taylor@hplabs.hp.com (Dave Taylor) (08/31/88)
This is a copy of the presentation made made by Kenneth M. King, the
president of EDUCOM, to the Science, Technology and Space Subcommittee
of the Senate Committee on Commerce, Science and Transportation on
August 11, 1988.
Summary of remarks (not presented):
There is a broad consensus among government, education, and
industry leaders that creation of a high-speed national research and
education computer network is a critical national priority. The benefits of
such a network will be realized in greater research productivity and
improved access to specialized knowledge databases and experimental
facilities, such as supercomputers, accelerators, and telescopes. The
network will accelerate the transfer of research results to the private
sector and substantially benefit our international economic
competitiveness.
Critical to the success of this effort are government leadership
and a stable source of funding during the development phase of the network.
Effective leadership from the federal government only partially exists
today, and funding for major elements of the effort is either nonexistent
or undependable. Federal participation is essential to ensure compatible
network development as well as access to the network by the broadest
possible education and research community, including industrial researchers
and researchers at smaller colleges and universities.
EDUCOM, an organization with more than 550 university members as
well as 100 corporate associates in the information technology area, has
been working to achieve a national instruction and research network as a
major priority. Our members, from higher education and industry, are eager
to work with the federal government to realize this important goal.
*************************************************************************
Statement of Remarks by Kenneth M. King, President, EDUCOM
to the
Science, Technology and Space Subcommittee
of the
Senate Committee on Commerce, Science and Transportation
August 11, 1988
*************************************************************************
I am pleased to address the subcommittee on the subject of computer
networking, an area that EDUCOM has been working on for almost 25 years. As
President of EDUCOM, I represent a consortium of more than 550 universities
engaged in cooperative and collaborative programs involving information
technology. In addition EDUCOM has more than 100 corporate associates with
a stake in information technology in higher education that help support our
activities. The higher education community has defined two major priorities
for cooperative activity in the arena of information technology over the
next decade. One is creation of a national instruction and research network
and a knowledge management system on that network. The other is to develop
and infuse into the curriculum information technology programs to improve
instruction and learning.
I am here today to address the first of these priority areas. Our
goal is to connect every scholar in this country to every other scholar, to
national research resources, and to knowledge databases. This goal includes
researchers in industrial and governmental laboratories. Ultimately we
strive to connect every scholar in the world to every other scholar in the
world, so a component of our effort is international in scope.
Networking is making a major impact on the lives and productivity
of scholars. First, it has reduced the isolation of scholars at small and
remote institutions by enabling them to collaborate with colleagues and to
have access to specialized research resources such as supercomputers,
accelerators and telescopes, as well as information resources such as
national bibliographic databases. Second, it has assisted in bringing these
scholars more fully into research activities in their disciplines, which
will not only enhance the rate of research progress in many fields but also
help to keep instructional programs at small institutions current and help
them recruit able faculty. An important role of government support for
networking is to ensure the broadest possible connectivity and access to
scholars at small and remote institutions. Without government support these
institutions will be left out.
Academic networks are essential to the conduct of big science. For
example, progress in understanding the fundamental nature of matter
involves collaborations among hundreds of scientists at many institutions.
Major experiments on accelerators require several years of intensive work
by geographically distributed specialists who must coordinate their
activities. This coordination is enabled by effective electronic
communication. When a rare event in the heavens occurred last year, a
supernova seen over Chile, an international effort to get experiments in
place around the world in a matter of hours was greatly facilitated by the
BITNET network in the United States and Asia, and its European counterpart,
EARN.
Studies have shown that productivity is proportional to access time
to information. By greatly reducing access time, networking is a critical
element of the national competitiveness and productivity equation. The
explosive growth in overnight mail services, fax services, and corporate
and university electronic networks over the past five years is testimony to
the importance of reducing access time to information. Our vision of a
national network would enable text, digitized sound, and pictures to be
transmitted from one place to another in seconds rather than days. Scholars
at colleges and universities, industrial laboratories, and government
laboratories would be able to work simultaneously on a common body of
knowledge. Thus the network will accelerate the transfer of research
results to the private sector. The impact of this technology on scholarly
productivity and national competitiveness will be profound.
