Technology demonstrations provide an opportunity to make applications development visible to the general public while driving progress by establishing specific deadlines for operational capabilities and calling on resources-such as OC-48 network connectivity-that might not otherwise be available. They enable developers and scientists to see what kinds of tools, services, and innovations may be useful in their applications discipline, foster new interactions and collaborations among researchers, and allow the public to participate in IT R&D research advances.

Federal agency IT R&D demonstrations are held several times a year to reach different audiences. In FY 1999-2000, an IT R&D Expo was held on Capitol Hill to foster greater understanding among legislators and their staffs of the progress on applications development. IT applications demonstrations, including NGI applications, were showcased at SC99, the national high performance networking and computing conference, to expand awareness in the science and commercial communities of Federally supported R&D efforts in high performance computing and networking. And IT R&D agencies were invited to conduct applications demonstrations in Asheville, North Carolina, in December 1999. Select applications demonstrations from these three events are highlighted in this section.

SC99 demos included:

DARPA's SuperNet (part of NGI, page 87) was a critical elemnt in setting the 2.4 Gbps packet-over-Sonet world land speed record. The network infrastructure for the long-distance, high-speed HDTV demonstration linked DARPA's National Transparent Optical Network (NTON-II), the University of Washington (UW)-led Pacific/Northwest GigaPop (P/NWGP), and Nortel Networks. Together, the neworks delivered unprecedented levels of standard Internet capacity from the Microsoft Corporation and UW campuses, through the P/NWGP in Seattle, and on to the exhibition hall. There, Microsoft, the National Computational Science Alliance, Sony (in support of the ResearchTV consortium), and UW concurrently demonstrated two realtime gigabit applications, setting a record of more than 2 Gbps in aggregate throughput--by a wide margin the fastest real-time applications yet run over a WAN.
 
The demonstration was the equivalent of simultaneously transmitting the entire channel lineup of a 150-channel cable TV system, or 50 channels of broadcast quality HDTV, or five feature movies, or interactions among a large number of shared virtual realities.

DARPA
Microsoft Corporation
National Computational Science Alliance
Nortel Networks
Pacific/Northwest GigaPop
ResearchTV
Sony
University of Washington

NASA's VLAB is a virtual reality environment enabling researchers working at desktop computers anywhere in the U.S. to monitor and actively participate in real-time experiments and training simulations at the Ames Research Center Flight Simulation Laboratories. The emphasis is on the perspective of the user, who can define the specific data and display configuration for an experiment, and multiple "players" at different sites can each customize their displays for the same simulation. For example, the VLAB interface lets engineers at the Johnson Space Center, Boeing, Rockwell, and Lockheed Martin access tests at Ames' Vertical Motion Simulator, where researchers conduct astronaut training and study the physics and engineering requirements of rockets and space vehicles such as the Space Shuttle. New capabilities being developed for the VLAB system will allow researchers to create math models, displays, and control systems, validate models for higher-quality experiments, and do virtual prototyping for cockpit design and lab data system layouts. NASA researchers have developed a prototype Mars VLAB client and have deployed a full VLAB client at Johnson Space Center to support shuttle entry simulation.

VLAB concepts are broadly applicable to remote virtual control rooms such as those associated with wind tunnels, flight test facilities, and multiple inter-operable laboratories. The NGI's high bandwidth and multicast technologies make it possible to provide the digital audio, MPEG-2, audio/video-conferencing, whiteboarding, and client/server support, as well as the real-time interactivity, needed by VLAB. The QoS afforded by the NGI ensures that visualization streams reach their destinations on time when networks are congested.

