|
Abstracts
Advances in Parallel Metacomputing
of Solid-Fluid Interaction (SFI) Problems
The Computational Fluid Dynamics (CFD) Laboratory carries out research
on parallel computing of solid-fluid interaction problems that require
extensive computer resources. A coupled computational fluid dynamics
and computational structural dynamics code has been developed for solving
aeroelastic flutter phenomenon. The major difficulties involved in the
implementation of the code are the volume of data to be handled and
computing times needed which make it necessary to have huge amounts
of processing power and memory.
Development of parallel and metacomputing tools has made it possible
to overcome these difficulties and made the solution of such complex
problems a reality at the IUPUI CFD Laboratory [1, 2]. An unstructured
unsteady CFD code for Euler equations, with a cell-centered finite volume
approach is coupled with a computational solid dynamics code for dynamic
structural response. A domain-decomposition based parallel algorithm
is employed. MpCCI, an MPI based code-coupling library, is used to loosely
couple the solid and fluid meshes. The problem is decomposed into several
blocks and is solved using parallel and metacomputing via I-Light over
multiple domains and on heterogeneous systems. Aeroelastic flutter calculations
for an aircraft wing will be presented as a test case. Comparison of
the elapsed and communication times with and without I-Light will be
presented for various block allocations. A virtual reality animation
of the results of the test case will also be demonstrated.
References:
- Hasan U. Akay, Xiaoyin He, Resat U. Payli, "Parallel Metacomputing
of Solid-Fluid Interaction Problems via I-Light", I-Light Applications
Workshop, December 4, 2002, IUPUI.
- Hasan U. Akay, Xiaoyin He, and Resat U. Payli, "A Code Coupling Application
For Solid-Fluid Interactions and Parallel Computing", Parallel CFD 2003
Conference, May 13-15, 2003, Moscow-Russia.
Introduction to Virtual Environments
- A Collaborated Classroom Offering between PU and IU
The Access Grid (AG) is an emerging technology that allows for unique
learning experiences. Currently, we are conducting a class titled "Introduction
to Virtual Environments" (Purdue TECH 519V) via the AG in collaboration
with Indiana University (IU Informatics I590). The class meets twice
a week for lectures over the Access Grid.
Conducting a class via the Access Grid offers several advantages over
traditional classroom instruction. Because the topic of the class (virtual
reality), is relatively new to Purdue and is quite specialized, class
size is rather small (approximately 10 students at Purdue and 6 at IU).
Holding the class lectures using AG allows us to hold lectures which
combine the students from both the Purdue and IU offerings of the course,
creating a larger virtual class size. This allows both students and
instructors to interact with a larger, more diverse, student population.
Conducting joint lectures also gives students the benefit of learning
from two instructors with different backgrounds and experiences providing
more information and viewpoints than a single instructor alone could
present. Additionally, the instructors can share the lecture load, allowing
each to create better lectures and lab materials, and spend more time
interacting with students. Because AG technology is cheaply and easily
accessible at a large number of locations, expert guest lecturers can
also join the lecture sessions remotely, eliminating the time and expense
that would normally be incurred, and allowing students to hear from
leaders in the field that might not otherwise be accessible to them
The topic of virtual reality presents unique challenges for traditional
distance education methods such as videotaping and web streaming of
lectures. In particular, it is important that students not only learn
about the technology of virtual reality, but also experience immersive
and interactive virtual environments as part of their class activity.
Live and Computational Experimentation
in Bio-terrorism Response
The study of bio-terror threats requires a significant improvement
in our capabilities to analyze human responses to an attack, both at
the level of citizens (individual freedom of mobility, infection rate,
and the collective feeling of well-being) and at the level of responders
and policy makers (coordination, control, planning, and policy formulation).
