Since the hyperbole of
the title has attracted your attention, I want to take you on a stroll down
memory lane to the year 1984.On
January 1,1984 the Bell System divested itself into a flock of Baby Bell
companies.In a 1984 Super Bowl
commercial, the Apple Corporation announced a new personal computer, which they
called "The Macintosh".Ronald
Reagan was running for his 2nd term as President against Walter Mondale.The 1984 Summer Olympic Games were held in
Los Angeles
and were boycotted by the
Soviet Union
.Carlo Rubbia and Simon van der
Meer won the Nobel Prize in Physics for the discovery of the Z0 and W±
bosons, and quantum mechanics pioneer Paul Dirac died.Tom Clancy published The Hunt for Red October, and William Gibson published
Neuromancer.The Bhopal Disaster
at a Union Carbide pesticide plant in India
killed 8,000 people outright, with the death toll ultimately rising to over
23,000.There was no
Internet as we know it in 1984, and the World Wide Web and web browsers would
not come on the scene until the 1990s.
In 1984 I was a
Professor of Physics at the University
of Washington
in Seattle (UW).At that time, I
was engaged in research in experimental nuclear physics with heavy ions, using
accelerators at the UW Nuclear Physics Lab, Oak Ridge National Laboratory in
Tennessee, and the Lawrence Berkeley Laboratory in California
.I was among the first generation
of physicists who leaned to use computers as a principal tool in physics
research.I had developed one of the
first online computer systems for data collection, an SDS-930 system at the UW
Nuclear Physics Lab, and I had written some important nuclear modeling software.In the physics community I had acquired something of a reputation as a
physics computer jock.
In the early 1980s Cray
Research, Inc. of Chippewa Falls,Wisconsin
began producing wonderful supercomputers, in particular the Cray X-MP
supercomputers costing about $4-8 million.These
supercomputers were being sold to the U.S. Department of Energy weapons labs at
Livermore
and
Los Alamos, to the National Security Agency and other secret intelligence organizations,
and to private industries with sufficient cash, like big insurance companies.
TheCrays were beautiful objects, the likes of which the computer industry
had never seen before.Form followed
function to make them tall C-shaped towers hiding long equal-length wire
interconnect bundles in the middle, with low boxy projections around the base
containing cooling units and power supplies that formed padded bench seats
around the C-ring.Programmers in
need of physical contact with the object of their worship could sit there on the
Cray's bench and lean back against 400 mega-flops of warm, high speed computing
power.
Unfortunately, because
of the cost, these supercomputers were essentially unavailable to university
researchers.A group of academic
theoretical physicists, like hungry urchins in the snow, staring with envy
through frost-streaked windows at the feast of massive calculations that their
national-lab competitors could do, petitioned the National Science Foundation
(NSF) to do something about this situation by setting up supercomputer centers
at which university researchers could have access to these new powerful
machines.In response, in late 1983
the NSF created an ad hoc advisory
committee to study the problem and make recommendations.Several of the computer-hungry theorists were appointed to the Committee,
along with a Livermore Cray-operations manager, and two experimentalists.I was one of the experimentalists.The
chair of the Committee was the late Prof. Herb Chen of UC Irvine, a neutrino
experimentalist who was the inventor of the SNO heavy-water solar neutrino
detector (which he did not live to see in operation at a deep mine in Sudbury,
Canada or share in the Nobel-Prize-winning work that was
done with it).I had the distinction of being
the only member of the Committee who was NOT based at or near a location that
might become a NSF Supercomputer Center
.
Our Committee met at a
variety of locations in the early months of 1984, at Washington, DC,
at O'Hare Airport, at UC Irvine, and at the University
of
Houston.We had a long agenda.The other members of the committee were concerned with the mandate of the
supercomputer centers: how many Centers with what geographical distribution,
what kind of supercomputers to specify, how to apportion the available computer
time to users and hosts, how to handle tapes and data input/output, how to accommodate
on-site visitors and remote users, how many staff members the Centers would have
and how much time for personal research they would be allowed, how tightly the
Centers would be coupled to a sponsoring organization like universities or
national labs, how they would be governed, how much overhead they would pay,
etc.
On the other hand, my
concerns and my input at the committee meetings were focused on how remote users
like me, far away from all of the proposed sites, would be able to access the
supercomputers, submit their jobs, and receive their results.The University of Washington in Seattle is in the remote northwest corner of the nation, and I would be thousands of
miles away from any of the proposed Supercomputer Centers.I envisioned calculations that needed to receive the input of
gigabytes of experimental data and to generate outputs of gigabytes of graphs,
plots, figures, and multi-dimensional matrices.A dialup modem on a long-distance telephone line was not going to do the
job for that kind of computing.
So how could fast
supercomputer access be provided to remote off-site users like me?The US
military had already established ARPANET, a TCP/IP-based computer network that
our physics colleagues who had ties to the weapons labs were already using for
e-mail, computer access, graphics, and high speed data transfer.I proposed and promoted the idea to the Committee that, along with the
establishment of five or so supercomputer centers, the NSF needed to create the
equivalent of ARPANET for the supercomputer users, so that mere mortals without
weapons-lab connections could use it to access the new supercomputer systems.
There were, of course,
many points in need of extended discussion by our Committee.Our meetings droned on, hour after hour, presentation after presentation,
meeting after meeting, until the final meeting at which the chairman had to
write the consensus committee report.Rather
to my surprise, the creation of an ARPANET-clone called NSFNET emerged as a
bulleted item in the committee's report.Herb
Chen did an excellent job of writing the draft report, we made comments and
corrections, and it was submitted to the NSF.I cynically believed that it would vanish without a trace beneath the
murky waters of Washington bureaucracy, but that did not happen.
The following year, the
Supercomputer Center Initiative appeared as a line item in the NSF's proposed
budget.Senator Al Gore of Tennessee, who was then on the right committees of the U.S. Senate, was interested in
cutting-edge technology.He was
nudged by his constituents from Oak Ridge National Laboratory and the University
of
Tennessee, who told him and his staff that this NSF Supercomputer Initiative was a very
good idea.There was some resistance
in Congress to the idea that the supercomputer centers needed a network, but
Gore and his staff got behind overall initiative and pushed very hard.The result was that five new NSF Supercomputer Centers were created in
Princeton
,NJ,
Ithaca, NY,
Pittsburgh, PA,
Urbana-Champlain,
IL, and San Diego, CA, and the NSFNET connecting them to universities and other users became a
reality.
As it turned out, I
never actually used any of the supercomputers at the NSF Supercomputer Centers
that I had helped to create.After a
few years, with the ongoing exponential growth of microcomputer speed and
capacity, the 400 megaflop 105 MHz Cray-XMP supercomputers, which had looked so
fast and tantalizing in 1984, lost their speed and power edge to wave after wave
of faster microcomputers with Intel and AMD processors.The desktop computer on which I am writing this column has four
processors, a clock speed of 2.6
GHz, and 8 GB of memory, and it could run rings around a Cray X-MP, if any were
still around.
But unlike the
supercomputers that it connected, NSFNET did not reach obsolescence so rapidly.The NSF wisely decided to open NSFNET up to all academic users, whether
they were supercomputer users or not, and NSFNET, now linked to ARPANET, became
a principal backbone of the new Internet that came into being in the late 80s
and early 90s. Thus, NSFNET turned out to be a perfectly-timed key federal
investment in infrastructure that fostered a whole new generation of economic
growth and technical developments for the nation and the world.
Particle-physicist-computer-jocks
at the CERN Laboratory in Geneva
built on the TCP/IP protocol of the Internet to create the World Wide Web
(WWW), a system of computers and servers that could rapidly exchange and display
information.The arrival of WWW
further boosted the role of the Internet, giving it a new focus, and
revolutionizing computing and communications.After a visit to CERN, in August, 1993, I used the UNIX workstation in my
UW office to establish what was turned out to be the 3rd WWW server and web site
in the State of Washington, beating out Microsoft, Amazon, and the UW Computer Science Department.(However, my physics colleague, Prof. Vladimir Chaloupka, beat me by a
few months, setting up web sites one and two for himself and the Particle
Physics Group at the UW.)My AV
column "The Bandwidth Revolution: The Internet and WorldWideWeb" that
appeared 19 years ago, in the March-1994
issue of Analog, described the web, my new server, and the Internet in
detail from the perspective of the time. At
the 1993 World Science Fiction Convention in San Francisco, the largest meeting room of the WorldCon was filled to overflowing with SF
fans who wanted to hear the panel discussion about the Internet (chaired by my
wife Pauline) and to learn what this new WWW thing was. By
1994 the web had become truly world wide, and it has continued to grow ever
since.The deluge of
Internet-related developments that has followed: search engines, internet
retailing, the dot-com boom and bust, music file sharing, You-Tube, the Arab
Spring, crowd sourcing,social
networking, etc., were both inevitable and unanticipated.
Ultimately, in 1995
after 10 years at the cutting edge of digital communication, NSFNET was retired.
It had done its job, and it had been
superseded by an explosion of commercial networks using newer and more powerful networking
hardware.And the rest is history.
So that's how Al Gore
and I invented the Internet.Many
other individuals and groups must have similar stories to tell about how they
contributed to one piece or another of the vast Internet system, and their
contributions may well have been more important than ours.But without NSFNET as a part of the NSF Supercomputer Initiative, the
initial stages of the Internet would have lacked a major backbone, and
development would have been much slower and perhaps would have proceeded in a
different direction.
So I'll be glad to
autograph the backs of your Internet-intensive iPhones and iPads and Android
devices, if you ask me nicely.
John G. Cramer's 2016 nonfiction book (Amazon gives it 5 stars) describing his transactional
interpretation of quantum mechanics, The Quantum Handshake -
Entanglement, Nonlocality, and Transactions, (Springer, January-2016) is
available online as a hardcover or eBook at: http://www.springer.com/gp/book/9783319246406
or https://www.amazon.com/dp/3319246402.
Alternate View Columns Online: Electronic reprints of 212 or more
"The Alternate View" columns by John G. Cramer published in Analog
between 1984 and the present are currently available online at: http://www.npl.washington.edu/av
.