NASA Ames Flight Simulation Laboratory
NASA Johnson Space Center
NASA Research and Education Network (NREN)
Apple Computer, Inc.
Lockheed Martin Corporation
Rockwell International Corporation
The Boeing Company

NASA

DARPA's SuperNet is the set of interconnected NGI testbeds dedicated to prototyping the engineering, hardware, software, and connectivity of future networks capable of 1,000 times current Internet transmission speeds. Over the first two years of the NGI program, Federal, academic, and industry research partners developed and put in place the pioneering components necessary for a wholly new network infrastructure with capabilities far beyond those of today's Internet. At SC99, these partnerships reached a major developmental milestone, demonstrating for the first time full coast-to-coast interconnectivity across the network of SuperNet testbeds. The following components of SuperNet will be operational in FY 2000:

• NTON-II, DARPA's high-speed optical network on the West Coast, with four wavelengths each with 10 Gbps capacity
• High Speed Connectivity Consortium (HSCC), connecting Los Angeles and Washington, D.C., sites at 2.5 Gbps end to end
• ONRAMP, a Boston-area testbed operating at OC-48+ and fielding advanced metropolitan area and regional access technologies
• BOSSNET, a highly experimental WDM network developed by DARPA enabling physical layer networking and communications experiments over dark fiber, with connectivity from the Boston area to Washington, D.C., at OC-48+ speeds
• Advanced Technology Demonstration Network/Multiwavelength Optical Networking (ATDNet/MONET), a 20 Gbps dynamically reconfigurable double ring network
• Collaborative Advanced Interagency Research Network (CAIRN), a coast-to-coast testbed for networking research with an infrastructure that is heterogeneous in hardware, interconnections, and speed

More than 40 research institutions use SuperNet components for network and applications research in areas such as HDTV, border gateway multicast protocol, reliable multicast, domain name system (DNS) security, DWDM switching, wide area gigabit ethernet (GbE), and active networks. Robust, scalable versions of these technologies will be needed to build out the NGI prototype networks into a full-scale, very high-speed infrastructure for research, commerce, and communication.

Demonstrating DARPA's SuperNet at SC99. More than 40 research institutions use SuperNet in applications areas such as HDTV, reliable multicast, domain name system (DNS) security, wide area gigabit Ethernet (GbE), and active networks.

Carnegie Mellon University
Corporation for National Research Initiatives
Defense Information Systems Agency
Defense Intelligence Agency
GST
Laboratory for Telecommunication Sciences
Lawrence Livermore National Laboratory
MIT Lincoln Laboratory
NASA Goddard Space Flight Center
Naval Research Laboratory
Nortel
Qwest
USC Information Sciences Institute

Darpa

Dynamic construction
of virtual overlay
networks interconnecting
NGI testbeds

The nationwide SuperNet interconnectivity demonstrated at SC99 was made possible by a new technology called multiprotocol label switching (MPLS) and a publicly available prototype MPLS research platform created by NIST called NISTSwitch. MPLS dynamically constructs virtual overlay networks that enable QoS-controlled data transmission from network to network. Virtual overlay networks are possible only with wide-bandwidth network infrastructures that permit WDM. In systems with WDM such as the SuperNet testbeds, the NISTSwitch platform operates as a kind of rapid-transit service hub, setting common communications and routing protocols for all carriers on all routes and working to maintain performance standards across the system as a whole.

MPLS will enable NGI researchers to expand their experimentation with virtual overlay networks for QoS, multicast, and security assurance on NGI testbeds. MPLS also holds promise in signaling and routing technology as a means of optimizing traffic engineering and constraint-based path selection in NGI networks.

NIST
University of Southern California/Information Sciences Institute (USC/ISI)
NASA Goddard Space Flight Center
NASA Ames Research Center
DARPA

The ImmersaDesk is a projection platform that "immerses" the user in a virtual reality environment using stereo glasses. Wearing the special glasses, a user can look into the ImmersaDesk's 4'x5' angled screen to view a computer-generated 3-D image simulated with near-perfect accuracy from complex statistical data. At the SC99 demo, visitors witnessed a tsunami approaching and crashing onto a coastal Japanese city, with realistic rendering of the wave, sea floor bathymetry, land topography, and the city itself. Using the device's interactive, animated, 3-D stereographic visualization capabilities, NOAA scientists are modeling other environmental influences such as the activity of hydrothermal vents, the effects of oceanographic forces on fisheries, and the characteristics of regional weather systems. The 3-D visualization of complex data sets reveals subtleties that scientists cannot see in 2-D graphics. For example, a 2-D graphic of an Alaskan fur seal's hunt for food shows only the seal's linear route through the water. A 3-D animation of the same data reveals that the seal makes numerous sharp upward and downward movements through the water along the way.