We address these issues using a geography based synthetic environment
with artificial and human agents. Computational models of artificial
agents' positions, mobility, infection-susceptibility and the state
of well-being are developed. Intuitive interfaces are provided for human
agents to experiment with complex coordination roles. Together, their
behaviors are used to analyze how a bio-terror attack may spread through
the population and how its impact may be mitigated by different intervention
strategies. The behavior of the artificial agents is calibrated in accordance
with standard models from the fields of epidemiology, psychology, and
economics and the agents' movements reflect the actual behavior patterns
in the cities concerned. The representation of human behavior in artificial
agents reflects human capabilities, cognitive processes, limitations,
and conditions that influence behavior (e.g., morale, stress, panic).
We use individual-based epidemiological models for person-to-person
contamination scenarios. An important aspect of this work is the development
of a scalable architecture for distributed, tera-scale, grid computing.
Our results indicate that if the intervention is done early, city block
vaccination is most effective; in the intermediate term trace vaccination
will be most effective; and if the response is delayed, then mass vaccination
is the best strategy. In terms of quarantine strategies, early extreme
quarantine is more effective than the city block quarantine. However,
for delayed intervention, there is no significant difference between
the extreme and city block quarantine strategies.
Ingestion, Analysis and Distribution
of Real-time and Archival Satellite Data for Indiana: The Purdue Terrestrial
Observatory, the Laboratory for Analysis of Remote Sensing and IndianaVie
The immanent acquisition in real-time of panchromatic, multi-spectral,
radar and, potentially, hyper-spectral data by the Purdue Terrestrial
Observatory (PTO) is described. Subsequently, plans are detailed for
near-real-time analysis of both archival and newly acquired remotely
sensed data through Purdue's Laboratory for Applications of Remote Sensing
(LARS), soon celebrating its 40th anniversary, and the distribution
of such data, facilitated by I-Light, I-Light II, the National Science
Foundation (NSF) Extensible Teragrid Facility (ETF) and by the web-enabled
IndianaView affiliate of the U.S. Geological Survey (USGS) supported
AmericaView Program. Opportunities for collaborative multidisciplinary
research, distributed instruction, economic development and community
outreach utilizing these technological resources are delineated.
Variations2 Digital Music Library
Variations2 is a digital music library system developed by researchers
and IT staff at Indiana University. Variations2 provides online access
to digital music in a variety of formats, including sound recordings
and musical scores, along with software tools for searching, playing,
displaying, and interacting with music, with a focus on the use of online
music resources in teaching and learning.
Variations2 utilizes I-Light to deliver high-quality streaming audio
from the collections of the IU Bloomington music library to students
and faculty participating in pilot projects at IUPUI and additional
satellite site locations in the United States and United Kingdom.
The Variations2 project is supported in part by a Digital Libraries
Initiative research grant from the National Science Foundation and National
Endowment for the Humanities.
Building Community Grids: Collaboration
and Distributed Computing
The Community Grids (CG) Lab team investigates problems in scalable
messaging for real-time collaboration systems and in Grid/Web Services.
The Community Grids Lab's collaborative computing technologies are based
on a general purpose messaging infrastructure (NaradaBrokering) that
may be used to transmit messages ranging from real-time audio/video
feeds to XML events to multiple participants. We further investigate
the problems of scalable shared images that may be delivered to mobile
clients. In the field of Grid technologies, the CG Lab's SERVOGrid effort
combines earthquake simulation/modeling codes, computing resources and
databases from several universities and labs. Computing Web portals
are a standard way of interacting with Grid services, and the CG Lab's
NSF National Middleware Initiative project for reusable portal component
middleware allows Grid developers to combine user interfaces to "best
practice" service components from several collaborating universities
into a single portal framework.
Using I-Light and the Purdue Nanohub
Computing Resources to Run Computationally Intensive Codes in Nanotechnologies
The NSF funded Network for Computational Nanotechnology has a dual
mission of research and providing computing services for the nanotechnology
community. The research performed at the NCN is focusing on three areas:
Nanoelectronics, Nanoelectromechanical systems (NEMS) and Nano-Bio.
In each of the research themes, projects are defined that are ready
for a coordinated approach from a team of experts that can tackle the
atomistic level, the device level up to the system level. These computational
challenges create useful computational tools that are made available
to the community through the NCN computing portal the Nanohub.
