PUBLICATIONS Archived Press Releases

Preeminent Researchers Using National LambdaRail Dominate the SC06 Bandwidth Challenge

Visionaries use NLR and its International Peering to Capture All Three Awards November 29, 2006 Cypress, CA - National LambdaRail played a key role for the winner and two honorable mention recipients of the annual supercomputing bandwidth challenge at SC06. All three teams worked with NLR to show how real applications could fully utilize one 10 Gigabit path, end-to-end, disk-to-disk, from SC06 in Tampa, Florida back to their home institutions, using the actual production network back home. Each participant was challenged to demonstrate and publish all the configuration, troubleshooting, tuning and policies, not only to show off at SC06, but to leave a legacy at their home institution showing other scientists how to achieve the same results. The winner of the Bandwidth Challenge was “Transporting Sloan Digital Sky Survey Data using SECTOR” by a team from the National Center for Data Mining at the University of Illinois at Chicago. The team was led by Robert Grossman and included Yunhong Gu, Michal Sabala, Shirley Connelly, David Hanley, Joe Mambretti, Alex Szalay, Ani Thakar, Jan VandenBerg and Alainna Wonders. The team achieved 8 gigabits per second (Gbps) of sustained data transfer on a 10 Gbps link and saw a peak transfer rate of 9.18 Gbps. For the Bandwidth Challenge, the National Center for Data Mining transferred astronomy data from the Sloan Digital Sky Survey (SDSS DR5) between the SC06 show floor and the University of Illinois at Chicago campus via StarLight. This was an end-to-end, disk-to-disk data transfer using the NCDM-developed UDT and SECTOR on a 10G link. “Winning this year’s Bandwidth Challenge shows that with the proper software and network protocols, a working scientist can now easily transport terabyte-size e-science data sets from disk to disk over wide-area 10GE networks, such as provided by the National LambdaRail’s PacketNet,” said Robert Grossman, Director of the National Center for Data Mining at the University of Illinois at Chicago and Managing Partner of Open Data Group. The Bandwidth Challenge also presented two honorable mentions. The team from Indiana University was honored for “Spirit of the Competition” and Caltech’s team was recognized for “Heroic Effort.” Indiana University’s effort entitled “All in a Day’s Work”, emphasized the fact that IU’s advanced IT infrastructure provides powerful resources to a diverse research community every day. Led by Stephen Simms, the project featured four IU research teams simultaneously running scientific analyses using Indiana University’s Data Capacitor, a 535 TB Lustre file system used for temporary storage, and a remote Data Capacitor constructed on the exhibit floor. The pair of Data Capacitors were cross mounted to span the distance between IU’s local cyberinfrastructure and the exhibit floor in Tampa. Using modest compute resources on the exhibit floor and IU’s Big Red supercomputer, IU researchers ran a diverse set of applications which captured crystallographic data from multiple distributed sources, mined streams of weather radar data for tornado prediction, analyzed disease-forming protein mutations, and simulated subatomic particle interactions. “A network like NLR’s PacketNet allows IU’s Data Capacitor to play a role in all steps of the data life-cycle, from acquisition or creation, through computation and visualization, to archive storage,” noted Stephen Simms, Data Capacitor Project Manager, Indiana University. “By enabling the Data Capacitor to bridge geographically distributed resources, NLR’s PacketNet will facilitate remote data collection and analysis, open the door to new workflow opportunities, and most importantly empower researchers to achieve their goals faster.” The Caltech high energy physics team used a single 10 Gbps link provided by National LambdaRail that carried data in both directions to achieve a disk-to-disk throughput of 17.77 Gbps between clusters of servers at the show floor and Caltech, using a new TCP-based application called FDT (Fast Data Transport) recently developed at Caltech. The HEP team also carried out several other demonstrations, making good use of the ten wide area network links connected to the Caltech / CERN booth including two National LambdaRail FrameNet links, one to Los Angeles (the official BWC wavelength) and one to StarLight, two NLR PacketNet links used from Korea over GLORIAD and University of Michigan over MiLR. During the test, these network links were shown to operate at full capacity for sustained periods. “Our demonstrations showed that by utilizing network links like NLR’s, the difficulty of transporting Terabyte and larger scale datasets among multiple sites is substantially reduced, thus enabling physicists throughout the world to have a much greater role in the next round of physics discoveries expected in the near future,” noted Harvey Newman, Caltech’s team leader. About National LambdaRail National LambdaRail, Inc. (NLR) is a major initiative of U.S. research universities and private sector technology companies to provide a national scale infrastructure for research and experimentation in networking technologies and applications. NLR puts the control, the power and the promise of experimental network infrastructure in the hands of our nation’s scientists and researchers. Contact: Concordia Chen National LambdaRail editor @ nlr.net 760-510-8406 x5#