Data Centers' Growing Power Demands
Technology Review (02/15/07) Greene, Kate
A new Lawrence Berkeley National Laboratory study concerning the amount of
power used by servers should help the industry decide how best to go about
creating power consumption restrictions and solutions. "Estimating Total
Power Consumption by Servers in the U.S. and the World," found that server
power consumption doubled between 2000 and 2005, and that servers and their
auxiliary elements make up 1.2 percent of the energy consumed in the United
States and 0.8 of the power consumed worldwide. In 2005, U.S. servers used
more power than 20 different states. The report is based on IDC research
concerning the number of preexisting servers and shipments of servers,
measured data, and estimates of the power used by servers of different
classes. The quantification provided by the report can serve as a starting
point for increasing efficiency, says report author Lawrence Berkeley
National Laboratory scientist Jonathan Koomey. "The industry sees that one
of the first things you need to do to address the problem is to figure out
how big it is," says Koomey, who sees "a lot of opportunity for
improvement" in data centers. Companies have been aware that servers and
data centers used more power than they should, and the report "corroborates
our thinking and gives us quantification that might have been lacking
before," says Hewlett-Packard fellow Chandrakant Patel. Researchers could
use these results to contribute to a congressionally-mandated study on
power consumption. Google's servers, which are actually motherboards made
into custom servers, did not count as a server for this study, but Koomey
found that if included, Google's servers would increase the amount of power
consumed by 1.7 percent.
Click Here to View Full Article
to the top
Carnegie Mellon Software Steers NASA's Mars Rover
Carnegie Mellon News (02/13/07) Spice, Byron; Watzman, Anne
Initial tests have proved a success for autonomous navigation software
created by Carnegie Mellon researchers to control the NASA Mars Rover
"Opportunity." While previous autonomous software used to control
Opportunity was only able to turn the rover when it found itself up against
an obstacle, the CMU software creates a map of the terrain encountered by
Opportunity and enables the rover to retrace its course or chart a new
course if it needs to. To test the software, researchers directed the
rover to an area on Mars that was known to be free of obstacles, where a
"virtual keep out zone" was set up. The robot was able to make its way
around the zone without going into it, showing its ability to maneuver
autonomously around the craters of Mars. Earlier tests had only used the
software to create paths for the robot to travel along but did not actually
control the robot; the recent tests shows the software's ability to do
both. "Much more work and testing remains to be done, but we are thrilled
to see our software operating on Mars and we believe it will ultimately
expand the capabilities of this and future planetary rovers," said CMU
robotics professor Tony Stentz. Stentz had developed this particular
software, Field D*, for use on the CMU "Crusher" experimental unmanned
ground combat vehicle, but it had to be adapted to the weaker processor on
board Opportunity. Field D* is very computationally efficient, able to
create new path plans many times a second. The current rover's longevity,
four years compared to a life expectancy of a few months, has enabled the
testing of new capabilities.
Click Here to View Full Article
to the top
Blogs to the Rescue!
Nature (02/15/07) Butler, Declan
A policy paper written by two University of Maryland professors recommends
that the government incorporate Internet "community" tools to better deal
with disaster relief or similar situations. The online community, using
blogs, wikis, and other tools, could provide and share valuable information
that would improve the effectiveness of professional emergency response
efforts, say computer science professor Ben Shneiderman, founding director
of the Human-Computer Interaction Laboratory, and Jennifer Preece, dean of
the College of Information Studies. After the 2004 tsunami, the most
common way to coordinate damage assessment and support was through the
information being provided by volunteers using Web tools. Similar efforts
were seen in the aftermath of Hurricane Katrina, when many Web sites
emerged to keep track of missing people and relief efforts. The UM paper
calls attention to the lack of online reporting and networking incorporated
into Homeland Security's new Information Network for Disaster Response as
well as its online volunteer forum citizenscorp.gov. "If such systems were
formalized in whole or in part, the impact could indeed be enormous," says
the American Association for the Advancement of Science's Lars Bromley.
But, "It's entirely possible that [the plan] is simply too decentralized
and technically advanced for the relatively moribund .gov sector." A
project known as Instedd intends to create a decentralized global reporting
system for disease outbreaks. For such programs to have an impact, "A
sympathetic balance between local and central will be necessary," says
Preece.