Networks have made it possible to assemble all the current
information on a research topic in one place and to make this information
available electronically. Examples include databases dealing with toxic
spills and toxic substances, land use management, chemical and molecular
structures, as well as emerging databases on AIDS and high temperature
superconductivity research results. Being able to access almost everything
that is known on a subject electronically and developing a consistent
scheme for assembling information from many sources on the network in order
to develop new information are major goals of the national networking
effort. The impact of this capability on all sectors of society will be
extraordinary. Thus all of us have a major stake in the development of this
technology.
The importance of this technology coupled with the absence of an
effective coordination mechanism has resulted in a rapid proliferation of
networks that do not effectively communicate with each other, but that
nevertheless have had an enormous impact on scholarship. EDUCOM manages the
network information center for one of these networks, BITNET (Because It's
Time network). This network, started in 1981 by the City University of New
York and Yale University, has grown to connect more than 400 colleges and
universities in the United States and more than 800 colleges and
universities worldwide. To join BITNET, an institution must lease a
communications line to an institution already on the network and agree to
provide at least one connection for another university. Thus the network
has grown and is continuing to grow at the rate of more than 100
institutions per year worldwide. This network costs each institution about
$25,000 per year to sustain. The success of BITNET was greatly facilitated
by a grant from IBM to EDUCOM and the City University of New York. A
similar IBM grant helped develop BITNET's Canadian, European, and Asian
counterparts. These networks are now self-sustaining and connect thousands
of computers from virtually every manufacturer.
The low speed BITNET network supports the transmission of
electronic mail and small text files. It does not support interactive
connections to knowledge databases, nor does it enable the transmission of
digitized sound (voice, music, etc.) or pictures. An additional problem is
that the BITNET technical protocols are incompatible with most other
academic networks, which requires users to negotiate complicated and often
unreliable "gateways" between networks.
The current situation is not unlike the situation that might have
developed if the states independently had responsibility for planning and
developing their highway systems so that major highways crossing one state
might not connect to highways in adjoining states except through side
streets. Signage might be different from state to state. Interstate traffic
and commerce would be stymied.
In an effort to create coordination at least in higher education,
forty institutions of higher education have contributed funding and people
to an EDUCOM Networking and Telecommunications Task Force, which has been
working on all of the problems of creating a national network. This group
quickly realized that our dream of a true, high speed, national network for
higher education (a digital superhighway) could only be realized through
cooperation among higher education, government, and industry.
In April of this year, EDUCOM, the National Association of State
Universities and Land Grant Colleges, and the New York State Educational
and Research Network sponsored a major national networking conference, held
here in Washington with the assistance of the National Science Foundation.
The proceedings of this conference, contained in the current issue of the
EDUCOM Bulletin, provide the most up-to-date statement of government,
industry, and educational perspectives on the national network. Copies of
these proceedings have been made available to the subcommittee members.
Speakers at this conference included representatives from industry,
such as Ellen Hancock, Vice President and General Manager, Communications
Systems, at IBM; A.G. Fraser, Executive Director of the Research
Information Sciences Division at AT&T Bell Laboratories; and Robert E.
Kahn, President of the Corporation for National Research Initiatives.
Representatives from government included Congressman Douglas Walgren; Erich
Bloch, Director of the National Science Foundation; and Henriette D. Avram,
Assistant Librarian for Processing Services at the Library of Congress. In
addition there were many key people in higher education, networking
specialists as well as senior administrators, such as Dr. Donald
Langenberg, Chancellor of the University of Illinois-Chicago and Dr. Harold
Raveche, President of Stevens Institute of Technology.
These representatives shared the view expressed by Ellen Hancock in
her keynote address: "We believe that national research and education
networks, based on the cooperation and interactions between government,
industry and higher education, are vital for the future. First and
foremost, a national research network is strategically important to our
country's economic future." She went on to say, "We have a national
transportation system, a national postal system, national broadcast
networks, even a national anthem. It is time to get our scientific networks
singing the same tune, too."
The theme of cooperation among government, industry and education
and the challenge we jointly face is depicted by the picture on the cover
of the EDUCOM Bulletin. It shows three men trying to ride a unicycle down a
communication line to a stable platform. The man driving the unicycle is
from government. On his shoulders, and thus able to see farther, is the
representative from industry. Riding his shoulders, and thus able to see
farthest of all, is the representative from higher education. Success
depends on coordination and cooperation of these elements. The key to
stability and the driving force is government leadership. Unfortunately,
the current state of affairs is that government support and leadership are
unstable, with the possibility that the whole cooperative effort may fall
apart. I would like to focus now on the critical support and leadership
role of the government in creating a national network that will enhance
scholarly and industrial productivity.
The National Science Foundation has taken one major step toward a
national research network by its funding of a contract with MERIT, Inc.,
for the NSFNET. This high capacity backbone service network is managed by a
joint venture of IBM, MCI, and MERIT (which is an academic network
consortium in the State of Michigan). This backbone connects to mid-level
or regional networks at 13 places and costs about $15 million per year. NSF
provides $3 million per year and IBM, MCI, and the Michigan Strategic Fund
jointly provide about $12 million per year for the project. This
constitutes a major example of government, industry, and university
cooperation and of government's catalytic potential for leverage. The
backbone, a successor to a smaller, lower speed backbone network, became
operational at 1.5 million bits per second on July 1, 1988.
Mid-level networks connecting more than 200 colleges and
universities have formed and are being formed. These high-speed networks
have been supported in part from the National Science Foundation and in
part by state, corporate, and university contributions. It costs of the
order of 50 to 100 thousand dollars per year per institution to support
communications costs to connect to a mid-level network, not counting campus
networking and support costs. As yet there is no stable, continuing source
of funding for these mid-level networks and, beyond the 200 best-funded
colleges and universities, no identifiable source of funds to connect the
next 200 colleges and the next beyond that. Together, the backbone and
mid-level networks constitute the interstate network system.
Building a campus network connecting scholars to the mid-level
networks costs about $500 per year per faculty or student network user.
Many campuses are building these networks largely with their own funds,
because they have faith that funding for the interstate networks is stable
and will be available to upgrade them as demand increases. They must have
faith also that the policies and programs permitting the development of
important services on this network will fall into place. The aggregate
annual cost of connecting the scholars at 500 universities to the emerging
network is estimated to be $250 million per year. The EDUCOM Network and
Telecommunications Task Force has just published a case study book on
successful campus networking projects at 10 campuses. Copies of Campus
Networking Strategies have been provided to the members of the committee.
In order to upgrade the interstate networks as institutions develop
their local networks, research and development in networking is required.
No source of funding for this effort has been identified. The 1987 Office
of Science and Technology Policy (OSTP) report, A Research and Development
Strategy for High-Performance Computing, estimates the annual cost of this
R&D for communications at $50 million per year. In time users of these
networks can be expected to support their cost, but this will certainly not
be the case during the next phase of development. In fact when the network
is operational and reasonably mature, the users should pay its costs. We
estimate that this will require five to ten years; BITNET took five years
to be self-sustaining.
Thus, federal government funding requirements may be of the order
of $100 million per year for five to ten years and should serve to build
the interstate portion of the national education and research network. A
major component of government support is required to ensure broad access
and the stability of the network during its developing years. It took
BITNET seven years to get at least one institution in every state
connected. I would hope that government support would ensure this level of
connectivity for a higher-speed network in the next two years. A major
current problem is the lack of stability of funding and a plan to insure
that if campuses put their funds into a local area network, a connection to
the national network will be available when they need it.
Aside from funding, a critical requirement is government
leadership. Five federal agencies NSF, DOD, NASA, DOE and HHSare involved in
supporting networking and are members of a federal research interagency
coordinating committee (FRICC). I do not believe that a committee can
provide leadership, although it can inform and support leadership. I
believe it is critical that the Congress and administration define an
effective leadership function. One possible mechanism would be to ask OSTP
to define this leadership function. OSTP could ask the FRICC to define a
leadership mechanism and protocols defining the role of the key agencies in
this effort. Without leadership and planning and stable funding, the whole
effort could collapse. It is urgent that the leadership issue be resolved
immediately.
There are many other important issues, but the key ones are
leadership and funding during the development phase of this effort.
Leadership from the federal government is as essential to this effort as it
was to developing the federal highway system. In the past year, I have
traveled to Asia and Europe to talk to university people building
networking support. The stability of funding and the leadership of
government in this area are not an issue in Japan and in many countries in
Europe -- Germany, France and England in particular. I believe that the
first country to provide an effective research and education network will
have a significant advantage. To achieve this advantage, we urgently need a
commitment from the federal government.
For more information, contact EDUCOM, 777 Alexander Road, Princeton, NJ
08540 U.S.A. telephone 609 520-3340; telefax 609 520-3975; BITNET:
INQUIRY@EDUCOM.