor, enabling remote, collaborative visualization of large combustion data sets. Achieving this vision requires a collaboration of researchers in combustion modeling and networking technologies as well as both DoE and University networking testbeds. This project proposes to bring together leaders in each of these fields to develop and demonstrate that such a Corridor is in fact feasible.
CorridorOne: Integrated Distance Visualization Environment for SSI/ASCI
Rick Stevens1, Ian Foster1, Mike Papka1, Deb Agarwal2, Nancy Johnston2, Bob Lucas2, Jim Ahrens3, John Reynders3, Kai Li4, Jason Leigh5, Tom DeFanti5, Chris Johnson6, Chuck Hanson6
1Argonne National Laboratory, 2Lawrence Berkeley Laboratory, 3Los Alamos National Laboratory, 4Princeton University, 5University of Illinois at Chicago, 6University of Utah
This project proposes to develop and deploy the most advanced integrated distance visualization environment for large-scale scientific visualization and demonstrate it on applications relevant to the DoE SSI and ASCI programs including: applications from the Utah and ANL/Chicago ASCI ASAP Centers (Combustion and Thermonuclear Flashes and LANL benchmark ASCI codes), and SSI applications datasets (Climate Modeling Data from LANL, LBNL, ANL; Fusion Device Modeling from LANL and Princeton; Combustion Modeling Data from LANL and ANL). This will include the development and deployment of network-aware tools and techniques for parallel rendering, deep image based visualization, tele-immersion, large-format multi-projector wall-based displays, fast volume and surface visualization algorithms, and streaming media.
Tools being developed for EMERGE or that will use EMERGE services include:
CAVERNsoft - a toolkit for developing collaborative immersive applications.
QoSIMoto - a tool for visualizing realtime network data.
CIBR View - a tool for collaboratively visualizing volume data. Developed in collaboration with Lawrence Berkeley Labs.
ASCI Data and Visualization Corridors
EVL/LAC, Princeton, Utah with ANL
The DoE/NSF Data and Visualization Corridor (DVC) initiative addresses critical computational, data management, and visualization requirements of urgent federal missions. In particular, the Accelerated Strategic Computing Initiative (ASCI) is pioneering teraops architectures and creating scientific datasets of unprecedented size. The word "corridor" implies the opposite metaphor, a wide path through which data flows freely, and through which users, both local and remote, can readily interact with each other and with massive scientific datasets. There is, at this time, an exciting confluence of new ideas on data handling, compression, telepresence, and scientific visualization. The combination of these ideas, which we refer to as "Data and Visualization Corridors," can raise scientific data understanding to new levels, and can potentially transform the way science is practiced well into the next century.
For this reason, Dr. Gilbert Weigand, leader of DoE's ASCI effort, launched a series of workshops, and backed them up with a commitment to support this line of research and development. [DVC] One often thinks of the input/output subsystems of high-performance computers, as well as the networks and user interfaces, as thin pipes through which data must be forced. Follow up "Corridors at a Distance" workshops focused on remote visualization and telepresence capabilities, metacomputing environments and software architectures, and networking protocols and QoS. These activities will greatly benefit from the EMERGE testbed efforts.