Supercomputing
While you were sleeping...
01/00
IBDS
It sounds ridiculous, but it's already
becoming a reality…
The sheer volume of data and calculations
now required for science projects have either proved too great
for the most powerful computers available, or smashed the
computing budget of the organizations concerned. The solution
to this problem - and the reason why you and I come into the
equation - is pure genius in its simplicity.
For many hours of the day and night millions
of computers in homes and offices around the world remain
switched on but essentially idle, with nothing better to do
than run a screen saver of flying toasters. So why not make
use of all that spare capacity, by joining as many of these
computers together as possible and creating immense computing
power in new internet-based supercomputers?
Under the not overly imaginative name of
Internet Based Distributive Supercomputing (IBDS) that's
exactly what's begun to happen over the last few years. The
huge amounts of data are broken down into small packets or
work units which are then sent, via the internet, to
individual computers around the world which have installed a
simple piece of software called a client.
When these work units are downloaded they
can be determined by the user through the client - the
computer then analyses them when it is idle (and probably when
you're in bed) before sending them back to the central
computer where they're collated with the other finished work
units.
Often the client will incorporate a
screensaver interface which keeps the user informed on its
progress and how much time and work has been contributed. In
some cases it could be an individual computer which
effectively does the crucial piece of analysis and solves the
problem.
ALIEN HUNT
The vast majority of internet users have yet
to come across IBDS but in the next year or so it's set to
really take off, unleashing the collective power of the
internet. The handful of IBDS projects (all US-based) that are
around at the moment can roughly be divided into two groups.
The first group have been around for longer and essentially
deal with abstract or number problems - often for cash prizes,
although that belies a more serious purpose.
The second group - and those that will be by
far the biggest supercomputers in the future could be said to
have the 'vision thing' - science projects which capture the
imagination of the public by the sheer wonder of what they
promise.
The one IBDS which has so far achieved this
more than any other is the SETI@home project. Researchers for
the Search for Extraterrestrial Intelligence (SETI) http://www.setiathome.ssl.berkeley.edu
at the University of California in Berkeley are receiving
signal data from the huge Arecibo radio telescope in Puerto
Rico in the hope that they might pick up messages from little
green men beyond our solar system. But to have a chance of analyzing the data in depth and in a realistic time scale
they're breaking it up and sending it to computers all around
the world for analysis.
The project therefore offers the small but
tantalizing prospect that one of its computers - and that
includes yours - could confirm the existence of
extraterrestrials.
This unlikely final goal of the project
already seems to have become a side issue for participants.
They seem more interested in fighting out for places in the
project's league tables of who has contributed the most
computer time and work units. In fact, many individuals band
together into groups to improve their chances of recognition
on SETI@home's wonderfully informative website.
TINY POWER
The SETI@home project will have convinced
many that science and IBDS are perfect bedfellows and as a
result there are bound to be many organizations which seek to
replicate its achievement. One of these is a group aiming to
build a computer ten times smaller than a human blood cell
that will perform one thousand trillion computations per
second - in other words a billion times faster than the
average desk top. Although this sounds a little unbelievable,
it's still probably a lot sooner in coming than contact with
aliens.
The computer will probably be one of the
first devices built using what is being described as 'the
technology of the 21st century', which proponents say will
bring a revolution as massive in its implications for society
as the industrial revolution. In other words nano technology -
building goods molecule by molecule.
The NanoComputer Dream Team - a non-profit
group - was born out of a $100 bet made in 1995 between nano
enthusiast and publisher of Hawaii-based Nano Technology
Magazine Bill Spence and nano skeptic Dr. Brad Cox. The bet?
The likelihood of the world's first nano computer being built
by 2011. Dr. Cox believes that nano technology will go the way
of predictions made about the imminence of artificial
intelligence a few years ago.
In the same year Spence picked up the
gauntlet and through appealing on his magazine's website soon
amassed a team of over 500 volunteer specialists. But he realized that know-how alone would not be enough to develop
the nano computer. His plan therefore is to attract several
million computer users to help build the world's largest
internet-based supercomputer.
"I see no reason why this idea could
not prove more popular than SETI," enthuses Bill Spence.
"Who needs aliens when you consider the possibilities of
nano technology? I believe it's going to allow everything from
widespread material opulence to blowing open the colonization
of space.
$100 BET
"I made a bet for $100 dollars with Dr.
Cox that the nano computer would be built before 2011 but I
expect we will have a working nano computer before then.
"A few years before 2011 you will be
seeing a NanoComputer Dream Team conference and we will parade
Dr. Cox on stage as guest of honor - all expenses paid of
course!
"You will be able to fit several
hundred rod logic switches (the envisaged building blocks of
the nano computer), each as powerful as a mainframe computer,
into an area the size of a red blood cell. In the space of a
desktop computer using nano technology you will have more
computing power than is presently on Earth."
He adds: "This project is also about
how the average Joe Bloggs without a science background can
contribute to nano technology. You don't have to be an
Einstein to contribute spare computer cycles. I'm proud to say
I'm the least qualified member of the Nano Computer Dream
Team!"
To find out more about the NanoComputer
Dream Team and the weird and wonderful world of nano
technology, visit the home page of Nano Technology Magazine at
http://planet-hawaii.com/nanozine.
CASH PRIZES
Back to present reality - and what of that
first group of IBDSs, essentially dealing with abstract and
number problems? One thing which can be said about them is
that at least they have achieved some of their stated aims.
One of these http://www.distributed.net,
or D.net for short, says the ultimate aim of what it does is
to further the cause of IBDS, as the name might suggest.
"As an organization our focus is on the
technology of IBDS and our goal is to figure out the most
effective way of using it and then making this information
available to people," says McNett. " We were the
first group to really approach distributive computing from the
technology standpoint, not the goal standpoint, and we are the
largest in that category."
D.net was established in early 1997 and
currently has around 50,000 users. Until recently it has
essentially been concerned with cracking encryptions or coded
data. Although this might at first sound dubious, it isn't.
Basically companies release the encryptions
and offer cash prizes of thousands of dollars to break them,
the point being that they want to prove that stronger encoding
methods should be allowed by governments to make transactions
and communications more secure online.
But while D.net is continuing in this vein
it is also branching out into other areas. It is currently
working on developing something called Optimal Golomb Rulers (OGR)
- very precise forms of measurement which can be used to
develop things like radio telescopes. Also this year D.net
will be contributing its computing power to the Human Genome
Project, the global effort to identify all the genes which
make up human DNA, and which is essentially driving all the
debate about genetics at the moment.
GIMPS
Perhaps the oldest ongoing (established in
1996), and the most abstract IDBS is GIMPS or Great Internet
Mersenne Prime Search, http://www.mersenne.org/prime.htm.
Again the stated ultimate aims are along the lines of
advancing IBDS technology and data privacy, but this time the
subject is to find incredibly long prime numbers - and this
time the cash stakes are higher - hundreds of thousands of
dollars.
Finally, what of the of future IBDS? D.net
spokesman David 'Nugget' McNett:
"In the future distributed computing is
going to become more and more powerful, flexible and useful as
bandwidth becomes more plentiful. It's almost unfair to be
asked how distributed computing will develop - you could say
its only limited by the bounds of imagination. Going back to
1978 when Steve Jobs and Steve Wozniak were building the first
Apple II computer in a garage, they would have had no idea of
the sort of tasks computers are doing today.
"Distributed computing is a tool that
computer geeks like me are making possible, but we don't know
what people are going to be building with it. Eventually I can
see everyone on the Net running IBDS software of one sort or
another. The more painless that software becomes to run, then
the more people will be attracted."
With the numbers on the Net seemingly
doubling year on year and with current estimates upwards of
300 million internet users, that's quite an awesome prospect.
by
the BBC
TOP