The Nanohub uses a middleware infrastructure that allows users to
run simulations seamlessly without knowing where and how the program
is running. The front end is a standard website but the back end is
made of workstations, the SMP machine and Linux clusters. Connecting
the Nanohub to I-Light and the Teragrid is of prime importance to the
NCN in order to make available simulators that require a considerable
amount of computing power to run in parallel.
Computer Simulation of Neutron Rich
Matter Using the MDGRAPE-2
Very dense matter in the crust of neutron stars and supernovae is predicted
to have a complex structure called "nuclear pasta". Properties
of this matter are important for X-ray, gravitational wave and neutrino
radiation. Using a semiclassical model, where neutrons and protons interact
via a classical two-body central force, we model the formation of pasta
by molecular dynamics (MD) simulation. From the resulting configurations,
we calculate the static structure factor, and then the mean free path
of neutrinos scattering coherently from the pasta. To get physically
realistic results, we need to evolve systems of 40,000 nucleons for
about 60,000 MD time steps. This would take about one year per simulation
on a Power3+ machine. Using the two MDGRAPE-2 boards attached to IU's
IBM SP, we can do the simulations much faster. A complete simulation
takes about one week.
Conducting Remote Sensing Seminars
across the United States
In 1999, I developed a Remote Sensing Seminar course to bring a focus
on the latest information about remote sensing topics. It was very popular
and during the next year, it was decided to offer the course to Indiana
State and Mississippi State University at the same time via IHETS. This
meant that I only needed to arrange for about 1/3 of the presentations
and that we could learn about remote sensing work at other Universities.
The purpose of the seminars is to provide a forum for graduate students
to: 1) interact with academic, aerospace and government personnel working
with remote sensing, global positioning systems, and geographic information
system called "spatial technologies", 2) hear the latest information
about spatial technologies and how they are being used, 3) provide opportunities
to explore future career paths involving remote sensing, GIS and GPS
technologies. During 2001, 2002 and 2003, we added the University of
Nebraska, the Delta Research and Education Center (Stoneville, MS),
the NASA Stennis Space Center, the EROS Data Center, and the South Dakota
School of Mines as participants. We also changed the approach and are
using the InterNet and Purdue's Information Technology Program (ItaP)
to provide higher quality seminars. Now we are able to ask remote sensing
scientists from across the US to present a seminar to our students by
transmitting the seminar from their location. ITaP transmits the seminar
to the other locations. The non-university locations are participating
to provide their professional staff with the latest information on remote
sensing. Each of the participating Universities sign up students using
their course registration and students receive one hour credit for participation.
The students earn their credit by attending the seminars and providing
a 5 minute presentation on their research with a limit of 5 PowerPoint
slides. The quality of the seminars has improved every year with changes
in the technology of presenting and transmission of the seminars. During
this current semester, Purdue is going international with this approach
and we are offering a special remote sensing seminar jointly with the
University of Leuven, Belgium. This is part of a joint Master of Science
degree program on Earth Observation with Belgium.
Large-Scale Distributed Rendering
of 3D Animation and Imagery
The rendering of 3D animation and imagery is often a computationally
intensive task. Rendering an entire animation may take many days to
finish and with the refinement process that is usually done to correct
mistakes or make changes, many more days are needed to get the final
results. Distributing the rendering process to all available computers
can dramatically decrease the time it takes to get the final result.
Purdue University's Distributed Rendering Environment (DRE) initiative
uses as many computers as available to process 3D animation. In order
to grow the number of available computers and extend the functionality
to other campuses in Indiana, IU and IUPUI have begun to use and help
develop the system over I-Light.
iVDGL Grid Operations Center and
I-Light
Based at Indiana University, the international Virtual Data Grid Laboratory
(iVDGL) Grid Operations Center (iGOC) serves as a single point of contact
for iVDGL information and operations. Grid3+ consists of 28 sites with
over 2400 CPUs.
The IU ATLAS cluster contributes 60+ nodes to the iVDGL and Grid3+
infrastructure. Grid3+ traffic runs via I-Light to the ATLAS cluster,
which comprises 3% of the only sustained computing grid currently in
production. Resources used by 11 different scientific applications,
including 3 high energy physics simulations and 4 data analyses in high
energy physics, biology, astrophysics and astronomy are supported by
Grid3+, and includes more that 100 research scientists. Peak throughput
of 900 jobs running concurrently with a 75% completion rate has been
sustained.
The iGOC maintains critical services used by Grid3+ such as Virtual
Organization Membership Service, Replica Location Services, the Globus
Index Information Services, the Ganglia and MonaLisa monitoring tools,
and the iVDGL web content at www.ivdgl.org.
Digital Library Resources Over I-Light
Abstract Since the I-Light network was launched near the end of 2001,
the production and use of digital library resources has grown significantly.
The IU Digital Library Program has developed a number of digital library
services and resources that deliver texts, reference works, images,
and audio files to users throughout Indiana and the world. These collections
are used by researchers, students (at the university, high school, and
grade school levels), and the general public. The I-Light network ensures
that these resources and services reach our users quickly and reliably.
Our poster highlights three representative projects:
- The Charles W. Cushman Collection, which includes over 14,000 images
photographed from 1938 to 1969 by amateur photographer Charles Cushman.
- Letopis' Zhurnal'nykh Statei, an online version of an important
Soviet periodical index.
- The Swinburne Project, an XML-based text collection of the works
of Victorian poet Algernon Charles Swinburne.
The MxN Problem: Coupling Parallel
Components
The original vision of the Grid was as a platform for running a scientific
simulation consisting of large numbers (on the order of millions) of
parallel communicating processes. That vision has since evolved to one
of Grid as a fabric for tying together heterogeneous resources: data
servers (sensors, instruments, data bases, and large archival storage),
computation resources, and visualization systems. Increasingly important
is the problem of connecting parallel programs, possibly running on
geographically distributed machines, each using a different number of
processes. This problem occurs in climate modeling, space weather systems,
and magnetically confined fusion energy simulations. In each case, large
legacy parallel codes from different scientific communities are being
connected to create new multidisciplinary simulations. This presentation
will describe the problems involved, and present some middleware solutions
for combining parallel programs even when they use incommensurate numbers
of processors.
Introduction to Virtual Environments
- A Collaborative Classroom Offering between PU and IU
The Access Grid (AG) is an emerging technology that allows for unique
learning experiences. Currently, we are conducting a class titled "Introduction
to Virtual Environments" (Purdue TECH 519V) via the AG in collaboration
with Indiana University (IU Informatics I590). The class meets twice
a week for lectures over the Access Grid.
Conducting a class via the Access Grid offers several advantages over
traditional classroom instruction. Because the topic of the class (virtual
reality), is relatively new to Purdue and is quite specialized, class
size is rather small (approximately 10 students at Purdue and 6 at IU).
Holding the class lectures using AG allows us to hold lectures which
combine the students from both the Purdue and IU offerings of the course,
creating a larger virtual class size. This allows both students and
instructors to interact with a larger, more diverse, student population.
Conducting joint lectures also gives students the benefit of learning
from two instructors with different backgrounds and experiences providing
more information and viewpoints than a single instructor alone could
present. Additionally, the instructors can share the lecture load, allowing
each to create better lectures and lab materials, and spend more time
interacting with students. Because AG technology is cheaply and easily
accessible at a large number of locations, expert guest lecturers can
also join the lecture sessions remotely, eliminating the time and expense
that would normally be incurred, and allowing students to hear from
leaders in the field that might not otherwise be accessible to them
The topic of virtual reality presents unique challenges for traditional
distance education methods such as videotaping and web streaming of
lectures. In particular, it is important that students not only learn
about the technology of virtual reality, but also experience immersive
and interactive virtual environments as part of their class activity.
Science Education Using Networked
IT
We consider some projects and ideas for projects that benefit from
networked information technology. Being a relative newcomer to this
endeavor (and to Indiana), the speaker welcomes and hopes to hear follow-on
ideas from the audience.