NGI networks interconnecting heterogeneous computing resources and large data archives via a shared interface allow geographically distributed climate and oceanographic researchers to work collaboratively in such 3-D virtual environments.

NOAA's Pacific Marine Environmental Laboratory (PMEL)
Old Dominion University

NOAA's HPCC program

OceanShare

When a major oceanic oil spill occurred-before the advent of high performance computational and networking technologies-scientists routinely spent days and perhaps weeks gathering and analyzing the information needed to guide damage control and mitigation efforts. NOAA's OceanShare is a powerful new collaborative tool that enables researchers at NOAA laboratories to quickly access and work with large-scale oceanographic and meteorological data sets from multiple geographically distributed archives. This means that scientific analyses requiring heterogeneous information-weather conditions, ocean temperatures and currents, topographical data, and marine ecosystem data-from a variety of sources can be rapidly integrated and analyzed collaboratively by researchers, no matter where they are located.

OceanShare's data-intensive and bandwidth-intensive capabilities for collaborative research require the speed and QoS characteristics of the NGI, gu aranteeing rapid response rates, a sophisticated shared interface, and uninterrupted flows of large-scale data. OceanShare combines interactive Java graphics, Java remote method invocation/common object request broker architecture (CORBA) network connections, and NCSA's object-oriented Habanero developer's framework to create collaborative networked access to distributed data sets. The new research tool enables NOAA/PMEL's Fisheries Oceanography program, in collaboration with NOAA's Alaska Fisheries Science Center and the University of Alaska, to study fish migrations, environmental conditions affecting fisheries, and related oceanographic influences on fish populations. NOAA's Hazardous Materials Response and Assessment Division (HAZMAT) will use OceanShare to assess response needs when hazardous materials are released into the environment. The new tool will also provide networked access to climate data archived at NOAA's PMEL, Atlantic Ocean Meteorology Laboratory, and National Ocean Data Center, as well as at the University of Hawaii Sea Level Center and other research sites.

NOAA's PMEL

NOAA's HPCC program
NOAA's Environmental Services Data and Information Management

 

A virtual tour of 3-D oceanographic data
sets using Virtual
Reality Modeling
Language (VRML)

From the TAO (Tropical Atmosphere-Ocean) Array-a group of some 70 moored buoys spanning the equatorial Pacific-NOAA scientists gather oceanographic and surface meteorological variables critical to improved detection, understanding, and prediction of the significant climate variations such as El Ni–o and La Ni–a that originate in the tropics. Today, using VRML software, scientists can turn the large-scale TAO data collected around the clock into 3-D objects that precisely mimic their real-world counterparts. These objects can be rotated and modified; the data display can be reorganized to show contour slices, color-coded poly-filled contours, surfaces, vectors, bathymetry, and topography. With the addition of Java scripts, these objects can be combined and animated in a 3-D VRML world that the user can move through and manipulate.

Researchers are able to "see," for example, what happens to ocean temperatures over time and at varying depths under El Ni–o and La Ni–a conditions. Scientists in NOAA's Carbon Exchange program use VRML to better understand another fundamental climatic influence-the fluxes in the vital transfer of carbon dioxide between the ocean and the atmosphere and their relationship to ocean temperatures. The tool enables the Fisheries Oceanography Combined Investigations (FOCI) program to examine the complex biological and physical oceanographic environments of fisheries.

VRML is a low-cost, Web-accessible application that can be widely used in education and research environments to model and interact with complex data in 3-D and stereographic 3-D. Broader deployment of these capabilities, however, will require the consistently high levels of reliability, QoS, and network speed now being prototyped in the NGI.

NOAA's PMEL
Old Dominion University

NOAA

 

Computational challenges
in climate and weather
research

NOAA scientists increasingly use high end scalable computing systems and component-based software to improve the climate research and weather prediction needed by Government officials and policymakers, industries such as agriculture, transportation, and re-insurance, water managers and public health officials, academic researchers, and the general public. Advanced IT capabilities have improved scientists' ability to predict hurricanes and other severe weather events; to examine long-term influences on environment, such as the effect on hurricanes of increased carbon dioxide (CO₂) in the atmosphere and the ocean thermal response to climate warming trends; and to understand macro-features of the earth's biosphere, such as eddies in the North Atlantic and atmospheric mixing in the Sourthern hemisphere.

Improved network speed and bandwidth will enable rsearchers to use increasingly powerful scalable systems and software designs to work collaboratively with more detailed and complex statistical models shared over high-speed networks.

NOAA

 

NIH biomedical
collaboratory testbed

Researchers in structural biology and computational pathology working at geographically dispersed institutions are using collaboratory tools--including electronic notebooks, source code repositories, and data-sharing and teleconferencing technologies--to examine and evaluate the effectiveness of these tools to improve their research. One group of collaborators is studying the biological functions of proteins using AMBER and CHARMM simulation software that depicts the structure and movement of proteins and DNA molecules. A computational pathology group is developing automated techniques to improve the speed and accuracy of tissue diagnosis. In collaboration with NCI researchers, they are working on 3-D visualization of prostate tissue samples and using DNA and RNA sequencing to correlate tissue characteristics with genetic tendencies.

Multicast technology and uniform routing protocols on the NGI will improve the high end forms of research collaboration including audio, data, and teleconferencing applications. Faster and more reliable connectivity among collaborating sites will make it possible to use and exchange large data files, 3-D volume-rendering graphics, visualization files, and source code.

NIH's National Cancer Institute
Pittsburgh Supercomputing Center
Scripps Research Institute
University of Pittsburgh Medical Center
University of Pittsburgh Center for BioMedical Informatics
University of California at San Francisco

NIH's National Center for Research Resources (NCRR)

 

BioCoRE and
interactive molecular
dynamics (IMD)

Using this portable collaboratory software for structural biology, researchers at different sites can start, visualize, and interactively steer molecular dynamics simulations on remote high performance computing platforms. BioCoRE, which stands for Biological Collaborative Research Environment, is a network-centered meta-application that has four components: a workbench that provides analysis tools, data sharing, resource allocation, simulation control, and interactive molecular dynamics; conferencing that includes audio and visual communication, visualization, and training; a notebook for record keeping; and a documentation capability. In the BioCoRE environment, researchers use its IMD molecular visualization program and simulation engine to display 3-D models of molecular systems such as proteins, nucleic acids, and lipid bilayer assemblies. In addition, a force feedback tool enables researchers to, in effect, poke at a system or change its shape to see how it responds. The application currently runs over NSF's vBNS network. The full package of BioCoRE tools is scheduled to be available to researchers in 2000.

BioCoRE's data-heavy and bandwidth-heavy applications require high-speed data transmission over networks and the high end storage capacity and operating speeds of supercomputers. The NGI will provide low latency to respond to user input and progressively higher bandwidths needed for increasingly complex simulations. The NGI will also enable transparent use of available distributed hardware, software, and databases.

NIH
Beckman Institute for Advanced Science and Technology
University of Illinois at Urbana-Champaign

NIH's NCRR

 

Terabyte Challenge
2000: Project Data
Space

This collaboration among IT research institutions around the world is designed to establish standards, new network protocols and services, and performance monitoring tools for distributed data analysis and data mining across high performance networks. "Mining" is the work a computing system does to find and organize relevant data. The larger and more widespread the archives of potentially useful data, the more processing is required. The vast quantities of information stored on computers around the world make the capacity to mine it one of the most important technical computing challenges of this era.

Terabyte Challenge 2000 participants have developed a distributed testbed and knowledge network for experimentation and disciplinary studies in managing, mining, and modeling large (too large to fit in the memory of a single workstation), massive (large and distributed on both tape and disk), and geographically distributed data sets. The testbed, comprising clusters of workstations connected with a mix of traditional and high-speed networks, uses several software infrastructures--including PAPYRUS, a data-mining and predictive modeling system developed by the National Center for Data Mining at the University of Illinois at Chicago--to aim for the processing speed and data validity benchmarks required for high performance networked applications. The wide area data transfers called for in distributed data mining become practical only with the faster speed available on the NGI.

