IBM goes grid

[Eugene Leitl]

Buzzword content: high.

-- Eugen* Leitl <a href="http://www.lrz.de/~ui22204/">leitl</a>
______________________________________________________________
ICBMTO  : N48 10'07'' E011 33'53'' http://www.lrz.de/~ui22204
57F9CFD3: ED90 0433 EB74 E4A9 537F CFF5 86E7 629B 57F9 CFD3

http://www.nytimes.com/2001/08/02/technology/02BLUE.html?pagewanted=print

AUG 02, 2001

I.B.M. Making a Commitment to
Next Phase of the Internet

By STEVE LOHR

I.B.M. is announcing today a new initiative to
support and exploit a technology known as grid
computing, which the company and much of the
computer research community say is the next
evolutionary step in the development of the Internet.

The grid vision is that everyone at a desktop machine
or hand-held computer could eventually have the
power of a supercomputer at his or her fingertips, by
amassing the processing power and information
resources attached to networks. Although the idea
has been around for some time, the types of computer
hardware and software to achieve it are only now
coming within reach.

I.B.M. is placing no dollar figure on its grid initiative.
But the company is comparing the program to its
earlier move to support the Linux operating system,
an effort that it announced at the start of 2000 and
later said it would spend $1 billion on over the next
couple of years, and its backing is expected to push
interest to new heights.

As part of its campaign, I.B.M. is also announcing
today that it has won two national grid projects in
Europe, one in Britain and another in the
Netherlands. The I.B.M. grid initiative will be led by
Irving Wladawsky-Berger, an executive with a
research background who has close ties to university
and government laboratories. Mr. Wladawsky-Berger
is also heading the company's major support for
Linux, a freely distributed operating system that is
increasingly used to power data-serving computers on
the Internet and inside corporations.

Grid computing - a concept that originated in
supercomputing centers - holds the promise of
transforming the Internet, according to some
computer scientists. At present, the Internet is used
for communication, mainly e- mail and instant
messaging, while the Web is the Internet's
multimedia retrieval system, enabling computer users
to have access to text, images and music.

The grid would add a new dimension."The goal is that
grid becomes the computing engine for the Internet
in the way that the Web is the information engine,"
said Ken Kennedy, a professor at Rice University.
"The real long term is that this becomes the
problem-solving mechanism for society."

The dream of computing power as an electricity-like
utility, available anytime and anywhere, to help solve
all manner of human problems is both decades old
and not likely to be fully realized anytime soon. The
grid takes its name from the utility analogy, which
first surfaced in the 1950's.

The notion of computer resource- sharing to
augment human intelligence dates back at least four
decades to J. C. R. Licklider, who in 1960 wrote a
classic paper, "Man-Machine Symbiosis," and to the
time- shared computing experiments of the early
1960's at the Massachusetts Institute of Technology
and elsewhere.

But what is new about grid computing is that the
vision is now within reach, at least for some
government and university research labs, because of
continuing advances in processing power, network
capacity and software.

One grid technology is distributed computing, whose
best-known application is probably the SETI at home
program, begun in 1999, which harnesses the power
of a couple of million personal computers worldwide
to help look for signs of extraterrestrial intelligence.

But grid technology extends beyond harvesting
unused computing cycles. It also involves sharing big
database files and application programs across
high-speed network connections, so that researchers
in far-flung locations can collaborate on complex
projects including climate modeling, high-energy
physics, genetic research and earthquake simulations.
The early test bed for grid projects is in research labs,
as it was for the Web and browsing software.

In the United States, the Departments of Defense and
Energy, the National Science Foundation and the
National Aeronautics and Space Administration are
all financing grid projects. In Europe, a flurry of
ambitious grid programs are under way or in the
works.

Britain is building a national grid, linking nine
research centers from Southampton to Edinburgh to
Belfast, and I.B.M. is supplying key parts of the grid
including a sophisticated data storage operation at
Oxford. The British government recently allocated
roughly $170 million for its national grid, and the
applications will include exploring the mysteries of
particle physics, genome research and medical
informatics.

The biology applications, according to Tony Hey, a
computer scientist and architect of the British grid,
range from exploring the phenomenon of protein
folding to devising less-invasive methods of surgery.
"The last thing a hospital wants to do is run its own
supercomputer center," Mr. Hey said. "But the grid
allows you to set up dynamic virtual organizations to
move quickly to solve deep problems requiring a lot of
computing resources."

Perhaps the most important enabling technology for
grid computing - and its biggest hurdle - is the
software. It must manage and coordinate the sharing
of databases, applications and computing power
across the network, and do so reliably and securely.

What is emerging as the de facto standard for this
layer of communications and collaboration software
comes from the Globus project, a development effort
led by Ian Foster, a senior scientist at the Argonne
National Laboratory, and Carl Kesselman, director of
the center for grid technologies at the University of
Southern California's Information Sciences Institute.

The Globus project, begun in 1996, is developing its
software following the open-source model, in which
computer code is openly shared, allowing
programmers to modify, improve and fix the
software. It is the same approach to software
development used by the Apache Web server project
and Linux. "We are rather like the Linux community,
open source and open standards, with more and more
people cooperating," Mr. Foster said.

As a business strategy, I.B.M. has become an
enthusiastic supporter of open source projects like
Linux. The company sees its competitive advantage in
computer services and specialized software instead of
at the operating system level, where Microsoft
(news/quote)'s Windows and Sun Microsystem's
Solaris are the leaders.

But Microsoft and Sun are embracing grid initiatives
as well. Microsoft has just contributed $1 million to
the Globus project for research on putting grid tools
in Microsoft's software like Windows and its .Net
Web services software. And Sun has a grid software
offering for use within corporations, and last week it
announced that it was making the software an open
source project.

Speaking of grid technology, Peter Jeffcock, a Sun
manager, said, "The productivity gains are huge and
we think it's inevitable."

Indeed, companies like Pfizer (news/quote), Ericsson
(news/quote), Hitachi, BMW, Glaxo, Smith-Kline
and Unilever (news/quote) are experimenting with
internal grids at the moment. Only big companies
with deep pockets and high-speed Internet
connections are likely to be interested at first. And
they are awaiting a more robust grid infrastructure,
which will require more software development in
particular.

But I.B.M. certainly is betting that will happen, and
with its initiative it is hoping to take a leadership
position. "As grids go commercial use, we think
everyone will jump in," Mr. Wladawsky-Berger said.

The grid community, it seems, is welcoming the
attentions - and money - that business can bring.
"The grid concept has really captured mindshare in
the academic science and engineering community,"
Mr. Foster said. "But we have to get the commercial
interests to get involved with resources and
investment."