The Biocomplexity Institute at IU-Combining
Experimental and Computational Systems Biology
The Biocomplexity Institute at IU, in collaboration with Purdue and
Notre Dame, addresses issues in post-human-genome-project interdisciplinary
biosciences, particularly the integration of quantitative experiments
and computation to provide predictive multiscale models of the development,
which the National Institutes of Health have identified as a priority
for the next ten years (see http://nihroadmap.nih.gov/). This goal requires
both large scale simulations (currently based around the Cellular Potts
Model (CPM)-http://www.nd.edu/~lcls/compucell/) and the development
of new experimental techniques. The College of Arts and Sciences and
the School of Informatics at IUB are greatly expanding in these areas,
with five hires completed, five planned for this year, and similar numbers
for the next two years. Collaborations include projects on vascular,
heart and limb development and regeneration (at the IU School of Medicine
and the School of Science at IUPUI), on neurobiology, microfluidics
and biofilms (at IUB) and on algorithm development (at Notre Dame).
Conducting Remote Sensing Seminars
across the United States
In 1999, I developed a Remote Sensing Seminar course to bring a focus
on the latest information about remote sensing topics. It was very popular
and during the next year, it was decided to offer the course to Indiana
State and Mississippi State University at the same time via IHETS. This
meant that I only needed to arrange for about 1/3 of the presentations
and that we could learn about remote sensing work at other Universities.
The purpose of the seminars is to provide a forum for graduate students
to: 1) interact with academic, aerospace and government personnel working
with remote sensing, global positioning systems, and geographic information
system called "spatial technologies", 2) hear the latest information
about spatial technologies and how they are being used, 3) provide opportunities
to explore future career paths involving remote sensing, GIS and GPS
technologies. During 2001, 2002 and 2003, we added the University of
Nebraska, the Delta Research and Education Center (Stoneville, MS),
the NASA Stennis Space Center, the EROS Data Center, and the South Dakota
School of Mines as participants. We also changed the approach and are
using the InterNet and Purdue's Information Technology Program (ItaP)
to provide higher quality seminars. Now we are able to ask remote sensing
scientists from across the US to present a seminar to our students by
transmitting the seminar from their location. ITaP transmits the seminar
to the other locations. The non-university locations are participating
to provide their professional staff with the latest information on remote
sensing. Each of the participating Universities sign up students using
their course registration and students receive one hour credit for participation.
The students earn their credit by attending the seminars and providing
a 5 minute presentation on their research with a limit of 5 PowerPoint
slides. The quality of the seminars has improved every year with changes
in the technology of presenting and transmission of the seminars. During
this current semester, Purdue is going international with this approach
and we are offering a special remote sensing seminar jointly with the
University of Leuven, Belgium. This is part of a joint Master of Science
degree program on Earth Observation with Belgium.
A Global Grid Analysis of Invertebrate Evolution
In recent years the evolutionary relationships of invertebrates with six legs (insects and their relatives) have been a topic of considerable debate in the biological community. There are two steep challenges in pursuing this question: assembling enough genetic sequence data to answer the question effectively, and assembling the computer resources required to analyze the data. The Center for Genomics and Bioinformatics and IU, along with experts from UITS, assembled a large data set of genetic data. A team lead by University Information Technology Services and the High Performance Computing Center of Stuttgart lead an international team that created a global grid to analyze this data. More than 600 computer processors were applied to the task of analyzing this data within a global grid spanning every continent on earth except Antarctica. Indiana University supercomputers and Teragrid resources were key components of this grid. This talk will describe this study, our conclusions to date, and also discuss the future of biological computing and computing grids. This project was awarded a High Performance Computing Challenge award at the international ACM/IEEE SuperComputing 2003 conference held in November, 2003
Beyond Computing: The Search for Creativity
Modern computing arose from the interplay of science, engineering and defense needs, and hardware and software technology advances. Computational science began, like most science, as a small and localized group activity. These islands of research are increasingly connected — the truly grand challenges require the skills of multidisciplinary groups, often internationally distributed, working collaboratively as global renaissance teams. Transformative collaborative, networked technologies, coupled with high-end computing and large-scale data archives, enable success through enhanced human interaction, community building, collective problem solving and innovation.
Recent studies on human innovation span the sciences and humanities, and they point to a paradigm shift in the way that we think about creativity, they demonstrate the power of understanding one knowledge domain in terms of another, and they reveal the contingencies of our perspectives. Fresh perspectives result from cross-domain interactions through common themes.
These common technological themes across the sciences and humanities include high-performance data collection, retrieval and integration, text and image data mining, software fusion, visualization, collaborative tools and human-computer interfaces. Enlarging the scope of high-performance computing to include a broad range of disciplines will enrich discovery and expand the applications of computing technology.
Both creativity and courage will be necessary to address the really big questions: understanding the matter and structure of our universe; modeling life and its complex processes; and enriching the human condition. From mapping the cosmos to mapping the brain; from interactive sensor databases to real-time severe weather prediction; from bioinformatics to situational awareness, "Thinking out of the Box" requires interdisciplinary, diverse and global interactions. How we enable these creative practices will shape the future of our HPC community.
An Integrated Environmental Monitoring
Network
Ecological studies depend on the ability to monitor an environment,
collect data at appropriate spatial and temporal scales, and analyze
that data from the diverse viewpoints of many relevant disciplines.
Historically, environmental studies have been conducted by small teams
of researchers, usually hand-collecting data at set but low frequency,
and organizing it according to ad hoc, project-specific goals. Recent
years have seen dramatic advancement in the ability to gather environmental
data remotely and therefore at much higher frequency.
We are working to create a dynamic and integrated network of environmental
sensors in natural environments to acquire real time data and create
tools for visualization appropriate for different audiences to advance
both ecological research and educational exploration. Visualization
of real time data from remote sensors distributed throughout Central
Indiana provides numerous challenges. The benefits of successfully integrating
remotely deployed environmental sensors in a post 9-11 world are obvious.
Our work will bridge both the extremes associated with the frequency
of data collection and the lack of data coordination by creating techniques
for data networking and retrieval, and data management and analysis.
We are working to integrate multiple data streams into a coherent database
and create applications that allow users to view data from multiple
instruments at different sites.
A key outcome of this program is creating visualizations of real time,
dynamic data from the everyday world and delivering it via web applications
as well as through innovative display spaces. These visualization capabilities
need to operate across a range of computing platforms to make this data
immediately accessible and useful to a range of interested parties,
across multiple disciplines.
Our goal is to use the instrumented sites to create analysis and presentation
applications to foster a community of learners interested in understanding
these complex ecosystems, and the larger environmental issues that they
represent. This broad-based community will include environmental researchers,
university faculty in lecture halls, K-12 math and science teachers,
university and K-12 students, civic leaders, and educators at informal
learning centers.
Using I-Light and the Purdue Nanohub
Computing Resources to Run Computationally Intensive Codes in Nanotechnologies
The NSF funded Network for Computational Nanotechnology has a dual
mission of research and providing computing services for the nanotechnology
community. The research performed at the NCN is focusing on three areas:
Nanoelectronics, Nanoelectromechanical systems (NEMS) and Nano-Bio.
In each of the research themes, projects are defined that are ready
for a coordinated approach from a team of experts that can tackle the
atomistic level, the device level up to the system level. These computational
challenges create useful computational tools that are made available
to the community through the NCN computing portal the Nanohub.
The Nanohub uses a middleware infrastructure that allows users to
run simulations seamlessly without knowing where and how the program
is running. The front end is a standard website but the back end is
made of workstations, the SMP machine and Linux clusters. Connecting
the Nanohub to I-Light and the Teragrid is of prime importance to the
NCN in order to make available simulators that require a considerable
amount of computing power to run in parallel.
Live and Computational Experimentation
in Bio-terrorism Response
The study of bio-terror threats requires a significant improvement
in our capabilities to analyze human responses to an attack, both at
the level of citizens (individual freedom of mobility, infection rate,
and the collective feeling of well-being) and at the level of responders
and policy makers (coordination, control, planning, and policy formulation).
We address these issues using a geography based synthetic environment
with artificial and human agents. Computational models of artificial
agents' positions, mobility, infection-susceptibility and the state
of well-being are developed. Intuitive interfaces are provided for human
agents to experiment with complex coordination roles. Together, their
behaviors are used to analyze how a bio-terror attack may spread through
the population and how its impact may be mitigated by different intervention
strategies. The behavior of the artificial agents is calibrated in accordance
with standard models from the fields of epidemiology, psychology, and
economics and the agents' movements reflect the actual behavior patterns
in the cities concerned. The representation of human behavior in artificial
agents reflects human capabilities, cognitive processes, limitations,
and conditions that influence behavior (e.g., morale, stress, panic).
We use individual-based epidemiological models for person-to-person
contamination scenarios. An important aspect of this work is the development
of a scalable architecture for distributed, tera-scale, grid computing.
Our results indicate that if the intervention is done early, city block
vaccination is most effective; in the intermediate term trace vaccination
will be most effective; and if the response is delayed, then mass vaccination
is the best strategy. In terms of quarantine strategies, early extreme
quarantine is more effective than the city block quarantine. However,
for delayed intervention, there is no significant difference between
the extreme and city block quarantine strategies.
Security Challenges for the Future
Network technology advancements like I-light have provided greater
network connectivity and increased bandwidth. This greater connectivity
together with advancements in hardware technology will bring us to a
point where we will be able to compute when needed and with whatever
computing devices that are available (no matter how limited). Computing
will be everywhere. In many cases this computing will be unseen and
invisible to those around. As these advancements bring small lightweight
devices to the computing mainstream, new applications that were once
foreign will be unleashed on these devices. Consequently security must
be addressed. In this presentation we will survey some of the more challenging
problems that are being examined in our Information Security Lab.
Ingestion, Analysis and Distribution
of Real-time and Archival Satellite Data for Indiana: The Purdue Terrestrial
Observatory, the Laboratory for Analysis of Remote Sensing and IndianaView
The immanent acquisition in real-time of panchromatic, multi-spectral,
radar and, potentially, hyper-spectral data by the Purdue Terrestrial
Observatory (PTO) is described. Subsequently, plans are detailed for
near-real-time analysis of both archival and newly acquired remotely
sensed data through Purdue's Laboratory for Applications of Remote Sensing
(LARS), soon celebrating its 40th anniversary, and the distribution
of such data, facilitated by I-Light, I-Light II, the National Science
Foundation (NSF) Extensible Teragrid Facility (ETF) and by the web-enabled
IndianaView affiliate of the U.S. Geological Survey (USGS) supported
AmericaView Program. Opportunities for collaborative multidisciplinary
research, distributed instruction, economic development and community
outreach utilizing these technological resources are delineated.
Large-Scale Distributed Rendering
of 3D Animation and Imagery
The rendering of 3D animation and imagery is often a computationally
intensive task. Rendering an entire animation may take many days to
finish and with the refinement process that is usually done to correct
mistakes or make changes, many more days are needed to get the final
results. Distributing the rendering process to all available computers
can dramatically decrease the time it takes to get the final result.
Purdue University's Distributed Rendering Environment (DRE) initiative
uses as many computers as available to process 3D animation. In order
to grow the number of available computers and extend the functionality
to other campuses in Indiana, IU and IUPUI have begun to use and help
develop the system over I-Light.
Methods for Network-based Graphics
and Visualization
Recent increases in network bandwidths and graphics capabilities promise
significant improvements in the way that researchers, artists, educators,
and students are able to visually interact with their data, concepts,
presentations, and colleagues. We will provide a brief survey of methodologies
for harnessing graphics resources over a network, focusing on visual
tele-collaboration and remote visualization and rendering. We will illustrate
these methods with a number of projects carried out in conjunction with
IU's Advanced Visualization Lab, and will describe potential enhancements
to the research, education, and creative activities at Indiana University.
|
