Science & Engineering Showcase Archive

Integrated Robust Assured Data Services (IRADS)

The distribution of NEXRAD weather radar data via the Internet was pioneered at the University of Oklahoma in a project known as the Collaborative Radar Acquisition Field Test (CRAFT). CRAFT was the highly-successful prototype for the real-time transmission of meteorology data from multiple radars that led the National Weather Service (NWS) to adopt an Internet-based data transmission methodology.

Integrated Robust Assured Data Services (IRADS) is an extension of Project CRAFT that is now distributing high-resolution NEXRAD Level II data from the network of over 140 Doppler radars operated nationwide by the National Weather Service. Data from each radar are transmitted to four NWS Regional Headquarters (RHQ) sites. Each RHQ server is linked to the Internet2's Abilene Network. From there, the data are transmitted directly to one of several top-tier sites, including the University of Oklahoma. Level II data provide the highest spatial and temporal resolution information available from ground-based atmospheric observations and are used for real time warning of weather hazards such as hurricanes and tornadoes; for initializing numerical weather prediction models; for verifying past events, such as the location of damaging hail; and for non-meteorological purposes, including bird migration studies, bird strike avoidance, and urban pollution transport.

Common Instrument Middleware Project

The Instrument Middleware Project is funded by NSF's Middleware Initiative and has been developing the Common Instrument Middleware Architecture (CIMA). CIMA improves accessibility to instruments by integrating instruments and sensors into a grid computing environment with web services interfaces. Researchers at Internet2 member universities Indiana University and SUNY Binghamton, along with collaborators at Purdue University, University of Minnesota, the ChemMat-CARS group (University of Chicago at APS, Argonne National Lab), and a number of institutions in Australia have used CIMA in several scientific applications. They are also developing a standard methodology for grid enabling instruments and collaborating with scientists across disciplines in academia and industry who either develop instruments or whose work depends on the details of using them.

One of the challenges addressed by CIMA is abstracting instruments such that not only can a single management framework manage multiple instrument types, but also that a single instrument can deliver data for processing and storage to multiple grids, each with unique resource management policies and procedures. CIMA project members are developing ways to interact with instruments in real time via grid applications by moving some of the instrument functionality into the grid application, and away from the instrument itself. Researchers are also exploring ways for virtual organizations to meet location, authentication, and authorization challenges. In related work the CIMA group is working closely with another NMI project, the Open Grid Computing Environments Collaboratory (OGCE), to develop ubiquitous user interfaces for CIMA instruments using OGCE portal technologies.

Currently, the most well developed application of CIMA is in X-ray crystallography. This application provides remote instrument monitoring and remote access to data from several crystallography labs in the U.S. and abroad, including a synchrotron beamline, to acquire data from diffractometers and related sensors. Users can access live or previously collected data from participating laboratories through the CIMA Crystallography portal. Additional applications to which this technology is being applied include remote interaction with robotic telescopes, wireless sensor networks, environmental monitoring, and others.

MIT Demos Live e-VLBI Transmission

Scientists from MIT Haystack Observatory, along with several collaborators, gaveattendees at SC2004 (held 6-12 November 2004 in Pittsburgh) a first hand look at the application of high-speed networking to the practice of radio astronomy with the electronic transmission of Very Long Baseline Interferometry (e-VLBI) data. The live demo featured the real-time transmission of e-VLBI data from the Haystack's Westford Observatory in Massachusetts and NASA's Goddard Geophysical and Astronomical Observatory (GGAO) in Maryland, which were streamed over Internet2's Abilene Network to the Haystack correlator at 512 Mbps. The live results were displayed in a 3D plot (correlation amplitude, differential Doppler, differential delay) in Pittsburgh as the data were correlated. During periods when the antennas were not available, the team transferred pre-recorded data from Westford, GGAO, the Onsala Space Observatory antenna in Sweden, and Kashima Space Research Center Observatory in Japan. The team at SC did an immediate correlation on this data, again showing results in Pittsburgh as the correlation proceeded.

Dr. Alan Whitney, Associate Director of the MIT Haystack Observatory, who headed the demo team at SC, would like to acknowledge the following collaborators:
Haystack Observatory: Roger Cappallo, Kevin Dudevoir, David Lapsley, Mike Poirier and Mike Titus; NASA/GSFC GGAO Observatory: Jay Redmond; DRAGON collaborators provided much support both before and during the demo.
Kashima Observatory: Yasuhiro Koyama and colleagues; Onsala Observatory: Rudiger Haas and colleagues; NASA's Goddard Space Flight Center: Pat Gary and colleagues; MIT Lincoln Laboratory: provided Bossnet network and network support; National Science Foundation and NASA provided the funding to support the underlying work that made the demo possible and are helping to move e-VLBI towards routine operational reality.

