Distributing Science

Document Details:
  • Uploaded: May, 1st 2015
  • Views: 53 Times
  • Downloads: 0 Times
  • Total Pages: 7 Pages
  • File Format: .pdf
  • File size: 735.93 KB
  • Uploader: amanda
  • Category: Miscellaneous
 add to bookmark | report abuse
ince late 2002, a distributed computing
platform called the Berkeley Open
Infrastructure for Network Computing
(BOINC) has been enabling computer
users to do all of those things. BOINC allows a user
who runs its software to divide his or her computer’s
processor time among numerous scientific projects.
If you want to predict the temperature in 2050, one
popular project Climateprediction.net allows its
users to do just that. Over 100,000 screensaver
scientists have aided professional climatologists in
predicting future weather conditions in an effort to
realize the impact human beings have had on nature.
Interested individuals can download, install, and run
the software (see screenshot below) on their
computers and instantly begin predicting climate
changes 50 years into the future. The project and its
ability to involve non-scientists in important research
are made possible by the process called distributed
computing.
DISTRIBUTED COMPUTING
Distributed computing is the method of splitting
a large problem into smaller pieces and allocating the
workload among many computers. These individual
computers process their portions of the problem, and
the results are combined together to form a solution
for the original problem. Together, the aggregate
power of these computer systems can greatly surpass
the computing power of the world’s fastest
supercomputers.
The largest projects, such as
Climateprediction.net, rely almost solely on the
Internet as a way of recruiting thousands of interested
computer users to run the software. After installing
the program, users are sent their first set of data to
process, typically called a workunit. The processing
happens almost invisibly. The client software has
been developed to only crunch data when the
processor is not busy doing other things which
means users’ computers are always available for them
to listen to music, check E-mail, or perform any other
task. Once the workunit has been processed, the
result is returned to the projects main server, and a
new set comes in to restart the cycle.
Over 100,000 people have begun modeling temperature graphs like this one, running simulations from 1800-2050 in an effort to
try and predict trends in the global climate. The screensaver shows the current progress to the user, including a color-coded map
of the Earth
(red indicating over 100
º
F
,blue
under
-
40
º
F)
, and the exact date and time being predicted at that moment.
Through the volunteering of computer
resources, participants are helping to direct the future
of distributed computing and pointing scientific
researchers towards areas that interest the community
as a whole. Through distributed computing, BOINC
is enabling people to get involved. Cutting-edge
science no longer has to feel out of reach for a large
percentage of the population.
BOINC
BOINC was created at the University of
California-Berkeley by the Space Sciences
Laboratory, a group involved with the SETI@home
project. It is the first computing tool that allows any
scientist to use high-powered computing to run
computationally expensive projects. Some of these
projects such as modeling Earth’s future climate are
so computationally intense that without this system,
they would not even be attempted. Thus, research
problems that were once considered impossible can
now be investigated wherever there is a computer.
From research computer grids to the laptop in your
bedroom, every computer connected to BOINC will
have an impact on the direction of science.
Currently there are 5 main projects using the
BOINC infrastructure. Among these, the ones
logging the most computing time are SETI@home
and Climateprediction.net. Together, these projects
have run millions of hours worth of computing time
that would otherwise be unused on idling computers.
The results of many distributed computing projects
are immensely important. From predicting the
climate for the entire 21
st
century to searching for
cures to protein-related diseases such as Alzheimer’s,
Parkinson’s, and Mad-Cow disease, BOINC projects
have the potential to make a real difference in the
world. Most of these projects have attached screen
savers. These screen savers are not only interesting to
view when a computer is idling but also show what a
computer is doing with the model or project it is
running. With fully interactive virtual globes,
multiple climate maps, and satellite views in
Climateprediction.net and a real-time view of the
radio signals being analyzed in SETI@home, users
are always able to see exactly what type of work their
computer is performing and can better understand
how their contribution benefits the project as a whole.
Even when project results leave a single
computer and are processed together with the results
from other computers, users will be on the cutting
edge of learning about their impact. From
participation in global computing teams to personal
discussions about the many different science (and
other) topics, the BOINC community will keep users
in touch with the projects in which they contribute.
COMMUNITY
Prior to the advent of the distributed science
allowed by the BOINC architecture, a large project’s
accessibility would be limited to a select number of
researchers, most likely from established universities.
External access would be limited and consequently so
would outside opinions and ideas. By placing the
discussion and projects in the public domain, BOINC
promises to overcome this limitation; thus, research
communities on a larger, global scale have the
impressive power to shape the course of a project.
This community based around the use of
BOINC is another incentive for users. The
relationships that the members within each project
develop are similar to the professional camaraderie
that would develop between the “classic researchers”
described above, but this community is on a much
larger scale.
"There’s been really an incredible amount of
virtual community formation,” says Dr. David
Anderson, BOINC’s founder and architect. “It’s clear
that a lot of people are really interested in getting
involved in ways other than just running software on
their computer.”
The community is what provides value and
excitement during the running of the distributed
science project. In this sense, BOINC is transforming
the concept of a research community. Through the
development of these larger, global scientist
communities, another concept of community develops
based solely on the use of BOINC. The open, global
communities of BOINC have the sole requirement of
interest in a project.