Terabyte Challenge members, whose demonstration at SC98 won the Most Innovative of Show award in the High Performance Computing Challenge category, are currently working with testbed applications that include:

• Anomaly prediction in brain scan data
• Digital Sky Survey: a classification of cosmological objects in widely distributed astronomical data
• Data mining and visualization of grid data in tele-immersion environments
• Impact of data-intensive applications on next generation networks
• Increasing the availability and usability of network traffic monitoring data
  
  

National Center for Data Mining at the University of Illinois at Chicago (lead)
National Computational Science Alliance
National Laboratory for Applied Networking Research (NLANR)
New York University
Rensselaer Polytechnic Institute
University of California at Davis
University of Illinois at Chicago
University of Pennsylvania
Washington State University
Magnify, Inc., Chicago
Imperial College, London, England
National University of Singapore
University of Hong Kong
University of Toronto, Canada
ACSys, Canberra, Australia

NSF
DOE/ASCI
DOE's Office of Science
NASA Ames Research Center

 

Asheville, North Carolina,
demonstrations

Engineering and
science tools of
the future

The Next Generation Revolutionary Analysis and Design Environment (NextGRADE) and Immersive Visualization software suite provides a GUI that enables rapid assembly and analysis of both aircraft and spacecraft, allowing structural designers to quickly modify components and analyze design features. After the analysis is complete, the engineer can examine the results using an ImmersaDesk 3-D stereographic environment to view and interact with the structure from various perspectives in real time. These research tools will enable multiple geographically dispersed users to collaborate within multiscreen virtual reality environments.

NASA Langley Research Center

NASA Cross-Enterprise Technology Development Program
NASA Intelligent Synthesis Environment Initiative

 

Digital library
technologies for
education

The Alexandria Digital Library (ADL) provides access to geospatial data such as maps, aerial photographs, and remote-sensing images taken from satellites, and to geographically referenced digital information. The Alexandria Digital Earth Prototype (ADEPT) project provides an interactive instructional environment that combines geo-referenced data with modeling analysis tools (for more information, see page 65).

University of California at Santa Barbara
University of California at Los Angeles
University of Georgia
Georgia Institute of Technology
San Diego Supercomputer Center

NSF
DARPA
NASA

 

Tele-nanoManipulator

The UNC-Chapel Hill tele-nanoManipulator system allows scientists to see, touch, and directly modify viruses, nanotubes, and other nanometer-scale (one-billionth, or 10-9 meter) objects. Using remote microscopes and graphics supercomputers over advanced networks, researchers can, for example, measure the rupture strength of DNA, build nanometer-scale circuits with carbon nanotubes (Buckytubes), and measure the strength of the adenovirus capsid that is used in gene therapy.

NIH's Resource for Molecular Graphics and Microscopy
UNC's Departments of Computer Science, Physics & Astronomy, Chemistry, and Psychology; Schools of Education and Information & Library Science; and Center for Gene Therapy

NIH's NCRR
NSF
Army Research Office
Intel Corporation Office of Naval Research
SensAble Technologies, Inc.
Silicon Graphics, Inc.
ThermoMicroscopes

 

Telescience for
advanced tomography
applications

This project builds Web-based collaboration tools providing remote access to high resolution, 3-D microscopy in biomedical and neuroscience research. Using these tools, scientists are able to access and work with data from specialized imaging instruments such as high and ultra-high voltage electron microscopes at remote facilities. The telescience tools provide transparent access to supercomputing resources required to produce, refine, and analyze complex 3-D images of cellular and molecular structure and function. With geographically distributed image databases enabling comparisons among archived research results, the collaboration tools can help molecular biologists nationwide study the basic structures of cellular organelles like mitochondria and debilitating disorders such as Parkinson's and Alzheimer's disease.

NIIH's NCRR
NSF's Partnerships for Advanced Computational Infrastructure Program

Enhancing search
engine effectiveness

The Text Retrieval Conference (TREC) project is described on page 61 of this book.

Academic, industrial, and government organizations worldwide

NIST
DARPA

 

Computational
challenges in climate
and weather research

This NOAA demonstration is described on page 106 of this book.