Click Here to View Full Article
to the top
$82 for E-Voting Secrets
Wired News (02/16/07) Zetter, Kim
Princeton computer science professor Andrew Appel was able to purchase
five Sequoia e-voting machines from a government auction site for $82 and
has already demonstrated the ease with which the machines can be broken
into and compromised. His work is the first time a researcher has examined
one of Sequoia's machines without signing a non-disclosure agreement. The
20-year-old machines, known as the AVC Advantage, have ROM chips that are
in sockets, rather than soldered to the board, and while Sequoia claims
tamper-evident seals allow officials to make sure no tampering has
occurred, Appel's machines had no such seals. The manufacturer also claims
the machine itself knows what software it is supposed to run, and that
election management and tallying software at election offices would spot a
change in software. However, Appel claims that the only connection between
the machine and the district server would be through a cartridge where vote
totals are collected. Even if the machine cryptographically signed the
information placed on the cartridge, this signature would be stored in the
machine's ROM, making it accessible to a hacker. "Whatever the legitimate
software does to take checksums of itself can all be simulated by the
fraudulent software," says Appel. Despite the ease with which he gained
access to the machine's sensitive information, Appel believes the AVC
Advantage is more secure than the Diebold machine his colleague Ed Felten
was able to break into and compromise last year. Appel admits that hackers
would need to access tens or hundreds of these machines to impact an
election, but he points out that they normally sit unattended in churches
or schools the day before an election.
Click Here to View Full Article
to the top
Lawmakers Gear Up for Patent System Overhaul
CNet (02/15/07) Broache, Anne
Although patent reform is a rare issue in which where lawmakers from both
parties find themselves largely in agreement, the effort to reform patent
laws still faces obstacles. Several industry disputes have held up patent
reform recently, including that between the IT industry, which argues that
damages awarded for patent infringement should be based on the value of the
specific patent in question not the entire product, since most of their
products involve multiple patents, and the pharmaceutical industry, which
wants damages paid based on the value of the entire product, since most of
their products involve a single patent. House Judiciary subcommittee chair
Howard Berman (D-Calif.) has unsuccessfully proposed two different bills in
past years that would have improved the quality of patents awarded, and his
new bill is expected to be submitted within the next month. The bill is
designed to ensure that the patent office has the funding to hire a
sufficient number of examiners and staff, require patent applications to be
made public for a number of months so evidence of predating inventions can
be submitted, and to scrutinize business method patents, although not
eliminate them. Effective patent reform legislation would make it possible
for patent examiners "to get more appropriate information from people who
really know what is novel and what is obvious," says Brandeis University
economics professor Adam Jaffe. However, not everyone agrees with
strengthening patents. The Public Patent Foundation, made up of the
computer industry and free software representatives believes, "Software
should not be patentable, and neither should business methods," says
executive director Daniel Ravicher.
Click Here to View Full Article
to the top
ASU Helps Create the Real Face of George
Washington
Arizona Republic (02/16/07) Ryman, Anne
Researchers from the University of Pittsburgh and Arizona State University
were able to create life-size figures of George Washington using a
combination of anthropology, 3D scanning, and digital reconstruction. "The
whole idea is to put science, history, and art together and come up with
the absolute best guess of what he looked like," said James Rees, executive
director of George Washington's Mount Vernon Estate. ASU was responsible
for the 3D scanning technology that was used on objects such as a bust and
mask that were made of Washington when he was 53. Computer software was
developed with the help of an anthropologist to reverse the aging process
to generate a replica of Washington's face at various ages. ASU computer
science professor Anshuman Razdan, who lead the school's nine-member team,
says, "This was ... a once-in-a-lifetime opportunity." The digital images
were sent to a studio that made clay models that were then painted and
dressed. The 30-month project, whose idea was to put a human face on
history, was made possible by ASU's Partnership for Research in Spatial
Modeling (PRISM) and a grant from the Mount Vernon estate. Razdan says the
technology used to recreate Washington could be used on other famous
figures. He says, "You could do Lincoln, Jefferson, so many of our
Founding Fathers."
Click Here to View Full Article
to the top
Senator Introduces 'Disappointing' E-Vote Reform
Bill
InternetNews.com (02/14/07) Hickins, Michael
A Senate bill that would require a paper trail for all e-voting machines
is not receiving support from many voting activists who feel that it does
not go far enough to protect against fraud. The bill has been introduced
by Sen. Bill Nelson (D-Fla.), who said, "If Congress doesn't get this done,
I'm afraid our democracy could die from lack of legitimacy." Along with
mandating paper trails for e-voting and requiring random audits of paper
records against electronic counts in every voting district, the bill would
prohibit state election officials from working on candidates' campaigns.
It is virtually identical to a bill filed in the House last week, except
that the House bill allows exemptions for states that have mandatory
recounts in specific situations, or a "'Get Our of Audit Free' card," as
VotersUnite executive director John Gideon refers to it. Although
VotersUnite supports the elements of the bill that require hand audits,
disclosed source code, and the use of testing labs that are independent
from vendors, "we believe we have a duty to call attention to the bill's
unacceptable shortcomings" says Gideon. The group would prefer to see the
use of direct record electronic machines (DREs) banned in favor of paper
ballots. However, the CalTech-MIT Voting Technology Project has found that
most voters fail to check paper print-outs for accuracy. Advocates for the
blind are against banning DREs, because they allow the blind to vote
confidentially, but e-voting activists point out that it is a computer
interface, not the direct recording of votes by a computer, that allows for
this confidentiality.
Click Here to View Full Article
to the top
Photo Software Creates 3-D World
University of Washington News and Information (02/14/07) Hickey, Hannah
Computer scientists at the University of Washington and Microsoft are
developing software for organizing digital photos and compiling images from
common scenes into 3D virtual environments. Beyond organizing photos, the
software, dubbed Photo Tourism, could be used for virtual tours, recreating
events, or even developing a map from all the photos on the Internet. "You
might look at a photo and say I wonder what's just to the left of it, or I
wonder what's just to the right of it, or I wish I could expand the field
of view," says UW doctoral student Noah Snavely, explaining the frustration
that accompanies searching photo sites such as Flickr. Photo Tourism
utilizes recent innovations in computer vision: By examining details
shared by images, the software can tell that images are of the same scene
and fuse them together in 3D; the individual photos are then represented as
small squares placed where the picture was taken from in a 3Dsketch of the
complete scene. Looking through the images is meant to be game-like; when
viewing one image a user can click on a section of the screen to see images
of this area, or can choose from images tiled along the bottom of the
screen that contain parts of the current image. "I think it has the
possibility to be much, much richer than just a static 3-D model," Snavely
says. If the technology could be applied to Google Earth, users could
potentially zoom in continuously without resolution decreasing. However,
scaling the software to handle millions of photos would require a
tremendous deal more work.
Click Here to View Full Article
to the top
ACM Digital Library Now Exceeds One Million
Entries
AScribe Newswire (02/14/07)
ACM's digital collection has surpassed 1 million entries. ACM reached the
milestone in February with the entry of "Spam and the Ongoing Battle for
the Inbox," the cover story from the February 2007 issue of Communications
of the ACM, the association's flagship publication. ACM's digital
resources consist of the Digital Library, which includes the full text of
all material published by the association over the past 60 years, and the
Guide to Computing Literature, which offers a bibliographic database of the
key publications in the industry. "For people who design, develop and
manage information systems, the DL and Guide provide easy entry into the
massive amounts of scientific and technical information that drive
innovation," says ACM CEO John White. "With 1 million entries, ACM's DL
really is the definitive digital collection for computing." Each year,
50,000 items are added to the Guide, which includes journals, proceedings,
books, technical reports, dissertations, and requests for comment.
Click Here to View Full Article
to the top
Making Operating Rooms Safer With Open Communication
Among Equipment
UNH Media Relations (02/13/07) Potier, Beth
Researchers at the University of New Hampshire are looking to decrease
medical errors caused by miscommunication between operating room
instruments. "We're trying to get pieces of equipment that don't normally
talk to each other to do so," says project leader and professor of
electrical and computer engineering John LaCourse. Although major pieces
of equipment are computerized, they cannot share information. For example,
when a bed is raised a patient's blood pressure fluctuates, but the
monitor, which does not move, provides a faulty reading. Humans can
usually calculate a more accurate reading using mental calculations, "But
we want double-fault controls because there are peoples' lives at stake,"
LaCourse says. His team has been exploring the use of CANopen, a
communications protocol that uses a common hardware and software package
and is able to maintain the accuracy of the different proprietary
electronics involved. CANopen has been used in the automobile industry to
make computerized parts of a car that were manufactured separately work
together. For LaCourse, the biggest challenge has been to get information
from equipment manufacturers who are trying to protect their intellectual
property. The team is now focusing on what they call "closing the loop ...
We're trying to see if we can not only get the bed and the monitor to talk
to each other but also control each other," says LaCourse. He hopes that
manufacturers will eventually install CANopen in all operating room
equipment, which would alleviate the need for personnel to calibrate the
instruments as they must do now.
Click Here to View Full Article
to the top
Haptic Glove to Touch on Virtual Fabrics
New Scientist (02/13/07) Simonite, Tom
Researchers in several European countries are collaborating to develop
"virtual" fabric that combines a "haptic" interface with a "touching"
interface. The "haptic" interface is an exoskeleton glove with a
mechanical control system that simulates manipulating fabric by exerting a
force on the user's fingers, and the "touching" interface is an array of
pins located under each finger that simulates the tactile sense of a
fabric's texture. "Nobody has ever linked a 'haptic' device with a
'touching' device," said project Nadia Magnenat-Thalmann of the University
of Geneva. "It should lead to a much more realistic experience." A
virtual fabric operates on two levels: A "global" model of its properties,
and a more detailed model of the area in "contact" with the glove. To
fulfill the second level, a realistic touch interface would ideally be able
to change 500 times per second, but the researchers have only been able to
make one that changes 40 times per second. The tactile arrays consist of
24 pins in a space of one square centimeter that are moved up and down by
piezoelectric actuators. "We're not trying to replicate the topology of
the surface, but to provide the right stimuli to the touch receptors in the
fingertips, which are about 0.5 mm to 1 mm apart," explains Exeter
University's Ian Summers. Tests have proved successful, allowing users to
differentiate between virtual fabrics.
Click Here to View Full Article
to the top
'Virtual Dublin' R&D Project Gets Euros 2.5m SFI
Grant
Silicon Republic (02/14/07) Smith, Gordon
Computing gaming, assistive technologies, and urban planning are expected
to benefit from the creation of a computer model of Dublin, Ireland, that
will be made to scale and offer a level of life-like street scenes, crowds,
and traffic noise that has never been rendered before. The 'virtual
Dublin' project will make use of a scalable simulation server that is
capable of streaming the virtual environment to a game console, mobile
phone, or other consumer device, as well as the IBM CELL Broadband Engine
processor that is found in the Sony PlayStation 3 console. Researchers
from Trinity College Dublin (TCD) are pursuing the project, which is called
"Metropolis." The team consists of computer science professors Carol
O'Sullivan and Dr. Steven Collins, neuroscience researcher Dr. Fiona
Newell, and mechanical and manufacturing professor Henry Rice. "The aim of
the research is to simulate large crowds consisting of millions of people
and to introduce a high level of variety in animation, appearance, and
sound," says O'Sullivan. In addition to IBM and Sony, Europe's Team Soho,
Creative, Havok, Demonware, OC3 Entertainment, and the Environmental
Protection Agency are involved in the initiative. The Science Foundation
of Ireland has provided 2.5 million euros to fund the project.
Click Here to View Full Article
to the top
IC Designers Need to Focus on Security, Panel Says
EE Times (02/14/07) Mokhoff, Nicolas
A panel of algorithm experts and hardware specialists at the International
Solid State Circuits Conference agreed that circuit designers must do their
part in securing digital systems. "Security is only as strong as the
weakest link in the system," said Katholieke Universiteit Leuve (Belgium)
professor Ingrid Verbauwhede. "Mathematically very strong algorithms have
been and are being developed. However, if the key leaks from the
integrated circuit, this will be the weakest link." Algorithms such as
AES, DES, and triple-DES must be implemented into ultra-low power platforms
such as RFID tags as well as high-throughput platforms like Gigabit IP
routers, but area and power costs must be kept under control. Given recent
innovations that are enhancing the speed and power of cryptographic
hardware, these systems are virtually impregnable if they have a long
enough key. "However, digital systems are vulnerable to side-channel
attacks that deduce information by monitoring side effects of the
encryption process," said IBM engineer Norman Rohrer. Public-key
cryptography needs to be further developed, "requiring advanced algorithms
and design techniques," said Oregon State professor Cetin Kaya Koc. One
possible solution is to create a secure ASIC, but the challenge is that
while developers improve defense measures and make use of new technologies,
hackers share information and learn from past attacks. A successfully
secure ASIC would have to stay ahead of hackers while meeting the needs of
manufacturing, maintainability, and cost.
Click Here to View Full Article
to the top
Teraflops Chip Points to Future
BBC News (02/12/07)
Intel has developed a new Teraflops chip, but in order for the chip to
lead to more powerful processors programmers will need to provide
instructions for a large number of cores that work in parallel with one
another. The chip, which can perform more than 1 trillion calculations per
second, has 80 processing cores, which would be difficult for desktop
applications to handle. In comparison, desktop computers have up to four
separate cores, and the Cell processor in PlayStation 3 game consoles has
eight. "It's going to require a revolution in software programming," says
Dr. Mark Bull of the Edinburgh Parallel Computing Center. Justin Rattner,
Intel senior fellow and chief technology officer, says the chip has huge
implications for multi-core and parallel computing. "It points the way to
the near future when teraflops-capable designs will be commonplace and
reshape what we can all expect from our computers and the Internet at home
and in the office," he says. The chips could be used for artificial
intelligence, instant video communications, photo-realistic games, and
real-time speech recognition.
Click Here to View Full Article
to the top
Quantum Computing at 16 Qubits
Nature (02/15/07) Brumfiel, Geoff
Many experts are doubting that the 16-qubit quantum computer demonstrated
by D-Wave this week can be scaled up to a size that would make it useful
for real-world applications. In its present form, the machine is thought
to be "completely useless from an industrial perspective," says Waterloo
University computer scientist Scott Aaronson. The quantum computer was
shown to be able to solve a sodoku puzzle and look for matches in a protein
database, but it is about 100 times slower than conventional computers,
admits D-Wave chief technology officer Geordie Rose, who calls the current
machine "a very early prototype." He believes that scaling up the
technology will be possible, and claims to have several parties interested
in the machine. D-Wave is not the first to have qubits interact to execute
simple operations, but the company has made more qubits interact that
anyone so far. Although the system should be easy to scale up because it
is constructed using techniques from the semiconductor industry, "the issue
isn't how many qubits, it's how many well-controlled qubits," says
University of Oxford quantum computing expert Andrew Steane. During
operation, the qubits are not individually controlled as they are in other
machines, and if left alone the current of each one can simultaneously flow
both clockwise and counterclockwise around a circular superconductor,
producing two values at the same time. Many worry that if D-Wave's
prototype is released too early and does not live up to expectations, the
field of quantum computing could lose credibility.
Click Here to View Full Article
to the top
Spin Doctors
Engineer (02/12/07)
European researchers involved in the Nanospin project see spintronics as a
way to improve the speed and memory storage capacity of computing devices.
Experts from eight academic and industrial organizations plan to use
gallium manganese arsenide, a ferromagnetic semiconductor, to develop new
kinds of spintronic nanoscale devices. They will use the low-temperature
ferromagnetic semiconductor to test the spintronics approach, but the
technology will ultimately need to work with room-temperature
semiconductors. With spintronics, the idea is to bring the properties of
electronic and magnetic devices together, which would enable permanent
memory and its control electronically, explains University of Wurzburg
physics professor Laurens Molenkamp. "Programmable logic applications
could enable you to have your computer applications instantly available,"
says Dr. Charles Gould, a post-doctorate researcher involved in the
project. "You wouldn't have to wait for it to boot up, you just store
where you were in memory at night, and the next morning you start
immediately where you left off." Aside from spintronics, carbon nanotubes
and organic electronics are other technologies being considered for making
computing devices that are faster and use less power.
Click Here to View Full Article
to the top
Printed Electronics Set to Be Next Big Thing
Electronics Weekly (UK) (02/14/07)
New markets are expected to be birthed from printed electronics, and most
of their potential lies in smart media products (SMP), according to Toppan
Forms. SMP will boast intelligence and mass production while also being
customizable, says the IDTechEx technology consultancy. Printed
electronics will most commonly take shape as tape, posters, packaging,
patches, and "wallpaper," and will be put to use in situations where
traditional technology is infeasible. Hybrid organic/inorganic structures
hold the most potential for printed electronics because they combine low
cost with rapid printing technology. The success of most printed
electronics applications will ride on their ability to perform new
functions and create new markets rather than supplant current solutions.
The biggest opportunity for printed electronics is for devices on flexible
paper or polymer substrates, not just for their low cost but for their
physical suitability for largest-volume applications such as smart
packaging, smart labels, books, signage, newspapers, billboards, and
posters. Low installation cost is also an advantage. The most significant
announcement this year has been the raising of $100 million by Plastic
Logic to establish a plant to manufacture flexible electrophoretic displays
for "take anywhere, read anywhere" electronic reader products, employing
Plastic Logic's process to assemble active-matrix displays using thin,
lightweight, and robust printed transistor backplanes.
Click Here to View Full Article
to the top
A Race to the Technological Frontier
Federal Computer Week (02/12/07) Vol. 21, No. 3, P. 28; Rogin, Josh
The Defense Advanced Research Projects Agency's (DARPA) robotic vehicle
competition, now entering its third year, has shifted its focus from
autonomous vehicles racing through the desert to an urban setting. The
goal is to make unmanned driving a reality so that one-third of the Defense
Department's military ground vehicles can go robotic, thus reducing
battlefield casualties; a side benefit of this technology is systems that
can lower the 43,000 yearly civilian deaths attributed to vehicle
accidents. The 2007 Urban Challenge will set up a model downtown area
complete with paved roads, traffic circles, stop signs, and sharp corners,
and all participating vehicles will be on the course simultaneously instead
of racing in turns. Each vehicle must execute a set of military supply
missions using only sensors and software for guidance, while also dealing
with each other, DARPA-supplied traffic, and other moving obstacles. A $2
million, $1 million, and $500,000 jackpot awaits the teams that finish the
60 miles of assignments in first, second, and third place. Some teams
participating in the Urban Challenge are not competing strictly for the
money, but for the opportunity to cross-pollinate other research
applications with the lessons they take away from the contest. Still other
teams have entered the Urban Challenge with new product development in
mind, and at least one team is being sponsored by every major auto
manufacturer. Stanford University researcher David Stavens says the
biggest challenges of this year's competition will be compensating for the
loss of GPS reliability in an urban environment and recognizing moving
obstacles.
Click Here to View Full Article
to the top
Can't Touch This
Fast Company (02/07)No. 112, P. 86; Penenberg, Adam L.
New York University Courant Institute research scientist Jefferson Han has
devised a touch-screen interface that is potentially revolutionary, because
it carries out actions according to movement and tactile pressure from
multiple inputs. Han founded Perceptive Pixel to market his invention.
His prototype system supplies a light source using a piece of clear,
retrofitted acrylic with LEDs mounted on the side, while an infrared camera
is positioned at the rear; when one places one's fingers on the surface,
some light bounces straight down and the camera captures the light image.
Applying more pressure causes the camera to capture more data, and Han
designed software that measures the shape and size of each contact and
assigns a set of defining coordinates. "One thing that excited me about
Jeff Han's system is that because of the infrared light passing
horizontally through the image surface itself, it can track not only the
position of your hand but also the contact pressure and potentially even
the approach of your hand to the screen," notes Douglas Edric Stanley,
professor of digital arts at the Aix-en-Provence School of Art. Han has
developed mapping software and an application for manipulating photos as
proof-of-concept demonstrations. "Touch is one of the most intuitive
things in the world," explains Han. "Instead of being one step removed,
like you are with a mouse and keyboard, you have direct manipulation."
Click Here to View Full Article
to the top