These worldwide communities are each
centered on the common interest of the goal or topic
of a project. Whether it is an interest in searching for
extraterrestrial life or a desire to contribute to the
investigation of global warming, each community
bonds over its common goal, through message
boards, individual profiles, teams, and a credit
system.
What does the future hold for the weather on Earth?
Climateprediction.net is attempting to study the
possibilities that are in store for Earth’s climate in the
next century. Running a model thousands of times
with small variations allows scientists to study the
sensitivity of the environment to changes in variables
like carbon dioxide or the sulfur cycle.
Participants are given their own climate model to
run on their computer. With the screen saver that
comes with the software, users can watch their own
version of Earth as precipitation forms, temperatures
change, pressure rises and falls, and many other
characteristics evolve. At the completion of the run,
results are returned to Climateprediction.net for
analysis.
So far, Climateprediction.net has allowed scientists
to see that global warming may be more of a problem
than previously believed. Simulations of ocean
churning have provided more insight on the effects of
thermohaline circulation as well. As more
participants join the project and complete model runs,
climatologists will learn more and more about what’s
in store for the weather on Earth. More on
Climateprediction.net can be found at
http://climateprediction.net/.
Proteins are building blocks of life as we know
it. In an attempt to learn more about the structure
of proteins, Predictor@home is searching for and
testing new ways for linking the sequence of
proteins to their structure. Ultimately, they hope to
use the knowledge they gain to answer questions
regarding protein related diseases.
The work unit for Predictor@home consists of
data for a sequence of proteins and an algorithm to
attempt to predict the structure of the protein. The
user’s computer runs the algorithm on the structure
data provided. Upon completion of a run, the
results are returned to be compared with the other
predicted results made by other users.
Predictor’s work is providing new insight
towards the study of proteins. As methods for
predicting the structure of proteins improve, we
will be better able to design drugs to treat new and
existing diseases. More information on this work
can be found at http://predictor.scripps.edu/.
Is ET really out there? SETI (Search for
Extraterrestrial Intelligence) is attempting to find him.
Experts have reasoned that the most probable way alien
life would attempt to contact us is through radio waves.
The SETI project harnesses the collective computing
power of its 128 thousand current users to sift through
massive amounts of data collected at the Arecibo
Observatory in Puerto Rico.
A user’s computer will receive a 100 second block
of radio data recorded at Arecibo. The computer’s task
is to analyze the data, looking specifically for pulses,
triples or other patterns in the signals that would not
occur in space naturally. Any signals that seem out of
the ordinary are sent back to Berkeley for further
analysis.
Most of the data - 99.9999% - are dismissed as
either noise, interference from terrestrial objects, or
equipment malfunction. However, the other very small
percentage of signals does not currently have
explanations, but none have provided conclusive
evidence of other civilizations. For more information,
visit the SETI@Home web site at
http://setiweb.ssl.berkeley.edu/.
It’s been almost 90 years since Einstein first
predicted that ripples exist in the fabric of time and
space. However, concrete proof has never been
found. Einstein@home is an attempt to find these
gravitational waves.
Work units received by users for Einstein@home
consist of 12-megabyte data chunks from the US
Laser Interferometer Gravitational wave Observatory
(LIGO) and the British-German GEO-600
gravitational wave observatory. This data portrays
only a very small fraction of the celestial sphere, thus
allowing a user’s computer to rigorously and
thoroughly scour the data over several days in search
of a gravitational wave.
The discovery of gravitational waves would be
monumental for both theoretical and experimental
physics if found. Their existence would confirm
much of Einstein’s theory and provide us with a
better understanding of our universe. Visit
Einstein@home at http://einstein.phys.uwm.edu/ for
more details on the search for gravitational waves.
Proteins are the building blocks of life as we
know it. In an attempt to learn more about the
structure of proteins, Predictor@home is searching
for and testing new ways for linking the sequence of
proteins to their structure. Ultimately, they hope to
use the knowledge they gain to answer questions
regarding protein related diseases.
The work unit for Predictor@home consists of
data for a sequence of proteins and an algorithm to
attempt to predict the structure of the protein. The
user’s computer runs the algorithm on the structure
data provided. Upon completion of a run, the results
are returned to be compared with the other predicted
results made by other users.
Predictor’s work is providing new insight
towards the study of proteins. As methods for
predicting the structure of proteins improve, we will
be better able to design drugs to treat new and
existing diseases. More information on this work can
be found at http://predictor.scripps.edu/.
What does the future hold for the weather on
Earth? Climateprediction.net is attempting to study the
possibilities that are in store for Earths climate in the
next century. Running a model thousands of times with
small variations allows scientists to study the sensitivity
of the environment to changes in variables like carbon
dioxide or the sulfur cycle.
Participants are given their own climate model to
run on their computer. With the screen saver that
comes with the software, users can watch their own
version of Earth as precipitation forms, temperatures
change, pressure rises and falls, and many other
characteristics evolve. At the completion of the run,
results are returned to Climateprediction.net for
analysis.
So far, Climateprediction.net has allowed scientists
to see that global warming may be more of a problem
than previously believed. Simulations of ocean
churning have provided more insight on the effects of
thermohaline circulation as well. As more participants
join the project and complete model runs, climatologists
will learn more and more about what’s in store for the
weather on Earth. More on Climateprediction.net can
be found at http://climateprediction.net/.
Displaying 5 of 7 pages, to read the full document please DOWNLOAD.
Tags: