Virginia Tech will present advances in supercomputing and networking at the Supercomputing 2005 (SC|05) Conference in Seattle Nov. 12-18 at the Washington State Convention and Trade Center. In particular, the university will partner it's Apple-based supercomputer, System X, with other supercomputers to respond to a challenge involving high performance computing (HPC) storage capability.
Under the theme, "Gateway to Discovery," SC|05 will showcase how high performance computing, networking, and storage lead to advances in research, education and commerce. After a successful debut at SC2004, the StorCloud challenge will continue as a special initiative for SC|05 to build HPC storage capability showcasing HPC storage technologies and applications.
Virginia Tech's System X supercomputer will be combined with cluster computing resources at Los Alamos National Labs and the University of Utah as part of a StorCloud Challenge at SC|05. During the conference, the combined computers, collectively dubbed the "GreenGene Distributed Supercomputer," will attempt a sequencing operation for the GenBank NT Database. This the genetic nucleotide database maintained by the National Institutes of Health (http://www.ncbi.nlm.nih.gov) as a resource for scientists. Jeff Crowder, program director for strategic initiatives and information technology at Virginia Tech, explained that researchers can compare discovered sequences against the NIH database to learn whether there are similarities with sequences stored there.
The team proposes to compare the largest query available, which is the entire NT database, against the largest database available, also the entire NT database. "A problem of this size would be difficult for BLAST, the tool researchers typically use to search the database,” said Crowder. “mpiBLAST segments the database and distributes it across multiple cluster nodes, permitting BLAST queries to be processed on many nodes simultaneously. In this case, the query will be running on dissimilar clusters tied together over great distances. It’s an ambitious attempt and it will be a real feat if we can pull it off."
System X will be connected to Seattle and to the other computing resources using the National LambdaRail (NLR) and the VORTEX Virginia optical network. Virginia Tech's Network Infrastructure and Services group is leading VORTEX in cooperation with other Virginia universities and Verizon. "A new 10 Gbps link connecting Blacksburg to the LambdaRail node in McLean, Va. was lit in September," Crowder said. "Virginia Tech will operate a major hub node in McLean to connect VORTEX and other networks in the region to the National LambdaRail, Internet2, and other research networks."
From McLean, System X will reach Seattle over LambdaRail's Layer3 network. “The new VORTEX link augments the university's production connection to NetworkVirginia and increases Virginia Tech's off campus capacity 1,700 percent,” Crowder said.
Virginia Tech is the first university in the mid-Atlantic region to connect to National LambdaRail.
If the challenge project succeeds, the result of this comparison will provide critical information to the biology community, including insightful evolutionary, structural, and functional relationships between every sequence and family in the NT database. The experiment will generate an estimated 100 terabytes of output to StorCloud. For an abstract of the challenge initiative, click here.
Virginia Tech's exhibit will also feature a demonstration of DIVERSE (Device Independent Virtual Environment: Reconfigurable, Scalable, and Extensible), an open source, virtual environment software developed at Virginia Tech to transport research to the CAVE.
Virginia Tech's booth number is 470.
Learn more about the StorCloud Challenge at http://www.vtksolutions.com/StorCloud/2005/index.html.
Learn more about the conference at http://sc05.supercomputing.org/.
Learn more about System X at http://www.tcf.vt.edu/systemX.html.
Learn more about National Lambda Rail at http://www.nlr.net/members.html.