FAST Protocol

The development of robust and stable ultrascale networking, at 100 Gbps and higher speeds in the wide area, is critical to support the new generation of ultrascale computing and Petabyte to Exabyte datasets that promise to drive discoveries in fundamental and applied sciences of the next decade. The goal of CalTech Networking Laboratory's project FAST Protocols for Ultrascale Network is to develop theory, algorithms and prototypes to design, demonstrate, and deploy protocols that are scaleable to arbitrary network capacity and size in order to fulfill the vision of ultrascale networking. Its FAST TCP kernel has stably achieved multi-Gbps throughput with higher than 90% utilization over extended period using standard MTU, during SC2002 Conference in November 2002. FAST is sponsored by the NSF, Department of Education, Army Research Office, Cisco and the Caltech Lee Center for Advanced Networking.

NEESgrid Early Adopter Sites

The first connections in a network infrastructure that will link earthquake engineering sites across the U.S. and create a national virtual earthquake engineering laboratory are now active at Oregon State University, Rensselaer Polytechnic Institute (RPI), and University of Nevada, Reno. The three sites are the first recipients of NEES-Points of Presence (NEESpops) on the integrated network that will support the National Science Foundation's (NSF) George E. Brown, Jr. Network for Earthquake Engineering Simulation (NEES) project. Called NEESgrid , this infrastructure will soon link earthquake engineering sites across the country, provide data storage facilities and repositories, and offer access to high-performance computing applications used for conducting simulations. The National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign (UIUC) leads the NEESgrid system integration effort. The three early adopter sites will test capabilities of the NEESgrid as they are developed and help NEESgrid researchers create a common infrastructure that can be used across sites and for all NEES applications. The sites were chosen because they are home to the three main types of equipment used in earthquake engineering experiments: centrifuges (RPI), shake tables (University of Nevada), and tsunami wave tanks (Oregon State). The sites are connected to each other through Internet2's Abilene backbone network at 155 megabits per second. The early adopter sites will test collaboration tools, local storage systems and data repositories, streaming data and video services, and tele-operations of experimental equipment.

Virtual Rooms Videoconferencing System (VRVS)

The Virtual Rooms Videoconferencing System (VRVS) provides a low cost, bandwidth-efficient, extensible means for videoconferencing and remote collaboration over networks within the High Energy and Nuclear Physics communities. VRVS is a web-based system for interoperable videoconferencing and collaborating. VRVS supports multiple platforms – Windows, Mac, Linux, Unix – and diverse collaborative applications – Access Grid, H.323 videoconferencing, QuickTime, chat, desktop sharing, and, soon, Microsoft Messenger. Ninety-five percent of the code was re-written for the spring 2003 release of VRVS 3.0, which includes an advanced booking system, new virtual rooms for meeting spaces, a streamlined web-based user interface, firewall and NAT solutions, AccessGrid tunneling, self-selection of video streams, user authentication, and synchronized time zones. No port reservations are required in order to initiate a videoconference – simply book a room in advance for any number of participants to join. Using Internet2 high-performance network infrastructure, deployment of the Web-based system has expanded to include more than 13,300 registered hosts running the VRVS software in more than 60 different countries. More then 45 VRVS network servers (or reflectors) are installed throughout the U.S., South America, Europe, and Asia. VRVS hosts an average of 350 multipoint videoconferencing and collaborative sessions worldwide every month.

Scientific Imaging and Computing Institute

The Scientific Imaging and Computing Institute (SCI Institute) at the University of Utah integrates three major facets of scientific computing--visualization, simulation, and geometric modeling--into computational problem solving environments. SCI Institute researchers and technologists have contributed to solving computational problems in fields including combustion, fluid dynamics, petrochemical reservoir simulation, cardiology, neurosurgery, radiology, nuclear fusion, and the atmospheric diffusion of pollutants. One of the key missions of the SCI Institute is to provide integrated solutions and problem solving environments to help scientists in different disciplines solve their own computational problems. According to SCI Institute Director Christopher Johnson, “Internet2 high-performance networks provide a direct data channel to our collaborators throughout the country. From researching Grid applications to leveraging the interactive videoconferencing capabilities of the Access Grid, Internet2 provides the networking we need to pursue collaborative Grid-based research.” SCI Institute researchers are using networking protocols such as multicast and IPv6 to accomplish their research while incorporating new and innovative networking solutions in a wide range of engineering and scientific application areas.

The SCI Institute also houses two national research centers: the NIH NCRR Center for Bioelectrical Field Modeling, Simulation and Visualization and the DOE Advanced Visualization Technology Center (AVTC). Additionally, the SCI Institute houses the NIH-BISTI Program of Excellence in Computational Bioimaging and Visualization: Tools For Image Processing and Fusion, Interactive Visualization, and Inverse Problems. (Image courtesy of the SCI Institute.)

SARA Observatory

Astronomers at a six different universities in the southeastern U.S. are able to take advantage of the clear desert skies in Arizona to perform their celestial research without having to leave their campuses. Members of the Southeastern Association for Research in Astronomy (SARA) consortium use Internet2 high-performance networks to remotely control the 0.9-meter Ritchy-Critchen reflecting telescope at Kitt Peak Observatory, located southwest of Tucson, AZ at 6,800 feet above sea level. According to Dr. James Webb, Director of the SARA Observatory, “In order to perform remote observing in real-time and get quality data, network speed, reliability, and bandwidth are critical.” Webb, who recently gave a presentation on remote observing at the Internet2 Joint Techs Workshop, added “Before Internet2, network latency and dropouts severely impacted observing. With Internet2, latency is small enough and dropouts rare enough to have essentially no impact on observing.”

A typical remote observing session starts with a check on the weather in Tucson. The astronomer then logs on to the telescope control computer using Virtual Network Computing (VNC is a platform independent, client-based system written by AT&T Laboratories), creates directories to hold the images, opens the dome on the telescope, observes, closes the dome on the telescope, and log off VNC. The actual images from the observing session are simply transferred back to any home or office computer using FTP. Webb summarized the benefits of remote observing, “It allows us to save travel money, keep our on-campus teaching commitments, and still collect scientific data from the best possible sites.” (Photo by Dr. Terry Oswalt)

GriPhyN

The Grid Physics Network (GriPhyN) collaboration is a team of experimental physicists and information technology researchers who are implementing the first Petabyte-scale computational environments for data intensive science. GriPhyN will allow geographically dispersed extraction of complex scientific information from massive datasets, provide access to large-scale computational resources, and enable collaboration among worldwide scientific communities. GriPhyN will initially give scientists access to the vast amounts of data that will flow from four large-scale physics and astronomy experiments including the CMS and ATLAS experiments at CERN's Large Hadron Collider, LIGO (Laser Interferometer Gravitational-wave Observatory), and the Sloan Digital Sky Survey. The data analysis for these experiments presents enormous IT challenges. Communities of thousands of scientists, distributed globally and served by networks of varying bandwidths, need to extract small signals from enormous backgrounds via computationally demanding analyses of datasets that will grow from the 100 Terabyte to the 100 Petabyte scale over the next decade. (GriPhyN will initially be able to store 10 petabytes of data.) In addition to the Abilene network, other high-performance networks, including ESnet, NREN, I-WIRE, and international connections through STAR TAP and StarLight, will be used to share data throughout this globally distributed community of scientists.

GriPhyN project PIs are Paul Avery of the University of Florida and Ian Foster of the University of Chicago and Argonne National Laboratory.

America View

One of the difficulties involved with using satellite imagery for time critical applications is the delay inherent in the acquisition, processing and delivery of these relatively large datasets. Through a cooperative agreement, the Internet2 Geospatial Working Group and AmericaView (a U.S. Geological Survey sponsored consortium of remote sensing researchers) are using Internet2 high performance networks to help minimize this issue. The Earth Resources Observation Systems (EROS) Data Center's Rapid Application Prototype (RAP) and Abilene network connectivity are used to enable the delivery of MODIS, Landsat, and other data sets to researchers. As a result, a wide variety of applications are able to capitalize on the near real-time satellite imagery. Several of the applications were demonstrated at the Internet2 Fall 2002 Member Meeting. According to PR Blackwell, Internet2 Geospatial Working Group Chair, “Internet2 high-performance networks are critical in realizing the goal of real-time, or near real-time delivery of satellite imagery. Internet2 networks enable these large data sets to be transferred from ground receiving stations, such as the EROS Data Center, where they are captured, processed, and delivered to university researchers and other users in minutes, rather than hours.” (Composite image of South Dakota courtesy of USGS.)

The Web Lecture Archive Project is a web-accessible collection of lectures and training courses focusing on the high-energy physics community. The main purpose of this joint venture between the UM-ATLAS Collaboratory Project, the University of Michigan Media Union, and CERN is to test, build, and evaluate an archival system as a tool for collaborative communication and learning. The current project builds upon the work initiated in a 1999 University of Michigan pilot project supported by the National Science Foundation that was targeted toward the archiving of lectures given in the prestigious Summer Student Program at the European Organization for Nuclear Research (CERN).

A web-browser and RealPlayer software are all that is required to use the archive. The archive itself contains over 300 entries—searchable by organization, author, title, and date—and has student presentations rubbing shoulders with lectures by Nobel Prize laureates.

In addition to serving as a repository, the archive has also been tested as a tool for distance learning—a key concern for the large and globally dispersed Large Hadron Collider (LHC) and high-energy physics communities. In collaboration with the ATLAS Experiment and CERN Technical Training, several of the ATLAS Offline Software Training courses were archived and are being reused for training.

Dr. Homer A. Neal, Director of the UM-ATLAS Collaboratory Project, notes that the work accomplished to date demonstrates the power of web-based archiving both in university level instruction and in large globally-dispersed collaborative research projects. According to Neal, "We can forsee making available to students around the globe, at essentially zero cost, lectures by the relevant world experts on any current scientific topic. We also see how web archiving can permit techniques and methodologies developed by a few members in a several thousand member research group to be quickly shared with colleagues who need this information to make coordinated progress in their own work on the same project." Neal concluded, “The original impetus for creating the web, also developed at CERN, was to support collaborative research. Web archiving as one of the ways that this goal is being achieved.”

The WLAP Group is continuing its work in creating archives for advanced training courses for ATLAS detector description and software architecture applications. The group is also pursuing networking and other enhancements to insure the robustness of the recording and retrieval processes.